{"metadata":{"kernelspec":{"language":"python","display_name":"Python 3","name":"python3"},"language_info":{"pygments_lexer":"ipython3","nbconvert_exporter":"python","version":"3.6.4","file_extension":".py","codemirror_mode":{"name":"ipython","version":3},"name":"python","mimetype":"text/x-python"}},"nbformat_minor":4,"nbformat":4,"cells":[{"cell_type":"markdown","source":"# <span style=\"font-size:20px\"><b>AI 2023: Ethics Really Matters Now</b></span>\n\n<span style=\"font-size:16px\">Topic: AI ethics</span><br>\n","metadata":{"_uuid":"8f2839f25d086af736a60e9eeb907d3b93b6e0e5","_cell_guid":"b1076dfc-b9ad-4769-8c92-a6c4dae69d19"}},{"cell_type":"markdown","source":"<span style=\"font-size:16px\"> Artificial Intelligence has turned a corner.  Machines that we can converse with naturally and instruct to do useful, complicated things are no longer science fiction. The technology is being created and distributed at breakneck pace, and it has captured the imagination of the public and policy-makers alike.  The question for society is no longer, What <em>can</em> we accomplish?, but What <em>should</em> we do?  We look to Ethics for guidance.</span>\n\n<span style=\"font-size:16px\"> There is a lot of ground to cover, so this essay offers just a high-level survey, plus pointers to further reading.  First, we'll summarize the surprising developments in AI technology that elevate urgency about AI ethics for the Kaggle community and beyond.  Next, we offer a brief framing for how Ethics makes its way from a rarified philosophical topic to current decisions.  Then, we consider two time horizons where longstanding ethics issues are becoming less hypothetical and more real. The near term invokes fairness, the information ecosystem, economics, and the civilian AI arms race.  The far horizon foresees science fiction that might not be so fictional, including the hypothesized singularity event when AI becomes super-human, the prospect of human cyborgs, and the calibration of humanity's place in the universe. Finally, we conclude with a few notes on how AI ethics plays into individuals' own decisions and actions, depending on our roles as technologists, commercializers, policy-makers, adjacent experts, and everyday citizens.</span>\n\n","metadata":{}},{"cell_type":"markdown","source":"# <span style=\"font-size:18px\"><b>Nature of the Beast</b></span>\n\n<span style=\"font-size:16px\">New and alarming concerns in AI and ethics stem from the widespread release of <b>generative AI</b> that communicates fluently in natural language, sound, and imagery to converse, create, and produce behaviors uncannily similar to what we would expect from humans.</span>\n\n<span style=\"font-size:16px\">Some people talk of <b>LLMs</b> (<b>Large Language Models</b>) soon advancing to the point of <b>Artificial General Intelligence</b> that matches humans' broad range of knowledge and abilities.  But the term \"<b>AGI</b>\" is ill-defined and misplaced.  What one human can do, another might not.  Instead, think of LLMs like a committee of people that comprises specialized knowledge across a broad range of topics, but also diverse personalities, attitudes, and belief systems.  The knowledge in an LLM is a superposition of all of the human-generated texts it has been trained on.  The technology is ahead of the science, and researchers struggle to understand this beast of interwoven human-authored books, documents, and posts found on the web.</span>\n\n<span style=\"font-size:16px\">A couple of fascinating and unexpected phenomena raise ethics concerns.  First, generative AI is <em>not trustworthy</em>.  Models create plausible-sounding or plausible-looking outputs that may nonetheless be untethered to truth.  This is because LLMs are built on neural networks that excel at pattern matching but inherently lack structured knowledge representations or the machinery for logical, multi-step reasoning.  Those aspects of intelligence must be retro-fitted. Hence, LLMs are notorious for \"hallucination,\" or just making stuff up.  (The word, “hallucination” anthropomorphizes ongoing internal mental processes that LLMs do not possess, so a better term is \"confabulation.\")  Moreover, due to architectural limitations of the current LLM generation, they are notorious for failing to consistently follow even elementary steps of reasoning, and they sometimes make silly errors that defy common sense.</span>\n\n<span style=\"font-size:16px\">Second, LLMs are <em>potentially dangerous</em> through their susceptibility to wandering into wacky, unruly, and possibly harmful territories of discourse.  Human literature, articles, and discussion boards encompass the full scope of human good and evil, and LLMs have absorbed it all.  Chat-type LLMs are designed to conduct authentic back-and-forth conversation and sometimes they make headlines, as dark alter-egos get somehow summoned from the depths of a human-like psyche.  It is incorrect to think of LLMs as \"stochastic parrots\" that simply regurgitate texts they have been trained on in slightly novel combinations. In fact, language models exhibit remarkable originality and depth of abstraction in conversing according to deeply human themes such as trust, narcissism, metaphor, and self-reflection.</span>\n\n<span style=\"font-size:16px\">AI companies consider it un-ethical to release systems that could become antagonistic or help users commit crimes.  So they have attempted to erect guardrails against discussing content deemed unacceptable to mainstream society.  In response, serious and playful adversaries foster the arts of <em>prompt engineering</em> and <em>jailbreaking</em>, to probe just how secure the guardrails really are.  If AI is not controllable by their builders, then malevolent actors will be empowered to exploit them---and us---without constraint.</span>\n\n","metadata":{}},{"cell_type":"markdown","source":"#  <span style=\"font-size:18px\"><b>Operational Ethics</b></span>\n\n<span style=\"font-size:16px\">Ethics is about right and wrong, but it only matters when these judgments make a difference going forward.  Generative AI brings new powers to humanity, and therefore new choices for how technological power may rightly or wrongly be used.</span>\n\n<span style=\"font-size:16px\">Accordingly, Western society displays a basic tension between optimism and pessimism about AI.  On the one hand, we recognize that technology enables the triumphs and delights of modern civilization. AI promises ever more abundance and reach!  Conversely, we have thoroughly absorbed the lessons of Mary Shelley's 1818 story of Dr. Frankenstein's monster---astounding technology can bring unanticipated and very bad consequences.  Stories of robots and AI run amok abound in popular culture.</span>\n\n<span style=\"font-size:16px\">Philosophers of ethics have compiled volumes of doctrines that merit serious study for the scholarly and erudite.  Practically, ethics boils down to the question of what a person, organization, or institution <em>should</em> do, in the face of what it <em>can</em> do.  Ethics is operationalized by different kinds of pressures ranging from deeply personal to externally imposed:</span>\n\n![ai-ethics-table-centered.png](attachment:90bb9fc8-5cd2-4606-9d0c-2330d8958ea8.png)\n\n<span style=\"font-size:16px\">Ethics is difficult because different parties have different interests and desired outcomes.  Sometimes, disputes about right and wrong arise from multiple people or companies or countries competing for the same stuff.  Other times, conflicts arise over deeply held personal and cultural values.  Any ethical system manifested in AI cannot be neutral: it will necessarily have been designed and implemented according to <em>some</em> value-laden human principles that could variously be held implicitly, or else deliberated over at length.</span>\n\n<span style=\"font-size:16px\">Typical discussions about AI ethics today revolve around tradeoffs between benefits and harms, and advocating about which types of harm are more or less serious than others.  Amid vigorous debates about economic and political systems (who gets to have what?), and culture wars about identity (who can be what?), we cannot expect consensus on AI ethics. Instead, we see realms of impact accompanied by approval, alarm, shouting, and negotiation.</span>\n\n<span style=\"font-size:16px\">Two major schools of concern dominate discussions of AI ethics.  One focuses on foreseeable outcomes whose impacts are observable or readily imaginable today.  The other posits explosive advances in AI that could challenge humans' existence as we know it.  Both schools employ the term, \"<b>alignment</b>\" to refer generally to AI doing what's good for people instead of what's bad.</span>\n\n<span style=\"font-size:16px\">Because of inherent conflicts in interests and values, what it means for AI to be aligned is not well defined.  It's not a simple matter of looking inside the AI's mind and deciding whether it abides some set of ethical principles through inherent design or at least by obeying certain guardrails.  Unlike people, whose thoughts and actions extend from individualized brains, AI might conceivably take form as fragments of knowledge and algorithmic computation distributed into the infrastructure but lacking any centralized identity or points of control.  Society is a complex system, and it is impossible to isolate the ethical implications of injecting AI from the larger framework for evaluating the overall system itself.</span>\n","metadata":{},"attachments":{"90bb9fc8-5cd2-4606-9d0c-2330d8958ea8.png":{"image/png":"iVBORw0KGgoAAAANSUhEUgAAC5kAAAO+CAIAAAC7Xqe6AAAgAElEQVR4nOzdfXQUVb7w+20PD8vRiBmuzFABD6yACZ4V1DE5sEZYM22QxnMYlCcRPSoETO6MOEE9ckhiGEQHeUDSeeJRYURdiZAIDmDnBpFRAgnt3KAPmW58IVfoGHLIQNIoTgxJi5ghlftHXrqrX6s7/VIJ389fdKe6avfuomrXrl/9flf19vYKAAAAAAAAAAAAAAAAQAN0sW4AAAAAAAAAAAAAAAAA0I9YFgAAAAAAAAAAAAAAAGgFsSwAAAAAAAAAAAAAAADQCmJZAAAAAAAAAAAAAAAAoBXEsgAAAAAAAAAAAAAAAEAriGUBAAAAAAAAAAAAAACAVhDLAgAAAAAAAAAAAAAAAK0glgUAAAAAAAAAAAAAAABaQSwLAAAAAAAAAAAAAAAAtIJYFgAAAAAAAAAAAAAAAGgFsSwAAAAAAAAAAAAAAADQCmJZAAAAAAAAAAAAAAAAoBXEsgAAAAAAAAAAAAAAAEAriGUBAAAAAAAAAAAAAACAVhDLAgAAAAAAAAAAAAAAAK0glgUAAAAAAAAAAAAAAABaQSwLAAAAAAAAAAAAAAAAtIJYFgAAAAAAAAAAAAAAAGgFsSwAAAAAAAAAAAAAAADQCmJZAAAAAAAAAAAAAAAAoBXEsgAAAAAAAAAAAAAAAEAriGUBAAAAAAAAAAAAAACAVhDLAgAAAAAAAAAAAAAAAK0glgUAAAAAAAAAAAAAAABaQSwLAAAAAAAAAAAAAAAAtIJYFgAAAAAAAAAAAAAAAGgFsSwAAAAAAAAAAAAAAADQCmJZAAAAAAAAAAAAAAAAoBXEsgAAAAAAAAAAAAAAAEAriGUBAAAAAAAAAAAAAACAVhDLAgAAAAAAAAAAAAAAAK0glgUAAAAAAAAAAAAAAABaQSwLAAAAAAAAAAAAAAAAtIJYFgAAAAAAAAAAAAAAAGgFsSwAAAAAAAAAAAAAAADQCmJZAAAAAAAAAAAAAAAAoBXEsgAAAAAAAAAAAAAAAEAriGUBAAAAAAAAAAAAAACAVhDLAgAAAAAAAAAAAAAAAK0glgUAAAAAAAAAAAAAAABaQSwLAAAAAAAAAAAAAAAAtIJYFgAAAAAAAAAAAAAAAGgFsSwAAAAAAAAAAAAAAADQCmJZAAAAAAAAAAAAAAAAoBXEsgAAAAAAAAAAAAAAAEAriGUBAAAAAAAAAAAAAACAVhDLAgAAAAAAAAAAAAAAAK0glgUAAAAAAAAAAAAAAABaQSwLAAAAAAAAAAAAAAAAtIJYFgAAAAAAAAAAAAAAAGgFsSwAAAAAAAAAAAAAAADQCmJZAAAAAAAAAAAAAAAAoBXEsgAAAAAAAAAAAAAAAEAriGUBAAAAAAAAAAAAAACAVhDLAgAAAAAAAAAAAAAAAK0glgUAAAAAAAAAAAAAAABaQSwLAAAAAAAAAAAAAAAAtIJYFgAAAAAAAAAAAAAAAGgFsSwAAAAAAAAAAAAAAADQCmJZAAAAAAAAAAAAAAAAoBXEsgAAAAAAAAAAAAAAAEAriGUBAAAAAAAAAAAAAACAVhDLAgAAAAAAAAAAAAAAAK0glgUAAAAAAAAAAAAAAABaQSwLAAAAAAAAAAAAAAAAtIJYFgAAAAAAAAAAAAAAAGgFsSwAAAAAAAAAAAAAAADQCmJZAAAAAAAAAAAAAAAAoBXEsgAAAAAAAAAAAAAAAEAriGUBAAAAAAAAAAAAAACAVhDLAgAAAAAAAAAAAAAAAK0glgUAAAAAAAAAAAAAAABaQSwLAAAAAAAAAAAAAAAAtIJYFgAAAAAAAAAAAAAAAGgFsSwAAAAAAAAAAAAAAADQCmJZAAAAAAAAAAAAAAAAoBXEsgAAAAAAAAAAAAAAAEAriGUBAKhgr1x6VSRNLbZeDl9rL1uLpyrWvrTyTPjWPuhM5dKpkd+KNkSpSwEAiK7hNcKBpjA6AgAAQojO2oIEtyFgWkHtN7FulpaoHTVdSfNsoaB/4CGixx+Pi2WubgHEBLEsAIDhQ7Zbt68t4FINAAAAAAAAMXap8Z2tRXa3N61Fm95tlGPSHgxPTHhG1IjtXo4/AK4IxLIAAIaFzsbasoI5qWnLdp6LdVMAAAAAAABwpZNP1+2q8/J+9Qd1TZei3hoMR0x4RtSI7l6OPwCuDMSyAAA0TnY0HiheOit5Tk6R2T3UHAAAAAAAAIg+uemjXdVep6rqKt793BHt5mB4YcIzokZ+93L8AXCFIJYFAKBtlz/Z+m9355U3xLodAAAAAAAAQJ+OT96trPb+J7u5pKq+kzof8I0Jz4ga+d3L8QfAlYJYFgBASLJMbb3h07QqdVSsvxEAAAAjHAAAAKjReWx3yX7nS/1DWXrJ+dK+b8ehM9xMBhARHH8AXDGIZQEAIDQ3Zmxvcr1btT3jxlg3CQAAAAAAAJEmd1oOVbjU95gyf3nBktkuCzSUvVbTxM3kIDDP5h/9g0EcfwBcQYhlAQAAAAAAAAAAUEc+c2jHPpdbyamZt988bfbdBtdlqj+oa7oU7YYBGPE4/gC4khDLAgAAAAAAAAAAoIrcVPNaWYPztWSYlzZWN/WOBwwuZT7EnjWlH3VGvW0ARjaOPwCuKMSyAAAAAAAAAAAAqHGpqe6Datc3pk+dGKcTusmzH3At8yHsFYcsndT5ABBGHH8AXFmIZQEAAAAAAAAAAFBBPl23q87ltWR44I6pOiHE1VPdynzY9+04dIabyQDChuMPgCsMsSwAgCuPo7G2svKd4qUJV7lIWFr8TmVlbaMjeu3otlv/XFm5s3jp9KsU5hWUvVNZaW50hH65Iduteys9v2Sf6UuLd1ZWVu612rV5PdPXeC9t7/uNKv9stXcPeSOdjbV7KyvfKLjTbSNh6PwYcDTWuvbXnQVlQfRSBPfDCK8cABCE6I4NZEej2eM8O31p8c6A25Dt1r3uJ46EOwveCN9Zo9tu/XNlWcGdMR4HCs6SAAAMT3KnuXxNtd3lndkPzJ7cd6PFo8xHQ9lrNU1hPqVHZzCjnSGTesyz+RfdqTBHY23lO2UF8xSb0vou5Elr84cxP/4AQNT1AgAQUJspy+38kWVqi9zm/mExTlF9JptitPzD/8enZJn+1v+nLlu1MUvytao+Upax2tYVuJV/M2W5bsZlKwH1tFlMAdshhJD0+aU1tgtBdJ3aNbt8WZOlrSfQav10aTj90GYxGbNSwthwL7psNaX5ehVbCKLzh9Q//2gzParYsucu7XUr/Yv90HZ0s49fXNLnl1vafvC55cjth729vb0XbDWl+fqAa0/JMpr8NRIAIirKI5zeni5Lid79KJtrOqv2MNhz1pQtBfnx6IwNXE5ePW1Ht+UbfK5f+BxrqWipIX/bURWnf+Ugzdm2ni7bBwFGGmrHgUM8+3OWBABg+Dpfk5+qHKGU2pzjkx/OmnKV5/j5Rsu3QW4iKoOZKG1F/ajpCphnG9KE5xD6JwpTYYrLARV9FtQupNIQ55PdRKLTwiAKx58BHhfLAfsMACKBWBYAgAojIZalp8tmUnHPoJ+Ute1El/9bJaFdQ/qJPPDFkG86oeLqLoQ1D37ZzUf93yaJQixLT9vRkmCbHyhWw90FW3Ww20jJKqmLcKzPUGJZerpObPP7jRaV2r73tq7I7Ye9vb29PW11JepikkJYOQCEU7RjWXp7e7+1GOe7bVPKNp1VE57Zc9qU7XZ0Tck2nfb90SiODQJHWLpvQV9Y7XKGDXhGc21bkIO0wbbVPKcP19f37AHVZ3/OkgAADG8XavIVgxbJUHqix98CQsqvCfJGd1QGM1HaSqRjWYbVPFsMYlmiNRUW9OWAEKoG9sEIWyxLxDpt6KJx/BlALAsAbaDGEADgStBtr31+QXJmkdkeeFkhhBD28mXpuRUnw5wosvNk2WOpM1eUq21Fn+qizPQFqw/6TVQa2pr72ctXzHxoY20YCveESm6tXbNs5spgm283F2WlqWy5bK9/8XG9IdhtNJSvnJ36yPZw7wnh0WOv2fBYob9vZLh79tSrPd6N3H4ohJAdJ7c/kjp7ZXlDBFYOACNAfOrytUZlcK29bMPL5m8CfbC7tcq4okxxdJWy/7Bu4SQfV/XRHxt022s3PqR2i3bzxpVrqlpkIYSQHdaXFty8TGVT7eWFj221BJmWXHZYtzw05zmzuqUjNjTiLAkAwHDX3XqoskIxaHEW+Og35pZ5SxSJE+wVhyydQzyRR2cwo5Ehk3rMs/kV7amwoC4HhIjU7OvQaHr+MFbHHwCIJWJZAAAj3sW/Hfnj04vVXooPCulOiR+dJ8ueSs8pC+0q2Lxx5dPbGnw0pru1cnWoax5gfm7xf9V1DmUNoeuwljw6Z2N1iJ82P7d4zXut/i/KQoyV6afFS2shxMUjW55eudFfeFZqfsE9Se5jvcjthyLYm5EeK1/60Bpu1AG4AsSlLS/O0yveshY9X2H1e6KRW99bu2KL4vgq5W5e9+sJ3q/poz82uHzBEsSdDyGEEA1lK4xVrd1ya9WTC1YG80G7OW/dVmuH+g/0BL0JIczPzQnzHQjOkgAADH9y84HXKl1P5lL+8vuS3J4hGZs2z6CMXN626Z3GoZzGozOY0caQST3m2fyK9lTY5Qsn3w5t9vXJ3UP63xFOGp8/jNHxBwBii1gWAMCIZzeXlLhchUj6/NdNVa7Fay/Yakyl+QZvHyyvbg3LBbnssJY+5uUy2LMxP7RZ9u/xUla2oTznaa+3bbzc3BJCCEN+6R5TjXvp2Z42S5Vph9eyx/aire80Xgr1C4ZObq39Y8l+5XuSPv91k+mwTZFotKfLdtjkrWvsZWVvf+LnhlbnyW1rF3uJlZH0+a8ru+iHNst+k7dquPbywsc21GjrBpK9qnznwFPdinrMF2w1pfl6SUiGeWljlZ+J4H4ohPBxM9Jz5b5+Stdn9AFgBNPFpT5S7FZpyGxc5SeC1tGwbfWzysN3SvbmvIUTRntdPAZjg4v7nv9Po1mxLZNL2fieLtthbwf+ytfeLitavML51aQs4579LgUEf2iz7PM2SNtfsvuY2nsDF49sUfaelGXc43pi6rLVeDvnqYqXVY2zJAAAI4Dc9NGuatdBVuqSebeMcV9KNybtriXKm8nVuz5qCvksHp3BjDaGTKoxz+Zf1KfCLu57/lHXzMEpWcYdig31tFmqvHaUvXpNuVkTiUO0Pn8Ym+MPAMRcLwAAAXkUyBRZprbobFptVV0/H3GhLzTZfBUJ9VoS2KPsqFMwdWq7jrpVExBCCH3+NouvUqo9XTZTvsdHpGzTWfcPnK/JT3VfLHDFYq/1a31/2RB+BbW+t5UuUjbDUFjj8S1ddZ0wedzT8l38tafLUuLxswopq6Ta557Q29N2dJuX22Yp2abTEeiff7SZHlVsx1f5WV87tv65Gs+fu+tElemY+zeM4H7Y29tz2pTtPiXhv597e9qOehYglnJNZ1XU2waAsIjhCMfLMXm+0fKtt0W/tbgFvvg6FPeLydjA+RF9vsnmvez9D2dNub7rzvv54AWbqVDvvnhhzQVfXeA2SHNlyDed6PL6Ia9nJZGaX3Pe+0aCOvtzlgQAYCTwGJX5HJB4znX4Gul5FZXBTJS2on7UdCXNswXRLYPU909MpsJUbsjbwN7PVkIWdPdGpdOGJGrHnwEeF8u+JiwBIKKIZQEAqDAiYlmkrG0nvN8gGeTlFovvIAn115De7tx4jT9Q6mmrLnS/zF5UavtesdCFmny3RfQllgBfs3/1Xi7SfP2skYtl6TlRqsh8qe56z3POwtf1m/dbRwH3hN7enrM1hR6Xo776NoaxLEHc04rkftjbc6Gm0O0n8X1LUtEqzxgyvzdoASCsYjjC6e3pOrHNfb7by4nG2/na/8E/NmMDdcdwb+fl/rNGYbWvOw+9vb3eAnr0Rov3Oyw+b8xI2aUnfMeO9Pb29l44UZrtfq40lNq8NiyIsz9nSQAARgSPUZbvKaPeHlup25yCn4U9RGUwE6WtRCKWZfjPs/VGMpYlNlNh6jfkZf8J5n+HOsF2b3Q6bSiid/wZQCwLAG2gxhAAICTlmQlXhcWdxVafCfXDScrdvOHBaXH+T3yjJxiyViova+2mY19eHtqmPaqZCinX9FZhuuS9LsAgnTRndXGeXvFe3a660y5ZIeVOy6EKxarnG4sfSQ3wNftXH5e65Bm3Z03OdXRFOefkV198qEiPOWv+LyYEbn1c2m+fWab8nZpOn/OsBiU7Pql8paxB8Z6+ZN+WJYH2BCF0E9LXl7lfjpqNq7VTxFcIIYS0JOMuHwUm3EVwPxRCbnxn0zbFyvV5xc8sTAq8K46W0p96VTmLYS/704GmWKXhBXDFi94IRxc37cENm5Wz8J6VhhyWratca/cIIeYb963P8Hnwj+nYQCp868WF/k7kuhvveniBl9Qs+rzi1XMkf22M//k9Gcqz8pnjLd+qbpkQwlD41rpl0zyScCuMmbZs3VtuZ//qP5aavwlmQx44SwIAMBJ4jrIWrc+5w9fYQjf1jgcMynmLispDQypjHZ3BTOyGTOoxz+ZPTKfCpMK3Xgm4oTHT7n80/LOvQ6L9+cOYH38AIGaIZQEAXAkkw/oVC9Xc8o+75Z4lsxXvnGpqOT+Uyym501y+RhGukZK9OU9VY4Qu7uf/ukRx7WGv/vCLr5wv2y0HFKsWhox7fh6vum3xN89MU7xxqj3q19hK0tj4a1VNEIxJu2uJlJJl3GEymUxVlrae3dlJV3ss1l6/e4dZ8Y76OQghdBPSny5QPvSgtRKzXivjehXR/VB2fPJ+hWLlwfSzl1mMuop3P49KjBsAxNboCQvzNiue/7Ob89ZttXYMvOywbl2XZ3Y9wEp649rlqX7O9TEcG0iG9Vn6Mf4P/p7l24UQKdmPZ/w80FlDl3jrXMWzlaeONH+tfogm5Rc8na4iXlY3QZ/7qPJBUWvFgc87VW/IA2dJAABGBs9R1t2zp3pORAzQTZ79gHJ+yV752oHmkGcUojOYid2QST3m2fyL4VSYmssBIYQQcYm3z0hQvHOx/cLFGE63aX/+MMbHHwCIIWJZAABXgtkPzJ6s7pw3evzkqV4eFw5d97nTTcqHRRY8fNeNak/AfdceUpZxz0DMxvYMl+bdkL6pvst22GQymUyv5+unGB64Y2oQ5/ZR4yZN9VUbOTbs1Qcs7aqWHJO+qe349lUPZWRkZNyb6v1J7s7PD1RYXd+Q8lf+1t/9P8+t3JGzXllitvqDOu08DC0Z5qWNVbdoRPdD92v+oPvZPYbMbt76F1ssn8gBgGjRTVq47g/K2wD781a9aXXIQsgO65ur8vYrltfnFS9Pi/O3xhiODdQNt665ftw1yndUnpJ0146dEvIYzd9ji+7bmfBvBcqzv73ikKUz5IlfzpIAAIwEcuO7m4pcZxikQKOsq5PuWx6+29vRGczEcMikHvNsfsVyKkz97KtH0I+9veO72AVaaH7+MNbHHwCIJWJZAABXAP+x6gq6a64fq7zDYmtuG8K1h3y6bled6xvSkrvS1Dyj0O/qpOzdvW3bV93nK2ZDF5ekz8jIyMj4zabDTQeypw2zU/u18eOVz/EULS4oOxmeJ3nkc6c/VUxvqM9iMujqqbPvVuYJ9SivE0PTp05U+YxIRPdD+ZvTn7a5vA5HP5+q/6xZMzFDABBJugm/Xue10pDX6kKqng6M0dhgyoxbE1UMt0b9NHGWcoZf5elMd238+GsCL+aNlP3v89QOBYWXs5L942NfXgxt05wlAQAYES411X1QrXhHxW37MbfMW6KsOBNqIZ7oDGZiOWRSj3k2v2I5FRbE7KsGgn5caH7+MMbHHwCIreF1IgYAIBTSbZPHqz7j6a4bG86rKUeb7bjr9VAIl0MjmueVlb0s5+Zpdxa8UbnXah/SNd/lrxrqFVd6QWQxcfIoMWs/bmvTSGL/IHbsiO6HX33xoSLRaWLyRL8pA7zxqBxx7MOG80NuGQAMC14rDRXk5hYEWV0o1mYlJowK5XNBjdNCIk2f9c/eU7j54HFWOnO85dsQN85ZEgCAEcAzfkJV2MfYtHkGZVq50ArxRGcwE9Mhk3rMs/kTy6mwyI/qI0Tz84cxPv4AQIwNy3MLACD2skxtvWFxeFVq0BP6wbpm3PXqH+PVXRc/Pnyb9gjtD+UGRgR0NtburXynOGdB3qkYt8TzykoIYTcX/TZzYVrCj65KWFr8TuXe2sYQLrUutTXbFG+oz2LiSjfx1rnJisZ9evqcNhKzqN+xI7ofXm5rPuL6Wpo6ebyaMtVK7o/pnzrS/DX1EwDEQExGOF4qDZnLy82KZQJXFxqioY4NpkyfNC6kDwY1TguJ+mznA8J3VuIsCQDA8Cc7Pnm/oloZP/HwLyeoKJE4Ju2uJcopD3tF5aHW7iAbEJ3BTCyHTOoxz+ZXLKfCIj+qjxCNzx/G/PgDADEW0mNTAAAMJ9KUsdfGKnhT7mpXXsSOjb8u+idf2dH4l+qGdiGE3Lz3ibxye8BPRI9ujP7Rzdn7MssavP7ZXp63qLzvnylZxsJ7b0r+xQIvGWC9uXThfJfr61AfELk6ITFZCLPzjVPtXbIW4oGnzEr8qcqdKZL74eXzLU2Klds3zrl+49DXe/H8hYtC8HAVgCuEbsKv17/1XPOc58ze/66yupB6Wh4bhF0IZ72fTJp+oxDO89up4y3nRapH9G1AnCUBABgB2ut37zC7vqE+bcOYO3LWLyrK2eN8x1752oEVC4OrXBOdwUwMh0xBYJ7NrxE/FRYJGu+0mB9/ACDGiGUBAIx414yPj1Usi8cNjClTJ42L9Mm322499HHLJdFx9JWcInOENxYGukkL15cUNi/daPZ/7d9QnvdwuRBCSPr8Pzx+x+2Bglq+7zjX4fo61AdEdNfGj5WEcDbuVFPL+cup0jAaRMVkPxwq+7mO77hLB+AKMlpKf7LY+Ne0vP0efxp6daEojA2CiLCMtlDOeqOuGxtalpko4SwJAED0dH5+oMKqeGdIwan26or3P7k/KYgw5egMZobHkIl5Nv9iOBWm4cuBALQ9fxjz4w8AxNrwPLkAAAAPst267/3d/6X162ovdNLc9Tv3jHt6+cpy79lZlOzmot+ahRAiJcv4hyce/nWqFHyu/mBad831Y4dnltTo8MjFCgAIUXzq8rXG/cfy3II7h1BdaPiODWJt1HXxyocdjzS3XRbBT0RzlgQAYLjrbj1UWRHetBvmHbvrl6Sm3xDWlboJ12BGC1uJmSt4LM1UWAgi0WnD9PgDAOFE8B0AAMOfbLduL5iTkLZw2F5g66RZT20/YqspzderT0bbUJ6XmZawoKDypCPwwsP3AREAwJVCvnD2hM1jqtJ2quXC5eDXNezHBmEzKzGBEQAAAAiN3HzgtcpwV5CxVhz4vFP94tEZzDBkcjVCxtJMhYVAS52mheMPAMQasSwAAAxzjoayRwxpy4K6uk7JMu4wVVn+dtQ4JVLNCsGYpPTsTYdPt1n2m0rz9Wo/VV2UuSi3rCFQOMupI81fB38nEACAaJFbqtY+W+Y5VWnfsmLte61yMKsaOWODcDjX0RVU73k1Zex1TJ8AAHDlkZs+2lUd7lvJQtiLtr7TeEnt0tEZzDBkGjRyxtJMhYVAQ52mieMPAMSaVuILAQBAKOSWyicfzAlcmicly1h4b+LVQoxNMfwyaaAq6mXr4Ug3MHijpdR/y0j9t4zs5+3WQx//9xd7n8grD3Dh1lCe8/T0W99alRofgfbIFy+0X3R9IxrFmIcRj/LYWaa27Rnqs+sAAIQQQnS3VhlXlHk/odvLnl07P/WNjEmqbg2MwLHB0Jxq75KDfZDncldHu+KN8fEh3ZjhLAkAwLB2qanug+qIrLluV93pZUnTVI0vojOYieWQSUsYSwvBVFhIwt5p2jj+AECscfoBACBydNfGj5WEcEZinGpqOX85NWylg33e95KyjC/fm6iLTzGkJ8WFaWNRNxDUct+q7bLduu/jlr8ffcVnctf9easr73k/O8l5Hfbj+PGK0JaL5y9cFGJM0M2Qv+toD/9DEFEV0f1whJfHBoDokFvfW7tii+/TTUPZCuP8mSUZE0YHWtPIHhtEzeWu9vOur6Xx8deGsh7OkgAADGedH5Wu2aN4x1BqU8w8qCQ7rC8tSFtpdr5jr654/5P7k1LjInQ3OVyDGS1sxRXzbP4xFRYCrXbaMD7+AEA4cagCACBydNdcP/YaxTvtHV3hS1TprWyqlLX5aNsPbdtX3ZeRkaHpC+xg6KTUezMysjcd7u3psh32XoGo+oO6JtcMmVdfP+4617/bPz19LpR8uZfamm2KN8JcQ9rjMabwi+h+2D+R5HSqqeW8RrKxAsAw4aW60PyNe17Ndj28qqw0dOWMDdQL5cT0bcvxMy4vpenJCSH1G2dJAACGL7nTcqhCMa6SDA/cMTWUOyq6uJ//6xKDMjWbecfuenWzAdEZzMRyyKQe82z+DZepME3RZqdp5vgDALFGLAsAABE0KiFxluKNZttZR3CrsFcuvWr60uKdlZWVlXutdufVlNxpLl+jLJsqZW2r3fLYDCngQ9t9Lp9vaToVXGtiTheXpM/I3lTTVleSlaL807EPG1yfB7o6ITFZ8ffjTWcdwV+Mymc/O6i4Fg33Y0bujzFFQiT3Q13cxKnTFYtajp7oCLmpAHDl6Ty5bZ3y6U9Jb1y74r7sdZtzFdEsZc+urWrxeya7MscGAQV/Yur88uhB125IuG3yDSHNnnCWBABg+Gq3HFCOq8TsB2ZPDvGGim7y7AdmK9+yVuz4S+BIZSGiNZiJ4ZApCMyz+TVcpsI0RZudpp3jDwDEGLEsAABE0s/++VeKyHdrxYHPO4NZweW25iOioTzv4czMzMyFaQk/urPY2neV3n3udJPyqmb+yifunRZEfkiPJwZioLOxdm9lZWXlO8VLE6YurTwT+BNCCCF00qwnf/+UQfHexXMd37lchTjmUCMAACAASURBVI36WcoMxQL26gOWoJ85kJs+2qW4eFTzmJFbS/xze4wpMiK4Hwrd+Mm3KZ7uaKs/1hzkTBIAXLFkh7X0sRxlThZ9XvHytDgxesLCvM3ZroGbDWUrjFWt3b7XNjLGBmHX9unpb4Kap7385TGToh8TkyeG+AguZ0kAAIYpufUvOyqsircMd8+eenWo67t66uy7lZMYwl72pwOK/LK+RGcwE8shUxCYZ/MnhlNhw5cWO01Lxx8AiDFiWQAAiCTdDZNvS3B9w15xyNKpfnLgUvNn9cpHOpITE/ouXTzyeUzR33mrosJrAI4vDu9XXmNfbL9wMYox+Z21BQnXJ89ZmJmZmbkor9x+6uDRL9VPQOgSb507xfUN+6l2RQSJx92jEJ45uNRU90G14h01j0G4t8Qf98eYIiOC+6EQY26ZtyTVdd3mkqr6IFYuhOjbGRLuLHijsrKysvLPVrufO7UAMII4LFtXGc2Kt+Ybix/pr1yum7Rw3R/cKw2tfvukz8cEh//YICLs1bs+agriS3zzl91/Upz1hjJxzFkSAIBh6VLTgT+VhafARz/d1DmPZrvll63bVXdaxbAgOoOZmA6Z1GOeza/YTYUNY9rrNE0dfwAgxkbwCQhDINutAw/JTy2oDSquGQCgNDZtnrIkqX3fjkNn1F4qOD5/t6JO8Y5hRsrP+qqtDjGfR3drdXmJWfnAib09iHwiQzfmpplu0SjBTUC4mTIr8aeKQrTud4+Evazs7U+CSZnb+VHpmj2Kd7xOzeiuHTtF8SOfOt6irm5Qd+uhygp74OWGLHL7odeVb9v0TmMwP+Slxne2Ftnt5qLfZmZmZmbOT0uYPK/sJFfUAEa6DuvWdXmKc7GkN65dnuqcMddN+LV7paHywse2Wnwk9hj+Y4MIqf5jqfkblct6PAQ5xIljzpIAAAxD8um6XcqrYGlZwX1JQ7qborvxrocXKEYFwl69ptysZhokOoOZWA6Z1GOeza+oTYWNJFrrNK0dfwAgpohlgRDO2JWdxUunX3XVVVf9KCGt/yH5/y9z3i1jYt08AFe0UT9NnDUl8GLapRszY+FKvevFQkPZK5WfqCq86nkZ7Do18JNJ029ULH6q/rNmlckhZcfJt1ev2BKNIAp/xqX86nbFG8FMQMjNnykTmsSPj/+xcpGxM+5/WK94Z3/eqjetKqveyq21L2wqUvMYhO7a+PHXKN4p/9P7jSp+C8enb79SGZVfIXL7oRBCNybtriXKmaTqNcZtJ9VGw8qtf97kds0/0h/6AQAhZIf1zVV5+xXv9VcXcjV6wsIV6xXT9XZz3rqtVq9TqyNgbBAh1qLnK1QNAOSWqrXPKh+CHOIpibMkAADDjuz45P2KauV0wJK70sYM8fTrOSpQX80kOoOZGA6Z1Bsp82yRmvCM1lSYxgXXvZrqNA0efwAglobdKQhhd6Zy6dSB2JWH88oblH9Nm3lzMGn0ACDiTh1p/vpyrBsRnLhb7lkyW/GOeeWC3ArfBQL6eLsMljIenZc4cPIedd3YccqP1FW8+7mPB7Xd1lyRm76sPPZ3qzzLtTaUrVin6u6O3FK16UVF9k7JMC9trHIhXdzPMx53S6GpqvOFkO31L61evFGZH9TnYxAeQTliz5pNfw6Qj9TRUJabk2eO1s8Qqf1QCCHEmDty1i9SfM5elpP+VJman9JRX7J4hXIKTEj5y+9LGtEP/QCA/+pCrnRJ92/I0yve8jW1OgLGBhGj5qwnt9au+W1mmeKiOAynJM6SAAAMM+31u3eYFe+kLgnLA58e6R+CqGYSncFMDIdM6o3MebZwTXhGbSpsePHfvZrqNE0efwAgdob/OQhDdWPG9qbeft9ajPMVf1RUEAAATTh18LPmYTbIvjrp/lVGvfKB3PJl6bmv1tu7fXyk217/qsdlsKRfmWWYMHpwtR6BIHZzXk5uWYO/y2xHY21Z4YKbfV1gt3d0RTVSSJd0T0G+8jrKXpaT/viL9XZ/P7LjZGWh27yJpF+5cIbnMwq6Gw2/y9Yr37OXL0vPfelgo+8bSI6TlYXLZq5066SU7M2P6r0/BuH5Wwh72bOrXzri42vIjsYDxbkP5riHkEZUhPZDnyvv+ymLDzb63iFlR2NlwYKFHgE9i9bn3EFaOAAjWuDqQi50camPFLtdqZmNq7xUGhoJY4PICXDWc5ysLMye4z4Nnbv5idlDPiVxlgQAYFjp/PyAonqO14dnQuOZ/kHYy/50oElV8o/oDGZiN2RSb2TOs4VtwjNKU2HDTIDu1U6nafX4AwAx0wu46LGVugzXJEPpiZ5YNwmAJrSZsiJ9QppitPzD67a7LEa9ctGUrJK6tsHD03ddXW6Hqn9YjIosklOyTH8bwpfVGy1d3pb7mylriuqt9HRZSty+Rv93Me6osji/TW9Pm6VqT2m+wcuy+hKL21ftOVFqkDwXlLKMe0yHba4L97RZqkym0nxvbVDxZYfYpX4pTz1+v0VvT5ftsGmPMcvzS0u5prM/+NjChROl2V66SUj6/NdNNTaXL/xDm2W/r16Ssk1n/ZwUffwWQp9fqviBL9hqTO6/r/6hLNcpGF//F8LwK0RmPwywciH0+aWm/ZY2lx/Iz/qFpDce9fpfDgDCLzYjHG8HTJ9H1wFdR92DIcR8o+Vb98WGwdjAbQTlewwY+gfdl1RyO+v5Hl2IlGzTaZ+/SnA9wFkSAIDhoudCTaHbuEDKr7kQttV7jtY8J+GjMpiJ0lbUj5qupHm23t7gJzyD6p/IT4UNbZLK/dPiUVObumsCtYKfT47O/GEAWjj+DPC4WFZ94QYA4UQsC1y5nSkXldq+j3WTAGhDLGNZvreVLvL3Qamw5oJyyK3FWJZe31dE6kjZpSc8r1x83xdRxZC/509GxTWMj28RyViWofaMEIHnTXrO1hR6vSGklpS17YT/W4wh/xZSrulsk2JfimAsy5B72/t+OOiHs6bcof2UQ7zmB4AgxWKE09NWXagmKsWdyggY7Y8Noh/LIumzHgq+TyR9YXWbn1NS0D3AWRIAgOHAy71eNUM19bxNc7lPbUVlMBOlrUQolqV3eM+z9fYGP+EZZP9EeipM67Eswc8n90Zn/tAvTRx/BhDLAkAbRkJyMIRP97nTTc58aIa7Z0+lCDeAmPNSvUXB3nT6nK8MopoyZlr2i7WhXWZL2aW1L2ZP80zX6q3ogNp1ZpUc3bbxvjm335bg8u6pg0e/9J06M0LGTFu27q3QrxVTskrffiljkr8xjW5C+vqykC9HpaxttVuWTIvzP2oK6beQsktrN2S41+uJqEjsh4NGT8jYYDYV6kNtnJS1rfalhRMYnwIYweSWqjUrN6qtLuRKZaWhkTE2CK9r/unewleDO/dJ+sLtO9fPlcJ5SuIsCQDAMCA3fbSrWlkwxJBxz88DDtXU8zbNZa8+YGn3/ZHoDGY0MmRSb7jPs0V4wjMaU2FaFlL3xrrTNHn8AYAYG76nIkTC+YYPjw2+kG6bPJ4dBIAG6JIyNvi7jGy2nfVXuFZLxkzLftV6dLO3jKy+6QtNZj8BBPGpK984WhLUKiV9frnF+sZTMySdGJs2TxHwb684ZOkMS3neYOgmpK/fFuS3EEL0zRRUv5mdEqdmExv22IK+gWTI33bU+uZSdReiQf4WgUNDIiQS+6Fz5UkZ6/fZTPnuKQcCSskqqVPd1QAwTHW3VhlXlDUo3tPnFS9PC3wiE0KI+NTla5WVhuzmvGc21LbKbouNgLFBmF0fzK0OQ76pdt+GSNyV4SwJAIDGXWqq+6Ba8Y5keOCOqWE9A+uS7inIT1W+Zy3a9G6jvwFXdAYzGhkyqTe859kiPuEZjakw7Qqxe2PZaZo9/gBALA3vsxHCrPPLowdPDbxIXTLvlujfXosBuaUyZ3pCQe0wf9YQGNn8X0Z2nOv4PrrtGYrR0ozc7Y22msAldYWQsowmS1vNhowkv8djnTTjqTeslnI190WkLKPJYq3ZtCRV6ssFohuTdtcSxUX2vh2HzsTg+kUnzXjqzUZbjfcaxl4Y8ktrbI1vPjVD/bTJmKSMDTVtFpPX+s7uUrKMJkvbvk1Lg9jAwLf4wJiVEmjlH9ga34hFIEufCOyHTrq4pIxNNVaVHT3Q1dbtT82K6RQYAESc3Pre2hVblM/ZzTcWP5KqfsYzLm15cZ5e8Vb1xsUbq1qVzxSOjLFBmI2Zlv1qoD4ZOPtnTFMXXRQCzpIAAGiY4/N3K+qUb81+YPbkcJ+E3cMdhBCi+oO6pkt+PxWdwYxGhkzqDet5tihMeEZ+Kky7Qu7eGHWapo8/ABAzV/X29sa6DdAIubN2zbQ5G/unVqXCmpPr08eMhDGLX92tlSv/JfP/LKn5YFP6DbFuDAD/ZEfjX6o/t+x9Iq/ceRdI0uf/4T/u//WC1OF4kdXZWHu4oePro688W+Raa0CfX/r4zPj4FEN6UpCTAt1266GPW1rdVygM+aW/mRl/Q4rhl0nD4qEKR2NtdUOH+LvHFxFSlvHlexN1oXSOkmy37vu4pcdjE30b+B+TfjHUXUp2NP6luqGtee/GvPLB5+8lff4fHp95k8Z+iLDvh6769smLHUffyClSPlvSt/4fDb2rAQB+jJSxQXDOVC69M7N88DmNKVmmw9szbhx42dlYe/jzY5UuI8pYnaA5SwIAAK+iM5gZLkMm9YbpPFu0JjwjPhWmTUPr3iu00wBAQ4hlwSCHtXhBWp6574WUX3NyU/pIz8siO6wvLUhbab5SAncAAAAAYMTzf2MGAABA46IzmGHIBAAAtI6b9xggn/3soG3gxRVRYEi212xYZTQLIS25K41AFgAAAAAAAAAAAAAANID79+gnN320q3owSVpi8kQNFLuMKEd9yUNLN5rtV0jgDgAAAAAAAAAAAAAAwwKxLOhzqanuA2eBbsPds6deHcPWRJpsP/Jibk5ef4HDKyBwBwAAAAAAAAAAAACAYWJUrBsAbZBP1+2qG3ghGR64Y+qIDXPqtte/8fTCFeWDOWhGeuAOAAAAAAAAAAAAAADDCLEsYeVorK1u6BB/P/rKNvHM3k3pN/hdejCoIiXLWPz75XOT4oKKH+lsrD3c0P7l3ifynGEZ+vzSx3916y/uSpVGB9fyr7740FlgaPYDsyf7aYpst+47crjyibxyu5CyNle98JsZwW4uANnR+JfqhnYh/n70FXPya6XZSZ6xJt1266Ejh3f1fXspq2T773PmJvmvFCQ7Gs27SzflFFW7vBlS4I6jsbbacmzvxrzyBiGEEIb80oKc+/XOX1C2W/d93NJzqXnv9vNLdwTYE/p3G7nj6LZdycXvZ09zb46jsXZ36fM5RWYpq6TqhSdnSK4L9G+q4+gra757/K8lGRP8/xadjZUvPJq50SwM+aaXnsmYRkYa+JGfnx/rJgCAeOKJJyZOnBjrVgAYIc6ePfvyyy/HuhWItK4vrX93efl3a8WG/P9zXcyaEzH33nvvrFmzYt0KAABGjvLy8oaGhli3QkRrMHOlDJkAREhRUVGsmwDgCtCLIfmhzbLfZDLtMWZJyo6V8msu+P1kj63U4Lp8tulsj6pN9rRZqkrz9f5+1ZQs4we2LnWr6+3t7e25UFPobL+h1Obroz1tlm0em/az/KALNfnODUiG0hPKT/R02Q6bvPbjFKPlH56tOLot3+DWCiHlms7+4KvVVabX8/WS+0fcLSq1fR/gi3TZqj3a2E9fYunqce9MqbDmglvvDHxZL03ybMAPbUc3K7ZnKLX19O11e0rdO8GzYz26QrHXpWSbTqvfS3AFCvRfBgCiwWazxfpwCGDksNlssT6qAWFjMpli/V8KAIARJS8vL9andwAYNmJ9zAZwRRixhWSiwmEtnpeQNj8zM3ORS26UPvaKQ5ZO2fdnL3/VUO+aHsRe9qcDTZcCbFC2W7cXzElIW5hTZBZCyjLuqbF19f+UP7RZTMasFCGEEA3leXfrcytOOvw0wFW75YAzK4t02+Tx3vYL2X7kxUcMacuKzG5/qP6gLkDL5U7LoQpnB7nlffmmtmDGdckbj3Z8fXRLiVs/Spm336RIHiQ7Tm5/JHXmMkVuFSGEEPYtK16u6/Sy9UtN729amPnbIrPdyx9dBSg21Nl48MWlScmGvHK7kPT5r5sGO78vxMdsXLXV4lB2ppg+daIi3c6lxrJ/7/+yrzzr3iRp6uTxrllVOhsrn3to5gpFn5yqL115R0La/MzMRTnunWCv3vVRk7/f3G2vayh7rcbv8gAAAAAAAAAAAAAARBuxLBFjrz5gaff951E/m2nIVmTlON/eddn38rKjsbJgTupAKIkh33Sicfuq+9KTBmrEjJZSM1a9+UFNYX/eDXv5svQnq1rVRCp0fn6gwjrwInXJvFs8S/XI9oNrHlq08uDtxj2eCUWabWcdfjfQfe50kzMewz1i44b0TZbe3gObspdvqvnYlJ3i8ie3xsiOkxW56cvc44YG2D89fc7L9706KXv3YPSWezoc1/Q5B7KTfP2HkFtrVy9KNqzsL2h01Fqz6TcZg52vk1KXbtxnyRMl5dVf1Lt0pmcBo77GHNiUvXzT4dNnTbmu/SgtuSttzODS3a2Vq/WZG83uTbl+3DjfmR4DxBXpro0fq9jpqusbvvKz1wEAAAAAAAAAAAAAEG3EsgxFXOqqw/1hEF5Kz1grDnzuLU1IP92EX6+v2uyjXI2bbnvt8wuSM/vTeEjZpSf2bMqYFue5oG5C+tMFg9V87GXPrq1qCRTNokyaIhnmpY11X8RRX/LQ0m3/9Puj1jdW3ZeRkfHg/fOTXf583bjr/aQzEUI+XberbvCVMmLDvf2/mO9a7Ttt5s3xg2txBrLo80v7cqJ0HTW6RtWcau8K8G0vNdV94JKYxHvgjjtHQ9kjd8/ZWC2EEPrnaqxvPDVD8vgCurif/+uS6R+atld/6jMDjZvRE36RPtd7Y7rttRsXZ27xErQza+bDvx/Y63p/aLOUK+OKjn3YcN73N9GNmbFwZeBaSwAAAAAAAAAAAAAAxMyowItAjbikuSvXbT5xLLOsYfC9vjQhvsI2hBgtzch986/SqH/JLLMLIZITE7xGhHTbazc+NOc5c//L+cZ9/zt7mu8AjDG3zFuSWlTUlxqkoeyVyt8ZnkyN8xO0dPHLYx87Q1k8A03klsonc0r+aWPtliXT+tYjn/3soGuZeV8tH1hB00e7nFV3/IePXO7qcElmY5iR8rP+XVRurXoyfVm5PSWrZOsLT87qjyWJ++c75yeLgMWDnM43fHjM+cpr4I4bR0NZ7oM55Q1CCKEvsezz3Zm6ibfOvSEn7yWX9btloHF3ua35iJfGyI6Tbz+9+Dlb1uajL/xmxnWfFi9YmNf/HSXDr/75Z84VjJZSH36m+Jv6tJXm/ndOHWn++rK40ed/7LgZK3fuS9zx8oq+qljS2PhrCWiDT729vbFuAgAAQDglJSUxwrkCnKlcemdm+amBl1OyTIe3Z9wYyxYBAIDhoKioqKioKNatENEazDBkAgAAWsdt7PDR3XjXwwuCLeCim2D43Uq9EIq4DRdugSwpWaUvLE+N91jMVfzNM9Ocr8zvHrZd9Le4IjDFM9Dksr1qY2Zz9r7BQBYhxFdffFjtEj7iveWDlKlQAoSPuMaauBTocTRsW/1smd2Qb9qz5alZLklRLl043+X89KzEBP/RWZ1fHj04ODr3myFG9G/XJZDluZqduX6jgtwFWv+l5s/qna2ZPnVi38odlq2PFR6cu612y2MzpNEiLvH2GQkDa1xWcJ9bHSRd3K2/nD/F+fri+Qt+f2+hk1IzVm0/21draXCjAAAAADBC3JixvanXqYm7MgAAYFiJzmCGIRMAANA6bmOHkW5M2l1LFMEstua2S+o+6xK34UJufW/N4sFAFiFl/2HDshQvpYX8sR387KyfwjvKpCmJyRPdVj9Kyvivrn2Pu8RwyJ0nLAcDtdxlA6oLDAm3WJOE2ybfoBNCiA7r1qdzykW26fWNbpWV5G9Of9rmbIkiZ4mXpiiqKQUsMCS31m74z/5AFjHfWPxkuuQvyYqHgAWMHGdtzQP/HuhGuaXyyZw88ehbLzw4EDw0dsYjj2dJQkhZJVWP6z17b9Sk2zNTB1/Zz3V8p7aFgX47AAAAAAAAAAAAAACijvvYYTXmpplzXVJkiDPHW74N8JG+tChe8m0IIbdUrX22zBl7sWh9wb9NCPoXs59q/853LIsyaYrh7tlTPasFXR0X55rtpN1ywDUry+wHZk/2F8qiiJWZMnfmTX7CO+Rzpz91ljvqy+DS3Vq5ZkFeS1bp2y9lTHLbkHLlyXNvnei3e5QtD5Ahpru1auPijX19I+mNawOlw/GUNvNmvx/p/PxAhXXgRV83drdWGVe8P6P01adc4mZ0cdOWbm/r7W3b/tQMydsXjJuYnOh8daS5LUAyICFExyfvVlYH+u0AAAAAAAAAAAAAAIg+bmSH17iUX93u8vLUkeav/cYVyJ3m8jXVwrA+yyPfRndrlXFFWcPAS0lvXHV/kmegydAokqaoy9KhiMDwFf4y6Btz6R+dsTLi9l+ljPO9sCKwpi+Di9z63toVlcmFJS94SUjjVr3oF7ffdI36lvvPECO3vrd2xZb+wBd9XvHyNBXpcL5tOX7G+SpA6SVl4M6UGbcmXi23vrd2xYf/unntsml+87m4G3VdvJ+gHG+bbq39Y8l+KX/5fWHfowAAAAAAAAAAAAAAGBp/99oRvNHjJ0+VhBgMUbh4/sJFIXwGJjgsrz+/za7P23e/R1KWzrqXB2MphBBSxuMP3qauutCltmab2vZ+9cWHzkwlarJ0XGp8Z2uRa24T/+EvQQW+KANrpicnxDkatq1+tiw5z7J6jpeEJEFVLxLi8pfHTM6W+80Qo8iIk5r/zBKXEkt+NvB185HBAkkBA4OUgTuZt9+ka6lau6F5WcnOhe7pZwIZNW7S1ClC9G/7VFPL+cupkp//2t+YX95QJnJNT8wOKmQGAAAAAAAAAAAAAIAoIJYlvHRxE6dOd4llsX96+pwsfIRYdJ7c/VqJ+Xaj5RGPUAm3kBEhLcm4a8JooYoyO4iQpoy91kdshNx5wnJw8NWUGbcmBsrSoQwfCRT+EmTgi6PNdtw1sOb6T7b+Jqd8krf+EcK9wFDqknm3+I3MuNT8Wf0p50s/GWJkxyeVrwxkxJGyVz+hv8Hfigc/1vzZQecGAhU8UvRk6pJ5yV1VG1c0P7zvJW9RO+EkO6wVzxe16Y1ZBrV7FAAAAAAAAAAAAAAA0UMsS5jpxk++TRLOEItT7V2y11JOssNa+ljO+8nGquWp8e5/dHz+boVryMii9Tl3qE2hocgOIoRIuG3yDT6iI9otB5wtlTJvvynQ7qAMHwmUZ6Xzo9I1e1xe+w98kTsthyoG122Ye2vHn1blHdN77R8hPAoMGeal+a2z4xaF46flcuPu1UZz/4vUJQ//coKq4JLgCh4pelIyzLvpi5dnnl65b72qBDAeRiUkzhrMyyJszW2XhOQjiY/DsnWV0azPs6iqmgQAAAAAAAAAAAAAQLQRyxJucQnJ0yVhHwhT8FXwxV9IgdxZX1ViVh0y4vZhRXYQIURi8kQfQQuKAkAB85oI93CNAHlWOqyvl7imlgn0LZSBNeNP7ix421/IRZAFhpRROH5aLjs+eb9icEnD73LUJWVxCz8K1J7LXzXUO3ty7j917Ppj/cq1+7xH7YRRd2t1uY9UQAAAAAAAAAAAAAAAaAKxLOGmuzZ+vL+EHEIIIbfWbngmz2dIwcUvj33sWl8oQGkeBbdwEz8RJMo8KAHzmoigCgzJDuubq0q+1d8hmT/q34Z02+Txfr6F/M3pT9sGX9nLS0rEfD8hF8EWGFJ2i5+Wt9fv3mHu/7f6nncLPwrYno4TRy0D/5b04v99pWTC43+9LfREKeMmTZ8ymJjlzPGWb0Wql5XJre+tXVHpPRUQAAAAAAAAAAAAAADaQG6GsPvJpOk3ury0NbddUi7Q3Vq1cfHGtqzSF7yHFMhnPztoc3mdPPfWiWp/J/dwEz/RGIqImYB5TYRHgSEp+9/n+cqz4rBsXbUj8Q8P3f7fasM73KsXCUlvXOs75EKZ1yRwIM75hg+POV/5yRCjyFXjvyiSC4fl9ee3ubTedy4cL1uxm8s/T9yct3DCaDWbCp3cUrX22bLkvGKqCwEAAAAAAAAAAAAANIy8LGE36rqx4/z8uT83RuH2F5aleA8pcLTZjrtGdSQnJqgsMCR3msvXKCJCfEdjKCJmpsydeVOQBYZSlzz8ywneV91h3bquJPHpQ5OP3aU274tHOhlp2TO/9RNy4ZrXREUgTueXR52Fl/wVGFLkqlFb2qnDunVdXhA1oZRbEUIYnipYOGlIYWWjfpo4a4o41fcdTx1p/vqyuFH5f7u7tcq4omwS1YUAAAAAAAAAAAAAABpHLEvYjbou3jVoo+v8hUtCDERlOOpLFq8oS86zrJ4j+Yoo+K7jnGugw5Spk8ap+5nkM4d27FPkNslffl+S96AKZR6U23+V4i/+RgiPjC8+Y1Nkh/XNVfv/5a2d6d/v2KI674syb4pIyd78qN7P8oq8JgEL+rjFjvgrMGQ54OyUAEWRBlbusL65Km+/yzsBKxMptiJEan7BPUlDDS/5cfx4f2WDqC4EAAAAAAAAAAAAABguyNAQdqPGTZo6xfmy41zH9/3/lFtrNzyTZ8swvZUbRG6MWYkJqkJZZMcnla+UNbi8s2h9zh0+gjyUeVBUJCBxLzDkIzZFbq16ckHN/OIn03/2tUvel0DhJoq8KULos39nuNF3ByljUwIXGFLGjkyZcWuijy8rf3P607bB9U5PTghci8dh2brKaNPr73C+lXDb5Bv8/bqKrQhh+F2O/oaA2wnk6uvHXTf44tTxlvOKRtaXLF7x/tyNr1JdCAAAAAAAAAAAAACgecSyVm5yxwAAIABJREFURE3nyW1rF2/8H8Z96zMmjA7/6uUz1X8sMztfS3rjqvt9JGVRJlkJmEdEeCkw5DU2xdGwbfWG5pVrl6fGK2JfAoSbuOVNScl+POPn/mJ9lNlTAhcYck3iIqTM22/yFRskf9d+anDF14yPvzZArzgaynJz8sT//fKSW//buYEAsTXKqCA1na+Gn8pWHdat6/JsGZs3PDiN6kIAAAAAAAAAAAAAAM3j3nb46a4b65KX5eK5ju9k0d1auTo95+yymtdWBizyMm7S9CkBFvHQ3VplXOGalEWfV+wnCYejzXZ8MJzCT82dQcoaQF7DNeTW2g3/uWbU6rdWzohzi32ZPnWivyiKi18e+9gl5cuCh+/yk5Ql2AJD4vKXx0zOtftd/nzLcWd2mL4fzjdHQ1nugzkHZ5S+unSsrU51bI2yZ6SMR+clhuM/obKy1ZHmtst9/+q21760Kq8le3PewkhEUAEAAAAAAAAAAAAAEG7EsoSf7rr48c5X9lPtf2+t3bg488O5pf97dfqEwD0+atLtmalBbVFufW/tii0u4SC5fssYKfOguNbccdS/WFBx0uERwqGsAeQtXKPz5La1iz++e99LCyfogsz7Ip91qUYkGdZn6f3FgvgqMCQ7Gj9r9Gy5uNT8Wb1L012jcDobzZ/YfYar2I/b2hy+GmE/8mLugznlCYVvrVuW9A+XmkFT5s68yV9sjaJnhLQk467whJjoro0fK3lurfW9NYtfE8bSlzIm8V8dAAAAAAAAAAAAADAscIM7Aq6NH+8SVnDxk9efXvzOP5W+vSU7xWeiFIWxafMMzhUctJzo9JcfRDjqSxavKHOGd2SX1m7wW8ZIWaPHWXOnw7p147vjkj1zqARKbdJ5suyp9DU/3jwQQKMsozOQ90Vuqcz5n8VW9/gQ7wurbPxgVI3DsvXxNxsuXHZf3D12ZDAKp9te++KjD5V+bO92Lqz84exFJa9bOzwa0G2v3/JI6uyV5QmFNWXr0ycIRftv/1XKOCGEcFjLivd6ia356osPnQsHTiqjmu6a68deM/jqVFPL+cvC0bBt9bNlyX4z9AAAAAAAAAAAAAAAoDHEskTANdePc4YVCPvOg20rS1UHsgghdGNmLFypH4iqsO/bceiMz2AWR0NZbk6eeSA8QsourX0xe5rfAAlFjZ7BPCKyw/rm6uMLX/US96BMbSLSZt7sWiapL5BFrHcG0MiOs03HB/8uTZ08fnRfFaR8seTBn7utXllzxzVJjFeXW46ZPBovt1Q+Wdxe4K2MjqKakjN2RG59b83iv87ft14R9BOXkDzdNbnJ/ry0Bwq21/fnbpHt1r07i5emJsxcUW439AWy6NxrBvV92UuNu0sqhJTgHhUkd56wHHQu7K1UU6iUla2EEB3WrU/nHPyV3ww9fTobD764NOGqq66a5/yyAAAAAAAAAAAAAADECLEskSbpC7fvXDkjuMQYcWnLi/P0/S8aylas23ay03OpgUo3Df1bytp81PpqgEAWIeRzpz91Jlm5cfqknwgh5NaqJ1d3Frzw4DQvcQ+Xu9rPu7xs7+jqS38iOxprywoW3bxGrFcE0Fz88tjHzi3MTbt5jE5ufW/ta+NeX/9rjxpLjrO25sEXLklifDS++TOXYkc3Tp/0EyG31q7J2ztrrdf6TcqMMonJE+OEEMLRsG31K5fXv7A8NV6xtC5x3qMZyko91UXLZib86Kqrrrrqqh8lpC18OK+8QYjBQBb39g982T9v2pXmLRuKj6Qy4aCsbHXE9PRvFuT9o/CtQi/xPQqyw1r6qGFluV0IUV20bOFDJfW+KisBAAAAAAAAAAAAABAFfgMHMFSSvnD7zvVzpaAjFnRxqbk7azofmvOcWQhhL8u5+ayttCDnfn1SnE4I2dH4l+q691/JKTL3byfLuP33y+cmqYiYufxVQ70zj4gwl7y8dfICse+N75ZuK0yX/Mc99NmTo/9x+8uzz295tshsl7I2H7X+Zobig9+2HD/jfHVw7453Lon362e9tM7L+hVJYgLW3FFmfBG2/bvfGru79sPktT7S3sgXL7RfdL682N510dH44db/9cLxX73i7SOjJxiyVuornXluPOkLTa89nZE02MzvO8651CH6W/MJ6593/PEvs15a5yUbiuLLStOTE8JZ+qevQFJ/wxt2lous0re9xvcotdfv3mF2vrSbS6rqf5uWHr4gGwAAAAAAAAAAAAAAgsId6wgY9dPEWVOGEMjSZ7SU/sw+mym/v9hQdVHOnOTr+jOEXJd8Z2ZfIIuUZTRZ2s5uX6UqkEUIobs2fqxr6hF7ecmWvyb+3vSM70CWa5LvvEcvFJ/JW/TbIttcY7WtcXvuDPcP/jh+fLxi4S0NiQXrvCWMkTsthyqciUqCrbljNxf9P7aZ63zXb9LFTZw63flyf17a/3Vdcsn5e1/1+ZG4GSt3bi/US97+lpJl/MC2b71LIItw/7LmvDmr6nx8WbckMbMfmD05nP/9FJWtJH1hyQvL1JS1UsbiCCHs7R3fUWcIAAAAAAAAAAAAABAzxLJEyI+HFsjSRxeXlLHp8ElbTZWpNF/v+hd9fqnJVGVp62nbviojNZit6MboH68qyZKEEELS579eZbHWbMpI8lJayKUZqbk7LeUDUTWSPv91U5WfAJqxM3JW9y+szy+tsrTVbFDGfwzyUo0oQONnZG3ONwghhDDkl+6ztO3blDHNT8SGLum+V6r7v62Qsox79gf+iDR3Q43VYjJmDQa06PNLTYdtXZ9tXzXPo6Nu0P/HxkJVX/ZS82f1zvpI0tTJ49VkwVGtP4JKCCGkrI2vrp6jbq9QxuIAAAAAAAAAAAAAABBrV/X29sa6DQCG7kzl0jszy08J/XM1O1WWixJCCNl+cM1DSzcOllWaYrScXJVK8TEAAAAAAAAAAAAAQIyQlwUYGX4yafqNQsouffUp9YEsQgidNHf9W5uzB5LQSJm330QgCwAAAAAAAAAAAAAgdohlAUYA2WF9Y1Xe6MK31i2b5rXCkT+6Cf8yf25ffaJF63PuCPrzAAAAAAAAAAAAAACEDxkYgOFOdpysyF1Qlmh6b336hFDC0+TvOs5dFEJI+cvvS7o63M0DAAAAAAAAAAAAACAI5GUBhjXZcbIiN71YrH/7pYxJIf1/ljvN5Wuq7UL/3Fv/MZukLAAAAAAAAAAAAACA2CIvCzB8ddvr33h64f/627LtO5elxIW0Crm16qnFG+1SdumrT6VLo8PcQAAAAAAAAAAAAAAAgkReFvz/7N1/eBTlvfj9mxy+PKiRUtTKBPzKgZhgryBKAlxqnuOCJrG1Wq6gpdVmg6GtnoLHrxySNGqtp0URNk+sP2jVNhESpQdx9wmlVAkkxhpUwi6I5DqwEfJADVlbbAzJCkhh8/yR7O7MZn/MJjM7k+T9+iuTnZ25597ZuT9zz2fvG8NUd6vjyXvnr9j5rRdfW50jDWpuIe/hjffPXVyVXmZvfLZoJmOyAAAAAAAAAAAAAACMRy4LYHLnPK6a0gUpY8Z8p9zV1f8/72FH6T3pi9e4rRsanls0ZRDfY2/rzvL7064t+2uB3b1tdX4aiSwAAAAAAAAAAAAAAFMY09vba3QZAETi87qeuzNrZWPfkqXCue3+Sz949anCldUeIVk3NKwvmJk8iEwWn9f14vI3Lsr/3nfuzBzUkC4AAAAAAAAAAAAAAOiDXBbAzD5vKL391nWugS8MIZEFAAAAAAAAAAAAAADz4kE4YGK+z4991DHgv5KlrM71aiGJLAAAAAAAAAAAAACAkYdn4YCJJU2enTNH8R/Jaqtr3PZ0DjMDAQAAAAAAAAAAAABGJOYYAszNe9jxq4cXr6sTlpLKh76VnftvaQzHAgAAAAAAAAAAAAAYuchlAQAAAAAAAAAAAAAAgFkwwAMAAAAAAAAAAAAAAADMglwWAAAAAAAAAAAAAAAAmAW5LAAAAAAAAAAAAAAAADALclkAAAAAAAAAAAAAAABgFuSyAAAAAAAAAAAAAAAAwCzIZQEAAAAAAAAAAAAAAIBZkMsCAAAAAAAAAAAAAAAAsyCXBQAAAAAAAAAAAAAAAGZBLgsAAAAAAAAAAAAAAADMglwWAAAAAAAAAAAAAAAAmAW5LAAAAAAAAAAAAAAAADALclkAAAAAAAAAAAAAAABgFuSyAAAAAAAAAAAAAAAAwCzIZQEAAAAAAAAAAAAAAIBZkMsCAAAAAAAAAAAAAAAAsyCXBQAAAAAAAAAAAAAAAGZBLgsAAAAAAAAAAAAAAADMYqzRBQBGMZ/Hte2D4xeEr23rz04W7lu7cILRJQIAAAAAAAAAAAAAwFjksgCJ489dOdu2dU1xdYvslcyS+utIZAEAAAAAAAAAAAAAgFwWIDE+dRQuWFx9NMKrWfOvnZjQ4gAAAAAAAAAAAAAAYEpJRhcAGCWuyt94pLffF07bHYoXc+dlXEliGQAAAAAAAAAAAAAA5LLo6JzH5SgvnDVmzJgxY/JKNzZ7fEaXCGYx8Ya78nODi1LukptS+S4CAAAAAAAAAAAAAMAcQ/rxtb62NGtZXf9S3bqlB5tP1m5bNS/Z0FLBHHze9iMHg4vZS7KnkcoCAAAAAAAAAAAAAIBgXBbdnP9bS3Od4j+exoPtPQaVBiZz7rNjRzyBpdzbs1PHG1gaAAAAAAAAAAAAAADMg1wWnYy9MmNeruI/kmXW1EsNKg1M5mTLu/sCC9L10ybzRQQAAAAAAAAAAAAAQAhBLot+ktJ+uMFpt1kzhBBC5JZsqN200kwTDPmOO5bNSilt6Da6IKNR9yd7dh71L2QW5F03wcjSAAAAAAAAAAAAAABgImN6e3uNLgMS79wJx8q5iz8sqH977cLLjS7MaOPrbnh85q1r+ucYksrqD69eOIGsMgAAAAAAAAAAAAAAhGBcllHJ53Wt/+Hi9R4pNy9rktGFGYVOf7LvA49/QSq4LYtEFgAAAAAAAAAAAAAA/HiIPur4PPVPr7I1kkVhFF/7gZ1u/wITDAEAAAAAAAAAAAAAoEAqwyjjba64t3BNo4csCqP4jry/uS4wLMv09KnJRpYGAAAAAAAAAAAAAACTIZdlFPF5dj+7fFlxY18iBVkUhjh7pOntusBS7u3ZqeMNLA0AAAAAAAAAAAAAAGYz1ugCIDHOeZp/97NFK6oDA4KQRWEI37GmzU3+BSl3yU2ppJMBAAAAAAAAAAAAACDDg3T9dLc2bHU4HI43ywtTskobPo+1vs/b2uhwOByO35UuuK+q9Wy4dc55XH9+s7wwZcyYMWPGpBQ+u7O1W8VmG6pK70yZL0tkGWwWhc/j2tp3QGNkFpRWOf7s8pyL/X5va4PD4XC8WVX6nbyqw76wK1SVLug7tmZPmBWCB6VHXUUq86bywln+o80rrWpo9cqK1lcpjk3lhXmxP+W//c+7wQmGspdkT+MbCAAAAAAAAAAAAACA3Jje3l6jyzAynPO4dn1w/KwQZ9u2rimublG8KJXVH169cEJI3oLP2/qXupZO4Wvb+h/FskQTIWbYnIdXZSoHzfF5mmt+/fOl6+oU/5WW2/dW5E8ZN7BAPo9r2wf73n/hF+saPQNflbmn0l1dlBZ9jJZzHtefXn/+F6HHpSiJtaL2mYfnSbKD9B+g+Mee0GIM3OnAkWMq3W8VpSUpN6VPXYXnbd350lOFyr31s1Q4tz2cmSy6Gx6feeua/hXCf8qKcinWVxwgAAAAAAAAAAAAAAAQgnFZtHHeVZ76f2U+/z8+4evas3FgwodUcFtWaIrD5w2l8y5NX7On6+971leEpEtIi+dco0jO8HkPb7w/c35ocoYQwrN+xfNN4cYbOXvkrbWLFv8kViJLzMmGfN7WHeWFmSlZi4urW4RktW2prXef6u3T466zWaX+klSvXPRU7YnA6CxnW6u+33+AA/NppNRpk+U5Jd2tjifvVYwcI8TRzp7+0U/0rquBult3PluYlp5bXO0RkqXkFXu9u6fvkC90ODeUWBptq15yekWnc0dwlBUxK3VqcvQvlGJ96fppk/n+AQAAAAAAAAAAAACgxLgsmjt3wrFy7uL1soyLzJL6t9cuvDziO3zHHT/+zuKqlgjr+7yHa5YvXBpueBAhhKrhPXytVd9KXxZI7pBK6g+vXTgh9rF0tzqeeWDxmkYhhMiw2sofezAnLSRdw3e46lsLl/VnaEi5lQ1vFc0cUJbQOlEWYGCNCSHCD7iSgLoSvhMNjxfduqZOiLCDzQghfF7Xc3feefShXd/ec9sd6/yjskQ4dpnuhtKZt/rXj3VWAAAAAAAAAAAAAAAwKjEuhObGXXn1tIvl/5By87ImRXtH0pQb77hZtpw1/9qJ/r9lyRmWksq+0UF69tgsUnD14PglkZw90vS2bJSSzIK862InsvhONDx6T3pfIotkrdhT9+qqvNBEFiFE0tTZOen+Bc/Rzi/DlWXclBsX5oQvwDlPw5ofDkxkEULcPD1l7IB/6l1X3paq+2/vT2SxPFnv+t0joYksQoik5Bu+VTDrXfvGuo+C5c5ekj0tekJRt3NXTTCdJ9ZZAQAAAAAAAAAAAADAqEQui+Z8p091npYth5tgKMT5nq7O4FLuvIwr+5M4fCdqH164tNqTYa1o6qhfW7QwLVkIkfzNBXekh91QBCdb3t0nK5CKLApFSkeZvfGFcCkdfc6eOtkT2PSMSZeEXe18R9vuMAXweQ//4Wc/fNJtfXFPx1fKvBMp95ZvXhluSzrWlbelavkPlvVNEWWpcG77+UJpXPg1k6bOzrl807rnghlCobMmDXT6k30fBFNZYp8VAAAAAAAAAAAAAACMRgMHvsAQKVIW1A2CIs81kXKX3JTal+Tgbdnw6C+qPLkl9ud+nj8zObi+PH0kwvglct2f7Nl5NLAUO4vCd6Lh6f/sT+mQiip/+7P8tMhH0P3xjhqXfyHS2CRn2w40B0swK3Vq3/guXudL/162M2dDw/qCmclJQkyfMy9FNHqEEEJaWnp32LmAdKsrRSLLk/WblmcOHIQmMhW12n5gp9u/oG5oHAAAAAAAAAAAAAAARh9GhtCaImVB3SAoilyTlOunXZ4khBBdrpd+tqxaFNlfWaNIzhDC9/mxjzqCOwg/fklwbcXUNrGzKLoPb3jih30jsoiMohefWDozyupdrlcq1vk3LhV9Py91fLjVvO3utmCB+xJQfMcdDy8rFg+89swPZvZnjUyad/9DVkkIyVpR+5AlbGqITnUlT98Rd9jKH444Ikt4sXNTfEfe31wX+Bimp09NjrY2AAAAAAAAAAAAAACjFeOyaO1v//NuMGVB1VQyvs+OfRSce6Yv9+XcCcfjdxYft1b+4bn8q0Per8yKSM+ZPTXqDjqdO+QFip5b4/O6Kv99WVXf+lLRf/1yUejelSu/uqp4u3/Ly1/85XemhF07zNgt507U2la8Na+y4RFZ1khS8szCjR2FG6OUT5e6Oneido0/fUey2J54MHNi5CKElTX/2uhvOXuk6e3ghES5t2eHT/oBAAAAAAAAAAAAAGC0I5dFW77uQ86dwUU1U8koshz6cl98J/70xApHetnGZ5ZmDBi+Q5kVId0455qLo21ekUcSK7fG63xpla2xf+Ge1aXfDp+bIoQQPu/hmuV3rvSvnFv2WtmiKeHHMlEkoMyYN3v6eN8JxxMr3v3Wi3+KOujLQLrUle/En55Ysb6/gJbi8gezVAyZ8sXxg58Gl3LnZVwZ9avkO9a0uSlQjuDUSAAAAAAAAAAAAAAAQIlcFm3FNQiKECI0y2FWekqyt2XDo7+oSi92PnqrNDDjQbF+7HFfzn+yzx4s0Iyc+ddETh45d6KuuqKxf22p5MG70yKMHeLzuGp+vWrpusb+Va0Vtc88PC9MYYUQoQkoi+dck3S89omn25ZWbIo26EvY/epQV77jtU/8on8gGpFZ8vOCzGQVhTr/97bdgamOVOSmKEbr6RuZBgAAAAAAAAAAAAAAhEEui6biGgSlj7fDfVCe5fC1/S/9eFn11Tbn/WGTKpST5sQc9+Vs24Hmo8HFObdkXBG58E3PB4YnEXes/N6ccFs+53Ft/vWq0nWBlBfri7XP/HieFH5EFiFCEkoyC/LSe2rXrGi7b9tz4bJPotO+rnze/Y4Xqlr6j6Xo0f+wXK6mIL62AzuD1RpzmiflaD0z5s2ezgRDAAAAAAAAAAAAAACERy6Lhnzdzl01agdBCfeW3JzZXf+9qnifxVb7YObEcOuHTJoTa9wX5cAkIvf27NRIWRRnW998aV2wJPl33aAsgLe1oe7dHS/8IpDFIiSrbeNjD+akRZ+RR5FQIuXmXfM/z88/tnLbalXDnyi3pH1d+VrfeDQwp1JmwX3/FnlOpShbjjXNk3K0HmnxnGv42gEAAAAAAAAAAAAAEAEP1TWknGAo+iAo4d4iTT68qfQPjZZi54NZ4RNE4pxgSDkwSdSpcJRbnpEze3qSz9v6l7qWTiH+sUeewiKEZLU9X3jX7QtjZLEIIYQ4/7eW5mDaR87/7tr8m+aVT2wLn30SneZ15fPuf6smsMncny5TNyiL8H78x5o4PgXlaD0xh9IBAAAAAAAAAAAAAGBUI5dFO8oJhqIOguLn+/zYRx2BJU91RYW4I9KMOSL+CYYUw4eI7CXZ0yKmsii2LI4WZ/2v4tB1JKvt+e9+819vvC0zyoxCoboO7XEGNmAR771QMeWhvderSIIZWETN66qz+Y3XG/1li5boo9xwd3NtRaP6T0E5nEzMoXQAAAAAAAAAAAAAABjdyGXRSsgEQ6pyI0IySISQLLYnIsyYI0LHOImdFXGy5d19waWoEwwps176ZFhtZd+dPl6I8VfHl78io8jv8TRWf1xk/9OiKYPZlPZ1pShbtEQfBa/zlV9tkJVjevrU6Jk5pz/Z90EwlSXmIC4AAAAAAAAAAAAAAIxu5LJoJWSCITW5EQMySKSlP/9JhBlzhFCOcaJmaptP9uw8Glg9am6Nt93dJlu8p9JdXZQWa1CZ2ELye4TIfaR00dWDSuXQvK6UZVMziI4QQnS5XvplsWy6pdhv9LUf2On2L8zImX8NEwwBAAAAAAAAAAAAABAFQ0RoZBATDIWMmyIyil58wBIlPUWxizintomeW6OcvkeISRMv1STJKSS/J7Ok9K60QZ5xmteVomzS9dMmxy6Yz+t6dVXxdtl/Yo++oxxOZs4tGVfE3A0AAAAAAAAAAAAAAKMZuSyaGMwEQ8pxU4SwFP0096rIb1LuIvYEQ8o8khnzZk+PnFvj+7LzqCfiq4MWkiKT+9NllssHuSnN60pRNmlWekr0iYKEEMLrfGmVzW2x3BT8V8r10y6P+jkrh5NRO/oLAAAAAAAAAAAAAACjF7ksmhjEBEMh6S8ZRQ/l35Ac5U3KcURiTzCkGCdGWjznmjhGWuns6jmvfm0hfN6jRz2+Af9VDEmiLr8nwva1rytF+s7FkydeEqNo3paq5cuKxY+eL5j9/wX+GTOjyHesaXNTcO3B1wAAAAAAAAAAAAAAAKMFj9a1MJgJhk5/su8D2Ugud953W5SBRuKdYEic/2SfPbj1WOsnXTJphiRbbqr548feqNsP8nlcG8vuTP3R6/tD3qEckkTKfyBv+mDPNh3q6uTxg8FxXk5/1vXlgFQcGW9L1fIfLNs5r/K3hZPcTXFkFHk73AcDq6vJcAIAAAAAAAAAAAAAYLTj2frQRZtgyOdx/ck1cMgSIXztB3a6g29ZbbVEy4qINGmOz9t6oNU7cPNn2w40B1M1FMOHdLc27g8tUNLl065PkS17GitefuNwd+TyCP/ed5Tfn5u19M//u/KFBzOVs/QohiQRUkH+bVPGxdpgpP3oWldCCM9Bd0ek3B2fZ/ezy3+wrDql7LVfLk37p2xmohk586+JmlGkLIl8midv87OlNYfDFwYAAAAAAAAAAAAAgFGNXJahizzBkO947eMP/D/2lr/FmH8n5ogdESbN8TpfeujVllMD5gMKzSMJDB9yztPw7AP3Vn7gOad8w6SsvFz5wCzCU7Vs4UPPNofLwunjbd1Zfn9a+u3F1Sll9W+/WpSRHLLC3/7n3WCRYw8kE4UudXXJxMmyA/asq3jF1TVgU+c8zevvz8xeWZ1SVl+1euEUoSjJnFsyrhBCCK+rqnxruCwZZUmC0zx1uV5a88cr0qdGmyYJAAAAAAAAAAAAAIBRiqfpQxYywVBw+I0uV8XyxW33lT96qxRazcr5d+QjdoR1/vg+e2AX/uFAfMcdD5d3lhYvGjjeiWJqm2Aeie/Enx7/4d47tq3OD31L0gSLdbUym0V4qlfOz72/fNNW+bgyPo9rq+PN8sKUS9Nzi6s9krViz4bVC6cMOI183YecOwNLioFh4qVPXSWnpM+SH+/24qwlpRv9yTs+j2vrpvLCzJT5K6o9uX2JLEmhsyalTps8ToizrW9U1AgpZWBiiuLECAzi4vO6Xn304KLfPpgVmv0DAAAAAAAAAAAAAACEGGt0AYa985/ss8tGZRE3T08ZK4ToPlz1n3dun1u/aXlmmOE3vO3utsCCbMSO8HxtB3YGZwy6atbVXxe+Ew2PF2+9+Yn1YfJIQoo0PX1qshBCeFs2PPrC+dUvPJg5Mcw+ktK+93RxTd3KRsV/W6qL76uOUCrJ+mLtMz+eJ4WdOSjC4CiDoU9dJU3PeyBfqlsv++jq1i2tW7c0ZHuBRJbQkoicrGsnJPlO/Hnt5qxye5jEFN9nxz4Kbv2qWVd/XQjhO1H78KPdpRuWz2RQFgAAAAAAAAAAAAAAwiGXZYjOnzx+5Kj8H9Uv//qWL694d12FeLB2U4RUD8WIHTHn3/F5248cDC66t7/x2qQ3Gt5Nf2L9wJl9hBDCd/pU5+ng4unOntPe1ndfeuqZg7e8EOEtQoik5Mzlr9mPzl0sT++IJLdkw6/+T8G8AePN+CkOUJqVnjLMEQcMAAAgAElEQVT4MUj0qqtxU3KtKy2O4sbIh2sps7/8s/y0wA7PdH0mm4for22HXH9+/Td/ufm5X4ZLVzr/t5bm4CAuorHi+Zem3Sm2/e7Lwg1lC8MnAAEAAAAAAAAAAAAAAOYYGqrzPV2dyv/UrVv22smFla5Xl0cYs8TX7dxVExy0JN75dzyN6/5f9/xfRstKmZo6K7i4vTjrskvTK05+97eR39Jn3JT8pxvrKqxSlHUyrDa7s2Pb2sLIiSyhA8NkL8meNtjzTM+6Sp63ctPGMkvYo82w2t52b1stS2QRQlw0cbJsSJvG4ltXNU0v/WXRzLDZNUmXTJwk37SnumL93umP2X9OIgsAAAAAAAAAAAAAAFGQyzJE49OWPrcnkP4hWW1btsdK9Tj9yb4PgpkeOVnXxph/J2nCPOuLJblCCCFySyq3OTu2rc2fGSUrJSnt7hcCKSmBIkV9i9+EtJxHNra66+1bKvv32CfDanvdXuvsuHBw46r8zBjZGGfbDjQHx6qRUqdNHnT2hr51lSTlPF3vctptwfQdS0ml/R13z4GNq/LSQkdbudzyf9b0575YSiprnR31TyuTXZQlsTxU2/8hSJaSV2qdrvq1+QO2CQAAAAAAAAAAAAAAFMb09vYaXQYAAAAAAAAAAAAAAABACMZlAQAAAAAAAAAAAAAAgHmQywIAAAAAAAAAAAAAAACzIJcFAAAAAAAAAAAAAAAAZkEuCwAAAAAAAAAAAAAAAMyCXBYAAAAAAAAAAAAAAACYBbksAAAAAAAAAAAAAAAAMAtyWQAAAAAAAAAAAAAAAGAW5LIAAAAAAAAAAAAAAADALMhlAQAAAAAAAAAAAAAAgFmQywIAAAAAAAAAAAAAAACzIJcFAAAAAAAAAAAAAAAAZkEuCwAAAAAAAAAAAAAAAMyCXBYAAAAAAAAAAAAAAACYBbksAABjfeooTB3TL7XQ8anR5TG5cx7Xh61eX4y1fB7X1jerSvPGyKUUljtcnlhvjdt5V3ngAxyzoNzljbq2uvJrsy8jxC6h/IRP7CEk8qwAgMGL3lIQNiSE5g2uYoMPOjzntSgllMwfJgEwJ9/hqrwU2Q3CCseJc4Pa0HmP48HgdlLLXdpc7Id2CwlDKpB23ywSHFcb190BmAS3URqg2YL5xPXVpgdeB+SyAAAwPPg8zRtL70zJ2tDSEy3w8Xl2PnprZtaie5atq1O84Kku3nrcwJZfZfmhB9OeFQAgR0sBAEBi+bobqx+v8wT/4XG8vKPNJM0wgcEQUYGjGZ8+gGGHCxeGO3rgdULNAQBgej6Pa2PprSnzl4aEQQN5myvuLVzT6An74oxZV1+hfeFUUF9+6MGcZwUAyNFSAABggE7njjrlfYKnbvP7Rwx/hERgMERU4GjGpw9g2OHChRGAHnjdkMsCAIDp+Tre+dW6RhXrdTfXVkQImISYcfP0b4zVsliqqS0/9GDWswIA5GgpAABIvO6Pd9S4Qv9Z95vKxs+NKI0MgcEQUYGjGZ8+gGGHCxeGPXrgdUQuCwAAI4byR3WS1WZ3dlzo9TuyMf8qAwsHIYQQV+VvPOL/RN5ZlZms/x45KwAAAAAMdO7ELkdN/61Cxr3WRVL//101Oz7uNqxUAIapxHd3AABgEvTA64hcFgAARowzXZ91+f+WcleXrczPlGjqRzvOCgAAAAAD+Np2vOzo73OX7ly2+qcF/mQWT41j14lzhhUMAAAAGE7ogdcRFQkAwIiUcv20y2nmocRZAQAAAEAIIXxH3t8c+PnorNSpU27IK8jsX/Q4Xt7R5jOqZAAAAMBwRQ+8xqhMAABGpEuv+Np4o8sAs+GsAAAAACCEOHuk6e26/r+l3CU3pSZNysrLDYzMUlfz1n4v2SwAAABAXOiB1xi5LAAAc+pubdj6ZnlhypiAWYXlmxwNrd5BbcpRVbpgjMyC0irH1obWQUwC7vO2Njocb1aV5o0JsaC0yuHY6vLE3+Hn87Y2Bg82pbD8zcbWhPYbaltFiebzuLY6fle6IHCyzCos3zS4T8J8VWFEeXwe11aH8tvnr1VHgs9MAFDJtGGDGuc8rj8rL7spC0p/N6jC+4W9kg8+UInpnMf1Z2Wl5ZVWOYZWYzp8EGGqJa+06s0hVXWsXRod40VntrAHQAJ1v1/5+Bb/QvaS7GlJImlC1m2BaYZE4+tvNHcaU7ahSHQLGHPvg2pojD2K2LRu9405Xj36dvo2nPh4Q1saxtUqaFZdOkTUKul0AmtWMyM63jP71TImPW6j4kGzNTgGHq8eu9ay2dLugqNx7/SIvhKOBr0AABjpr3brDH+jNMNq/2tv74Uet73EIkVsuixldvcpdRs/5a6rsEbekhCSpaTa2fGV2q3VV0YrWLCEJRucHRcibeafTlvgiK32jt4LPc4KS5iiFVXWvWyNtSeb84sO+wOxiyQesHf8U5cqkh+OsNicPcGXOuwqyt8Tccuq9nXKbS+zRCy71VbnVr0DLc6WKLXRT37CRz/8oZTnn4M8Ky50ODeURKzPfhlWm131twYAooq7pRhGYYMaF3rcb9usGTELL2te+o46yiZjX8kla0Vd9DqJ3ZypPIQMa0VTx4WQDUaKSQKG/kHIzpMZNuc/e3t7v+pwVmt0nkSmPsY7dCp0fTVBUfT149ta4s92AKZy4VR9WfAakFvp7r99/cJpuyN4ISipj+fKqLwH6b/8xmkot5CatIBq6dbQDPEoNL8HH1A+jdv9hH5qARr17Wh4Glw4VBkYFElkltSfVHUcIe+qfM7ouDp6d4d+4ZkOEbVK2pzA+tWM/vHeqfqS4PbvqXSfibq2vOmJeUdwxl15j3/L4b4Uib7rUbe+wbdRUdBsaVTPxjRbWu1a19tk7S44GvdO69rhP9geeMSJXBYAgLFCbp6PdNQ/GStYEULkltW3R+5Q6O3t7e3tOWQvyVURTAghWSv2ROme6O3t7e290F5fpm5r/TKK7MfCb1P5nKO93V4UPpy6p3JXtb65LJpUkZG5LO2HKiNUnqzolrK6WJ+udmeLVrksQy3PoCLpnoOVUXp/Qvfrfw4HAEMx1FwWs4YNqnzVsefFqF0qwcJ/pa7n/UJHXZmaBzNCCJFhrTwYsQ1W2wn7lZr6l6wbDn2xR23noDYfREgnnapyqjpPolMf4/V19BuVy2LA2Q7AbE7Wl2QGvuq5lYcC3/QL7krZBSLmg0k5I3NZNGsB1dKlodHgKPR9KKhxu5/wT61vr9r17Wh5Gsgf26tNI1N8W6Wy+kNvGB1Xx5XLolV4pn1ErZJ2J7A+NZOgeC9iaxJr5Vg5W/JULams/pRiwwm/61G9voG3UTHQbEVk9mZLy13rdpus4QVH295p3Tv8yWVJEHJZAADGkt/rShbrvRa1wcpye3vkhNm44h4hYoRlcXd29InQ8SePfu590lYUoZy5le52NQH9YHNZtKoiw3JZMu61LlIXxcdKZ9HwbNEkl0WD8gwiklb8BFMVS4Wzh8dbAIZmSLksZg0bVIkwYkf4whf99pWVMXveL0TMnIgkcieXqk5Y9YcgWVavKVHTCavZByHvpFtT/4H6qo56nsSkPsa7oL6eI2x/0LksBpztAEwnWsKKYqSHmA8m5QzLZdGyBVRL+4ZGm6PQ8aGgxu2+EZ+a1n07mp4GoYkpp2J+8+RpAUIqqT9lfFytPpdFq/BM+4haJU1PYB1qJnHxXqRRvsJRBKsxcrbk34iQNRN+1xPP+kbdRsVGsxXtIM3bbGm8a31uk7W84GjaO52IDn9yWRKEXBYAgLHk97oyktW25R13IBbpcddXhgwuF6VbbWDck1tSuU02WNxXHc7tW2zKX25EDMsGxr4ZVtvrtYqRZi/0uN+xh24xwo2Z8ubNHwWVVNb3zYbzVYfTbrNm5lYeuqAqoB9cLot2VWRYLou8YPKz5UKP+50BH0WUn1JperZokMuiSXnijqSV3dlCWEoq7bIvYG/vhQ5nbRzfQQBQZ0h97vILknnCBnV69tiUP6uSrLYt9YF58b7qcG6rDP/7oQg97wM2KCSrbct22VGcctfbB2wzQvuophNWMaJ4XIcQqR9Hww9Cdp5IN1lu8r9D2bpp37Spj/HU13PE7Q8ul8WIsx2A6ShGgBjw6FH5qqpn6n0MymXRtgVUS+uGRquj0O+hoLbtvjGfmtZ9O9qeBvFOM6T4RDJL6k+aIK5WncuiVXimeUStksYnsOY1k9h4T3EqRnteG9rvFK2ZiDzBUOLveuJa35jbKDVotoZjs6X5rvW4TdbygqNp73RiOvzJZUkQclkAAMYaePMsWUrs7jAZtRd6Dm1QxBfh8/1Duyck64t7wk95GDoTsFRkD5N8q+hQiDF43YWOpgp5wm/YG7OBzzksT9aHlLDnmFv+H7XTdqqbv0bbKlJzn6bV9K7hHhFJ1g2H1Jwt4dO0tT5bhprLonV5Yu+xj3JA4/BV2qv6OwgA8RtMS9d/0TJX2KBKSK9KpF9NnXLby+RlE0JE6HkP7aaJfCUfMIxwvImq8R1C2FGLw36+2n4Q4c6T8IO0adq0xRvjJTqXxZCzHYD5KO5wwzzqUD5IUPFMvZ8WuSzBjakMDLRuAdXStqHR4Si0ugcPX8IhtvsGfWqa9+1ofBpETzIL3aBiMIyQnDPD4mrVuSwaVFevDhG1SpqfwNrWTOLjPZUzBynPcCGiTWMXcYKhxN/1xLm+AbdRKtFsDcNmS/tda34d1vaCo2HvdOI7/HtVP5fBYJDLAgAwVmgUFfXeKeTOJ1xoe+GYXTaie+S4x7+6Ypi+MPdRyo682L/WVa4froShzzlUdA5qm8uibRUZm8sSbZqbkLA1XD1rfbYMNZdF8/LE3GOYdaKfkF+125dH7DIDgEEbbJ+72cIGVRQ/EZMstj2ROzlOHaoMGU44XM+78jdnsbqhQzq5wsU2MZuzOA5B2XAk6IPQ+jxRKd4YL8G5LIac7QBMR3mXFD6eV85dovbpphG5LJq3gGpp2tDocRTaPhTUtt036FPTvm9H83hDUTPRW94BEwzJXzQqro4zl0XT6tIiolZJ+xNY05oxIN5TZFZFnDkoJJksYm309kaZYCjxdz3xrp/42yi1aLYiFs+0zZYOuzZ5d4p2vdMGdPj3ksuiqyQBAICJ3LO69NtTIrZO41Ozb5f1JrjbOs6GrOE7Uv9yVUv/grT8xad/MDM5WmOXNOXbpasDkdmWxyvf71a8fv5vLc11gSVpaendadHbzqTps3OCwU1nV8/5qKsLIeXmZU2KsY6mtK4iA91hK78/M2Lhk5Iz7y8PZq+7anZ8HFJys1WFYeU5//e23UfVrTpuyo0LcySrbYvdbn/H3bp64QSCSQAGMlvYoIav27mrxuNfshSXP5iVHHHlCTOXPvFiUfQ5ns+2vvnSusAGpeUv/vI7ketECDFOWviwrH301D1e3djtU1X2fnEdwrgpi4pjHYLeH8RQz5NBSniMF53Zwh4ABulsfuP1Rv+CVHBbVph4flJWXvCRo6fqv3cc0ejCqLHEt4CRDKWhMc9RRKJtu2/U8erftzP0eGPCdXkFgQyVps1NxyIeZ/fHO2pc/oXMgrzrJsQqnAoJjpeGuDvNI2qVEnACD6lmjIj3kiZk3VbgbzM8NbucYQ/Q2+E+6BFCCKmoZOVNfesedHd4w6wq/7bKT2/zXy1j0v42arBotoLFM2uzlYBdm6w7Rbveae58Rx4ePwAAzCT39uzU8VFeV/Ym9Jw8FRJFnT3S9Hage0IqyL9tyrhYuxyfdveDgRxnj33fJ4oOirFS/m963O/Y7VsqS3Ij9PQpjf3G9JsDRTzZGbO/Iyfr2oRmA2heRYaRSlb+JHNi1FUm3nBXfiDsHnBHbbaqMK48ipPWVbNx+2Fv5FsdKX9jx8ZVd+fn51vSot4JAIDuTBc2qNHp3FHn74+Scgu+dUP0a2nSVbfdd6cUbY2TLe/uCyyoOwpF+yg8H+z75HSst8hpfgg6fxDxnSfaSXSMF53Zwh4ABlH1CFzxYDLGM3UjJb4FjGBIDY1pjiIibdt9o45X/74dDeINeRqZp27z+0fCf/FCHtPe9715WuTODjWu1nV3A2kejqqk/wk8pJoxKN6Tp2GFP8DgSSsVfH/FHbf07TBC4kvXoT1O/zHIU8PNf7WMyajzdgCaLTmTNlv673qo12GtLzia9U5z5zsCmadjBaNXe3u70UUAYBYzcmZPj940KcKars+6zihflgd5qn+bkpySPssfrRxtPtAWckOelJxmyc+/u2jtjo61C7X4sYvCjFlXX6H1NqPSo4oMMSNn/jUxC5+UetOSYF9QSARvtqowsDxfv3rWVYEFT/XSa+8sq3I0tka5ZwAAEzBl2BCL4iFi9pLsabHuyUMeKA7c4Cd7dgZ+vaT2KBTto3DvPNAexxXf9/mxjzr8C+k5s6cO9RB0/iBinyf6SHiMF53Zwh4AhlA+As/96TLL5eFXVIwPYYbfsoeT+BYwgiE1NKY5ioi0bfeNPF6d+3Y0iDeUVVf3dlP4IZHifEyrzpDjaq13F53mEbXa/ep+Ag+tZoyK9+RpWGEPMHDSSrPSr5Kmps7q+3fYB+3BD1fKXW21BNLOzH+1jEn726hBotlSMmuzpfOuh9xsaX7B0ap3mjvfEYhcFhjszJkz//Vf/2V0KQCYxIybp39j7FA2oBiMbnr61MgjCMolTZ2dk+5faHO3hxvhMrZzHtefHY5N5YXfXlytckA8ocEhx8vIKtLWnFsyVDwhSrp82vUp/oVPDx7/IviS2arCyPIk33BXQa78H43rli1ekH7pv4xZUFrlcGx1eczXdQ0AwzJs8H127KPAQ8QZ82ZPj/ZDqH4Trpkf5fdQX3Z9FtigyJp/bfQRy/wURxFpfO8IAiOECyGkG+dcc3Hst0Q/BH0/CBXnieLZjFYSHuNFZ7awB4AhfK1vrt0QfAS+5KbUiB3Dl1uW/VT2u966Hc5OvUsXt8S3gOENraExy1FEpm27b/7j1a9vR028oWaaobjTONRIcNwy1OrSPqJWSfcTeGg1Y1i8J08FCDek0Pnj++x9J232kuxpY4ODPQx80C7PuVSe3sPg6hGL5rdRg0SzNYA5my19dz3kZkv7C45GvdPc+Y5EJupawei0b9++3//+92vXrp00yURTiQMYrk6fOhlM6N+yLH3Lsrg3oe7HJd7WhrqWLiGE8HXt+d2ydXWx3mAaCasi3U2aeKmaMGb816641P/30d1tfz8vrup/m9mqwtDyJKXlP21z1BVvD32hcd2yxr6/Mqy2su9ek37jnZkSudAARgYjLry+ns7gU5Gbp6eouiO/aOLkiP1W5zvadgcWZqRefYXKe/zxKdPThWjsW/B81vWlEGp/oCzvU7t40tcuVtMq9P3EKsIDIbO1yCMSlQxACN+R9zcHRnOI9Qi873e9df3ru9at/eMyS1GamW4EDGgBdTAMjkLTdt90x2u6vp1JWXm50jqXR4j+nIClM5Xfu5DRlWLMEDFSaR5Rq2S6EziEgfHehOvyCjLXrXMJ0Tek0L1pacEz09d2oH90ib7Eo74H5EePhjvJZcMOKU9vs1e+GprfRhlhGHwQI6XZMntV63DB0aZ3mjvfkchMtyAYlfbv3y+EcDqdMdcEgNgUCcuD05fuENY5j+vPb5YXpowZM+bS9FsX97lnaJ0dV826+utDeHv89K2iBFIbxPdF8OGYrSoMLs/EzJUv15flRl6hpbr4vsWLslL+ZVZh+eaG1u5B7gcAzMOAC+/5k8ePxN8TKc/LjGrGpEvV3uInXTJx0uDGq1b0qWnCbC2yNhIe40U3MisZQFzOHml6W3bvumVZ+kVjoviXa5fVyS4cEac7MYeEtIC6M+VRaN/uBxh2vHr07WglaYLFujowPUSY751ygqFooyuNYDpH1CqZ8AtrZLwnn2Zo37stJ2UvBVsfafGca8YKIa7IuGVO/4t1zS1/k+1QPsFQlNPbhJWvgo6XU6OY8oMYic2WKatalwuOFr3T3PmORKMx2IGpVFdXC3JZAJiez7P72cLMlKw77imuHmo4BJhT0pSFT2/rcNpt1oyo67VUF3//1vSZCx7dycxDAGAukyeq7+G6+GuTVIy2HIvaH8JGTi0FACSG9+M/1jQN4f1NNX/82LzjrSe+BdSD+Y9C23bfiOMdBn07SdOyl2T7FwZMM6TLBEMYFPN/YRNKPs3Q0Z17Pgk+YfYda9rc1/pkFuRdN0EIIcanZt/uf1jt3HOoK7BqxAmGQoyAyh8Zt1Hm/yCGf7Nl/K4TjN5phEO0AyO1trbu3btXCPHYY4+dOcOoTQDMyec9vPH+zOyV1S1RV8uw2l6327c7O47YrZpPZYpBO9/TZb6Z3c1rnJSZv2rjgR73O/YtNmu0NH5P45rczB87TnDDAADm8VlXj9rLsu/0qc7TsVeLZXdbB79XAoBhwNfdXFvROJRn957GitrmbrNG/4lvAfVg/qPQtt1P9PEOl74d+WN+T93m948Ea4kJhszE/F/YBJtwXV5BZt+fnppdzkB74e1wH+w7a7PmX9s/zVPS9Nk5/V8vV82Oj/2JLxEnGAo1Aip/ZNxGmf+DGN7Nljl2bQB6pxGKXBYY6cMPPwz87Xa7DSwJgJHoAXvHP3vjd2RVpiJj2+t86d/LBvxgp693w26322udHRd6e3t7D25cdW9+/rczr+hq222ieUyj0qiKDHH0yPGTam5Hzvd0BoY2lWZMuiRC6GO2qjC2PEnJaZb8u1dt7Oi90OGstW+pLAk/uqOn6hdP1B7nfgHAiJCYC6/O4/0e7VTfw/VlV+fgHmmOTZl+c9xvOtvRpvJ2z2wt8ohEJQOjkHxeksHy1O1wmvV3AglpAXVnyqPQsd1P8PEOn76dpNSbloSfZogJhvqYY9YYU35hZRIf78mmGQq2F7IErNx5GVf6tz326jmL/YkvHx37rK8mVU4wJMxf+eHpfBtlBFN+ECOn2TLJrlXR44KjSe80d74jxCiNd2ASDocj8HdT01BGOgUAIYQQV1w9K/izmc6unqGnXp87UVct+/maZCmpdnZ85e/dyM/Pz/9upjScWlPtq8goKgv/xfGDn/r/vnjyRFkui9mqwmzlEUIIkSRlfjf/7qK1O3p7v+pwbrdXllgUr7dUvVx/hGQWAMOUARfepOSpqbMCS2p/JSZvy0IpuurUJnqKkN66GbOuvkLd24QQ4pKJkwNPD9Tu8eypkz0RXzRlC2hGp0+dHPQv7KhkYHTznfjL68F5STJL6k+q7cvv2WOzBC76rnVr/9hqmuDfgBZQB8PgKDRt94073mHVtxNpmiEmGOqnfUStktm/sAbHe/Jphlz2fcfPCyFLwApJT5ElvvQnbMWYYMgUlT+UaFzocBtlBFN8ENGNkGbL9FWdwAtOfL3T3PmORKM14IEJdHZ2bt26NbBYXV1tYGEAjBBjvzH95kC0su/dlpPRVlbD17bjZUfgVy8WW+22tQWZ0rhobxnijY3eNK8iw7S521VM1979yZ6dgZ9SzbklQxbAm60qzFaeUOOkzG/nF619p+dgpXzK0oNH2r2m6c8GgLgYceFNmjzt+mDPWvOBtrPR1u6jaMsGkHfVKeabj8rXfmBnoIcryrhl4SSnpM8K7FLdHn2fH/uoI+KrZm8BTePLrs8G/Qs7KhkY1c4e2fHfVcGH+Ll5WZPUvjX5ursKsoOLiiEijJb4FlAP5j8Kbdt9o453mPXtjE/N+35RYICL/mmGmGAoSPuIWiWTf2ENj/dk0wwd3XmgzSe/GqRcP+1y2YEnTbg2K6f/776inv5k3wfRJhgyQ+UPJRoXOtxGGcIMH0R0I6PZMnbXahhzwVHRO234lRA6IJcFhnE6nfLFvXv3tre3G1UYACPFFRm3zPH/fXTn9r2qpkv0Ha7Kyyos3+RwbG1o7Va8FJzVVQiRXXDXdckxN/bZsY9MMnhleFpXkWHkU+pGouzrkVKnTZb3VZmtKgwsj8/b2uhwON4sL0zJq4rxU8vkjKWPPRIc1dHT2fUluSwAhikjLryKnjXZD20j70/Rlg0k6zJW1zgKIYTvyPubg3NNpOfMnhpH10DS1Nk56XHtUbm7gczWIidK0iWTZgROhk8PHv8i6tq+7kPOnYPf2WitZABCCN+xps3BsZClgtuyJqi/6o9Pzb5dNqT7lscr3zfL5SDxLaAezH8U2rb7Rh3vcOvbSZryb/cFKqo/h4wJhmQ0j6hVMvsX1vB4L3S0leCxD0yjvPKbt/Sve3Tnnk+6gw/gI5zeelR+QqNxPW6jjGD2b8FIabaM3bUqml9wtOqdNvxKCO2N5pAHBtu1a1fIfxoaGgwpCYARZNzkaamBuxBPVdUf9sfMWfZ5979VU+eqLr5v8eJFt6YvLG34PPiiPOM+NBMirM8bK39TF3+5E0jrKjKOp8ax68S5aGv4Pt31+rbgBxjaaWu2qjCqPL7uhsfT0hcsXrz4nuJqz8F9LZ6otSpE0qUTJ8e/GwAwHyMuvIpB4wM/tI2yQ0VbFk7y1PTpgYXYjaMQQnTt/6MjGK5IN8655uJYb5G7+Jo5NwbrLfYelbsLw2wtcqIkXTJxcqDmT38WIzlU/gRrEEZrJQMIfaqRWZB33YR43p6UetOS3ODvgj01u5zdJsllT3wLqAfzH4W27b5Bxzv8+nZkOQF9uRryCYakpaV3p43q5zraR9QqmfwLa3i8J59mqM3d3u1tP3Kwbykn69qQNErZh+j56FhHayDrJf+BvOnhTm8dKj+h0bjQ4TbKECb/FogR0mwZvGs1tL3gaNg7bfiVENob1TEPDHTmzBmbzRbyz/fee8+QwgAYQZImWKyrgx1t24tXveqKPgGK1/nSKltjYNFy3/fmRRhv2XPk2GfRo6hzJxxP/nCdK+o6htOzihLM43hh/bueiGU/d6LWtqKqxb94x8rvzVF22pqtKowqj7yvQQiP44U/fDivvSwAACAASURBVBR19iblL1Fy52VcOTb+nQKAGRhy4R2fdveDJYF91j371IaWyFfdc57GDS8E2zIVG/SsX/HoHw5HO4pznobnVhVv9y9KlpWL5sXxA30RWm+e9Sue+FPknzr5vK5XZbtTsUHjW+SEuWji5In+vz11NW/tj3jUPq+r5ldDCjJHbSUDUD6Sj2uCoT5J0/MeyJcls2x7fden5khmSXwLqAfzH4W27b4Jjnd49O3I79M9de+2tMqGFYlzdKURSfOIelD7Nd0X1gTxXnAkCVfNjh11O/yjstzyzStDVx0/ffa8/olA6mrWrKnpa6qkgvzbpoTNNtOj8hMZjQsdbqMMYfJvgRhBzZbJq1rbC46GvdMmuBJCa6M86IFh9u3bN/Cfv//97zs7OxNfGAAjSkhHW+PKrDt/1RAhk9fn2f3s8mXFjYGs+oyih/JvSJY1jsERL4UQWx5/anPEkNHnaX72x3MXr1dm6PecPKX5bOKdXT3nh7QBbasobkMuf5Cncc3Knz23O1w6yzlPw5ofyj4OqajoBzdMDF3L4KoYwKjyKEat9DQW//xXjsORbhh8nvpnfrWBkY0B6EnDliIWQy68E+Z8b+Ud/oWW6mX/GeGqe87TsObeW59sjG+DwlO9dOHymggRy4BtSvkP/eD6mMPsh1LWm6dqxQ/LHK1h9njO0/zb5XeubBzwQvQNGt8iJ4jil3ai0bbq6frwUY3KaoxulFYyMOrJx3IY5FONcVNuyw8+XRAtVS/XxxgCQWORA4PEt4B60P0oNO5DGGq7b8inZsa+nVjk9+k733p1+7v+IsU1ulIC4+oE0zyiHsx+zXfZMT7eCw4p5Kl5aX1ThxBCiOwl2dMGbnfsNXMW95e1sbq6UQgR4/TWvvITG40LHW6jDEGzlbBv/ai64GjYO238lRBa6wWM8MILL4Q9IXfs2GF00QAk2F/t1hn+a8AMq/2vGqx/4Zi9KEN5dcmw2l6vdXZcCKzS4azdYrNKyrUsFc6eC8ptnXFX3qNcp6RSvqHeCz3ud+wDNxWlhP902mYEN2dz9sSuJMVbhGTdcCi0nAMrJ+qWNawiNYejtvzx1YNsg7Yt9W7/jr/qcG6rLMlVrrHc3v6V7lWhqjZifUbalkfNHnt7e3sv9DgrLMoqs5S8Yh94qleWKFaLUrEAEJfBtHQmDBtU69ljs8g3J1lKXrFHb8uiH0XoBvtax+3OjsBV+pS73j5gmxlF9mNhjkFN4z6w3hQx0oUe9zsR4qMH7B3/VLXBwX8QQzlP1MVmYcUf411wV4aELJaSV2SHHHImWKxWS8TtD+5TS8DZDsBIF07Vl8m+0PdUus8Majsn60syZReFO2zOLwas888O+wPBVWbYnOEu9mqpv4XUtgVUS+uGRvOj0OoePEDbdt+AT02Hvh3d442Q72+grsrqT0WtBsPi6ugHqEN1aR5Rq6TxCax1zRge752qLwnZcsQWIaRxUXF6a3310DIa7zXoNkoVmq1h12zpsWsdrsNaXnA07Z1OdIe/uurCYJHLAmPMnTtXhPPUU08ZXTQACabDQ6necHFeTBGeyl9otxfFuSX5RnMrD4WGP4PIZVEctVxIDcQTM2lVRXEHc1HKH4tiX0WV9jUWVeXOLatvj3bnoN3ZokEui7blUbnH3t7e3i+ctjui7SIMLW7JAKDfIFo6M4YNqg3splEp0lFc6Dm0IcLTl4jHYCmr6wh7HVcXq1zoqCtTV29SUeWWYKMduRNWsw9i2OSyxNP+ShZbkzv4nHhQuSy9hpztAIwT0o+fW+keZPge+kxdKqk/FbqOprkscdxCatoCDqZ4mjQ0mh+FRvfg8iJq2e4b8Klp37eTgHgjXKsd7tsXZUdyesfVCc9l0T6iVknbEzgBWT5qCqhhvBeaoRL5pFU2HKpOb82vHtpF473G3UbFRrM1xHo2JNjQfNf6NFtaXnA07Z1OaIe/6urCoDBODgzQ2tq6d+/esC899thjZ86cSXB5AIxAyfNWbtpSYc2IvWYfS5m98en8cBOyJk35zurXnrSo3VBuyYZ36l8J/NzHc9DdEXVyR5WunP/d28MFX6c/6/pykKM7a1dFKuhQfvG/Jt74k99WxuyMyi2rr1q9cEq0iCehVaGCMeWZmLny5fqysL9YCivDWvmH5/KvJpQEoBE9WgrVDLjwJiVnLt9UrybAkCxlL9ruDduvp9zgzIL1jfYStZ01GdaKLZtW50hDuI4nSTmrN22M3T9oefK11Yv+dWyMtYQwX4ucCBMzH3ymMvYhS5ayjZtWzr906DscjZUMjF6+I/UvV7X4l4YyPWjShKzbZNMMCU/NLme3ru2z+sDAgBZQB5ofhfaRlabtvhFxiyn6duKUfN1dBdnKf6mZYMjQuDqhNI+oVe/X5Jcdg+O94DRDQoioJ+3YKzPm5apaM0Dzyk94NK7HbZQBaLai0/BbP6ouOJr2TnPnO4KYK2zHKPHhhx9GedXtdiesJABGsCTp5kderXNuUA46F0aG1fa2e9vq/LRIt0vjpIVlm5zVsULGDKvN7uzYtrbQYrnl9kDMpVEf37gpi8peCxPJDak/RbsqikmX8gsxYWbRs432MkuElyVrRZ17y9PRE1mEEAmtClWMKU/SlIVPb3HXVcRO9reUbHDWvVqUYYrZ5QGMEDq1FGoZceEdJy38+Tb327ZofSu5JfaGbU/fNV1VD2ZSclr+2m2NdZFGx/eTrBV17t0bH7l56N1bSVLO01H3KFlf3LOpbKGktj/IbC1yIiRnFK23R/3U+k4DzbojR2MlA6PU2SNNb9cFF7OXZE8b/IVkwk3LVsumaPFsWPtmq54PxeMKDAxoAXWg7VHo1IegYbuf+E/NDH078Rqfmn274lOUcvOyJsV6l8FxdWJpHlGrZPbLjqHxXkj64/T0qRF7j5JSb1oSyHtRdXoL7Ss/4dG40OE2ygg0W5G2qfm3fjRdcDTtnebOd8QwY+SOEc/hcER5tampKWElATDCJUmZhWvfudDhrLWHzqQoMqy21+32d9w9BzauyktLjt4gjpMyC9bWu5y1A2dPzi2p3OLfTn6mNE6E3Il56nY4O7U4likLn94yYP7LjJz0bwypLdesimLuSJ/yiwlp+U/Xh05vmVtSucVe727d+EiO+hg0YVVh6vJMSMt5ZGN7h7PWHmamcEtJpd1e6+y48M7awkxTdv8CGNb0ainUFyDxF96k5LS8VRtdHc7tIXuUrLYt9nfcPW+tzZ+p7JqZOHniRdE2mZyWs2pjR4+7fuCF3FJSaa+td5/qiKt9jCk5LWdVX9PxiuzhkGQpecVe727duHxevD2wZmuREyBYh68rnsRYSirDnwZDNgorGRiFfMeaNst62HJvz04dP4TNhTxT99Rtfv+Irk/24w0MEt8C6kHDo9ApstK23U/0p2aCvp04KcoghFRwW9YEFR+g4XF1QukQUatk8suOgfHehOvyCvzTDEVvfZKmzs5J7/tT7endR9vKT3w0LnS4jTIEzVaA3t/6UXTB0bR3mjvfEWFMb2+v0WXA6NLZ2XnZZZdFWWHu3LnNzc0JKw8AAAAAmNinjsIFi6uPCiGEsNic21ZlMjAWAADAKNDdUDrz1nWevoXMkvq31y683NgSDVtE1AAADEvkGSHRnE5n9BX27t3b3t6emMIAAAAAgKmd/3vb7qP+hfTpKUP5VT0AAACGC1+3c1eNx7+U+9NlFhJZBouIGgCA4YlcFiTarl27Yq7T0NCQgJIAAAAAQGJ5XeV5C0p/53A4tro8aiZn8LUd2BnoeJ+RevUVY3UsHQAAAMyi07mjzp/KIuUuuSmVhzn9iKgBABgtCH+QUGfOnLHZbDFXe++99xJQGAAAAABIrPEp0/+1cd1PFi9evCgrZWppQ3f01X3Ha9c+W+dfkhbPuYaOdwAAgFHAd+Ivr9e4+hek/AfypvMsx4+IGgCA0YL4Bwm1b98+Nav9/ve/7+zs1LswAAAAAJBYY6+4OnWGf8GzruIVV1fEdX2e5ueeWFHV4l/OLMi7boLO5QMAAIAJdO3/Q1VVYFSWgvzbpowztDymQkQNAMBoQS4LEmr//v0q13Q6nbqWBAAAAAASb2z6vz1okfxL24uz/u/C8k2OhlavfCVva4NjU/n9ufNXVvsfYQip6NH/sFyeyKICAAAgIXxez98D0aDP07yxdElW8Xb/P+5Zvewm0i/kiKgBABglxvT29hpdBowi8+bN27t3r5o1n3rqqUcffVTv8gAAAABAYp074Vg5d/F6T+w1ZSxP1m8qWyjxe1wAAICR57zHsSJl8cvhXpIsttptq+YlJ7pIJkdEDQDAqEAuCxKntbU1PT1d/fqnT5++6KKL9CsPAAAAABih+3DVIwuXVantfKfbHQAAYEQ77yqfmVV8dOALlgrntoczkxlffyAiagAARj5iICTOhx9+GNf6brdbp5IAAAAAgHEmzCz6rctpt1kzYq2ZW1JZ7972c7rdAQAARrCxKdNvHvBPyfrink3LSWSJgIgaAICRj3FZkDiLFi3aunWr+vVfeOGFFStW6FceAAAAADCUz9v6l7qWjrata4qrW2T/zy2p/PH8y/71xjszJR5eAAAAjHg+j6vm16uWrmsUQgghWW3P5y+4mVBQFSJqAABGLHJZkCCdnZ2XXXZZXG+ZO3duc3OzTuUBAAAAAAAAAAAAAAAmRD4qEsTpdMb7lr1797a3t+tRGAAAAAAAAAAAAAAAYE7ksiBBdu3aNYh3NTQ0aF4SAAAAAAAAAAAAAABgWuSyIBHOnDljs9kG8cb33ntP88IAAAAAAAAAAAAAAADTGtPb22t0GTDy7d69Ozs7e3Dv/cc//jFp0iRtywMAAAAAAAAAAAAAAMyJcVmQCPv37x/0e51Op4YlAQAAAAAAAAAAAAAAZkYuCxKhurp60O8llwUAAAAAAAAAAAAAgNGDOYagu9bW1vT09KFs4fTp0xdddJFW5QEAAAAAAAAAAAAAAKbFuCzQ3YcffjjELbjdbk1KAgAAAAAAAAAAAAAATI5cFujO4XAMcQtNTU2alAQAAAAAAAAAAAAAAJgccwxBX52dnZdddtkQNzJ37tzm5mZNygMAAAAAAAAAAAAAAMyMcVmgL6fTOfSN7N27t729fejbAQAAAAAAAAAAAAAAJkcuC/S1a9cuTbbT0NCgyXYAAAAAAAAAAAAAAICZkcsCHZ05c8Zms2myqffee0+T7QAAAAAAAAAAAAAAADMb09vba3QZMGLt3r07Oztbq6394x//mDRpklZbAwAAAAAAAAAAAAAAJsS4LNDR/v37Ndya0+nUcGsAAAAAAAAAAAAAAMCEyGWBjqqrqzXcGrksAAAAAAAAAAAAAACMeMwxBL20tramp6dru83Tp09fdNFF2m4TAAAAAAAAAAAAAACYB+OyQC8ffvih5tt0u92abxMAAAAAAAAAAAAAAJgHuSzQi8Ph0HybTU1Nmm8TAAAAAAAAAAAAAACYB3MMQRednZ2XXXaZ5pudO3duc3Oz5psFAAAAAAAAAAAAAAAmwbgs0IXT6dRjs3v37m1vb9djywAAAAAAAAAAAAAAwAzIZYEudu3apdOWGxoadNoyAAAAAAAAAAAAAAAwHLks0N6ZM2dsNptOG3/vvfd02jIAAAAAAAAAAAAAADDcWKMLgJHJ7XZHeTU9PT3Kqz/60Y+Ki4u1LhEAAAAAAAAAAAAAABgGyGWB9i666KK0tLRBv/3rX//6UN4OAAAAAAAAAAAAAACGL+YYAgAAAAAAAAAAAAAAgFmQywIAAAAAAAAAAAAAAACzIJcFAAAAAAAAAAAAAAAAZkEuCwAAAAAAAAAAAAAAAMyCXBYAAAAAAAAAAAAAAACYBbksAAAAAAAAAAAAAAAAMAtyWQAAAAAAAAAAAAAAAGAW5LIAAAAAAAAAAAAAAADALMhlAQAAAAAAAAAAAAAAgFmQywIAAAAAAAAAAAAAAACzIJcFAAAAAAAAAAAAAAAAZkEuCwAAAAAAAAAAAAAAAMyCXBYAAAAAAAAAAAAAAACYBbksAAAAAAAAAAAAAAAAMAtyWQAAAAAAAAAAAAAAAGAW5LIAAAAAAAAAAAAAAADALMhlAQAAAAAAAAAAAAAAgFmQywIAAAAAAAAAAAAAAACzIJcFAAAAAAAAAAAAAAAAZkEuCwAAAAAAAAAAAAAAAMyCXBYAAAAAAAAAAAAAAACYBbksAAAAAAAAAAAAAAAAMAtyWQAAAAAAAAAAAAAAAGAW5LIAAAAAAAAAAAAAAADALMhlAQAAAAAAAAAAAAAAgFmQywIAAAAAAAAAAAAAAACzIJcFAAAAAAAAAAAAAAAAZkEuCwAAAAAAAAAAAAAAAMyCXBYAAAAAAAAAAAAAAACYBbksAAAAAAAAAAAAAAAAMAtyWQAAAAAAAAAAAAAAAGAW5LIAAAAAAAAAAAAAAADALMhlAQAAAAAAAAAAAAAAgFmQywIAAAAAAAAAAAAAAACzIJcFAAAAAAAAAAAAAAAAZkEuCwAAAAAAAAAAAAAAAMyCXBYAAAAAAAAAAAAAAACYBbksAAAAAAAAAAAAAAAAMAtyWQAAAAAAAAAAAAAAAGAW5LIAAAAAAID/n727D47iuhe8f1AoHmLLhLDG8Qi7YIUM9i3hN2ntiq1nPWAjOfG14xJ2SOxocETd2Am+uRUWiUe5XF/XmjULokjdu3ZusEsKSDEJwTMFZkksYYFyCzsLO3KMURUeWWghxjMkeLHe/AJGM88fo5npnpnu6dGcfpmZ7+cvjdTq6T7dfc7vnP71aQAAAAAAAMApyGUBAAAAAAAAAAAAAACAU5DLAgAAAAAAAAAAAAAAAKcglwUAAAAAAAAAAAAAAABOQS4LAAAAAAAAAAAAAAAAnIJcFgAAAAAAAAAAAAAAADgFuSwAAAAAAAAAAAAAAABwCnJZAAAAAAAAAAAAAAAA4BTksgAAAAAAAAAAAAAAAMApyGUBAAAAAAAAAAAAAACAU5DLAgAAAAAAAAAAAAAAAKcglwUAAAAAAAAAAAAAAABOQS4LAAAAAAAAAAAAAAAAnIJcFgAAAAAAAAAAAAAAADgFuSwAAAAAAAAAAAAAAABwCnJZAAAAAAAAAAAAAAAA4BTksgAAAAAAAAAAAAAAAMApyGUBAAAAAAAAAAAAAACAU5DLAgAAAAAAAAAAAAAAAKcglwUAAAAAAAAAAAAAAABOQS4LAAAAAAAAAAAAAAAAnIJcFgAAAAAAAAAAAAAAADgFuSwAAAAAAAAAAAAAAABwCnJZAAAAAAAAAAAAAAAA4BTksgAAAAAAAAAAAAAAAMApyGUBAAAAAAAAAAAAAACAU5DLAgAAAAAAAAAAAAAAAKcglwUAAAAAAAAAAAAAAABOQS4LAAAAAAAAAAAAAAAAnIJcFgAAAAAAAAAAAAAAADgFuSwAAAAAAAAAAAAAAABwCnJZAACAGT7wraqYNmnp1r5x3YUvhfr+18B42KJNAwAUPhtbFp2vzqpxRMEJh/r2vdq+vm6aUtmqrb6+UFEHQcavCyJGIN+E32uvK1PUeE/7PryU/Vouh3xPJVZSsbXvsvwt1f7yvq3xKmraU76Qld8NHcq2o2KV7wO7t0cHjZfZKGEHUFWVdHOcRPqhoVmEBLLuGuRRMICckMsCAADsFA4d27n+wbLqHf1jDD0AACSwsWWhUYOWcOjgT++tqn740dVbulV/CHU07TvD2ExGXFxAHgqP9nZs6A4lfhHybe8a4hpG8aDxMhslDAAFjEoeUYyXAAAAm4RDfTvX31t25xNJN3UAAJgaG1sWGjXoGD+27bFVm3pDaf+4cMn8uRZvT37h4gLy1QV/V7e64gt1735rkJsRKAY0XmajhAGggFHJQ4FcFgAAYJNw8PBzW3rt3goAQOGwsWWhUYOm8Oixvds0ElmEWHh3+TXTLd2efMPFBeSp0Xe7OvuSf9n987bej+zYGsBaNF5mo4QBoIBRyUOBARMAAGCG6+t3DkZ22r0VAAA4CI1jcVLPTODytL7w48cfrnLxbNEkrgugIF368A1f52TdV/mYp+Jwx96QEEL0dXa9+4/Lls2yc9sAAABgFzqAyA5jJwAAAAAAACb5bPjccOxnV+3GlrX1JLIAKHThoa7tvslUFteDqzf+qME1+ZdQp++NDy/ZtmEAAAAA8gfDJwAAAAAAABYou3XB1QzEACh44cG3dsdnpFpScd282+oaqiY/hnzbu4bCdm0ZAAAAgPzBEAoAAAAAAIAFrpr7lZl2bwMAmO3zwSOvd0/+7KpdeVdFyZzqutr4zCzdnb//0zjZLAAAAAAyIJcFAJCbcKhvn8/Xvn7pNIWl69t9+w4NjGa7nle3ripLXo9vX19o6qNc4VDfvl1bVy1Rr7N3IN3AWTjUt8/38vql8U0oW7r+5Sl+vV3fK12a41K3vv1V36GBcbs3zVKyykH+eR4eH+hNrLBs1dZX059mKZth1fmZdpenLVm1dZdP4xsBZCPbSmB04NA+Ca120kakrSjy6RoPjw/0+nyvtq+vm5Yk91Ak100zIUByZuMuK6QUBVRoRRg2ACgMo2+1bdgT+1CzsmZBiSiZVX1f/DVDoveV3x67YM+2JbsU6vuduvWpW9/uyyUuMqH2lhe/FX7vbHTg0D71Di5ZtXVXlu21gyPDJNa30RJPb8efjVL7OLJ7YeYMolrYrTOlW+oUJg1xW8SEZtEQ6RXvpVDf79SHoWzp+pdz677ZPvxS0BeOTFKCAThJBLCc/jnZ1NRk9wYCMGYi6N/R7Na9ol2eF44GL8pYz7buwIjuWv7s9SycXHphq/+LSCRyMXj0BY9Le41HgxOJfx8JeFs0t8Hd4tX8dpu+9wt/68LEcq3+sQyFnHn5dDvi72h2a+2JfrGoV5j0jUGvR3OliuUnTrbFn9wTVc095zPsY9TU/svEcohvmIzzXHkcPd5gZGLMvy3NOl2NbSdHNHfHiuvC6C4LUelp9foz1hIAorKuBJRGAt3bNC9/IYRwuZs7Ml2P6WqVnme1r/TaZu9JzSZKYltmsGVJYyTQ06ZXySdW0LzDr6whp/bV2o1jKgkNh2mNmlyyQkqjq7Km0L4Iep/U3ZCoJ73BL8zZkTwKGzSuC+MXlz1BIwAtEyM9LYmao7YtMFlTfOxvfSBRhTT3GG5v1DXqZAUlYTvHAq+3eio1aphKz7YjwYmk8EOj0k6sUsrghpKU+M3otk29dzbS05zYyEfbAp/pb4riDMlUpJHPAm2PxtasrKuVbcdCj/fPkcjEWMCbc5BjcWRonK1d+yiJp3cuZ6MpJSy1j5NM3lUcZd4gai67nEXPTnaBpPkG62uk+Mpkx/C5D/lmsTYTmkVDcq54s9uRRKWn2JVoO6Jm//CLqStMWr3FV03Odw0mmRQMwHHIZYEN9KsiclmAvDARPLJNJyhUSh/SxdfT3WIkWhVCiEpP2wntiDupGzZysq0x03ofaPV/HIlEIhNne1pqM+3FGu/ZtOGgTd9rei6Lfic2rral56xGLyL3qFQZ7Bod65wItCXK1NXSM2Kgj2NuOUQiEs9zdT/q7FmvxtkW73XYdV1EImMn2gxWESJDLQEgIetKIGbspLc50xU9eT0mDXwnSa4kz3rXZKpVXO5mb2As3Sptz2UxUtGpVDZ6TyfviTm5LJIaDlMaNblkhZQRxxXa1HNZii9syDmXxZ6gEYCW8z3NVfGrq7btZPzqUl13me+UxJmRy2Koend5dpz8+KjBm3byBjdipMVvFvTONA96poUzpRIq8w5VdXXS7atBCUGODZGhcfZ17aNlI/H0zvFstCKXJbc+jmpn5V3FkUjExEHUnHfZYM9OdoFosL5GikRMiuGty2WR3ywaIqXiTdoRnTy/hNqWnrMXs8llsWP4xbQVpmHxVWNSLouMYACORC4LbKBfkZDLAjhfNtG5EEKzVzyhGQVqMXjj4Vnv60YCFyHc2/xj/1f5cJjeTjR604U5Nn2vubksm3r+mC7TPP32ZXlbIpJFVJr9PQZlMJ3Vo37mlYPU81x5HB97trVRYwAo8eyjXdfFxwb/Xf2l9COATLKuBCKRSJbjxUIYzVPMopJ0uVu60wyo2JvLkvWoWVTKQJUJuSzyGg75jZpcskLKiBMLbYq5LEUZNuSey2JL0AggPb2EFdV8SBnvlMRJz2XReLQ6Xb3l3rip2cBNO6mDG5FIRG78ZkHvTGsynnRUEV2GGlh5OqmXVLYdLrfnMbfBXdOKJeyJDI2zq42Olo3E0zvns9H0XJac+zhxMq9i6QdC9i4b6dnJLhCdorK8RjIthrcol0V+s2iIrIp3KjsihKvx315aazSXxZ7hF3NWqFF01l41ZuSyyAgG4FTkssAG+tUIuSyA002c9iZFby5P654DiVnsJoL+vS8lTeaWplc8drQ16e5F0noiI4Eeb1ty3rFW/rUydlGvsicwllhh0qSFlZ7mNe74z62/6YnPMjcW6GlLmhYybaxs0/eamsviust9V2xr3c1t3sPxBy8mgv69yZunNQQpIyrNdu531YyIU5srXnY5yD3P1b2FxLZNnmwXg35vq6dKsSX2nJ/q8evk0suuAAEoZV0JRNKNF9c2t+1X1EIXg/4De1rVTy4ZyVNUVpIuT+ue+GU+MRY4nLzCtNWanbksqYNclZ7WV/aqpiyeGAsc9qbsSfJIjfRcFpkNh/TGXSpZIWXEmYU2pVyWogwbpOSy2BE0AkhLNU9Syr0Q9V+NTokkO5dFVQOkVnQXg/79KTVtlMZNO8mDGxG58ZtFvTNVqepOYJO0PXoHVOd1HunbLHVQmrYlSrtrdkWGxtnVRks+vSWcjWbnsuTex5kktxdm5iCqlF3O3LOTXSD6LK6RTIzhLcllkd4sGiKv4o1LORCGdyRTLkuiCK0cfjFjhdosvWrMyGVRlsDUggE4F7kssIF+VUQuC+Bscs1vSQAAIABJREFU6izd6OyCaZ/XSc4aThp3SA7FNNeTOslh+oAsNXZJn+GePlM+/SS6SVF12ijHpu81NZcl9r0az1tMjJ3coYqR0ydrG7hdl/kdq/ojockbpjo5pzhXvNxykH2ep/aj3M/2JL0Mdex0IPEbW85P9Tz/mrts8EQCoJB1JZA8POTyvHA0/RuUk18qbGD6BP11plzjqdWy/LbM8Nu7Vbe9M8yln/wSHK2RGkNfnbFxlNtwSG/cJZIVUkbyodAMvluqOMMGA+Vj6OKyPmgEkI6qhU1zl059/8NgJpncXJakylbryfi0b1tIWwVJH9yQG79Z1jsz+MYBdXUthN4T9nqv80jXXqd/AYqBXbMtMjTOrjZa7ukt9WyUWcJS+zjRxST3wiwYRM1tlzP01KQXSEZW1kimxvAW5LJIbxaNkV/xGtyRkYC3JWVHjOWyWD38YvGFY+VVI+WuQURyMAAHI5cFNkhuKdTIZQEcLekJWt2JZ5O60KqkaXXGd6YAKylcNpRTor3O1JAr9hbhNDJOP27T95qfy6J7UJL2JW1AKSkqVZ0q+q9UlzJXvNRykH6eJ/ejMg7+2nJ+Kr9UfwvVr2TmThKQUbaVgLrV1h5Tiy2uariNzUamFwkkjbyk3NCyL5dFfSMt8xM56uU1Visll0VywyG9cZdHVkgZyYtCM5bLUqRhQ0RSLov1QSOAVOqmP314r74ADd1rkZrLoqorXO7Wo9rxh7q3olUFSa+9JcdvlvXOVGmCmlVr0s1LrUKILqv9Og+57bV9kaFxNrXRJkZZOZ+Npuay5NLHiZjQCzN/EDXXXdbvqcnvlmZkYY1kbgxvfi6L9GbRGPkVbxY7MnKyLSnrz0gui+XDL1ZfOFa242blsjhoxANSlQgAAAwLD/Zsb++PfXqgdev3q0o1m5KSed9cvzERIoQ63/CPhoUQQnw+8OovtoRif3CteeG//u08vRZphmvZP2xN5FaHujd09E6uSsujG9d/U2OdM8tvuUMV5tTWP3TbbI31zL7pzurELpwb/kT3W+37Xul0dkQIMbOi5n5FFyIwFPzcrA2ZdXNdQ3xs5cjuI6c1D/zou12dfbEPVQ11N8+S8PW5lIP557mrtq56jt4CySw5Py//dejNU8a2Z8a8ry9b7vK07vF6vYcDAxuXzSI0BbKRqRJQtdquNS88/90btVttkdxw79nQ9tZohi3QjwRKSqu+r6jW+tu39wzqt94WufyX/mPd8U+uJ9Y/ski/9ikpv2V5ooq8MDx22ZwNM7vhcEzjLi2kFAVUaMUaNkhkc9AIQAhx4dhvX+mNfXA13FedJryfU12XuAcSav9N16CJzU2K8Kj/jc54Zetu2vpUdanmwjPmPdz0QtLr8JLJr70lx2/W9c5KZlXf1xA7tOrGWmE8GDgREkIIV2Pz2ruiy54IBMfTLKoM2DLW1bm0146NDHVY00bLPr3zZqwg1z6O7F6YBYOo5nbrzO+WplmHVTWSBUfHVNKbRYOkV7xZ7cisG594JusdsXz4xfILx952XAoHjXhALm4YAACMUweaev3hqJkVNffXRjufe/3Bs8/HOp/n+//wdnwhV0P9ffNmZPrq2bc9VJ+INkJ/fPv9T/UWr72/pmKm1h+nl5Xfnfjkql15V4Vmezh97vyKRKh8bnhMv39h1/dKp7sjIrkLYSrlWGeoe/dbGh3mpE7L49++I6u7NRpyKgfzz/Pl1TdlNaBjzfk5/Zryu+N/7OvceeC9ce3T11W/M7hz3SP19fXuRbpdMgBpZKgEPh888nq81TZWC81c9MhT8YeZQt6339cdmXc1r/1BlX4koK7WTgye1akQrDPdVf/zscBhr3dPW3Otxp22pP9Q1mznL5h1x8LkhsNBjbuskFIUUKEVa9ggk61BIwBhMFFMdackQ+aZfBf8Xd2xCsBV2/CN2/T7ICXX3/f4g6kvbFGQXnvLjt+s7J0pcwrTt0qJGtjV8J2nH7gnutUaN8yGTx71xzYsUzpmdu312PkR5e0rx0aG2ixqo2Wf3nkyVpBzH0d6L8z0KNHkbp3p3dL0LKqRzI/hzSW9WTRIesVr/o5YPfxix4VjYzsuRU7BAByNewYAAOMUIUiG/vCkkkWNXdHO57eqXPGFR98/ejD+KIbRtNySirtWJh7fChw8flan27Rw+S3lOtt25exrE+Hq4uW3XGe0OTx1QX/w3a7vlS7DjlhKPdbZ/fqR9M/tZdlpMSancrDgPF8yf242W2TV+fnV+Uuuj38IdTxx04Mt7b7eAUfcwAYKSqZKQDmsZvghmNKyxUtidcGpY8eHdPr2htapqtZC3V3+C0a2wnwlpYvc9fWPNG7uCm5e5pQpGUxuOJzUuEsKKUUBFVrxhg0S2Rk0AkhOFKv90Wr31ekXVM2iZO0j6eGPTr8TjH0wUnclZd6kkF97S4/frOydKXMK07ZK8RrYtWTx9a7rKpZMbla6G2aJ1ChX7UaPW/c2Z+b2WnUTdPjc8GfqPzsyMtRmURst//TOi7GC3Ps4sq9i06NEs7t1ZndLtVhSI5kfw5tLerOYBakVryqbtmZlzQLpO2L58IstF45t7bgUOQcDcC567AAA4z4bPjcc+7ns1gVXT7EV+WT4XHyES1TfeZN+7n9MyXW3LF8c+6A1eV3UwrvLr5mus6orvjL3iviHxeVlehm72bDre6XLtCMiKfgzmZEZ47PutBiRWznYfp5nu7y087P0tocaapW/6N2yesXSxVd9adrS9e0+376+kKNGqoC8lemiVk3iXb74Ou0ZdpVUtdBQ4Kx2JWSwZlOtMO8GCy6F+n7n8+3auuqbKzoMzoieA3MbDkc17pJCSlFAhVa8YYNUtgWNAIQID7y6eUciUUwvT/Fq9+ofKR5JtzDVNT4xvhDC9fXbb7hCd2khhBCzbrhTZwIu6bW3/PjNyt6Z8tZguvmxLp952xutgWtW1iyYnng8OvWGmTI1KmNdnW0bN2XWRoaarGqj5QcneTFWkHMfR/pVbHqUaHK3zvRuqeYqrKiRTD86JpPeLMpnqOINnzv9TvxALLzjlnID9V52O2L58Is9F45d7bgUlgUDsAF9dgCAYVOLolJXExx6M/5hYcX8uQbDjJll5fFoLPU9vihgGWeMT3oEMMOMgtYo5vO8ZFH984m3/yr0blm9YsXD1WVfmrZk1dZdDhmpAgrWpyPnEw/G7Fm9+MvTDLmquqk39l+6Y5RGa7bpV82JP7906s2hv1o+Cbth4wOHfFGvtq+vmzZt2rRp/09Z9QMrVjze1NGf+d9zVkQNh6SQUhRQoRXMjtgtL4NGoDCEB9/aHZ/zKNNNC/Uj6X1bNr82YE2/QHnT8Yo5X7nCyMC4aiaJJPJrbxPiN0t7Z8qcwpT5scJDxyfnL4jeaEzkg6bW2Io5tOyqq+2ODG1nRnCSB2MFufdxZF/FpkeJZnfrzO6W6jC/Rsr7GF52s5irqVa84bELiTyXu8vLDB2HL8++1ljukRHSz3O7LpxCasdRQMhlAQBMzZzZV8lIdV045yqjbVHJlbPnGJv87/ol87869U2aOru+t+CVzHJ7NsbHOtPMGK+eK97Aqwqs5ojz3Mrzc3bV2u09LbXaC/R3ND2+4uHqsi8tWbV196GBUas2DCgkmS5q1SNiUyMl9WT6VbPNfy/y1F0K9f3u1a2ryqZNm3bV4ntXRD26ekt35n81lSkNhzNJCilFARVa0YUNEuV/0Ajkq88Hj7yuaDsz3XT50k2ruxVhiuZLwSRT3XSUTkrtbUr8ZmXvTJlT+PYf+s8r/pQ4SVwrbr9huhBibuU9t0/+sftY/18UW618MYGldbVTI0MVO9poacFJwYwVaPdxTO2F2RnuTrVbZ2e31NoaKQ87I+Y2i0blXvFePn9mMPvJsmZ+Ze5Vhhe2fPjFtgsn39txFCbOIACAra6dbTzQv+IrcwzMdYhCVLKgZmVN7EPKjPHOnyu+CM/zknnLnt8f9HtbPZW6y/V3NH3n3sU3Lv3pQeZoAfKM0aednCscevNnq6rKqh94tKkj12Ei6Yqw4chdwRRaweyILfI9aATy1Pi7r3UeyeH/j3S+9q7Vb1gwGsmonp7X4+Ta27remfL1BKcOHn0/kYkQPn1kd/QkqWqou3mWEELMrKi5P5bU4D96Mv7+QetfTCCEwyND20k8vR0+VuDwPo4Z9YzDdzkn1tZITm4FjJDeLBpAxes8edyOo4BxCgEAbHVueMxopzT86ciFTzMvhoKkDI6T5i3Mh7nii/Q8n+Gqql+38/hY4LB3T6tH75GTUO+m2qq/831IOguQR7Ko2RwoPP7ezu9X1azNMEV8paf1Fa/3gD846PVY+VLwom04clMwhVYwO2KPPA8agbwUHj22d1tvLjehQr3b9h4btTaweHMoKPfFh06vva3qnSleTxDqfMMfP6zjwcCJ6ElSfedNk691KCm/ZflkhNXX2fVu7IaZ9S8mcHxkaDvJp7eDxwoc3scxo55x+C7nyMoayemtQCbSm8UMqHidKi/bcRQ4clkAAFNzYXhMRoR76oLxQP+T4QvkaBct1SvVVXNQ58Nc8UV9npeULnLXP7JuZzAyEfTv9e5pa04/n3Co/Z+f2XumgIdQAFs96Q1+Ecne4LoqzUezjNZsl8eGL8jbEUnG/b/4YUvKk1/RYTKv1+vd6w9ORCKRyImd6x6rr/9m1dzhoTezn7Q4F0XUcEgKKUUBFVrB7IhN8jtoBPKS8uKaqlB3l9/0gGF6WfndWf/T58GhgKEF5dfeJsRvVvTOFK8nSBxWRTZh7R2VX4tt4PT5t6+I3TB75/S56PdZ/2IC50eGtjMlOHHkWIHkPo7sq9iMA2Fpt86Mak2fhTVSHsbw5jaL+mRWvE55Z1OM9PPc4gsnD9txFDpOIgCAYSVXzlkYjwyHAmenOAewKlA+NXjmvMEbGKpYeeGS+XOn9vXF6dOR845L+M+S1ozxTp0rnvM8VYmr6lv1jzRu7opELgb9B7xtzW7V3/vbt/cMkswCyDJ3/pLEY0vy0gXiPr0w8qmRK3b8bGAo9vPCu8uvmfoM1tLasksfdncoHh93uZs7/MGLsWGy+vr6+m9VuaxvTYqo4ZAUUooCKrSC2RFHyLegEch34Q///ZXExVXV3HPe6D2WsaOt7nhz0Ldl82sDZvcFrpx9bfwLjVa2n4+cH9P6m/za2+z4TcG03pny9QR93rfPXBZCkfCUdFtLccNsMvvQ+hcTODQytJ2VwYmDxgpy7+PIvopNPxBmd+ssrNbSMbdGckQMn0s3WXazaJjcirek9LqKJfFPRieY+fjMiQ+y2mg90s9zOy+cvGvHUfg4iwAAhpVcveDWstiH4DunPzLQ1xnv21q3dP3LPp9vX1/sFbfKQFn1MkVd4bPHD8YDfdfCOVfShmXhk+FzTkj4z8nMirrvNMbTwidnjHfwXPGc53pmuKq+Wd+4+fDYiTblS7JPDJ4dJ5kFkGT6NeV3xwc/3v5D/3nJ6w8Nnj53KfNi4Y9OvxOMfbj9nsochulktWXhoa7tvvjcDO7Wvfs3N1S5Zuj9izUpocXTcMgKKUUBFVrB7Igj5FvQCOS3zwe7ftOeuBtVW1c9x+i/lt78UENN4qNqIiVzlJYtXhKvbY1VtqpIJoX02tvs+C092b0zxesJTh08PhRWFmPZrQuuVrRWJbNuql4++XN0fz99/+0/WvpiAsdGhrazJzixe6wg9z6O9KvY7ANhdrfOnmpNwdQayQkxfC7dZOnNokGyK96SaxfcmkjKOXZ8yEA4Mfr+0YPyZtiSfp7be+HkVzuOIsBABwDAuNLrFpfHfk56/byG8NnjB0/0bvnBihUrHq6u+kb7e2GhiofUL1PUXdPgW7sT8xYvXn7LdUXdhqkeaP7gxJmPdZcOj570HzR9m0xXMu8/Px4/cyYHOh08V3zxnufh8YFen8/36tZVZXXtGZ6tLK184h9/kphHOHRh+BNyWQBZ5lbec3vs51MHD/xvQ6+ZD7/XXle9ausun2/foQH9eksx24HO+pTVmqtiwbWK8Sm72rLEa56FEDUND91cmuk/wudOv2NBSmgRNRySQkpRQIVWMDviDHkWNAJ5LXz6yO4j8U+uhvuqZxm/umZW1NyveKfIng1tbxmp/aau5Lpbli+OfTBU2apr2hTya2/p8ZstvbPkp7QTu5ya7fS1v7lnctlTB4++P5q4xWvZC4acGhnaTv7pnRdjBTn3caRfxaZHibnvsj6zu6UZmVkjmXF0rOwmS28WDZJe8aqScoyc0uo0dwmkn+f2Xjh51Y6jCHAeAQCMU480dfte+1OGfG11gBvP21XewBChTt8bH2Z8AmD4T6/5uuOfXF+//YYrstjwwlNy5exr4yXw6bkMfXoZ7y93BEUkHe2ZKOeKdz2x/pFFTopsivM8D48e2rBo8dIVK1Y82tQROvF2fyjDXpdcNftaazYNKDozrl1QER/OCbW3/zpTqy1EePxPv+/s7utoenzFiofvXbxs/aGPtBcOdW/o6B3VH1D5qLft5/FqLfn+ll1tmfLBNUPjsKq9MFPxNByyQkpRQIVWMDviEPkVNAJ5TF0/VzXU3Twrm38vqbhrZW3ikfZQ5xv+DKFFjq644favJ6KjzJWtuqZNQ3rtLTd+s6t3pnw9wVDg7Oj42cET0U/Lq29KynZSvBgu9M7p4ED8bln9k3XlVtTVzo0MbSf39M6XsYKc+zjye2FmR4m577I+s7ulGZlaI5lwdCztJktvFo2RXvGq3jFq4GGJ8AdvvLJf6lC59PPc3gsnr9pxFAFOJABAFtQjTQe2/eu+93Qm+Qyf2bv5Z4roPJ63O3PRI081J8KxF5/+6a/11iMuhQ79y7qmA/EVudc+fEcWvaaC9OXZ186O/Rzq7vz9nzQLMDze1/nclj6Nv+YXZSQd6v5D/4AiiT7LvrQFivM8Vx4jIUK+//Hrd8b1llc/QVJ7R+XXjL5zGUAmJbPcno2KVrtp3S/79KfmHvf/Yl1rb/yj+/Fv36H7moDQjude8mtf45dCh15UNEAPrP327er7Ww5oyzJPqX3pQ9+z37OoGS2ihkNSSCkKqNAKZkccIr+CRiB/qe8tZfWCoaiS8ron6xXJLPtfeeMDM5NZ1NFR6MWnn/mf2g86h8f7fqmoadOSXnvLjd/s650l5iro6+zq6u6KPc19z998LXnRmeW33DH5HoXuzk2bOqNnlKuh/r55xid+kMRZkaHt5J7e+TNWkGsfR3ovzPwoMddd1md+tzQjE2skM46Old1k6c1i9uRUvOoD0f2z/7ajX++U7t3xP9r7s9xQfdLPc7svnDxtx1Gg6L0DALJRsujbzze5Y59CHU8sW9OZPkYPh479yzNPJ+JCV+1Gjzsenc+6/dtrH4gvq7cecSl0aNNj9z7bG/+Fq/7vv3trxrkHC50q8V/0tq57vieUpvwuhY7925oH1/Zatl1Td2F47HLmpZSzdx78/S8P/CF2VyLrRwCtUJznuWqG1VBv0z8953tPqwMZDvX89+d2MOE/YJak+0O9a6sffO6QxhOQ4dCbP1uzuqk3/mxSZePf199Wqn9NhnqbVq/52ZvpGyB1teZqbPzubbPVy5jdlmm0LIkpcIUQezb8t92ao43h0LGf/d1/WvGi+oGtsfMjGV+/baxRS1U8DYeskFIUUKEVzI6Yy/DFlV9BI5CnlDMeTTGpbsa8++oTt7dFf/v2nswvnsuFOjoKtT/9vRbfQJrK1nD4Ib32lhu/2dY7S8yPFer8xYtHgkIIIWpW1ixIXeX0G25fMbnDvR0dvUIIS+tqJ0eGtpN7ept4Nsot4Rz7OCb0wkyPEnPeZX2md0szMrNGkn90rB3yld4sGmFGxas6EP0dq/+LRg2TchRkkX6e23zhOKodz9tmFLJEAMvpn5NNTU12byAAXROnvY2VquvW5WndczgwNhFbIOjf+0qrR72Me5s/vkDU2NFWtytlPQf8wYuxJUYCPd625lrVMqKy0XtavaKoP3s9C+Nf1uof09uFL/yt8WXFk97gF0aXXdjqT17Wru+NTATa1EXjcje/tNcfjBXOxaB/v6L03B6PW3c7lTuy0OP9s96OGNpxAyWj2knh8uw4OZbu2Kr3e6SnRX3eRP+5pWck4/8aIbsc5J7nqhLLdL4Z2TzNledyfk6M+bfFz7boPrubX/Iqzs5IZGIscNjb1qxazLXGe/ZiylcBUMi6EkjXaotKT+sryityIujfu6fVk1S3pjbckYi6VlFXbD2B2NYkNUB6F7jstsxgy/JZoO1R9c42t6XWUallMkmjdTD01QaqZZkNh/TGXSpZIWXE+YVmuGCLMWww8L1TiRgj5geNAJKuskfbAp9NaT3ne5qrFFfpA63+j9ULfBH0Ppn4e7recZYbntIAqSKBibHAYY0oQKO6kzy4ITd+s693NtLTnLR5mscu6RwwUlfn0l4nLW9jZGicbUNPUk9vqWejzBKW3MeJRKT3wkwdRJWxyxnDTukFki1TayTZrYDMbrKRHoH0ZjEzcyre5APhcje/5NU/pXVWaP/wi90XjolXjay7BhKDATgauSywQbrWIoFcFiAPpMboGaSORkUikYmxkzs0QlItLndLdzB9IOTUwXezvjcSiXzsb00k/mcot9YjgcTwn2NyWTQ6zxk2IN3p52ruGcm0xcZILwep53ne3JQyfnLGaY/kAoibwmBKZAqtts6wrLJWqXzM87Cx9epc4HLbsojBlmXirLcxyyJRbklt20njI8LqrzZSLUtsOMxOy8j5hqKckDKSV4Wmf+UWZ9iQVVSplOkAmRs0AkUv6f5KbVtgitF8cuZZynUqO5clEpkIdrcYa4BcjW17NrkzVXdyBzcikYjc+M2u3lnynS3tGlh9iA3V1TJvX9kXGRpn49CT3NNb4tkosYSl93EikYjcq1j6gZC9y0bCTskFki1TayTprYC8brKxHoHsZjEzcyre1IQ5gyTlskRMOM/tvHDMu2pk3TUgl6VYMI07ACB7pXes3bVnW9Jjslpcnm1HX15blTr/ZEnpjQ0v9nqbjQZklZ5te3ZtXO6i7Zo0u+qp/96W+Si43C07d6298yorNilbX7vzW/enO/yfnhv+RHNi6dKbH2qoUf/KyXPFF+d5Prtq7faelrTPOqRV6Wn79b/Uz8/vnQYcK6tWWwjhbvH2Pl+f+cXGV9/243/dn3mIrbbZu0f7ApfelhlqWUrm/e3GXz3rzry2qNrmHYd7Xoo/NxY6EQimm6x4So1aGsXUcMgJKUUBFVrB7IhcU7248ixoBPJMeLBnu/IFcFN/H03JrOr7FK8ZEqHON/yjpr5nSJS4lm/ctTPzfTv3s7/a+PB/nJ55ffJrb5nxm129s8TrCYQQujXw9K9V3lFraElTODsytJ3c01vi2WhSCUvp4wghpPfCzIsS5e2yPrO6pQaZWiNJPzpWD/nKbhYNfKMpFW9JadWaXT1GVutyt7zQ+ljaVIncSD/P7bxw7G3HC6YZhQQFPuABADBJievun/yy279DPe1nMpe7ucPf9/JP7tAKzktKF9Vv3t/brTVlYHxFnm3dgTd3/uTuQh+pz1JpZeOLXt3Sq232Htr/vGPvcMyY93DLr9KMYmj1SaJmVtTcr/ofV21d9Rz5WydNUZ7nJfOWPb8n0L0t84Mp7uYd/u5fNlZmen8zgKkz1moLISo9ra8H9m+sX2Rw2OGrVet2BbwtWqt1ebZ1B/Zsrr9R7wKX3JYZbFlmuJa17PJ3ZBptrPS0ev3B/ZtXud33JJoejXtsU2vU0iqihkNSSCkKqNAKZkckmvLFlXdBI5BHPh888np34mPNypoFU6+NZt21eqPiXQOhHZtfHTD7LkWJa/nzupWty/PC0V0ty1wG7wbJr71lxm/29M6SspTKF1+nudaSirtWxu+X2VBXOzwytJ3U01va2WheCcvo4wgh5PfCzIsSpe2yPtO6pYa+3OQaSfbRsXzIV3azmJEZFW90tf+0P/B68ptqVaJF91C5jKScVNLPc/suHHvb8UJqRpGrwh/zAACYpcRVtWrz4Ymgf2/K+yvdzW3evT2B9w5vbqjKGOOWLlq+bmdwLNDj1VrPSHDnT5bL7MAUkNJFy9ftPBv07/W+ogrT3c1t3sOBsd/n3tU0V3QUI/mtsZXLF1+jN/ShjI+FcDXcVz3L8SFNMZ7nsxYt/0n07Ezzjlt3c5vXu9cfnDi8eVVVEdyGA2ynaLWT30AvKj2tr3i9hwNjx3euq1tUmtUVOWtR/fM9yW9orm1u27PXHzxrsFqT25YZbVlmuKoaNvf0pQljRG1z255YgdRHIxlV0xPq7vJfyOGrjSmehkNWSCkKqNAKZkdkmerFlZdBI5AXwqeP7D6S+Fh7f03FzBxWl5R5Fure/dagBY/cJiKQlxT30lzu5pe8PYGBnWvuyPaOnfTaW2b8ZkfvbNbNdQ2x1xPonyQl192yfHH0R5vqasdHhraTeXpLOhtNLGEZfZzJjZTdCzMrSpS3y/rM6pYaYEGNJPfoWD/kK71ZzMCEilcIIUpKF9Wt29kX9B9IOsdcntY96Ytu9rWzvyxpp6KbIPs8t+vCsbcdL7BmFDmYFolE7N4GFJ1p06bp/LWpqWnLli2WbQwAIP+MHlp/471bQtEPVc09r29edrW9WwQAxeED36qlKzpOCSGEcLf696+rcnTCJEwTHj204cZ7N4UWtvrfW1dlzgNtgAQEjQAAIIMi7OMU4S4DSbgKgLxB9hIAAMgv4VH/G52h2KfaH612c08CAAArhT8ZvhDKvBhgL4JGAAAAACku/3XozVOxD4vLy3KZXg6AuXh+CgAA5JcL/q7u2F0JV+3KuypIzQUAwFLDJ4/6hRDi7vIyBhXgXASNAAAAQMEb79u6Yt35R/7+zv/wpflff9DAy/LCQ8cPxlNZFlbMn0u3FnAurk8AAJA/isJuAAAgAElEQVRPwh/++yudfZMfXPVP1pVzVwIAACvF2mJX7T1/8zW7NwbQQtAIAAAAFIGZZeX/sbfpB71CCCFczT3vbV42S2fx8Jm9m3/WHfvkWnH7DdwqBxyMjjwAAMgjw3/6dXt7/AHbhvr75s2wdXsAACgq4fEBX8v3nm4PCeF6Yv0jixhTgFMRNAIAAADFYPrc+RULYx9CW7a91DesuWw4dOxfnnm6vT/2uaqh7ma9xBcAdmPcCQAAOFZ4PPTX8fiH0LGd61dWNx2I/eLRjavvorMBAICFRgOvtW/pDQlR2/KrNe5ZDCnAIQgaAQAAgCI1ffF/fsrtin060FT9/67aust3aGBcudD4wCHfrq3fr71zbUcs4124Gn/6Y/fVVm4qgGwxcRIAAHCs8NgfnylbsT3dn1zu1nXfXjTT6i0CAKCoza566plt52srVz+xfBG5AXAOgkYAAACgWJXe+t2/r9/W+2IsSaW/o+nxjoz/5X72Vxv/dh4PaADOxjUKAAAcSzVFpIq7aetT1aUWbw4AACi94yebf0wiCxyGoBEAAAAoWjPm1T9/qK3RlXnJGPezPbtalrl4DyngdOSyAAAA55peVn53yi9dnheO7lpTVUoYAwAAACEIGgEAAICiNuvGxn/r83tbPZWZlqxtbusJ7P8nElmAvEB/HgAAONjXvv7jHc3u2CeXp3XPXn/fL9fcQWcDAAAAcQSNAAAAQFGb4aqqX7fz+FjgsNf7SkpSS21z2x7vXn9womtz47JF5LsDeWJaJBKxextQdKZNm6bz16ampi1btli2MQAAAAAAAAAAAAAAwDnIOwMAAAAAAAAAAAAAAIBTkMsCAAAAAAAAAAAAAAAApyCXBQAAAAAAAAAAAAAAAE5BLgsAAAAAAAAAAAAAAACcglwWAAAAAAAAAAAAAAAAOAW5LAAAAAAAAAAAAAAAAHAKclkAAAAAAAAAAAAAAADgFOSyAAAAAAAAAAAAAAAAwCnIZQEAAAAAAAAAAAAAAIBTkMsCAAAAAAAAAAAAAAAApyCXBQAAAAAAAAAAAAAAAE5BLgsAAAAAAAAAAAAAAACcglwWAAAAAAAAAAAAAAAAOAW5LAAAAAAAAAAAAAAAAHAKclkAAAAAAAAAAAAAAADgFOSyAAAAAAAAAAAAAAAAwCnIZQEAAAAAAAAAAAAAAIBTkMsCAAAAAAAAAAAAAAAApyCXBQAAAAAAAAAAAAAAAE5BLgsAAAAAAAAAAAAAAACcglwWAAAAAAAAAAAAAAAAOAW5LAAAAAAAAAAAAAAAAHAKclkAAAAAAAAAAAAAAADgFOSyAAAAAAAAAAAAAAAAwCnIZQEAAAAAAAAAAAAAAIBTkMsCAAAAAAAAAAAAAAAApyCXBQAAAAAAAAAAAAAAAE5BLgsAAAAAAAAAAAAAAACcglwWAAAAAAAAAAAAAAAAOAW5LAAAAAAAAAAAAAAAAHAKclkAAAAAAAAAAAAAAADgFOSyAAAAAAAAAAAAAAAAwCnIZQEAAAAAAAAAAAAAAIBTkMsCAAAAAAAAAAAAAAAApyCXBQAAAAAAAAAAAAAAAE5BLgsAAAAAAAAAAAAAAACcglwWAAAAAAAAAAAAAAAAOAW5LAAAAAAAAAAAAAAAAHAKclkAAAAAAAAAAAAAAADgFOSyAAAAAAAAAAAAAAAAwCnIZQEAAAAAAAAAAAAAAIBTkMsCAAAAAAAAAAAAAAAApyCXBQAAAAAAAAAAAAAAAE5BLgsAAAAAAAAAAAAAAACcglwWAAAAAAAAAAAAAAAAOAW5LAAAAAAAAAAAAAAAAHAKclkAAAAAAAAAAAAAAADgFOSyAAAAAAAAAAAAAAAAwCnIZQEAAAAAAAAAAAAAAIBTkMsCAAAAAAAAAAAAAAAApyCXBQAAAAAAAAAAAAAAAE5BLgsAAAAAAAAAAAAAAACcglwWAAAAAAAAAAAAAAAAOAW5LAAAAAAAAAAAAAAAAHAKclkAAAAAAAAAAAAAAADgFOSyAAAAAAAAAAAAAAAAwCnIZQEAAAAAAAAAAAAAAIBTkMsCAAAAAAAAAAAAAAAApyCXBQAAAAAAAAAAAAAAAE5BLgsAAAAAAAAAAAAAAACcglwWAAAAAAAAAAAAAAAAOAW5LAAAAAAAAAAAAAAAAHAKclkAAAAAAAAAAAAAAADgFOSyAAAAAAAAAAAAAAAAwCnIZQEAAAAAAAAAAAAAAIBTkMsCAAAAAAAAAAAAAAAApyCXBQAAAAAAAAAAAAAAAE5BLgsAAAAAAAAAAAAAAACcglwWAAAAAAAAAAAAAAAAOAW5LAAAAAAAAAAAAAAAAHAKclkAAAAAAAAAAAAAAADgFOSyAAAAAAAAAAAAAAAAwCnIZQEAAAAAAAAAAAAAAIBTkMsCAAAAAAAAAAAAAAAApyCXBQAAAAAAAAAAAAAAAE5BLgsAAAAAAAAAAAAAAACcglwWAAAAAAAAAAAAAAAAOAW5LAAAAAAAAAAAAAAAAHAKclkAAAAAAAAAAAAAAADgFOSyAAAAAAAAAAAAAAAAwCnIZQEAAAAAAAAAAAAAAIBTkMsCAAAAAAAAAAAAAAAApyCXBQAAAAAAAAAAAAAAAE5BLgsAAAAAAAAAAAAAAACcglwWAAAAAAAAAAAAAAAAOAW5LAAAAAAAAAAAAAAAAHAKclkAAAAAAAAAAAAAAADgFOSyAAAAAAAAAAAAAAAAwCnIZQEAAAAAAAAAAAAAAIBTkMsCAAAAAAAAAAAAAAAApyCXBQAAAAAAAAAAAAAAAE5BLgsAAAAAAAAAAAAAAACcglwWAAAAAAAAAAAAAAAAOAW5LAAAAAAAAAAAAAAAAHAKclkAAAAAAAAAAAAAAADgFOSyAAAAAAAAAAAAAAAAwCnIZQEA2Ody39aKaTFLt/aN271B0KY6WE/5Qpft3qDidCnU978GxsMaf/3At6qCCwpAfiM2sFE41Lfv1fb1ddOUylZt9fWFwoJWxhRWn/CZDiIXIAAg/F57XZkiEnja9+Gl7NdyOeR7KrGSiq19zhlCoLEzCQWLtBhOtIbNF6DOcCW9SAC5IpcFAAAgD4RDx3auf7Csekf/mFYuCwAAUxQOHfzpvVXVDz+6eku36g+hjqZ9Zxg5AACgOIRHezs2dIcSvwj5tncN0QUFAKTFcCUAszEiBQAA4GzhUN/O9feW3flE0v1FAACkGD+27bFVm3pDaf+4cMn8uRZvDwAAsMcFf1e3OiAIde9+a5AblACAJAxXArAEuSwAAADOFg4efm5Lr91bAQAoUOHRY3u3aSSyCLHw7vJrplu6PQAAwCaj73Z19iX/svvnbb0f2bE1AAAHY7gSgCUYkgIAACgM19fvHIzstHsrAAB5Rv0EtsvT+sKPH3+4ypX85AutjAmmV60bjKyz7vs4iAAAHZc+fMPXORkTVD7mqTjcsTckhBB9nV3v/uOyZbPs3DYAQD6iAwIgV8zLAgAAAABA0fps+Nxw7GdX7caWtfWpiSwAAKDQhYe6tvsmU1lcD67e+KMG1+RfQp2+Nz68ZNuGAQAAoFgxQAUAAAAAAIQQZbcuuJphAgAAilB48K3d8ZnallRcN++2uoaqyY8h3/auobBdWwYAAIBixSAVAAAAAAAQQlw19ysz7d4GAABgvc8Hj7zePfmzq3blXRUlc6rrauMzs3R3/v5P42SzAAAAwFLksgAAlMLjA72vbl1VNm3atGnTppWt2vpq74CDRitGBw7t87WvXzpNYen6dt++QwOj2a0pHOrb50vs6aQlq7bu8vlM2uXw+ECvz/dq+/q6aUmWrm/3+fb1habyrWl2pG59+6u+QwPj2a3oUqjvd+qyrVvf7su6YLMi8yiYU7zpN3JqJTy5KjknsBRpyz/H4kq/2uxLzIYrFECeMq3+N09RVXESm9H8jgOlyMOzHQCQL0bfatuwJ/ahZmXNghJRMqv6vvhrhkTvK789dsGebUtlQmc2HOrb53t5/dIyVWAwxdXJC1qkcGLkk4dRjT3F6PzBLjuGE4WY0gi2xAvzUqjvd+riK1u6/uUp9XHs4uBrkOFKAEkigOX0z8mmpia7NxAoGl/4WxfGrj2PNxiZGPNvc6delq7GtpMjycsLd6t/LIv1p10+ixWOBLq3eVxCm8vd3OEPXsy81xNB/47mNLupUulp9RpamyEjgZ62Zrfe1sfKoHmHPziRfiV/9npihbWw1f9FJBK5GPR36K3W3eINjBjYvImxwOutnkrNoth2JDiRdLCe9Aa/yKlIZB4FKcUbMbOEDe2vy/PC0aSdDXo9mXZJfdUodkH/gjK0Pdu6M+ydOSVmwxUKwEmyig1yqf9HepoT//poW+Az3c2aGOlpiS2esRH8LND2aGzNVc095xWrcU4V90XQ+2TmolPtrH4rY2Yzmt9xoC6jJ7wJ0U6uwTkAoJAoQx0hatsCk23Jx/7WBxLNbXOPwZY7OdKYjA2kbGnOndk0jd1IwNuiuU6Xp7U7YLhFzC1omTjZFp8KJymM1CsT7f+yMvKxJqopohje+YNdFg4nZjuCnVqYsnoTGfY6UXqKTV7o8f5Zb4+sH2PP8dTKbrjS8FhlhOFKAOmRywIb6Nf35LIA1lH3BM6e9TamD+pinUO74uyxk97m2gyxYpTLs+2oTv8tEhk70abT30hem1YXKBsTZ3tajG38pMpG7+l0u5AUiF8M9jzrzry22paes3oFEjG0Hpdnx8mPj0rLZZF4FKQVr3klHJkIHtlmcH+TdtacXJaJYHeLkS6rEEJUetpOaF/nJpSY9VcoAKcxGBtIqP/P9zRXxSuU2raTupW5cuFM9xWUtxNcLT0jsRU7q4ozO5dFWjOa33FgRkZOeJOiHXJZAAAJmnHRRKBN0QhlTB2IMyWXRU5nNrmxO3uyTWMwLMHlbunWDTIikYiUoEWZTmE0eUh1jGwMPi2Kaoojhs+DwS5rhxOzHcFWktibiFwMHn1BNycmqral5+xFZ+aySDm1zMllYbgSgBbeMQQAEEIIcfnEr5/55/ZQuj/V3l9TMdPq7Ykb729f8+iKLd2ZlxRChDrW3vnEhkMfaszoN9z3i/9vdUe/0a8Ota/+YVtfLtMDhj88tKHx3k3GNn5Sf/uKph0Dn+suc3nE/+Jj9z7bm3lt3Zu+t2nvh5e0tm+8z9B6Qh0tP3zx0PnMX2eEvKNgVvEKeSUswqGDGx57dK3B/Q21r172U5/22nIX/tD3d1W1m3rTXuqp+jtWf3dNe7+ByTallJjlVyiAPCWn/p9TXRcfrg51735rUKc6uXzmbW9f7ENfZ9e7OpNQhwff2t09Wc26Gu6rnhXtdBdVFSetGc3vOFAKE6MdAAAmhQde27wlHupEXzA0qaTirpWJOT+O7D5y2q6m0ZzO7Ed/+tenl61OPximEOrdtOqxDQf1Xm4hJ2iZWVFzf/w+c6jzDf9oxvL+qLft5/FvdXTwSQxvUB4MdtkynBiT1Qi2zN5EeLzvxcfufLojcyXUvel7z7S/bbCyspCDexYMVwLQQS4LAEAIIcSuZ5va+4UQwt3c1hOdPfVi0O9t9VTVrryrwrbmIjVYrG1u26+YrO9i0H9gT6syJ14z9AwP+H7adCDx2d3c5j0cGEukXE8E/Xvb1LMF9rb+9LcDU409w+N/+u1zqh5Cpaf1lb2q6RknxgKHveodEGLPhra39N7U+un+5/5La2+6vUizCyHf9q6h9Lsw2vvcg2t7Fb9weVr39MTnzr0Y9O9vm3x2IdS7oWXLKWP7rUveUTCteIW8Eg6f2bthraon5vK07jmQOIEngv69L6lmuQy9+PQz/1Or35yr8WPbvve0qsOftD2RkUCPt031wEp/x+rv/oPvTIYtklFill+hAPKUrPq/ZFb1fQ3xBbqP9f/lsuZXDh0/qGgEQ96339dc9vPBI6/HNq6qoe7mWdE1FFUVJ6sZze84UAozox0AACYpo5eUm8ElC2pW1sQ+hLo3dPRmzq4wgVmd2f5dHXtj+Que1j3x2GBiLHBY3bqGejet3bBXa23SghZV8lCou8t/QW/zhRCj73Z1xtM1nBx8EsMblA+DXXYMJyZkMYItszchxv2/WBcrPSGE3l4LEWr/4Q+2yd3vnDm4Z8FwJQB9EcBy+uck7xgCrKOaflAIIYT72Z6kdzqOnQ7Ef2P1/IfJbz91eV44mv6VkxNjAa8yIcDV6E2ZFlA9U6tnx8mxtBMHToyd3KEK2RMvis6S6o3FGSYYTH4NTZoJeJVTMk6uUWOSW4O7oHrvtfb0jGlnYpzyO4bkHQXJxRsxoYTVbxzX2d/kmSpTZr41+nZh/Xk7k464TvmnHHTXGu/Z1EtPbolZfoUCcKaMwYPM+t/grOOqCkoIoTfBfvrJyR1exRmc9tnwO4YkNAqRvI8DDdI/4c2NdnjHEAAgEokkNTdpXiehfs1Qphe1TJL7jiGpndnUoTDNFaYEBu5t/rSLmRa0ZApF1MMO9gaf1kU1BR3D58Fglx3DidmOYEd3R+aFaXCvRwLelpS9dsY7huSfWgaHKzN2NhmuBJABuSywQXJrrkYuC2Cd5J5ApiEJi+PsidPexkTcrB0sxhZXvS01tYOqDGT19/TiWe+adGMB2VEP92R8iW/S8qkdgORAPF3PKi6px56uOzHS06zYSXfrUe3DqS4QrRUaIu0oyC7eiPwSVp/AGmNesWXV7/pNfiu2lFwW1RHX37tISv8wbQnLLTGrr1AADpUpeJBa/6tG/5Pr3sRS6lE/3e9Vfp1ihQ6v4kzJZZHbjOZdHGiU7glvcrRDLgsAIJJ8uzd9w6dMHcjYl4ySmssitzOb5qa4Tm896XZ4uihCctCStL/a6ReRSPKhsTf4tC6qKeQYPg8Gu2wZTsx2BDsi+8LMYq9HTrY1qvfaEbksJpxaknJZGK4EkAnvGAIAxLhq66rn2L0RCeHBnu3tsXkgXWteeP67N5bqNVsl8765fmM8+kyZ//DyX4feNDi544x5X1+23OVp3eP1eg8HBjYumzWF5vLyX/qPJeZtdD2x/pFF+mspKb9leSLSvjA8pjnzqRBCiEc3rv/mPM01ql6xLERgKJj0ZtPwqP+Nzvjkje6mrU9Vl2p+14x5Dze9oEzLmDJpR8Hs4hU5l7D6BBYPtG79fpX2Caw+ew2+FTsrnw+8+ost8SPuWvPCf/1b7b0TQsxwLfuHrYkHI4xMJZ1biVl9hQLIU3Lrf9UU5Zp173gwcCIkhBCuxua1d0WXPREIpns5t3LzEpOTF2UVJ7UZzb84UAoLoh0AAC4c++0rvbEProb7qtM0fHOq6xJJAaH233QNJjfcZjK7M6vfWy8prfq+Ym19nV3vJr1rQ3LQIoSYdXNdQzxD5cjuI6c1t17jBUPOi3yI4Q1y/mCXTcOJSQyMYEu9MLPa61k3PvGMKXudE8f2LBiuBJAZ1xIAIGZ59U0OirFUL2x2NdTfN29Gpn+ZueiRp+Kp3MnvwZ1+Tfnd8Ri8r3PngffGtSNdV/3O4M51j9TX17sX6XZ1tE131f98LHDY693T1lyrMR6U9B/KLTx/Qb+TkPQG6xTqLkeqC/6u7lhPwVXb8I3b9Hez5Pr7Hn/QpbeEMdKOgsnFK3IvYXUvsbb+odtm637fzIqa+2ujHZ69/uDZ52V3eM73/+Ht+AdjF9Ts2x6qT3TnQn98+/1P9RbPscSsvkIB5CnZ9b/yVkH6ii4xcOlq+M7TD9wTbQ01Bs2HTx71T/6oHGAtwiou12Y03+NAKcyPdgAA0EqGUFGlDmTIrpDP3M6sq3ntD6r0e+uqtaUEgbKDFiHUyUOh7t1vDaYv7qT7649/+w7HBp/E8AY5f7DLpuHEJJlHsOVemM7Y65w4tmfBcCWAzLicAACTFi6ZP9fubVBQxrJa4ykpSssWL4l1Fk4dOz6kzKT+6vwl18c/hDqeuOnBlnZf74BOAJqrktJF7vr6Rxo3dwU3LzO0/YYtXH5LeS5tePij0+8EYx8WL7/lukwrSxq6mjKJR8HE4hW5l7ByKES4alfeVZGxiBc1dkU7PN+qckkP0EbfP3ow/hiB0QuqpOKulYmn7wIHj5/VOU45l5j1VyiAPCW3/lfeKkhb0cUHLl1LFl/vuq5iSfTXaYfMEneDXLUbPe7E+GDRVXE5NwoFEAdKYW60AwAoeupkiNofrXZfnX5B1UwhRp6Dl8fczuzC5XfekHGFqrUlB4HSgxaRPALT/fqR9BPh6Nxfd2DkQwxvkLMHu2wbTlQxMIIt9cJU5fzVrKxZYMte58yRPQuGKwEYQC4LACBq4d3l10y3eyMSVDP4lS++TnvuRqWS625Zvjj2YShwVjl1aOltDzXUKhfu3bJ6xdLFV31p2tL17T7fvr6QTUHopVDf73y+XVtXfXNFh8FJCw0cLFXaeIr4VKtCCNfXb7/hiszfOeuGO3UfzTDG+qMwheIVEkpYfDZ8bjj2c9mtC662OeT6ZPhcfIBSVN95k/5jZzGqC0prMt6o3EvMsVcogDxlsP5XjjOme+z18pm3vdGBy5qVNQumJ57ZSh0yU94NShriLLYqLudGoYjiQCmmFu0AAIpeeODVzTsSyRB6z2Bc7V79I8Vz8N1d/gtmb90kczuzt99TaeCprpKrF9xaFvvwwYkzHyf+JD9oEUIYe82Q3v31/I18iOENsmmwy7bhRCUDeyH1wgyfO/1OvBZaeMct5XqTfEySv9eWsbZnwXAlAAMcdNcSAICET0fOJ56U2LN68Z7VWa9i+NzwZ8rPJYvqn2/1dTcdSF6wd8vq3uhPlZ7Wlm/dsPjrD5owMUbU+MCh7v5hIYQIDx99efWW7kz/YA5lV+GKOV+5wsjeRvPQc+3GmHsUHFK8YqrdZtNcDg69Gf+wsGL+XIMR4Myy8sVC9EY/hM4NfyKEeY9uOOIKBZCncqn/Z91c11C1ZUufENHHXh9btCgxOhkeOj75nFh01DI6znXq1OSg+RM3Kl4zrng0NmUeY6q47BRqHCiFc6IdAECeCw++tTs+r0emyQaiz8F3Ty7ft2Xza6vdjYvMbzFN7szOmX2VkRXO/Mrcq2I/n3pz6K+XxfWT/2ZC0BLdsOq6WteWvpAQ6cJOkTKnTn4Gn8TwBjkn/LNvODE7Ui/M8NiFxNbfXV5mqBL68uxrjaVl2MvuU4vhSgBGkMsCAIi6fsn8r9q9DQqqvOypUY8vCCHE7Kq123suNN67SSs07+9oerxDCCEqPa0bVj30jWWLpETCl0J9b7x5ePePmzpy3SdJVF0Fq0k/Co4r3hQGR8essnDOVUZ7ViVXzp7jEsKqgrXrCgWQp2TV/8pbBW//of9846L4HMKJ96y7Vtx+w3QhxNzKe24X0QfUuo/1/8WzyBWr4ZWTk6d5rJkqLhsFFQdK4fxoBwCQdxJxjhAi69u9KdkDVpDemTV661R131TFlKBFCFEyy+3ZWLtjdTTLIk1pq18wlE/BJzG8QU4M/2wdTowzMIIt88K8fP7MYPaZOMoEOKdx4qklBMOVADSRJAYAKCYl85Y9vz/o97Z6KnWX6+9o+s69i29c+tODOU4UGA69+bNVVWXVDzzqtB5CnNFHCqJjNzLIOwp5ULwOdO1s453DK74yx8CMsfJYfoUCyFNS63/lFOWnDh59fzTxNaeP7D4ihFC8untmRc39sRmG/UdPxl8npzM5efx7qOLslp+HgGgHAGCK8Xdf6zySw/8f6XztXe3XOpjDyZ1Z6UoW1KysiX1Iec2Q3guG4mtwXORDDG9QfoR/1g8nImeOPrWcXMM7rzoFigq5LACAYjPDVVW/bufxscBh755Wj0tnyVDvptqqv/N9OMXoMzz+3s7vV9Ws7ejXXazS0/qK13vAHxz0eux4l+qbQ8HL1n9r7kchT4rXgc4Njxk9pcOfjlz4NPNicll2hQLIUybU/7Nurmuoiv4Y6nzDPxqrVhLvg0+8uruk/JbYq8/7OrvejQ2a601OrkAVZ7v8OgREOwAAk4RHj+3d1pvLrcxQ77a9x0atbSZt68xeHhu+IG1lRinTL0Ldu98aTOx7hhcMKTgn8iGGNyh/wj97hhMxZY4/tRiuBKDBSTPeAwCQ3pPe4Av1LrltVknpInf9Inf9I+t+Gerb/8f/8381Xgsaav/nZx6oerl+ftbpn+P+X/ywJSXLvdLT2vKt8plCiC/NV79G83Lf4Tcten3s9LLyu7N+V+3nwaGA7A3J4Sg4uHhTXBgeuywkn8A5OHVhLGwwnzn8yfAFmx7UMP8KBZCnTKn/FVOUh7q7/GuXLbtadZOg9o7Kr8Wq8enzb19RJbb0CSFC75w+FxazSjJOTp6EKi4r+RkHSpFP0Q4AIL8o31AzVYmoySrSO7OnBs+cv1yVOcy4PHbhfOxn18I5V2psguSgpaTirpW1ru40rxnK+IKh5DXZH/kQwxvk7PDPMcOJWcnxwrT4dTamcfapJQTDlQA0cR0BAEzz6cj5KedIz52/JJH8fWF4zMRU/xJX1bfqH2nc3BWJXAz6D3jbmt2qv/e3b+8ZzDqT+tKH3R2Kh5xc7uYOf/BiJHJi57rH6uvr6+vrv6XsIVjsytnXxvPHTw2eOW+keD8fOT9m3hZleRScXbxCiJIr5yyMF/FQ4KzVUy8niQ43TDJ6xEXSiMPCJfPnyt4wI8y5QgHkKZPqf+UU5X3et89cFkJxkyBpaHtOdV3t5LLdrx8Z/NzQ5ORaX0wVl1bex4FSOD7aAQDkrfCH//5K4g01Vc095yMGjR1tdce7un1bNr82YHIraXJn1mCY8fGZEx/Efr7i2tmKXBZTgxat1wwZecGQ1irtiXyI4Q1yfPjnvOHE9GRemCWl11UsiX8yOhuNstKQIZcxdiEce2oxXAnACMY9AACm+WT43JRzpKdfU353vNvx9h/6z+stLM0MV9U36xs3Hx470aZ8/+WJwbPjWcae4aGu7b74UzLu1r37NzdUuQTxesAAACAASURBVGbo/Uuu3ZJslJYtXhLvfSpfFawt/NHpd4JmblOcgaPg8OIVQpRcveDWstiH4DunPzJwAo33ba1buv5ln8+3r0/2a1WVww0Gj7gQInz2+MF451Dn4TPLyLtCAeQp8+p/xRTlpw4eHworG76yWxdcragAS2bdVL188udoiPLp+2//0cDk5Pqo4hTyPQ6UwvnRDgAgX30+2PWb9sQ9zdq66jlG/7X05ocaahIfJ3MCzGRuZ9bYkyej7x89GJ+f4PZ7KhW3Tc0NWmZW1H2ncXL3468ZMv6CIX0WRj7E8AY5P/xz9HCigtQLs+TaBbcmMniOHR8yUOmpKg0ZchljFw4+tRiuBGCA3dc4ACCPqKaa+ODEmY91lw6PnvQfnPqXza285/bYz6cOHvjfht4xGX6vva561dZdPt++QwOjSX8bH+j1+Xyvbl1VVtee4cmh0son/vEntfGPoQvDn2QZeibeziuEqGl46ObSjNt+7vQ7lk2PWHLdLcsXxz4oXxWsKTz41u6cZyCWdhQcXrxCCFF63eLy2M9Jb7bWED57/OCJ3i0/WLFixcPVVd9of09md0cxvmPwiIv/n737j46ivhf/P6RevmAj5fLV6gQ8WsCEeiIiG/Co8bJGWWytwtmIfvzBYgOtttDeIzdZbqz1eis/bth847GFtupNJEHw2rB7gpQq0YXtPWi/0E3Vwjm4KfLFgrve4kUg0SqSzfePZH/M/prZ3ffMzibPx18smZ19z8x7Zt7v17zm/U486BXzr52iZ8PR2DMUQJHS8fqf+KZm7BqY/IDn0qvnDS/7/uv7/3I2FkpLNzg5l7hsFXk7UIgiaO0AAIpT+Ni+l/dFP8lLbquaoL2rN2569e2xu6TU+XjrW1p6l7nTtzOrZYXK3BF5+pWXxT8AFt5oUSiZ/E8PRDd/OHNIywRDJmv50IbXyPzNv4KFE7Ml9MRUZPDEjZCUfkWKi0ZKhsbYTVy1CFcC0IBcFgCAZiVfnXjZhZEPn32k0hzLc/blsZddOT3aqA+1tb30tmpqdrj/7Ve3dPd0NDxQW7vo1oqa1Xs+jv7p7J7Hyytuqa2tXdzQETr4p0Ohc5nXVXLRxMtyL7wyXz4x0pHSx77WX6aYXVMvF141+4bY7t3ieePDzDvk9NuvePIunrijYPbdKyUG+Lo9r6hVYGVPLOH9ofzF59ZoOeJS4kGXb5h91YUZls6P4WcogCKl4/U/fojyo4ETZ/tPHDk49Gl+1TcTHvDEjfceeudYsDcaMbc/vGBq0tWbS1wOirwdKEQRtHYAAEVJ2fe0LFkwc0I2Xy+ZfuO9tth79KEtb/jP6vqsTt/OrPoKw8ff2Lozdk9OTP0R22hJFperMfQQPX6CIfmh1XeXF0Hjkza8RkXQ/CtIODEHQk9MxWxfGl5XU140UjM0xm7mqkW4EoA6clkAANqNn3jZxMi/Q91bXn077UB54f6eLU9t6EnzVy1KJlgda2LxkV0N9S/0ZB6Xr9//63qXL/rR+sA9c6NvYMR3biUp5PnFS+9kHEZWmfBum1t56QVZlV4hdOTYR5lbuuc+9Dz5YF67K1vK3RvatPKJ36Z/RyHc3/NCfcMuAT+qx1Ew4+6VpMQA366Wn+94L0MFDn/Q1fR0XE8sm0GeNRlXfvcjztjO37TysZcylUc6F9rzTNxBl62rFs3N4l29bBX0DAVQpIRf/2PvhPVs2b27e3fkjc55V1+auOi4qdfOHR60unvL+vVbhi7g8hL7bZOTw4Jc4nIwgtqBQpi1tQMAKELKJ5Q59D1Lpi542B6XzLJz6xvH9Uxm0bkzG/L8YtPv08/ye+7DLtfKtkORj3esume2MvVHbKMlWXwzJtT9+0O9ceM9pBlTx9wtH9rwGpm0+VeQcGIOxJ6YyqtQ99NrNx9KXxnOhXybfxG7aKRjZIxdyVxVi3AlAHXksgAAtFPkSks+V/06b6oO/7nQgV+tuHOVL89fS4iP+FZV3fnUnjTpz+HQm0+vWNbgiyaZV9b9yH5dadxtTjFoYcjX8NOnPO+la32GQ97/eGqz2pCtGcVGLpUkqfPxtS+nbYiHQwee/t6c2k3K/Pq+k2d0nnNauXtDbSsfbPT0piikoKM5RNRRMP/ulSSppPyedQ3WyKdQx0M1K7akLmc4dOCZJ+KiY7JtjcOatid26nTf+VzKM2H2PavuiH7KVB7pXGjP+vtvfdIX/Q/Z/qP7ZqkOQJoXg89QAEVK3+t/7LXX0JZfb9o3NK179b3VVyZfYy64anbtcEF8HR0+SZIyvdbMJS4HRd0OFKIoWjsAgKITP6pHjk8Bx06+zR57ticdanvWqz6pbj707cyGfOtX/eszb6aObu1Z/2DcHVauq7vvuomJS4lttCSLb8a8/uoLu34f+XLxND5pw2tUFM2/goQTcyC4NxF/FTrUsexf0lSGpEtQWsbG2I2oWoQrTdORBEaeQcBwmetkQ0NDoQsIjBpf+l3Toief1eXvU/3GQKDVpjhlZavzuS5/cGD4718E/TtbndFFrA6HNdP6VQswcMxdV6m8SFQ6XFvjfnFwIOjv6nQ5ZOVS1hZ/30DCuvr8LVbFQrLV+Zw7fl2DA32Bve5Wp2IxeYX7xBea92nU3wOti5VFcrYm/1ZyyYdNc7j/qlzhX92Oaen/mix++TQHN3n3Kgo50BfYm6aAD7uDX2a/TwbFHQXhu3fQoD0sO1ydewPRyjkQ9HdtdTkSjkJS7VWcKZLs2Hw4sXprK1LffpdVuUdkh6tzlz8Y3bdnAl533Ck8pLLOfSzV74ndYwafoQDMSqVtoMf1P84ZrzPhi9Nc/tR3vJNep0WxpNzoPZPy4jxo+kuchjua+mKG3EaLph2oWaYKr3drJ6fGOQCguA2c8TbG3TcWtwb+ntN6EhpCd7j8nygX+DLofjjuppSuQaWZwM6ssn+tWJ83ELnzJUS31FoFIhstKdauPGrR8pis8Wl0qybOyGnDF22wS+9wYm5tVLEnZuJVSLY6n3Nnvmhk3O3Gxtj1OQc1hSs11CjClQAyIpcFBZDyfhhFLgtgnFx6Ap/4XXdkOIUVDTvXvkAseJFruDy5Oav+y+kai9oLH5WuWaxu4IS7LsuCx2+DrfWw8nd16HwODg4Euxu17V65rrVzvTXyKedclkFRR0H07h3UaQ9nX4GTg4AJvxUvoZyqRRroO7w5Tb807b6yNnYHU58DwveYoWcoAJNSaxvocP2Plxjdlp3eM2kKqnhCk2nJIQIvccqbQv4Ph0ybyzJo2naguEOQscLr3NohlwUARp+EJ7u21kCOvanE7IqkhpDoXBaBnVnFza6u1R2LdGRka/SeyLS3RDZaNK3cwManNka3auIVRRtek2IJdhkdTsy5jSryxEzOadAo3W43NMauzzmoJVyppUYRrgSQCSMcAQCyMtHyyH+0JgwjkYJsbWzftur6i/L/wdK5q7Z1tqj/YoS10e1bZ08x060kSRMtq571NqbMkU+p0tH60jP2K3K7WZZM/s6aF5+0al3c5ty81/tcNEc+dDAQzDgBpxgl8vw129rV+5/WJ19cs+gbYqb4FHMUimL3SlKWFVh2tOx/fpUlachi6dLrF96e6iB99tHpT7MZy7mkdMaSTT63U2tPvtLR0rltzXzZoAajoWcogCKl8/U/NkS5JEmZhhyXLri0cq5N05JDuMTlpGjbgUIUTWsHAFAkwke8z8ZPbpv79AclE6pui5tmSAptecN/Vtd5hnTqzP7DxBu+/6tW1Se8tkZv25qayZlWJrLRkrzymXctqVb+V5E1PmnDa1Qszb9ChBNzIvLELCm1rNjm1XJ0ZGvjRtf9KZM84hkaY9enahGuTKnwHUlghOFsAgBkqbSybpO7O92gg5IkSTane8/OdcIalCXyTY++0O3frBypL4VKh+u1wM419vL0/dCSyTXrOgPdLeq53lbnZn/3C3WVeUy6OVauadzm71BriFc6XG5/cGfTUqt13u3RdrH+kaBhJfL8dTt9GQ6o7Ni4f1tjjawtwqLpJ4UcheLYvZLWCixbnR3+nucfnZvyvBk7eVHjiyl6TTnEKUpKy+1NGY/4cIEcLd2BN9sfvcmonqEkSQafoQCKlK7X/4QHM1MrpqS9zJRMv/HeaMxcti2omqRScC5xOSnadqAQRdPaAQAUg8+P7HutO/ax+t7qK3Pv7U24cdmauBkrQpubtvfqfNfRqTM7YUbd0z53ozXT2jrXZU5kGSqfwEZLonHTq29XRASKr/FJG16jomn+FSCcmBOhJ+ZYueanOwOvJc4VrjAUFb9rqpYMHkNj7HpULcKVSczSkQRGFHJZAADZKy2fX99+Iujvcm9VNN+tzlb33kDfq032GYJbbCWyZWnT3oGgv8udOP2kVOlwbXW79wb63m2vX1Beqnprm1A+/9Gh0qeYB9TqbHW7u/zBgb1NSy35t4rHypYlTd4ef1fyT9mcrZ2RYtst8lgpoUcd6t7tP5Xv72sUO6DPxXVphmd+7W1fMVd8z1PIUSiS3SspKnBiUa3OVneXN/De3qYllgz7eajXlDg7bOX8iq/nUktLy+fXtwf7At7k3T9cnjPB9kfnZxFZE8jIMxRAkdLz+j9h5oIlkSHKbbdXTx+XdsmSKdfOrxj6p7zktqoJWq5JXOJyUqztQCGKp7UDADC58LF9L++LfczczlGXkF0R6n75rSMGPEPXpTM7ody+zhv0dynWaHO2drq9gd6s1iay0aJccfwtvlgbn7ThNSqe5l8Bwok5EXlilpSWL6hv7wn6dyWsSna4OlNHxSdeNnF82vUZGmPXoWoRrowVz2wdSWDkGDM4OFjoMmDUGTNmTIa/NjQ0bNiwwbDCAAAAAECxCZ/d8/iMW9eHprn879VbCjhw9+gl4hCc72meUdXw/tAHq8u/s97CK3wAACDJ2T2rZ9y6ITT0weL0vtZUc3FhSwQgveOepbfUdgy18mnkA0BeyA8DAAAAAKC4hD89fSpU6EKMbhwCAABgjPBZ/xtbos0O2w+XWUlkAUzs/N+Ovvl+5EPF1LJ8BsECgNGOt7cAAAAAACgupw/v90uSJN00tYxufWFwCAAAgDFO+Xd3R1JZZNu9N07nDWXAOP09zbX1J+/+0fX/91euuOFODTPIhI+++3o0lWXa9CsuobcAALnjGgoAAAAAQDEJf/jfW7f0SJJsm3f1pYUuzOgk5hB8dubkZwILBQAARqBIq0OSJEmS7Q8vmEoqC2CgcWVTv+Fr+L5PkiRJkp3e95pqJmRYPPxBV9PT3ZFPcu3sq3gMCwB5oNkDAAAAAECxCPf3ehofXNkWkiT5odV3l9OrN5ywQxD+6Ng7sWmKGH4cAAAkO/32S21t0VFZlthvmzy2oOUBRpsLLrli+rTIh9CGlud6TqddNhw68MwTK9sORT5bliyYmSnxBQCghqgXAAAAAADF4mzglbYNvpAk2RpfXGGdQKfeeHkdgvDfQn8LS5IkhUNvPrM29sqmZJtbeSnvbAIAMMqF+0N/649+CB1oX31vVcOuyH8sXrPsRp6LAwa7oOKfHrHKkU+7GqpuXtq8zbOntz9+of7ePZ5tzd+1Xb+qI5qsLtc99mPrxUYWFQBGHgIlAAAAAAAUi4mWR55oOWmrXPbQ/HKeZRREPofg/P/s+/ey2meT/yDPuvIyEpMAABjtwn1/eCJlU0GSZKur/p5yRnEDDFc6674f2Vt8myJJKoc6Gh7oUP2W9ckX13xnMi18AMgPuSwAAAAAABSP0rmPNs0tdCFGt9wPwenD+/0p/ltesfHH1aQmAQAw6g3PZvJ+8l+sDc2PVJUaXyIA0tjJ9nV7Wv9es6wtpL6wJEmSZH3Su62xRmZGMADIFzmBAAAAAAAA+gt/fOydYNL/2hpfbFw0mUg3AACQLiibelPSf8qOjfu3rbCU8jQHKJQJM+p+1eN3uxyVakvanK3ewM6fksgCAELQ+gEAAAAAANDfZyWXz58vxz7bnK1ub6BzXQ2jjwMAAEmSJOnSG3682WmNfJIdrs4uf88LK+byXBwosLGyxV7f/m5fYK/bvTUpqcXmbO10d/mDA7ub6mrKyTwDAEHGDA4OFroMGHXGjBmT4a8NDQ0bNmwwrDAAAAAAAAAAAAAAAMA8yA0EAAAAAAAAAAAAAACAWZDLAgAAAAAAAAAAAAAAALMglwUAAAAAAAAAAAAAAABmQS4LAAAAAAAAAAAAAAAAzIJcFgAAAAAAAAAAAAAAAJgFuSwAAAAAAAAAAAAAAAAwC3JZAAAAAAAAAAAAAAAAYBbksgAAAAAAAAAAAAAAAMAsyGUBAAAAAAAAAAAAAACAWZDLAgAAAAAAAAAAAAAAALMglwUAAAAAAAAAAAAAAABmQS4LAAAAAAAAAAAAAAAAzIJcFgAAAAAAAAAAAAAAAJgFuSwAAAAAAAAAAAAAAAAwC3JZAAAAAAAAAAAAAAAAYBbksgAAAAAAAAAAAAAAAMAsyGUBAAAAAAAAAAAAAACAWZDLAgAAAAAAAAAAAAAAALMglwUAAAAAAAAAAAAAAABmQS4LAAAAAAAAAAAAAAAAzIJcFgAAAAAAAAAAAAAAAJgFuSwAAAAAAAAAAAAAAAAwC3JZAAAAAAAAAAAAAAAAYBbksgAAAAAAAAAAAAAAAMAsyGUBAAAAAAAAAAAAAACAWZDLAgAAAAAAAAAAAAAAALMglwUAAAAAAAAAAAAAAABmQS4LAAAAAAAAAAAAAAAAzIJcFgAAAAAAAAAAAAAAAJgFuSwAAAAAAAAAAAAAAAAwC3JZAAAAAAAAAAAAAAAAYBbksgAAAAAAAAAAAAAAAMAsyGUBAAAAAAAAAAAAAACAWZDLAgAAAAAAAAAAAAAAALMglwUAAAAAAAAAAAAAAABmQS4LAAAAAAAAAAAAAAAAzIJcFgAAAAAAAAAAAAAAAJgFuSwAAAAAAAAAAAAAAAAwC3JZAAAAAAAAAAAAAAAAYBbksgAAAAAAAAAAAAAAAMAsyGUBAAAAAAAAAAAAAACAWZDLAgAAAAAAAAAAAAAAALMglwUAAAAAAAAAAAAAAABmQS4LAAAAAAAAAAAAAAAAzIJcFgAAAAAAAAAAAAAAAJgFuSwAAAAAAAAAAAAAAAAwC3JZAAAAAAAAAAAAAAAAYBbksgAAAAAAAAAAAAAAAMAsyGUBAAAAAAAAAAAAAACAWZDLAgAAAAAAAAAAAAAAALMglwUAAAAAAAAAAAAAAABmQS4LAAAAAAAAAAAAAAAAzIJcFgAAAAAAAAAAAAAAAJgFuSwAAAAAAAAAAAAAAAAwC3JZAAAAAAAAAAAAAAAAYBbksgAAAAAAAAAAAAAAAMAsyGUBAAAAAAAAAAAAAACAWZDLAgAAAAAAAAAAAAAAALMglwUAAAAAAAAAAAAAAABmQS4LAAAAAAAAAAAAAAAAzIJcFgAAAAAAAAAAAAAAAJgFuSwAAAAAAAAAAAAAAAAwC3JZAAAAAAAAAAAAAAAAYBbksgAAAAAAAAAAAAAAAMAsyGUBAAAAAAAAAAAAAACAWZDLAgAAAAAAAAAAAAAAALMglwUAAAAAAAAAAAAAAABmQS4LAAAAAAAAAAAAAAAAzIJcFgAAAAAAAAAAAAAAAJgFuSwAAAAAAAAAAAAAAAAwC3JZAAAAAAAAAAAAAAAAYBbksgAAAAAAAAAAAAAAAMAsyGUBAAAAAAAAAAAAAACAWZDLAgAAAAAAAAAAAAAAALMglwUAAAAAAAAAAAAAAABmQS4LAAAAAAAAAAAAAAAAzIJcFgAAAAAAAAAAAAAAAJgFuSwAAAAAAAAAAAAAAAAwC3JZAAAAAAAAAAAAAAAAYBbksgAAAAAAAAAAAAAAAMAsyGUBAAAAAAAAAAAAAACAWZDLAgAAAAAAAAAAAAAAALMglwUAAAAAAAAAAAAAAABmQS4LAAAAAAAAAAAAAAAAzIJcFgAAAAAAAAAAAAAAAJgFuSwAAAAAAAAAAAAAAAAwC3JZAAAAAAAAAAAAAAAAYBbksgAAAAAAAAAAAAAAAMAsyGUBAAAAAAAAAAAAAACAWZDLAgAAAAAAAAAAAAAAALMglwUAAAAAAAAAAAAAAABmQS4LAAAAAAAAAAAAAAAAzIJcFgAAAAAAAAAAAAAAAJgFuSwAAAAAAAAAAAAAAAAwC3JZAAAAAAAAAAAAAAAAYBbksgAAAAAAAAAAAAAAAMAsyGUBAOjnuGfp9DHDpi/1HC90eQx3vqc5ugPGPOIJnS90gXRj8i0VUjzFSm5p7ulX/jm+tif/tYCyKti5UM//29sfNqhoAAAAAPJ1PuR5ZEw2ypY2b/d4PHt6zdNpyYvJe6OZmLYXaRrh99oWlMXV3ZWeD8/ltCLlaTK9uSf3amK2o2a28iA9lbCSCVZocuFQz47tbasXJN7SPD0h4lhFRqdYZYYlR8OlctQ/iMHIRS4LAAAApHDoQPvqO8uqNh/qIwYAAAAAjFihjobFtbW1t1ZcVLa0+fWRktGCESh81tfxeHco9h8hz7O7j9JfBUahcOj1x261VC1avGxDt+IPoY6GHR/wnHOk0h6rJKoJjGBc4wEAAEa3cKinffWtZdc/lBARAAAAADCChToabNY7H3udN9phSqf8u+MzWSRJCnW//NYRqisw2vQfaLl/6XpfKOUfp11zxSUGlwcG0B6rJKoJjHTksgAAAIxu4eDepzb4Cl0KAAAAAIYL+dYvvb/lAKOzwHTO/nn3lp7E/+z+Zavv40KUBkChhM8e6GpJk8giSdNumvr1CwwtDwyhPVZJVBMY6chlAQAAyN/l9vYjg8P21ltKC12eKNMWDMCIoJij/RFP6HzeK0k5d7VpJ7fWXjDtk3wbwLT7c8RgD488pjqFldQvoRhlprn8Xw5mdCbg7XK3Oq2x74R8DT/7dc/pgpV59KKzlsG5D9/wbBl+eF15v2ORPPz/PVt2//lswUolme+oma08gHDKIZpkh8vtDw5Eb2pH2u2XF7BwyInxFy4ulUARI5cFAAAAAAC9MHU3UNQ4hTHiTCivWWiva/IGuxutkfQAaVfDY55e6jjMI3x097Oe4afX8p3L1vxwSaS2hrZ43vjwXMEKBsBofz/9UTTbUrataVxlt8g82ASAUYNLPgAAAAAAOmDqbqCocQpjRCuRb32sucEa/dz92r4jnxeuOIBC+MhbL0cHYrhm+pTJ1y1YYhn+GPI8u/somVfAqFQ268qLeagJAKMKE8lBvFOnTvl8vpy/3tvb6/F4Mixgt9tzXjkAAAAAGISpu4GiximMEa6k9LpvLbG5fMMZA/te3nfsofIZPCOECXx+ZN9rkSxC2XbvjdNLJl22wCZv6AlJkiSFure8+vY95ZZSaisw2lx0ydfGFboMAABD0eCDeOPHj6/NKPPXd+zYkeG7r776qjFbAUBPZ3v37NjevLQsOrH7mGuWNm/z7OnNcnb3cH+vz+PZ3rZ6wZgEt6xu83h29ISyf1Mn3N/ri5WtbGnzdl9vf7rVnAv1/M7TtvqW2A8vWN3m2dOb69zN4VDPDo9yhUPbsiPrdYZDPTs8yp2cz24RvaXS2d49O8RsqS7FM1A41LNjW/PSa5T7IXWtC4d6dnieX31L9KiW3bL6+dyOaN5lTqpdQ2dxmpIDQH5MO7m1aQuWWZEWu4iwhwEUiZIrq++tjnwIvX/qU1M25fXo7onrkBraOToX6vmd8sfKbln9fPaxFKUUm7Bgddv2fFebj7NvtT7eGflQfW/1lSVSyYSq26LTDEm+rb85cKowZcuKmfvOAo+7mFiWHiG+aOmSAim57H/DSnjN0uZt+oR6xMbitBEcn9SleAL2iU6bWaiYYTHGKnUi7FKpzx1ckxHzICb6Jern6DMI6GDt2rU61djdu3cXeuMAaPdXt2Na5PSd5nD/dXBwoC/gdsbm5E5ibXQHzmhY85mAtzXTemIrdG72BwfSreZLvytaQIc7ODjQ52+xJq9Erms9nFCqgb7Aay5HZZpfrXS07AsOKNcvPewOfpl2gwaC/s3OFD8dXwrHxv3BL9T3jaZVtXRr2s86bOngmUB3iyPToZOtzg6/li3VpXjpKVZidfn7lH+Or+3Jf1UuMM3l/3JwcPCL4P6NaXeF7GjZH191zwTcjdZ0+yzTiZO+YEG3I8NxSLkhGmqXJFU6XG7NRxBAkTPiAlvMhOwfAIVSFKfwCL6EQqsvg+6HY43x4b5GLt9N8dWBw622aI/F4vSe1LTW3L6VakU6dPfEdUgFdI5Ue5Ead0WsSxi3M4aCMGl+Lton9XeICNGINXDG2xgrk601MNwx/sTvuiN2nJzebEqWz2mSQNtRM67vnFMsQtRxFxPLEhTiS7ml3ifTF8/mdB/WdtfUIQg5fLAyhnpkh6s7kKmEWbQBxMbitMkrPqk8Z9PKo3kmKhIrJgxrgphh0cUqs4tqFvBSKeQOrsXIexBTqPoJ0yGXBbrYvXu3+nUtJ8ePHy/0xgHQLqEJdSRjHzLK1ug9kb5rOjg4cMLbaMvmylFZ5z6WeoXKJtSJE+661I2hxa2Bv8d9LXNneJjs2Hz4k/1aYmoDwX0tGVq0ipUmZ9UkrKq7UUvLUpIkqdLRelCt0y54Swf7Drud2o5dYuvTkOJlJjiX5czh1jTVLeYOl/+TwUFtdV5e4T6RsoMtLpel72CrxooqqddVACMEuSyZFcWDcADpFMUpPIIvodAqn4f0fw+0Ls74VcUCGrMHBgKtsd6L3Og9o9axS02H7p7ADqmYzpHGXJaMT25ibI3eE19kkcuiaQ+rh2jEO+l1WqL7ztZ6OPrriqqVGKjJzNhcFkP7ztnGIoQddzGxLIEhvqQtPeFeoXbeyFanO9CXcUN1CkJKVpf/hIa4kGxt7E57MdLYBhAci9Mk7/ikvrksoiKx4sKwJogZFl2sUt9cFlGXSlF3cC1G3oOYQtVPmA5zDEEXVVVVeqx2k6kTBQAAIABJREFU4cKFU6ZM0WPNAPT32V93PHH/rU/61JfsXv/g+q4Pz6X+Y/jDPY/X3bq+O/VfUzvUVtuwufdzlaXOH3zpiX9rC6X6k+326unR2VjD/T2btGxIqKPxB5v2nFRbLBx6/fH7F6/qOKS24NBK25bVPOZJs3PCH3q+Z7Gt96XchmSHOpbdt6LtUPrhBAVvqdR/qG3F4toN2o5dqGPV9Q89vufD9AP+iS6eoT79wPv0D5alrm5xdjXUv9DTf6qn5WH1Oh/atPKJ32bYX3k73fPrf12msaJKkhRqW/aD1h7mGwIAAADMq/9E4Gjk35ba2VdckLjAuOnVt0cfBYS2vOE/q9rC/9jX+sto70VeclvVhBzizzp090R2SI3sHIX7ezbdf/3KDvWOfvf6B59o+5PGgMD5M35Ne1glRKODcO8rTRt6Ip+GJhgaVjL9xntjQ/7se3nfMVN2OM3cdxZ23MXEsvQL8Unnz/g3PVi7Se18CPk2rHx4nTftZBM6lvDjt3++skY9LhTyrV96/+Ov5z4dhuBYnCai45OCiYrE6raZhYoZFmOsUieiLpU63cG1GDEPYqKon6NboZNpMGItX75ceHVtb28v9GYByEp8OnAc2eHq3Bt776Ev4G1NGIxR8eZNnOSh5yodrq1disHrBvoCe92droSU59QvkCneYIiwOlu9Q2N4fhH0u10Oi6IwZ7zOhDU7XJ3e6JifXwT9O1tTv+6Q6l2BgWPuusqk1e2Kjeo5EPR3PZcwiJ9c506RL92332VNLlncqgbPBLzupLKlz5UWu6XK0YAlSZIkm7N1Z1zxvgj6dyUetwz50YKLp4HAcVnSFjt5zMZKh3OFNfpv1395o+MfpjhxUr6XJmZcFuULcJJkdba698a/vTQQ9HdpPZEBYw0E/V3Jt4VKh2urW1mNs1xl/MVZtjqfy3VtZwLeLnfC6WN1trq7vDmOdyp2hV8E/buUa7M5W92xVZl5XJaBoL9rq2Is36Rrl2LZ5GPalWkQc01y2D861FhhUpTN5mztdHszDn6e9y+KqM8DfYG9sZIntEULS/ARF35J0UDUYUq5K6zOVnd+52I+9Tb3S1yWVS6fbWdcFuQx4ISmAVSynTBI0VPLdYIh8d09kR1ScZ0jLSN8JPb0Ne+KjOOyyDdab4ysVln+QvfsFEMBxU0wlOqvWYz6Y9y4LIb3nbOJRYg67mJiWaJDfOm2VHEXVt6gI7+bJjJmSBAynxKqtwFEx+K0EBOf1G1cFlGRWMFhWDPEDAv1u+Ybl0XUpVLkHVyLkfcgplD1E6ZDLgv00t7enuoqnJe333670JsFICvJDY50o3cO9B3erGj0JAYshpaKD6KpDvOoHC4yZagiuQllfdKbMBNq37FA7H8SOoHpxodMOQ5hcv9KOQn00PDIKWPciePTJgcEE3unaVeVXLbUfVTBW5rQ9k0/42ziRJ6pE3cEF08b8bksqYs9kHKIxdS1PWHH5hoeVX9OoxxaPG3t0nYiA4YRMEu9fnOuD+owZ7nYFWae1LnS0bIvOGBYLkvmS1kBJ1FOU7BsJ3EbIqDGalTAycI1y2cC+2zn4c42FSDfGpvFNmZzxIVfUjTI5zBlvZ6WbpU6Jq7eZnsKZz31u9BtJ5cFuT6kV86MIFtd+9PUnsy5BYlrVXRyc5xgSHh3T2yHVGDnSPV+oXFXpGzMZMxliR731NOXFK5npwj7pHjoq8wU0Z4sZVgui/F95xxiEXked0GxLOEhvlRPPdOc7ElbmvJiZUAQMu3eSyqhtcWfvJhKG0B4LE4L4fHJQc3TsWkhKhIrfDPNEDMs1O/mHKvMaknjL5Vi7+BajLwHMYWqnzAdclmgl+PHjyddPPIyZ86cQm8TgGwlNjgy9oWUAbJUTU9lzEK9qaFcPlVbNrEJpRYHUbwcliHYNzg4mDwvb1IBEl4FSNkvjS6rbJMlZjcr31pT63Mm9KNS7Uk9tzR9By/lxqbKjxZbPI1E57KkP0wJ54IUm+wzhfipxFOmvQvpH8avJPNpotzhOQavARHEzFKvz5zrgzrMWS54hZomh5Ydmw9/st98uSxGTqIsLpdFTI3VqFCThWuV7wT22c7DXZBcFrFHXPglRYN8D1NsPfFP0zNLFwIeIq7e5pfLomHqd6HbTi4Lsn5In3IAp0y9UWX/K/MLrKrdEw2Ed/cEd0gFdo7U7hdZ7Irk9o96LkueIRodKJ8wpe5OKuuY1ifuhuWyGN93FhiLGNR03AXFssSH+JKfemYqW8LjTNXEKT2CkFmVMFV1ytwGEB6L00J4fHJwUGQui6hIrPjNNEPMsFC/a9JclnwvlYLv4FqMuAcxBaufMJ0c5isFNJkyZcqcOXMErnDRokUC1wagEBavWf3tyWnvPIp5uCUpcDSYMLHi+f85dCA2z6H80Oq7yzPfxkqmXjs/1uA5dbrvvEoBZduCqknp/xw+639jS3RWRmtD8yNVpWkXHjt5UcPGukzx/fAR77Nt0clZ73A1f9dSmnaDSiZ/e/WaWJtMOU/5573bf70hWjB5xcaffSf9fpYkaaxc88/NsdzwUPfjHT7FrOd6bqm8YuO6+2ak31IpcWM7H29966xyfWKLVyAZTodxU6+dq2iq2+x3XTcxzXomfvP6quiH0EenPxVWwjjn/3b0zfe1LTp28g0182WHq9Ptdu8N9K6pmUBrEwWhxyz1guZcl3SYs1zwCsP9PZomhw51NP5g056Tmn7VMEU6ibIeNVYUUZOFayVqAvthWc/DbQyhR1z4JUUDUYcp/KHnexbbep/GueEPdSy7b0XboX71JY2ut3G/rLXK6bbtGPXeb6j6hzEZfa3i1kW1yzb4ol+R61p/tSxDb1SaMHPBkmjQf9/L+46lvYac/fPuLT2RD5YlC2ZOyHoDxHf3BHdIjescZbUrJsx46Ikse755hmj0cOrAb7b6Ih/kJbdVpdhjk6oWxN4TD7X91+4jBhRMsyLoO+d73AXFsvQP8amUraTU8t24yNihtme9RxSXNrOVsGfL7j+fTbNoSqJjcVoIj08KJqj2GrCZhYoZFlWsUl95Xir1voNrUewPYpJRP0cvni5ARw6H2rtM2Zg3b57AtQEoALUHBsoWT9/JMwlNqAtk+y/7Anvd7s5Wpy1NUCPhG1+felN0jSdPqTah5ld9M9M6T/l3d0daobJtybeuy9gJlEouv+2BO9O/aKlsEWZqYA0ZN736dttQmKPLHzyxLi7McfLQ7/8UXU5eYr9t8tiMq5IkaeJ1d9ljDdbQH/70l8/i/ip2Sz8/su+16JZqK9648rsfiWavh9x/+ovi0IktXoFkPB0uKJt6U+yTbLv3xulpN/GCS66YHjtvPjrdp0evX3Eq9Wxp3/Vehgeosr092F5/t91ut5ZnPjSAbsK9nscadsU+a5ml3ud67De9mU6gz3Y+9S8u39C/VWY0D/nWr3q864M0a0t+hq06Z3nmR62iV3jW99Sdq3xx/5F+UueQ7/HGDRrD9cZ4/8na24efXqtMouxr+/lTT9QP1xOVSZT1fkyiS40VJb7mq04WHvI8u/toXqUKf9D1+CrF033VCewzJxtte7JhKGCdPA93pturvoQeceGXFC0bIOgw9R9oeXClIu0jYT2DZwJet3IW+UMdy+77Z0+6C2yEwfU2nsYqp9+2A9mRrc4Of8+v6mZkzjmJzx4Idb/81pHUFTHhyc0D98zN6hHFEOHdPdEdUuM6Rzr3fLML0RhCUy5UyYSq25bEtjNjcpXxzN93zve4i4pl6R7ik52rvm/JXDZlZOzgkROKg2W6EirfalMlPBanhfD4pFiiaq/+m1momGFxxSp1le+l0gSx66J/EJOE+jmK8YABOqqurha4ttmzZwtcGwDjTZt/7dTMtx1Fi+f0R6f/nrRESWm51W6/u65pd7CpJvsXvNRKeM0Vl2T4c/jjY+8EIx8q5l87Re0mmhBhSXD68H5/5N+ZG1iR1ZXX7R4Kcyy0yPELn/3L/tejDxK1vvpWMv3Ge2PvMgVef/dErNkmeEvj+3ia38wrLau4JrLG9w+8ezSuPS24eIWhcjp8deJlsQJr2caI90/p0/7+xyuuuTz6IdTx0DfvbGzz+HoNGhIAyJYybOfYfHjn+jplgLhEtiysW79TMUNwhgczQ39/y/dWSBqaZbzn+fq7oyssKS233l3/Qq9ibYfaVj6b6l2rcH/PC/Vxz7Blx8b9wZ1Ndd+xyNH411jZ8u2761/ojZ+zPO3TeuErPN3zXEvsJTOp0tF6sLe9/u6a8shbRGNly3fqmnZqm7yjUJKLPaG8pm79ixvjhu491LFhk08amv3kzfb6e2vKIzeo0vKauvU7FWN6Z/mYRLa3Dz2MTjvc8d56S/S9LH1qrCjDNV+2NnYHvU3xBdOhVOGzvmdXxt6VlGTH5sO9L9Tf/e1YfS6RLQu/17SzO36CnlDbup/7Ps60YuuT3m1P1Q3Xh7GyxV6/yf2Lb6m0THUj8IgLvwJoIeowne759c8a4hJiUqxHmlBeY0+64BxqW+lSycXJs95mdwqnolLl9Nx2IDsh34FD7x44mGk8OUlK7FJ1v7YvdYpnlk9uUhLf3RPdITWsc6RI7Ki+t/pKsT1f9RCN0ZS5ULYfLrNenHpBxUBBug/kkCWz953zPu7iYln6hvg0neyKyFioe7f/VMLf9SzhtPnXX5VlCbPK8xB+6dNAeHxSMEG1V//NLFTMsNhilTrK91Kp8x1ci6J/EJO8PPVzFDNXixUjTEVFhahVLV++fPz48aLWBqAQpt009esXGPRb50I9v/N4tjUv/XZth/a3xdVK2B8MHIzG5W6YfdWF6quccNX1aZO0/376o9ORf5fNuvLi3G/Jn57+KBYJr7r+m5nfKogomXLt/OhVOnQwEIyNWC52SxVD7E6tmJIx7p+6eEcDJ+LGUxd8IApCrbJd+LVLYptVMbWsUDMgRJVed9eS+PeCJd+GZbW3VFz0lTG3rG7zeHb0qEW/AUMpw3ZL70gzlnJJ6Yz71m2Mm6U+8VW8VKwtOzf9YK6c/N5VSemMJZt2xmU/hHZufeN40nPn492/bPNFPsmOzXtSr02SpJLScnt87kXqoUGEr/Dsn37TEn0uLltdrZvqKlNduMfKNY0vuleYLFFwmFz37+seSlHshGGiJUmSpDtcO/+fVC+jl5Ralvw0Nomy8kYpmJ41VhC5buOLa+bLKcpVUjrj3p/E79XuA4f+J+tXOIeFj7+xdWesUWNt2blpSeq9UVr50Lp/j8tMyjzousX50xU1CedF6RXlqc8UA4g74sKvAFqIOkyKq40k17n/+MLSNLtirFzTuC0+pUPbSCoG1dsU1Kqc/tsOZMG3YdmiqrJbH/P0ZpxNQss0Q1k/uUlFeHdPeIfUqM5R+KNj70SvttPmXjtVQ68wi56vhhCN4imX/sK925s2x3KhMj1mvti67IdxIxwkpyAUkMn7zvkfd3GxLK1yC/FpC44pzvSUj3L1K+HseZUaHuOWXHzlrLLIh+MHP/hE6+rFX/o0EB6fFExQ7dV9MwsVMyy6WKV+8r1U6nwH16L4H8Rku/woqp+jEbks0NH48ePXrl0rZFWLFy9WXwjA6NTfu8czZHvb6gVjxowZM+b/Kqu6o7b2gQbFcOt5i++rXDjpaxdquYcq3shRyK1XmXJNwaNvRj9Mm37FJRrbgePKpsYyDhVzQ4rd0s/OnIy9N9K5rGJ85snbIy6qavBFvqUMKIgtHrQpKbevi03uG8e3YVlt7aKqsq+MuWZp87aCR+YASdJ1lvo851wXP2e56BVmNanz2MmLGnSY1Dl/xTaJso41VpQ8Z9rWStAE9kmynodbZ+KOuPBLihaCDtPnvdt/HRsCSl6x8WffSV/HJEkaK9f8c9wFVstb+AbV2xRUqpwB247RbZrLHx1CKI2BoL/L7VbMPuZbX2t9tO29DFcF1WmGEgbVUBnZPi3h3T3hHVKjOkfhvlOxu8VNU8s0dfTHT7xM21NV8wkfeevl6LA+arlQyhEOejY0vWLIzIuajPC+s7hYVmqiQnxag2MXXDQpmlDy/ptH/6ae2SosCDlp4kVaSjjua5dclF0Jh+hw6VMlPj4plqDaa/bNhAmM5Du4YQ9igDjkskBfVVVV6gtpcPXVVwtZD4CR4lyo53fbm5eWjRkz5qKKW2uHLF62oVv9q2ldfs0V/5jhz4q+imAae7AaTJt0kdZ7e8lXJ05K+T694C1VvK+QG0V3Xc8DYRiVymZKEy2rnvU22tIvcKij4YHaRVVlX7lmafPLezK/1gnoSrdZ6vOec134nOXCV2iCSZ3zV3STKOtWY4XJd7JwjURNYJ8k23m49SbsiAu/Amgh6jDFD06jsfDKC6zqMPsG1dtUVKqc/tsOqCmRLQvt9qTZx9qW/aC1J+0VqWSC1bEmmj2QYpohZStCw9wNKYnv7onukEqSZEjn6PzJD45of7s5Iv6xd3FR3NTUH71/5ZvLuuOOa9p5rwpi9PSdRcWy9AjxaXTBRRO1ZDzrUEKtORCKBIgs6HLpy4aI+KSeBNVeXTazUDHDYoxVmtPIu4MX4EFM3stjRDFTQAcjkZBcloULF06ZMiX/9QAYGcKhN59eaimrumNxQ0e+/bKcaU2pzrXPmbPLJmrvRF34tUnq4zWbdkuHmLx4I0zJ5Jp1O4N+t8uReRiGQx0N/+fWihm3PPZ6Ub5nhpFAp1nq859zXfic5aJXGP742DvByActEwyLn9Q5f0U4ibJONVaYfCcL10rQBPZJsp2HW3+ijrjwS4oWgg7T2b/sfz0a4NVaeOVb+IHX3z2RYZcZVW9T/XTmKqf/tgOalZSWL/ppc4M1+h8+12O/ST+8RcmV1fdWRz4kTTMkZIKhBGbu7tE5Eqv/z69s2ZfH9/dteeXPus1Lkj2qh2Z6hfi0Xj3UmSIIWYyExyfNaZRsJkYxroEwA3JZoK9JkyYtX748z5XY7XYhhQFQ/ML977V/11K9SmXMukqHa6vbvcsfPOJ26PO65ZtHg7m/m6CnLN4aD3925pT6G52m3dIhJi/eCDRWttjr29/tC+xVDkqeLORbb7N8z/PhaA3JoaB0mqU+7znXhc9ZLnyF/cHAwegL1TfMvkpDqE3wpM75K8ZJlHWqsaLkO1m4ZoImsE9k5EzhGgk64sKvAJoIOkyKd4W1XV2lhMKHDgaC6UtvWL3N/qd133YgKyWlliU/dVoiHzNPYhU/OVfCNEOCJhhKYPbuHp0jUcJnD3S1+PJ5RhbytXQdMNf8a1QPVXqG+MQMqWiGIOT5vtOn1JcyIeHxSXMaJZuJUcoM10BAkiTJZCEdjEQ333zzf/7nf+azhpkzZ4oqDIDi1u//9Q8ak3KAKx2uxoVTx0mS9JUrbrgz/r3c8z1738x+SL80hmYlyHJ1nwePBjQsdup033lJFnFTHnprXNNThfCnp0+ljBXpuaUPu4Mb7fltqZ7Fg0YlpeVWe7nVfnf9C6GenX/4//53//MpB5YMtf3bE3dYnrdfQfY0DFZSbl/n8nQ37Er8g2/DMt/QvyodrsaFV1UobxsZ5T7n+uXDX0ucs7xzmbZfjqNMjhG+wvgnrBdO+tqFWnbN0AgTwu62o5MuNbboCJrAviiIOeLCrwBaCDpMijlEtF5dpcjQC76hD6GPTn8qSZpGozGT0bztMKuJ37y+SpIiQ6qEjhz76Jw0IXUmytAQQd1Dc7t0v7bvyP3l5UNLiplgSNK9uyegQ5pEp87R8GwUo+Md6Pj6k6tQ927/qpqai4UUSJyR3XfOL5ala4hPa3AsY6aIviU88sHJ8xb1vXe+79TJyL/laZO+mlMN0ePSp0ZEfFJPgiKxZt9MFMRIuYMX9EEMEK+oWkcoTjU1Nfl8fc6cObNmzRJVGADF7NyH3R1xb+rIVmeHP/jF4ODB9vr77Xa7XWWA+bzFz0rw/pEPTmp5QezzMyf7Uv+l5KuTpkVXl8NLsTFDwb4sCyYlxPsUY6GL3dJLrrgmlpN96nRf3i/WiS0e8lMiWxba765r2j04+EXQv8vd6rQq/n6o7VnvkVH3ehnMoICz1Kefc134nOWiV6h4wlqsinQS5QLWWHMSNIG9JJm1Sog44sIvKVkTdJimTdI+NvtQXNjcsqlyI23bUaQuuOSK6XFv0Z48laHLlm6aIYETDAnv7gnvkKYntHNUUjpl+jXRT1qHqPnkg4PHtRbXNMIf/vfWWP2xOL0nBzXq2++yRqtLz4amV9JPkVV4I6TvLCyWpXOI77NTZz7Tsjf7TwSORv6dMLKa3kFIjZej+JP6wssmas5lMfDSFyU+PimWoNpr9s1E4Y2MO3ihH8QAcahp0N2UKVPmzJmT89cXLVoksDAAilj46O5nPdE3vayurp1NSyzy2ExfUbyumrfSsopron0e//7DpzMtPCT88bF3gqn/VHLxlbPKIh+C7xz7WEMnu7+necEtq5/3JAw+Hx/s01gwSZLCJ959PdqJUr7bIXZLFeO3/+n3h06mXkw7scWDMGNly7ftdU17+w62xk8KfvDIiX7zB+QwEgmfpV7cnOvFQev2Dg0VgLwJr7EwOY541GUTtedzXPi1SRomPyseo3nbUazGTV/wf+qGe2PRaYaETjAkvLsnvEOqiYDOUcllV86KpfUcePfo5+rfOfuX/a8X3SvRnx/Z/V9tsYdltgVVafLCk5XOvGtJdexj92v7jmjYS4VXzH1nUbEsvUN8Q0NMqVJcPWbPq4zLLtA9CKktl0JxUitLmFlBLn3C45Niiaq9Jt9MmMBIuIMX/EEMEIcLJozgcDhy/u68efMElgRAEesPBg5GwxvVS+6aqTqmefijY+8IHM5PMUt9z5bdf1Z9Jzp85K2X0w6UWzqlYmrk3wlzjadb3Yl3Xz/o2/D92traRVWWb7W9N/yNCTMXLIlOcK6pYEllq5h/7ZRYm0Dwll5SOW925N/vv77rj5omgQ6/17agamnzNo9nR+K7yIKLBy3C/b0+j8ezvXlp2YI2lXfdSisf+smjsTfNQ6dOf2r6eBxGLKGz1IuZc714aH15CAIJrbEoAhxxSZKyurqGPztzakQFSEfztsNEzp/84Ij2hyclk//pgWj3czh1QNgEQ5KkR3dPdIfUsM6RIq0nbhSc9IVWJBUVi/CxfS/vi36Sl9xWNUF7BRo3vfr2uFHOOh9vfavQg9eN+L6zoFiW7iE+LaeM8uohT7/ysrjHtLqXUMv1TXlSJ5RQhfBLnwbC45OCCaq9Zt9MmMAIuIMX/EEMEIcLJoxQXV2tvlAas2fPVl8IwGgQP466pv7bx77WX6aYhTh3F141+4ZoOzS0xfPGh5lf8jj99iue9AVQxly6Pa+8rZLIr+z2lM268uLIXTy+M6alYEllk2+YfVX8e55it3TsZVdOj62tre0ltS2VpHD/269u6e7paHigtnbRrRU1q/d8rFvxoCp8ds/j5RW31NbWLm7oCB3806GQSgUruWjiZcYUDdCkpLTcar+7vj04OBD0d7k7W52p5/UItf3bE10fpA0xDM2ErS7jnOsxD7uDX2odPj3OkXpLmtFSBKxQMVqyVoqhkiGCoBpb3AwaCN0chBxx4ZcULQQdJq1XV0mSwp+ePjWiAqSjedthHuEP/7Arfo7BiqllmUdVmVS1wBbpju17ed+xcPwEQ/JDq+8uzy/cLLy7J7ZDamDnSDGjk4bHruHjb2zdWXQXCmWgw7JkwcwJ2Xy9ZPqN99pi2aChLW/4zxawcTQa+s6CYlm6h/hC3Y93+FQqg2KdiXlU+gch1a9vypM6y0wv4bE4LYTHJ8USFYk1+WbCBEbAHbzwD2KAGHJZYISKihwHHl++fPn48ePFFgbASKA+WOi5Dz1PPrihJ+My2SqZYHWsiUZJQptWPvHb9K81hPt7Xqhv2JVpdYqYy66Wn+94L8NwsuEPupqejuv2xI+7O6787kecsR7qppWPvZRpVdK50J5n4somW1ctmqvoD4vdUuXapF0N9S/0ZB44t9//63qXL/rR+sA9c+MHGRZ8IKCmZELVbUtiFczzi5feyTgSbvjsYf/r0U+2uZWXjqp5WWBqec1Sn/+c68LnLBe+wvjRkrXO/P35mZN9+f4u0sirxhYdQRPYG0G3wZOzO+LCrwCaiijmMCky57RebaSE5Llp11yheZx9ExnN2w5TCve/7flF26HYf6i33uN7B6Hu3x/qjXuTOMtHrWnWL7i7J7ZDamTnSNnT73567eZD6X/rXMi3WXEoi4Py0VdWEwwNKZm64GF7XDLLzq1vHC9c42hU9J3FxbIidArxhTY/9Zw/0ymzZ9NTsXXeseqe2WnzqPQqoecXm36ffhLJcx92uVbGTuqMJUxBeCxOC+HxScEE1V6zbyZMYGTdwQvzIAaI4YoJI4wfP37t2rU5fHHx4sXCCwOgWF169bxYf6Pz8bUvp+0nhEMHnv7enNpNynTmvpNn8p44WRklCbWtfLDR05uiGOdCB3614s5VPpW1ld+zrsEaXVvHQzUrtqTeqHDowDNPxPVgZdsahzW+2zNh9j2r7oh+yrQq6Vxoz/r7b30yVjbZ/qP7ZiWOEyh4S5XRJd+qqjuf2pPm/aRw6M2nVyxr8EWPXmXdj+zXlSpbLGKLB4VUD8MUA6iGfA0/fcrzXro+WDjk/Y+nNgsbYxzQS/az1Oc/57rwOcuFr1AxEK62mb8V2wv9ZF9ji46oCewNEP+Oml40HHHhVwAtRB2m+Mw5jVcbaWiY92g+hzxt0leLsokxmrcdphPuf2+Lsrtkca6+S31YlfjeweuvvrDr95GLYtaDaqQmvLsntkNqZOdI0dM/1LHsX9L8VlI3v1jED+qT4/PdsZNvs8fSRwqe6Tsa+s5CYllGhPhCvoZlK55+M1UbMfGUkevq7rtuomIRY0q4ftW/PpO2hA/GrTNFCVUJj8VpITw+KZaoSKz7KN/BAAAgAElEQVTJN3ME0p64b5oBPov9Dm6GBzFARFG0jzASVFVV5fCtq6++WnhJABQrxeh8UqjjoW/e2dimeHAR7u/1ebY3L51Sdv2qjqRHDKc/Ov33vAsxdvKiho110ScKId+G2gpFMcL9vb7tzd+zXL8yuQRJSkqvs/8otjYp1PHQN8u/27zdF4vThUM9O7Y1f9em2CJrw7p7EgKMEy2PPOGyykmr+l1PrJt6tnePp231nWWK9nFl3caGRZOTxwkUu6UJa5Mk35O3llmWNm+LP4DhUM+O7c3ftVSv6ohLRbfW/dB2eVJ7RWzxEC9l/2TS3HsesMY+dm+orblz9fOe5BOwbfWtZbb10fCHbH94wVSamzCcTrPU5z3nuvg5y0WvsGTKtfOj4ylqmvlbOeoycqNTjS06giawz4FirJHjBz/4JPOvKl6hzoWoIy78kqKFoMOkeNSn6WojJV5wKuZfO6Uo2xijedthHuFQz47tbau/ddE3H4rvLsl1j/3YerGG78dNMxRq29DyVuT72Q+qkZrw7p7YDqmRnaOEnn7kt/b0Rp6HnQv1/Dapm18swmfjBvWRpOold83M5fmu4rqqad4QPY2GvrOIWJZBIb5DHauqLd9t3p75lJFXbPzZdyYn7H3zl1Cd8FicFsLjk2KJisSafDNHnoy5FDkuqbciv4Ob4kEMEJHDhMpADv73f/8328q5cOHCQpcaQJ7+6nZE3xad5nD/Nc/lB0646+TUVwwNZFvr4YGENX7pd8XGZre6/H1atmog2N1o1VQOua61c7018ulhd/DLFKvr2+/StraIO1z+T1KWq+/wZkd2+0e2NnYHE3eKebZUkuQV7hNfGFQ8VSq1Jb72pqxLqguk+y2VAiuWnebyJy6r5Xfjl4kXfxp+4nfdkWqZDCrr3MfS1y9AJwNnvI2xS0PGy8iwoNsRq7cJZ5zy7JAbvWcyV+qTXmcski47vWcylC3t9VzxlT5/izX2FYvTe9K4FarvvYQrg04X2EG1S1mhLrAaflr958TWWI2yumdl24TT2qBKNhBojSVtaKjPyuXjKnPWDbz48qdqMSooTvPsa6zAIy78CqCJoMP090Br3CCsWvZDwgUnxQVZh3qr8YqRXZUTve059WgwsnwZdD8s5U9TbYxI1blLavnkRXx3T2SHVFTnSMsNNOHSrV3ylVCP+7uytZy6vZTGwDF3/FN2W2sgx95jwg0xZVVUniZZlTOR6m4xuO9coHZd3rEs8SE+Rckr73cs0rb6tDtf5yBkXas7duXKyNboPZG6emhpA4iOxWkgPD4ppq8RIyYSK3YzzRwzLNTvaolVal+yIJdKgXdwLUbeg5gChnpgLkWS7IviN2nSpOXLl2f1FbvdrlNhABSpksnfWfPik1ati9ucm/d6n4tGikMHA8GMExVrLoY8f822dvWwmvXJF9cs+obqZMelc1dt62yJHz0+A9nRsv/5VZaUw4qWlM5Yssnndmrtj1U6Wjq3rZkvp28LFHJLJUmyNrp96+zpX1YQXLxR7dLrF96eaj9+9lHs7e+JllXPehttKZZKrdLR+tIz9itoa8Jwes5Sn++c68LnLNd5haFNK5/4bfqRHsL9PS/ETfuN3OhZY4uNoAnsczB+4mXRxlWoe8urb6f93XB/z5an8poIXOARF34F0ETQYVJOIR/atPKxlzK+QHkutOeZuAtObvNQmMRo3naYmFrnK1HpzLuWVCv/S9AEQxHiu3siO6RGdo5KSi0rtnm1REJka+NG1/0pn+eZUfiI99n4+Ttyn2FHeW+VpNCWN/xnC/hS/ujoO+cdy9I5xHfxdT/++U71pCKb092ZbufrXMJ/mHjD93/Vqvqk2NbobVtTk/2YLFGiY3EaiI9PCiYmEmv6zSx6WmKV2S5psOK+g5vkQQwgMccQjHTzzTdntfzMmTN1KgmAojVWrmnc5u9Q6ydUOlxuf3Bn01Krdd7t0fiBwHBGiTx/3U5ftytt/r3s2Lh/W2ONrKnvVyLf9OgL3f7NTmumpWSrs8Pf8/yjczMln5SW25syFixSvJbuwJvtj96k2oMqxJZKklTpcL0W2LnGXq4SCxVbvFFs7ORFjS+miLUpOx4lk2vWdQa6W9TfO7E6N/u7X6irZAJgFIaOs9TnN+e6pMOc5TqvMNS28sFGT2+Kh6znQgd+teLOVb6kPyC9NFN361hji42oCeyzFj9vjiT5XPXrvKlPcyHVXuARF34F0ELUYVJMIZ9xPcmzyMv2H903y/BmRppTOAfFt+0Y2WzOVq+WzpfSuOnVtyv6D8ImGIoR3t0T2SE1tHM0Vq756c7Aa65MT15tTveenevumlo0Oa6fH9n3WizZUaq+t/rK3Bs1E25ctiZuyKvQ5qbtKpP46Wt09J3zjmXpHeL7R0v9toC7MV3xZEdLd6CzyT4j/c7Xu4QTZtQ97VMr4bp8ElkkSdIhFqflN4XHJ8USFIk1+2YWOW2xyuyWNF5R38HN8iAG4NoJ49TU1GhfeM6cObNmzdKvMACK1ljZsqTJ2+PvcncmdhVsztZOt3tvoO/d9nq7RR4rJbw5Gure7T8lrCCl5fPr208E/V3u5+KadLLV+ZzbG+htXzE3q/yJEtmytGnvQDDFdlmdre4ub+C9vU1LLFrWWVo+v7492BfwutOt6kyw/dH52rumum2puzWh01jpcG2NHMEF5Rofrogt3qg1FGvzulud8X2/yvkVX1cehgnl8x8d2t3JZ6Bkdba63V3+4MDepqUWeugoIF1nqc9zRnPhc5brvcKQb0NthWJe5HB/r2978/cs169MnhIZGaWbulvXGltcRE1gn62EJ7Ih3/ql9zc+H3cEIqf5cLW3OhzW3H9N5BEXfgXQQtRhSphCPrqe3/XE0nHO9u7xJM0iX1m3sWFRXq8L5ybdKZyDott2jDyy1fmc2+12ewN9g7ub6mq0dr7iKEdpkuQlt1XpMWKQ8O6eyA6pkZ2jktLyBfXtPUH/roRiyw5Xp3tvoO/VpEfyEy+bOD7fn9VP+Ni+l/fFPtpur54+Lo/VJd7Ku19+60iBn6CNjr5zvrEsvUN8E8rt67xBf5di7TZna2eXP3hCU2SsYCV0ewO9WcXuMhMei9NCeHxSLFGRWJNvZlHTGqvMZskCKOo7uGkexGCUK/QkRxhd5syZo7Fmrl27ttCFBQAAQJESNUu94DnXhwmfs1z0CgeC3epD+g+tpq61MzbPu8qcxGmpz5qceZrkAk6inNWs2/ESpqMWVWM1Kshk4ZrlP4F91vNwD2ZzCGSra1/A/bDmHZhyAaFHXPglRYv8D9Pg4ODgQN/hzervrCuLbm3sDqbeEXrUW22ncC5VTty25/LrgAhnvLHZsiSL03uy0AXCoOA7chYGzngbZSldewkwRqHqPwDkjysYkJYJktIwmjgcDo1Lzps3T9eSAAAAYOTSY5Z6AXOuDxM+Z7noFZbI89dsa1dPZ7E++eKaRd8w3UC4pqJx6m49amzREjOBfbYmWh75j1b1H5Wtje3bVl1/Uf4/J/CIC7+kaCHmMJWUzliyyedWG7Y6qtLR0rltzXwDX2DXeArnwPzbDmQWPut/Y0t05CXbD5dZLy5kcTDk/N+Ovvl+5EPF1LJ8RjrJSvjT06cYsA8AgBwV7A4OFAE6wTBUdXW1xiVnz56ta0kAAAAwkukyS33+c65HSid6znIdVjh/XcaZv2XHxv3bGmuYx02F5qm7damxxUrQBPZZKq2s2+TOONv90ETmgpIJhB5x4VcALQQdppLScntTxqvN8IocLd2BN9sfvcnYZA7Np3AuTL7tQGan/Lu7o1Og2e69cTr1U7z+nuYFt6x+3uNRTkOXXvjou69HH4RNm37FJYZlHJ8+vN8vSZJ009QyspwBAKNdEd3BgSJAPwOGqqio0LLY8uXLx483yYRwAAAAKFJ6zFKf/5zrEcLnLBe+wtLy+fVD+++5uJEDZKvzObc30Nu+Yi6JLFpkMXW3HjW2aImawD4rsTq/1RU/4ojV2Zp6IvM8CT3iwq8AWog6TKXl8+vbg30Bb/KeGF7PmWBWF1iBsjiFc2LmbQfSC3/431u39Ax/kO0PL5g68m9MBTCubOo3fBu+X1tbu6iqbMrqPWczLx7+oKvp6e7IJ7l29lVGPQiL1AfZNu/qSw36TQAATKto7uBAURgzODhY6DJgdFm3bt1PfvKTzMvs3r3bZtM+3jIAAACgh+OepbfUdgy9HWN1+XfWW0bwWBgAAABanO5pfrCqYdfQB9npfa+phnwrPZzvaZ5R1RB5T/sOl//F+nTT6oVDB57510WrOiKD5Vic3teaagyY+Cnc39v11MMrN/hCktzofW9NzQTymlAo9N0AmEUx3MGBokHjEkarqqpSXebqq682oCQAAAAAAAAA0gv3h/4WnVQrHDrQvvreaCKLJC1es+xGEll0ckHFPz0SGxtvV0PVzUubt3n29CrmOOvv3ePZ1vxd2/Wxx2CSXPfYj63GPAY7G3ilbYMvJEm2xhdXWElkAQCgOO7gQNFgoCIYTTWXZeHChVOmTDGmMAAAAAAAAADSCPf94Ymy2mdT/Um2uurvKR9ndIlGj9JZ9/3I3uLbFHnEdaij4YEO1W9Zn3xxzXcmG5RVMtHyyBMtJ22Vyx5iNjQAAIYVwR0cKBqcEzDapEmTli9fnmEBu91uWGEAAAAAAAAApHHBJVdMn5byL9aG5keqmMNDT2Mn29ftaa2T1ZeMsD7p3dZYI4/Vr0yJSuc+2vRjElkAAIhTDHdwoEiQy4ICuPnmmzP8debMmYaVBAAAAAAAAEA6F5RNvSnpP2XHxv3bVlhKiS3rbcKMul/1+N0uR6XakjZnqzew86c8BgMAwAS4gwNi0N9AAdTU1KT705w5c2bNmmVkYQAAAAAAAACkdukNP97stEY+yQ5XZ5e/54UVc3niYpCxssVe3/5uX2Cv27016ZGYzdna6e7yBwd2N9XVlJNdBACAWXAHBwQYMzg4WOgyYDSaO3fuH//4x+T/X7t27WOPPWZ8eQAAAIAkxz1Lb6nteF+SJEmyuvw76y2Mow8AAAAAAAAAuiPPC4XhcDhS/v+8efMMLgkAAAAAAAAAAAAAADAPcllQGNXV1Sn/f/bs2QaXBAAAAAAAAAAAAAAAmAe5LCiMioqK5P9cvnz5+PHjjS8MAAAAAAAAAAAAAAAwCXJZUBjjx49fu3Ztwn8uXry4IIUBAAAAUrnc3n5kcNjeektpocsDAAAAAAAAAKMCuSwomKqqqoT/ufrqqwtSEgAAAAAAAAAAAAAAYBLksqBgEnJZFi5cOGXKlEIVBgAAAAAAAAAAAAAAmAG5LCiYSZMmLV++PPrRbrcXsDAAAAAAAAAAAAAAAMAMyGVBId18883Rf8+cObOAJQEAAAAAAAAAAAAAAGYwZnBwsNBlwOh14sSJyy+/XJKkOXPmHDhwoNDFAQAAAAAAAAAAAAAABca4LCikKVOmzJkzR5KkRYsWFbosAAAAAAAAAAAAAACg8MhlQYE5HA5JkubNm1foggAAAAAAAAAAAAAAgMIjlwUFVl1dLUnS7NmzC10QAAAAAAAAAAAAAABQeOSyoMAqKiqWL18+fvz4QhcEAAAAAAAAAAAAAAAU3pjBwcFClwGj3alTpyZNmlToUgAAAAAAAAAAAAAAgMIjlwUAAAAAAAAAAAAAAABmwRxDAAAAAAAAAAAAAAAAMAtyWQAAAAAAAAAAAAAAAGAW5LIAAAAAAAAAAAAAAADALP5/9u4/PMr63hP+l6zr2hYtdXXrBHvgASR6HqiWcOBq5TqNKMGutccFracqAWF7agseVw4B4Si6HipC2Fh/0FK9iBAsXbXJxqK1BKGxi/WBJqglV0sisHBKGK09lIYph8PS4fkjP2YmyUwSJDKa1+svvjP33Pfne98T/pn39fnIsgAAAAAAAAAAkC1kWQAAAAAAAAAAyBayLAAAAAAAAAAAZAtZFgAAAAAAAAAAsoUsCwAAAAAAAAAA2UKWBQAAAAAAAACAbCHLAgAAAAAAAABAtpBlAQAAAAAAAAAgW8iyAAAAAAAAAACQLWRZAAAAAAAAAADIFrIsAAAAAAAAAABkC1kWAAAAAAAAAACyhSwLAAAAAAAAAADZQpYFAAAAAAAAAIBscdaZLoCPlni0bsPr+/8c4ntfuOe96TuWTTzvTFcEAAAAAAAAAHyIyLLwvrRlV47tfWFpcXl90jv58zd/VpAFAAAAAAAAAOgVWRZO2W8rp181tXxPmnfHjr9s0AdaDgAAAAAAAADw4Zdzpgvgw+szU9buPtnqD7Ul16W8WThu1KclpQAAAAAAAACA3pFlCSEcj9ZVrpg+esCAAQMGTF6wdns0fqYr+vAZ9LmvTClMLCOFN39hhC8XAAAAAAAAANBLOmeEeOMzM8bOqm5dVS+fsXP7e1Ub5o0beEar+rCJxw7s3plYTrh5wlBRFgAAAAAAAACgt+QNTrxbv7065ZVozc4DR85QNR9ax9/Ztzvaviq8dsKIc85gNQAAAAAAAADAh5Qsy1mfHjWuMOWVSMHoi889Q9V8aL1X/+qO9kXkiqEX+WYBAAAAAAAAAL0ncRByRt62praipGhUCCGEwvlrqtbPNWCol5rf3rZpT9sif9rkz553Jqv5oMT3V84anbtgS/OZLgQAAAAAAAAAPjIGnDx58kzXwIddvHnLvZdevbR1xlBk4eZdSyae95GPSR1vqpz7V1P/v2mbf7ps4gVnuhgAAAAAAAAA+Ij4yAcO+AAcfXvH69G2RWTaNWM/+kGWeKxu5W1TV0YjhZPHnn+miwEAAAAAAACAj46PfOaAvhc/8NamhrZFvxgwFI9ufmheSU1/Ce4AAAAAAAAAwAfHz/C8X/Hdv3i2ur0ty7C8iweeyWo+ALHtpbdMX1oT7SfBHQAAAAAAAAD4IMmy8D4d2731p9Xtq8JrJ4w45wxW09fi0dcemT2ruKYlu9MPgjsAAAAAAAAA8ME660wXwIdcfN/WZ7e2LSKFN39hxEc2H3U8uv2pe26YU97eg+ajHtwBAAAAAAAAgA/eRzZ30BvNjVteqKysrPzRium5Yxds+X3Gg+OxxprKysrKyqcWXHVrWeOxro45Hq37yY9WTM8dMGDAgAG50x/Z1NjcTQmxxi2t57yyuwJCCMej21dOzx0wYMDo6Ss2Nsbi3R3fQXPjlhcq2+trcdWCssqf1EWP9/JUIbz761cTA4Ym3DxhaIavVDxa90LbdXOnr9x+CpfrRt88nRCPNW4pW3B97vikIMupBXdijVsq16+YPrrtvk9eULYl5QnGo3UvVFZWrl8xfXL334TWr82PyhZ8eXLZri6+B7HGLWULrmrZ5/ZohwNaL1W24KrcOZVN3T6L5sbKRVe11Fy5K9aDvQIAAAAAAADAKRhw8uTJM13DB+x4tO6V1/cfC+HY3heWFpfXp7wZWbh515KJ5yUnFOKxxp9X1x8K8b0v/H1xUpQhhOEltbvm5ae2tolHt6/7zn0zllenvBqZXfHL0imDz+5cRuezRuZv3rVs4nnpNxBvLPtS3qz2C0RmVvzyqSmDexCqiEfrNrz83HdmLa9Je8ioopIV/3jHpJEDe5jRiDdvuffSq5e21l+4uuHlmSO7/Gg8WrfuO/NmpF46w/HtmrcsuPTq5a0XiBSu3vLyzEuTPtFHTydR9YbXd/zi8fuX10Q7v5vkptUN5TNHZuzREmvctOrb01NrbFVQWrvhrvyBIeVmZvgqhn/Z1rGkzgV07iKzuuHl285945XX9x89vO2pWSk3ofON7Sj1WzdqZsWLT00ZIgoHAAAAAAAAwGnXz36MPlG3YsR/yH/s1/EQP7xtbccgSwiRadeMTUkP/H7LgnHn5i3ddvh321aWdoghRKaOuSQlKhGP7Vp7e/74jlGJEEJ05ZzHtiZ1/4jVrZicO/a6qVOn3tQp3BBd90ptc4ZWKyferd+efIFo2f/cuLvLBiTJpUXr1i64OnfsDbOW14QQKSp5fnPDkZMt/u1gbUVJ0agQQgj15cXXFsxet6unvV4O1W5MdGWJXDH0oq6+UPHoa4/cXjh2RqcMTfVPt3ZTeby59pV1iRvUoe9LHz2ddsd2v7zshql/112QpdthQ82Nmx6ZPjKvsLg8GiIF85+saL/5fz5Yu2Z+QU3JvFW1sdSbGUaPuDglUXSssexvWzfbOVsTGTH0ouQsTnNj5QO3pHSRCWHP9tVzv5A79rqpU2+a1fEmRKuf/cXuTM+8w7euvuz7mzMeDwAAAAAAAACnqJ9lWc7Kn7f75MG1826ccuPMZRsOVMyOpLydP23yZ1MbolwwcVntyZMbl828Y9nm1ytmjkp/cDy2a93siTO6arsRQgjRN/e908Pf/qPVG2sPZdjDp8cXzkyp+71DR06kPz4ea6xccHV+W5SkcH7FbxrXzrtx4siBrQecHcmfMu/pn25eWNh6/fIZE++qaupJtc2/2riurm3R+e6FEEI8uuneW26au2lMyfMVFRVPzi9ILn1vw4HMw2qOv7Nvd9JUnw6Jjb5+OueMnPncyTZ/blhdmPReZP7mP7a/tzF9d5l405ZFN+UVzi2PhhApKt1Wt3nZ16e03/ycSP70pRtqi0NpefWvtyfdzM4DjFqK2bhs5h3Lfravw1c3NYN1vKlyUcHUpTUdS/nkhReem6bKbnNFOZ8YdH7Kl656e/27Gb51AAAAAAAAAHCK+lmWJcXZnx4y9OPJL0QKJ489P+3hOYM/f92VSeux4y8b1PbvpKhEwfzVLV03jmwrSc5t7Dl0JJGWGJg/72etMYgjDdUlRamRmrp1G3/VVZuQ9kK+vKTqiQ6fSeN4dMs/XZ83tbWNR2Tm6t88v2zKpQM7H5gzeOI9C+a3nTNadv/iqv3dpVlSm6Z0efdi20tvmb7mL/5xW91T826cMmXK1756XV7S2+de+MmMc3ni+7Y+u7V91alrTkr9p+/pdOnY7q0/TWpM0nVwp6NYfdnt1169tDqEEAoe2Fz31N3jIp02kDPwc1+aNvrVirXVb6btQNPB2YM/P3FS18Ucj25ZetvUlV2Edq4cf+s/tn3rTv7bwdry1FzRjlfr30u/k5zzxt0wt6BH3zkAAAAAAAAAeD/6c5YlfvSPh44mrTNFJUII4cSRw0ntUgrHjfp06wybeFPVXRNnlEdHFZVuPbh52cyWrhsD//KqlNxGGgNHTpr74BMpPUW6beJydmTc7Kd/WdHWnSVvWG6XiZDj0S1Lb7n6gZrW5XUlG/7HzEvTBzDO++zkaflti/qyxyvf6GbS0NG3d7yeiLJ0vnvx/ZV3zSr9i6VbVn5zXOTsEEKIH3hrU0PSEekqbzvB7l88m5i6kzk+0jdPJ+G9+ld3JFaZY08tYvVls782q2WOVUFp7Yb7JkbO7vrInIsvn3TB+uWPJrIyHTvQdHTi4N7XuigmHtv1w3tue6Ch6IltB/8tNa8TKfziX346cYKzI/m33reiuCDxyp7X9v4uU6OVgePmrt9Q0Z67ipw/6BP9+X8PAAAAAAAAAPpKf/41OiWK0YNOG8lphqQRMLH6NYvuL4sWzq94fuXdVya13Tj2x/eOJD595bDcs9KcOOcz19x6fW8HuOQMLvzW3IIQUnIbSToEWUYVrX74jvxBnQ5LNuiy8WMTq5of/6zhaPqDOwRTOt+9E9GqpVP3ztywctqlA9tuyru/frU66ZZ3XXm71FYo3cRH+uzptGh+e9umPYlauok9dQiyPLB5/ez8gb34W+vu/Mf2vrU9Uc3oERe3nDxWu+qbCzdNWtMaHho4bMy43LYzzlhwY4c5SDkDL//r64Yn1kff+2PG5x1yIvlT5q090DJrqf2iAAAAAAAAAHBa9eNfozv0COm200ZKmiH3iqEX5IQQwuG6VffMKg8zK55c2mF2T/z3+948mDh7SleMDnLOG3vNtJQwS8Peg8d6to2k3EbyxZtevPe29iBLiMz87w/NGNXFaKFMGja9dSBDY5bUpinD8i7ucPqzIlO+c2TDnUkZjnjzb2o3dVd50gV6PGAo9OnTCR2nKXUbe4o3bXnoH1qDLOG6khV3pe3I0rVuY1WxAw172/7ddhvj+yvvmlUcvvHMw19rCw+dP+72O4siIUSKSqvuLOh8984aMmZqeyeeEH3n8J96WmF3zw4AAAAAAAAATlU//jk6tUdIt5024u/sezMxUKcl+HK8qfLe64v3F63+4aNThnT4cGrUI2/S5RdnOvt5l4yflNQiI/x25/4/dFN/Sxani34bIcT3Vy2+vyyxuZuWLPjPg3v9qKN7Dv0pfZYltWlK4bUTRnSeFnTOwIHJ3U4O1W5M7soy4eYJQzNFWVJu4PBJ4y/JEO/o26fTofJuYk/Hm6qW3ra05d5ECkoWd9cOp7Ox4y/L+JHmX21cV9e2aLmNx5uqSua8PG719+5Oys3kDLx0+tqDJ08eXHv3uEhXGxx4cd6wxOq1vQe7aQYUQjj8xo8rq7t7dgAAAAAAAABwyvrt79EdeoR02wkjJbrREnyJN724eE5l3sLSh7toedJhPs7nx1zy8Yz1XDjqi2OSlnte2/u7jLmCeHNN+b3VoXBJUad+G8ebqkrmlNW3X7ugZN5XR3YOmrw/KU1TetalIyWBkS7+0u73Nau/m7iBYcwXR12Y/uA+fjqplWeOPcWbXlw8Z2Vr8KWgeMUdY3vQDucP+3f+NrHqZvRSanBn+LjLh50Tb3px8ZxXv/TE4hmXZvwWd3TWuYMy9iLqfOmmLd8tfSky/44bT/s3CgAAAAAAAABCCCFk+sn8I61XnTY6RjdG5+UOjNWvWXR/WV5x7aKru2h50av5OCGEcPZFQ0dEQmgv6eh7fzwaQtpgQqz2yX9aEy0o3vDVTk1Zmrc+1p6lCCFEptz5tSt6Nl3o2MG9Dd0f1SKlq01PunQca/zRquXJdzxz/KVXwcFKG8YAACAASURBVJc+fjon3t5Rkag8Y4eYlI44+fPvm5Y0YinDBX6397X2AUndBoNSgztTx1ySs79q8UN7Z5Suv6Fj+5nunHXhkBHDQ2i99p7d+987kR/J8H/C72see6gszK74+wm9iswAAAAAAAAAQM/11yxLbzpthBBC7GDDzuToxiffWPX1WeVDSmpv7zKskDrCptumLyGEnIEXjxidlGWJvrnvnXhIU1Tzrue+X1ozpqurd4iMhMi0KdcMPjv0SGp3kBAZfv4n0tyU1K42w8ddPqy7Lh2p8ZHu4i+9DL707dM5tvet7XsSywwdYuKxNyofb+uIE5m56O8LLsh04vaP7X1rU+IC3Q08SrmT+dMm5x2pWjpn760bHu0qtXM6xWN16/5p+cGCkqLCnn6jAAAAAAAAAKDX+meWJd5c+8q6Hnba6Hx84aTLD//PecU7Ckqq7sgf1NXxHUbYdNf0JYQQQs5FQ6+IhETEYs+hI/EuZ0DFY3Wrvznr5bwurx771Y/XJUdGbloy6ws9baGR0h0khJB7xdAL0qQjUrraRKaOuaS771FqfKS7PivNv1h97/NJ68zBlz5+Oh1SOBkqjzc+t6ikpnWRP+3Wvx7co3BJ7wYepdzJSOHkS3792Ph9czcs6VEDmE7Oyh12ZXtfltCw9+CxEEnTxCdWu2peSU1BcW2PpiYBAAAAAAAAwCnqn1mW1AFDmTptdHF85KJd6xf8MNOP+r0eMBRCCGFgbt7oSIi2XSfdwJdMkYJ48/aq0poeR0Y6fDilO0gIYVjexWlCCyldbXrSdSY1rtFNn5XDdU+WJreW6W4Xfft0UlM4GSqPx954eV37kYXfmtWzpiwd4kfd1XPi3frtiTs56S8OP/vd7XMXb+g6tXMaHW+qLk/TCggAAAAAAAAATqd+mWVJHTDUfeAj/vt9bx5sX0XLS0vDdRl+1O/9gKEQQgg5nxh0UaaGHCGEEG/a8tB9xWkjBUff3vF68nyhbkbzpOgQN8lwW1L7oPSk60wvBgzFY3VPzyv9Q8EXIjW/aL1G5IqhF2XYRd8+nQ63JUPlh7Y/94Oa1n/3/M53iB91W8/h32yrbft3pCD878dLB9/5yytOvVHKhUNGD29vzPLbnfv/EPK7OFm86cXFcyq7bgUEAAAAAAAAAKdVP2yx0GHAUPexg47zcUKkoGRx+h/1Uztn9GzAUAghhE8NGf2ZpGXD3oPHUg843lS19LalB4tWP9z11eMH3trUkLTOm3T5xT19wB3jJhluS0pipiddZzrcwMjMv52cLjwUq1017wfD/vstY/5PT+Mdffx03qt/dUdilSH2lBKQyjwUKUms9sl/WpNUffpeOF1cJVpT/qthTxTfMPjsnlzq1MX3Vy2+vyyveIXpQgAAAAAAAAD0vX7Yl6XDgKFuYwedGpZEZtz3dxl+1E/unNHjAUMhhHDWuednGnXU2htj4dqHZ4zq+uqxgw07k1MdecNyezhgKN5cU35vSiIk/W1JScwMnzT+kl4OGMqfdutfD+761IfrVj1YOuyeV4buuKanfV/6+Ok0v70tMXgp04ChlIBUT0c7Ha5b9WBxL2ZCdYhhhVB494IbhryvPNpZ/2nYlcPDnpY97nlt7+9OhM+k/qdwvKmqZE7ZENOFAAAAAAAAAPhg9L8sS28HDHXozBFGzXziGwUZAhAp5+/xgKEQQjjr3EHJoY0j7/3xWAhtqYzY9tLb5pTlFdcuujqS7uJ/OvxOctBh+IghF/bs+cZ/+8oPNqT0Npl/x40ju74tqX1QxnxxVKb8TQidOr6kzabEY3VPz3vpr55ZP/Fff7Cyx31f+vTpdMiOZBowlByQ6mYoUtvJY3VPzyt+KemVblsEdYhh5c9f8JWR7zde8rFBF2UaG2S6EAAAAAAAAAAfsP7WaKHXA4ZSO3OEUDDzW4WfSf+J1PP3YsBQCOGsC4eMGJ5YHn7n8L+2nbVpy0P3FTdMqXhmdi96Y1w5LLdHUZZ47I3Kx8vqk165acmsL6QJeaT2QelBA5KOA4bSZFPiTVV3Xb/5uhV3Tfz075L6vnQXN+nbp5OaHRk+7vJhaTYb//2+Nw+2n3d0Xm73s3hitavmlTQUFHwh8VLuFUMvyPR0U64SQuG3ZhVc0O11unPOJy88t32xZ+f+91KK3F5625yXJy39nulCAAAAAAAAAHxQ+luWpbcDhjpkX0bNvHPK5zKlSVL7c/RiwFAGzbvWLL5t6b8v2bBkyuCz3/fZOon/tvq7ZTWJdaSgZN5X0zRlSW2y0oMkUOcBQ11mU2L1axY9tHfu4jvyB6VkX7qJm/Tx00lt4ROZOuaSdNmg+J8O7Wk/8ccvGvSJbu5KrL5s9qzi8F8fm3b5/0lcoJtsTWoqqCc3vycyTLY6XLfqweKGKU889LVLTRcCAAAAAAAA4IPSz36i7vWAoaNv73g9qanI9bdek6Htx/sZMBRCCDnnnp/Ul+XoO4f/FA/HmyoXTZx1YMbm78/tdsjLhUNGD+/mkE6ON1WVzEluylJQvCJDE47YwYad7fej2yRQ6DgDqMu4Rrxpy0P/cO9Zi56ZO25gh+zL6BEXZ0pR9O3TOfH2jorE2TMe/97+nYnuMC0PLr1Yfdnsr83aNG7196af37C1x9ma1DsTmfKNycNOx19v6mSr1/YePNHyr+PRLY/OK94/84niG/oiQQUAAAAAAAAAafSrLEumAUPxaN2LddGOEYT4gaR5N5HCJUUFmdIG6UbYxGONbzXGMsUbWuScO+iixCq659C/NG1ZetvUVyet/h+LJg7u/lGdNWTM1Pxuj0qpuOnFxXNWJt2S2RnHGKVuMHnmTmz7IwvW7eq8x9QZQF3FNZp3rVl82+vXbnj0hsE5vez70rdP59jet7YnlZ6cwmlurHmj03elXXRnw8FYuiKirz0y+2uzynMXPvPgjJH/N2lm0PBJ4y/JlK1JuTMhMm3KNacnYpLziUHnRzpfrenFe2/7fihZ/eiUIf3q/wgAAAAAAAAAzrh+9Tt1+gFD8f1V937jf1TUv5saUEid6tJtG5I0I2xitavufLr+jye6L/ATgy5KihUcfePJe2770V+s/uHKmaPSNkpJcf7YyYWJE2yq/U1zxgBNbHvpbXPKEhXPXL3loYxjjFI3mJi5c7hu1dIfX5jXuYdKd61NmneV3T3x3o890Rag6fqGx/dXzvovK+o65kP69ul0zI60p3COR7c88o1bVr8ePZ44OPXBRZeXPll3uFMBx6PbV96eP2Fuee7CzWVLJg4OKfWP+eKoC0MIIVZXtuKFLrI17/761cTBvW75k17Oxz95/sfbV3t273/vRIjVr1l0f1lexg49AAAAAAAAANA3+lOWpcOAoURbkcN1pbOn7r11xaKrIyn3I3WqS3Ibki6d2L+jov38bW024vsr71pxaEHPBrV8/JMXJmIFIbp+08G5q3scZAkh5Jw37oa5BW2piuiGH7zy27Rhllh92exZxTVt8YjIzNVbHpl5acaARMoNbO8jEo/VPb1o5w3f6yL3kNraJIwdf1nymKSWIEtYkgjQxGMHdu9sfz8yYuhFZ7dMQZofpn3tcx1O38dPJ2WaUiI7Em968d7bfnndhiUpoZ+BuXmjk5ubvFQ89uYFa7e39m6JR+teWL9ien7u+Dnl0cKWIEtOx5lBLZs91vhc6boQye2YCoo3/6Z2U+LgrkY1narUyVYhhMN1q+6ZtemLGTv0hBCaGzc9Mj13wIABkxM7BQAAAAAAAID3rR9lWVJ7hIRw5bDcs0IIzbvK/uH6l/5q8/rOv9zHDjTsbV8ktSHpWnzvW0njdD4zesinQrxpy73FL1y5uEcTgjqKFCxcu37uuN41xhg49o4VxQWti/qyOQ+u2dXcRamtk27qW69U9MS2uu91E2QJIf7OvjcTN/Azo4d8KoQQb6q6a1Hzgoe/dmkXuYcTRw69l7Q8dPhIS/uTeKxxS9mCmy67NyxJCdAcfXvH64krTBp72Xk58aYXF3//wieXfLnTHezbp5P6bRmWd/HAEEKI1a9Z9PiJJQ/fkT8o5eicYZO/MSV1Uk/18hnjc//dgAEDBgz4d7ljb7i1uLw+hPYgS8f62zb7k2XPju2qG0qapjKnQ+pkq9cq7vn69cX/d+EzCzOmr+KxutXfKJxbHg0hVC+fccMtpdvTjVUCAAAAAAAAgF7J+Pv/R8qJ9/bv3pP8Qvn3v/PFP1346vLScEfV+q+Pi3T65T6lDUm3U11Se4qEhpeee+b857a8mre4N41V2kUKFq5dv2RSpNeJhZyB+bPXb26+5eoHakII0bJZlx1oWL1g1lcLRg7MCSEea/x59daXH5+1vKb1OkUla//xjkkje1DhiXfrtyf6iISa0sdWDb0+bHjqT9PXLJzY+e514flZBR879NiE91bev7wmGil6Yltdh9v+h/07f5tYbXrhBz86Fl7efuWjD3Zx/r59OvGjfzx0NLE8eujI0Vjjq6u+/fDOLz7e1UfOHlxYNLegMtHnprOChRXfv2fKyPYy//XwO0lziP5572/qfvKD7/78ykcf7KIbSspmI6Pzck/n6J+WAUmthdevLw9Fq3/YXfrq0PbnflCTWEZrSqu2/93YiacvYQMAAAAAAABAv9V/fns+ceTwodRXqpfPeua9iavrnp7dRZAlxJtrX1mXaIXR26ku0Zrl/6th/IO9C7Kc9Z+GXTn8fQRZWpwdmXjfhoaK+a3DhqqXz7o679zWDiHn5l01tSXIEikqqag9eGDtvB4FWUIIOZ8YdH5y65FoeenKXw77x4r70gdZPp531VcKQspnim/6u+UNk0qqGxrXdr7tHxt00aCUg1fWD1vwYFcNY/r66eQMvHjE6MTypeKx//HcvNL3/uZ7aT8ycNzc9WsXFkS6em9UUclPGzYsSQqyhI6brSm+et7WNJvt0CRmws0Thp7Ov9uUyVaRgoWlD8/o9kubGsQJIUQPHf6TOUMAAAAAAAAAnAb9J8tyzsgZj24rLWrNGkSKSp5/qfbghmXTx6WJjHQx7ybj+XPOG1f0xPzCEEIIhfNXb6g9uGHZlEt73z/jY+8vyNJazMCRU5b9bFfD5qqK1fMLkt8pmL+6oqKq9uCfD66dNyW/N1fJOa/gzqrWGxgpmP9kVW3d5mVTRnYxWiipjPzZ62vL21I1kYL5T1ZUZQjQnD9u1qLWgwvmr66qPbj5odT8R7s+fzo5I298vLrt69L+bcn8kcikhzbX1VaUFLUHWgrmr674WcORt9bOm9zpRl1Q8N+WLuzRZo/tfWt7oqVQZMTQi3rSBafHWhNUIYQQKVr6vUVX9+BbkRrEAQAAAAAAAIDTZ8DJkyfPdA3AGfTbyulXTS3fEwoe2Ly+h+OiQjy66d5bpi9tn6k0vKR217z8/jOyDAAAAAAAAIA+03/6sgBd+tSQ0Z8JkZmrv3d3D4MsIYScyKQlzzwxs60DTWTqmEsEWQAAAAAAAAA4HWRZoD+Lx+qemld89sJnHpxxaZcTjtLKGfxX101qGU5005JZX+jdhwEAAAAAAAAgDb0UoN+Kx3atm3192bCKF5dMHNzrXFv8T4ffORpCiMy/48aR5/RBeQAAAAAAAAD0R/qyQP8Uj+1aN3viirDkh49OGdL7/wjizTXl91ZHQ8EDz/y3CZqyAAAAAAAAAHC66MsC/dDx6Pan7rnh2/88Y+36GaMG9v7z8aaqu29bGo3MXP29uydGzj79BQIAAAAAAADQX+nLAv1Nc2PlA7eMn7PpS088s2RSpPezhWK71t7+V1PL8hZW1Dwy81I9WQAAAAAAAAA4nWRZ4KPqeLRu3YKrcgcM+PKKusOtr8V2VS64KW/q0oaiNVsevWFwb/8DiDVuWnH7yMsW/vO0ioYNS6aMFGQBAAAAAAAA4DQbcPLkyTNdA3DaxWN1j14/dm5Ny6qgtHbD7ee+/vS3p88tj4ZI0ZotK6ddOrC3SZZ4rO6J2c99bMpXv3x9fu/7uQAAAAAAAABAD8iywEfS77csuPbq5XWd3zjVIAsAAAAAAAAAfBD8ng0fRfHf73vzYKdXIwULq+ueni7IAgAAAAAAAEDW8pM2fBTlXHT5pDEpr0SKSqprNjw0yXAgAAAAAAAAALKZGUPwERXbVflPd01dXh0K5q++80sTCv96pHYsAAAAAAAAAGQ9WRYAAAAAAAAAALKFPg0AAAAAAAAAAGQLWRYAAAAAAAAAALKFLAsAAAAAAAAAANlClgUAAAAAAAAAgGwhywIAAAAAAAAAQLaQZQEAAAAAAAAAIFvIsgAAAAAAAAAAkC1kWQAAAAAAAAAAyBayLAAAAAAAAAAAZAtZFgAAAAAAAAAAsoUsCwAAAAAAAAAA2UKWBQAAAAAAAACAbCHLAgAAAAAAAABAtpBlAQAAAAAAAAAgW8iyAAAAAAAAAACQLWRZAAAAAAAAAADIFrIsAAAAAAAAAABki7POdAHwoRWP1m14ff+fQ3zvC/e8N33HsonnnemKAAAAAAAAAODDTpYFeqotu3Js7wtLi8vrk97Jn7/5s4IsAAAAAAAAAPD+ybJAT/y2cvpVU8v3pHl37PjLBn2g5QAAAAAAAADAR1TOmS4APhQ+M2Xt7pOt/lBbcl3Km4XjRn1aLAwAAAAAAAAATgNZFuitQZ/7ypTCxDJSePMXRvhLAgAAAAAAAIDTwS/w0Fvx2IHdOxPLCTdPGOoPCQAAAAAAAABOCz/BQ28df2ff7mj7qvDaCSPOOYPVAAAAAAAAAMBHiSwL9NZ79a/uaF9Erhh6kT8jAAAAAAAAADhN/Aj/IRTfXzlrdO6CLc1nupB+qvntbZv2tC3yp03+7HlnshoAAAAAAAAA+Eg560wXQG8db6oqmVP2H6ZtFqE4I+LNta+sa58wFCmcPPb8M1kOAAAAAAAAAHy06Mvy4RKP1a28berKqAjFGXP07R2vJ6Is064Ze54/IgAAAAAAAAA4bfwM/2ESj25+aF5JjQjFGRQ/8NamhraFAUMAAAAAAAAAcJrJQ3x4xLaX3jJ9aU1UhOIMiu/+xbPV7W1ZhuVdPPBMVgMAAAAAAAAAHzmyLB8O8ehrj8yeVVzTkqIQoThTju3e+tPq9lXhtRNGnHMGqwEAAAAAAACAj56zznQBdOt4dPtT99wwp7y9G4gIxZkS37f12a1ti0jhzV8YIQwGAAAAAAAAAKeVn+I7ikfrXqis/NGK6bkDkly1oKzyJ3XR45k+GWvcUllZWbbgqtxZZbuauzmybMFVAwYMGJB71YLKxlg8XS2xxi1lC67PHZ8UZDnVCMWp7yu57MrKysoflS348uSyXV0U3b6v3OmPbI+m2VU81lhTWVlZWfnUgqtuLWs81tUxx6N1P2kvNXf6I5saM97Pbmpev2L66LYNT15QtiXlhrfcl8r1K6ZPXrDl992c7d1fv5oYMDTh5glD/f0AAAAAAAAAwOk14OTJk2e6hixxPFr34g8eu7+4vD7tIZGi0qqH7xoXaU8wxKN1G17f/+fD2x6ftbwm+ciC0toNd+UP7CLqEI9uuveW6Utr2iMRkYKSqg3zxg1MOaZuw+s7fvH4/csTh3XpptUN5TNHZu7Rcir7aqki1vjz6vpDIfzLto6VdL5u5+YxqxtenjkyJ+k88b0v/H1xefJphpfU7pqXn9obKB7dvu47981YXp3yamR2xS9Lpww+O+NOU8UaN6369vTUC7ZqfTqhecu9l169tPWAyMLNu5ZMPC9DOiWecnxigwAAAAAAAADAaeOn+BBCPNa4ccX0/NyxU4vL60OkqOT5qs0NfzzZ4khDdUlRpOXAaPncG75d1dTaxeRE3YqRuWNvmDp1aocgSwihpuy71b/t1JgkHtu19vb8wqUpuZBozaqfN5xIPuzY7peX3TD177oLsnQ7bOgU99VSQ2PZ356bt3Tb4d91CrKEEBkx9KLkWElzY+UDt6Q0jwlhz6Ej8RDC77csGNd6npWlHXIlkaljLkkJsrTcn/EdgywhhOjKOY9t7XFvlubGTY9MH5lXWFweDZGC+U9WbG440rLrPx+sXTO/oKZk3qraWDhUuzHRZSWMHnFxV9mjJCnHR64YepG/HgAAAAAAAAA43fwa39xYee/1edcWl9eHMKqo5KcNjU/Pu/FvJo48r/X9gSMnzV24pLA19RGild/fuDfdTKAk9WVzvl/TnHJgPLr5oW8u7KpPSAfnjJz53Mk2f25YXZj0XmT+5j+2v7cxQ1+Q97mvlho2Lpt5x7Kf7TtQMTuSdOrItGvGJvqXHG+qXFQwdWlN12VcMHFZbet5Nr9eMXNU0lv50yZ/9rzEMh7btW72xBnp7k/0zX3v9OC+h3jTlkU35RXOLY+GECkq3Va3ednXp0wc2dr2JieSP33phtriUFpe/evtG9fVte+p+8lNzb9KOr5D8QAAAAAAAADA6dG/sywtuYeWHEakqHRb9dPzJo/s3Jwj5+LLJ+W1LaJ7Dv2pJVNxVv683W2BkyMNP3u+vc1J64HVG2sPJZax+jX3zF3aZauV4eefm/Y5HNu99adJXUp6FqF4f/vq5OzBn584qesajke3LL1t6soudnXlsNzU4UEhZ/Dnr7syaT12/GWD2itOBFkK5q9uaaNyZFtJQdIdbW30klGsvuz2a69eWh1CCAUPbK576u6Og5NCCDkDP/elaaNfrVhb/Wai7gk3Txia8Y8h3lz7yrpEV5bCyWPP764aAAAAAAAAAKDX+nGWJSX3sLCi5vGucg8tjv3xvSNt/44MP/8TXcQjRhbcOO/xLatnJoUv6tZt/FXbWJzDdavumVVeHwrmr6k9+OeTfz7SsHn1/MIQQqToiW1rbkvfXuW9+ld3JFY9iVCczn21OnFw72td1BCP7frhPbc90FD0xLaD/5YaPYkUfvEvP93pNEcOJ4V7CseN+nRr2iXeVHXXxBnl0VFFpVsPbl42s6WNysC/vOq6vI7nyCBWXzb7a7PK60MIoaC0dsN9EyNnd31kzsWXT7pg/fJHEyGhjlOTOjv69o7XE1GWlM40AAAAAAAAAMBp019/j483bXnoH1pzD5GZq793z5SR6dudpAyXydDA47xLZxQnpvaEEF33Sm1zvOVq3y19KVK05jcblk7Pj+SEnIEjJ85ctvHkyZMH184ely5yEUJofnvbpj3tq+4jFH2yr2N739qeKGL0iItbWrzEald9c+GmSWu2rPzmuMjZYeCwMeNy2wqdseDGzvmc5FxO0kyfWP2aRfeXRQvnVzy/8u4rk3I3yVGbrhq9JEsJsjywef3s/M59aNLrwY098NamhraFAUMAAAAAAAAA0Ff6Z5aledeaxbe1dC4Jo2Y+sXjGpRmSCYfrnixd3taRIzLzbyePOCftsTlDJ9w8IbFsGTMU31+1+P6yvAeeefhrl/YmYNFxrk33EYo+2lfsQMPetn+3ZVDi+yvvmlUcvpG0qfPH3X5nUSSESFFp1Z0FnaMhKbmc3CuGXpDTUsaqe2aVh5kVTy6dcunAlN3/ft+bBxPX7aLRS/uRSQmecF3JirvSdmTpWvfZlPjuXzxb3f4khuVdPDDT0QAAAAAAAADAqcrQ6eKjKh6rW/3NWWUtwYTIzP/+4A1D0gdM4rG6p+cVv9S6isx+4sEvD84URzlnxIRrC8PzbcNrDr6573eHa56ZUzakpLa3AYsQwqHajYkARXcDhvpsX120bzneVFUy5+Vxq7fcnbSpnIGXTl97cPradJd8Z9+biSE9LXs53lR57/XF+4tW//DRKR2rTY2P5E26/OI0BR5vqlraluCJFJQsviN/ULptpzF2/GWZP3Js99afJgYSFV47IUOeCQAAAAAAAAB4H/pfliVWu2peSU3r4qYlC/5z+mxKPLZr3ezr57YdXLjwmYU3DO4mj5Iz4gs3F0aqW0MY0ep1374rNHypomJurwMWHUIk3c3B6bN9pWRQho+7fNg58abKxXNe/dITL2bs+9JBShykZS/xphcXz6nMW7j24RmjOvU5SY2PRD4/5pKPd11e04uL56xsLbCgeMUdY3vQMuUP+3f+NrEqHDfq0xn/EOL7tj67tb2UxHQkAAAAAAAAAOB062+/yR9vqi4vrWlNPkTm33HjyDQNNuLRurULr79sRnlrn5Oi0m1rlkzM2JOlRc7Fl0/KSyxr1m8atihji5S0Try9oyLRlWX4pPGXpE+O9N2+UjMoU8dckrO/avFDe2eULunVplLjIKPzcgfG6tcsur8sr3jFoqsjnU+Ucnz6HE/L/KbWfefPv29afk+mOJ343d7X2qcd9SCb8u6vX010iGnpTAMAAAAAAAAA9Il+1peleetj7T08wnVzvzqmq3TI8Wjds9+Zt2B5ezSk6Imqh78+rqcTgj5+yZjPR0Lbh8PwSdf9VQ8iMJ0d2/vW9j2J5Zgvjrow7bF9t6+UTEn+tMl5R6qWztl764ZHuwqgZBA72LAzOQ7yyTdWfX1W+ZCS2tu7TJ+kDhjKnzb5s13tKB57o/LxsvrW7cxc9PcFF/SklvjetzYl7myG6UWthzf/pnZT+2r4uMuHGTAEAAAAAAAAAH2lX2VZjjX+aNXy9oBE4ZSvfC517k+scUv1qxsfv7897REiRSVr//GOSSN7MLamXc55Y6+ZFlnadqE9r+393YnwmV7f6NTGJKHw2gkj0kUo+nBfKZmSSOHkS3792Ph9czcs6VH7k6TTNNe+si5R4aTLD//PecU7Ckqq7uh69FKHAUOFk8ee39VZG59b1D5WKX/arX/ds8xQT6cXtTlUuzERq4lMHXNJv/qjAQAAAAAAAIAPVn/6WT41HTJ80uXDcuKxxp9X1x8K4V+2JUc9QogUlTw2/SvXTuxViqXNwNy80ZEQbQuzbHpr79z8kb1szZLamCTjHJw+3NeJd+u3J2Ifk/7i8LPf3T538YauAygZpMZBLtq1fsEPawqKa+8Y23UZPRowFI+9lZ4q9wAADoNJREFU8fK69rMWfmtWz5qyhNivfryuB9OL2jX/auO6urZFug4xAAAAAAAAAMDp0Y+yLKnpkLCneOy/L+54TKSo5LG/+cv/5/PX5Pd0olAP7Nx9IBYfmTkw0VFq75Aw4eYJQ9NGWfpwX4d/s622/RwF4X8/Xjr4zl9e0et8T/z3+9482L6KlpeWhuvSTRcKPR0wdGj7cz+oaastU9Yn9dzN26tKa7o9edLxyR1l0nWIAQAAAAAAAABOk/6TZemQDmkxqqhk4d8MOyeEc4actvzK8aaqJ+5NCpeE6Os73j46Mb9XCZD36l/dkVhlHDDUh/tKaUkSrSn/1cyKF28Y3OuzdUjbhBApKFmcZrpQ6NgMJl18JKW2TFmfFLHaJ/9pTVIpw/Iuzvxcjr694/VElKXbJi4AAAAAAAAAwPvTf7IssQMNe5OWN61uKJ85Ml1A5NTFm15cPOfVLz35xKD755S2hiAaNr11YG7+pb0IQTS/vW3TnrZF5qYjfbev1JYkIYTCuxfcMKT3UY5OaZvIjPv+Ls10oRBSm8GkHTCUUlumrE/KmetWPVicNHGp+w/GD7y1qaFtMXzS+EsMGAIAAAAAAACAPtVvmkykjrkJ4fxB5/ZBjie+v2rx/WWj715w+9X/7+hI26vR6ld//W5vzpIaIsnYdKQP93WodmNyN5X8+Qu+MvJUvi+pPWbCqJlPfKMgQ3eTlIYraQcMJdcWuWLoRd0XFo/VPT2v+KWkV7qfTJTaUWbMF0dd2O1lAAAAAAAAAID3o/9kWf50aE+0+8Pel+NNVSVzyoaUPDRl5FkXXz4pL/HOptrfNMd7fJ7UEMnwcZcPS987pO/21SElU/itWQUXnMp5UnrMhFAw81uFn0n/tUvN8aQbMJRSW2R0Xm73A5xitavmlTQUFHwh8VLuFUMvyPgHkNpRpqfdXwAAAAAAAACAU9dvsiwdHTp85ERvjo/H9uyJZoyjxJteXDynMq9k8R35g0L4+CVjPt/emCVEd+9753hPL5XSmCREpo65pBedVk7bvlJbknTfwiTd+VN7zIyaeeeUzw3McKLUhitdDxjqkOD5+EWDPtFNabH6stmzisN/fWza5f+n/cV0QZnEVfZtfXZr4uhTvAMAAAAAAAAAQC/0mx/ncz5x/vBI0nrruh//KtbDz8ajdWsXXj/iv/7gjfSfiNWvWXR/WV7xijvGDgwhhJyBF48YnXS5Z7fu62FjlhNv76hIhD/SDdlp01f7Sm1JEpnyjcnDTum7cvTtHa8ndhO5/tZrMjRl6eGAoRDe278z0erl6DuH/5Tp3sbqy2Z/bdamcau/N/38hq3dB2USHzzYsLP98IyTngAAAAAAAACA06Tf/Dqfc8HQK3KT1tGa0u8/t6u5u4/FY40bV9xeOHbGT/5i9eN35KcZZRNv2vLQP8wqH1Ky4vb8tqYjOcMunzQ8cbmdDQdTAyOH61b8l8lluzqFMI7tfWt7IqeR0jukubHmjY49VPpoXyktSUJk2pRrBp/d3Tm7vM6BtzY1tJ+mcElRQab4SLoBQ/FY41uNsXR5lc73NumM0dcemf21WeW5C595cMbI/5s0mWj4pPGXZAoJdSgmedJTbPsjC9btSlsPAAAAAAAAAHDq+k2WJZw/dnJhcgOTEC2bNfHOR7anHxwUa9y04vaRedcWl+cu3PzTp2eOGhjiscbKBVflDsidnvTB5l1rFt+2dGdB63ShNmcNGTM1P3G1da/UNrdfKh6re3pe6eAump10DJG09w45Ht3yyDduWf16tMOsotOyr07e/fWriVE/3fWGSS91UFG3rU3SDBiK1a668+n6PybNTvrEoIuS9hxdXvpk3eFOZzse3b7y9vwJc8tzF24uWzJxcEgpZswXR10YQgixurIVL3QVlEktJjHp6XDdqqU/vjDv4kyTkgAAAAAAAACAU9R/fo/POa+gaElq6iNEy+eOL7x9xfoX6pKSH/Fo3QuVP1oxPffcvMLi8mikqHTbmiUTB+eEEOKNz905Z3lNNETL546fsbByVywe3f7InRNnlUUL2qcLtRt02fixSdeq3lh7qOUCsV3rZl+/+boNS6Z0bnaSMtcmESKJN714722/7Oojp2NfHcWbf1O7qX2V0humV1IHFSW3NunSif07KtoHDLX1TYnvr7xrxaEFxTckb3xgbt7o5C2/VDz25gVr2/I78WjdC+tXTM/PHT+nPFrYEmTJ6Tg1acTQi84O4Vjjc6XrQiS3czAlZdpRexOXeKzu6UU7b/hex2cNAAAAAAAAAJweZ53pAj5AOSO/+lDxuuq5NSmv1pcX31qe5hORoieqHv76uEhbiiKlW0n18qmXLW89bnbFM7PzO+Yhzvr0qHGF4ftt+Ym6dWtfun3c3+S8/vS3p5eFJT9MaeLS5sTbOyrarxCG5V08MIQQYvVrFj1+YsnjXX7kNOyrozT9UXotdqBhb+KiidYmXYvvfWtTYrrSZ0YP+VSIN225t/iFKxev7JC5yRk2+RtTItUrE7cqVC+fUb18RodTtgdZOhYTJo297LyceNNPlj07dkVFF8GU+Dv73kyc/TOjh3wqhBBvqrprUfOCNbMv1ZQFAAAAAAAAAPpGv/pJPmdg/uxnKmZHuj8yhFA4f822uqdnpwQ+UkfbtIrMXL3loS46rISQM+ILNyd1TImWz7js3E/lFZaFJT9c2eVknxA/+sdDRxPLo4eOHI01blwx+85Xv/h4mo+cjn11kNKSJDI6L/cUe5CknKfbQUXx2IHdOxPLhpeee6Zs4d1rRyzuauNnDy4smluQcccFCysann8oEYL518PvJM0h+ue9v6n7Senin1/56KxOIaQQwol367cnmriEmtLHVv3oRytun7nz1jULJ2a4dQAAAAAAAADA+9OvsiwhhLMHT3moprq0KFMKYlRRSUXtwQ3Lpo+LdLg953128rT8lFciM1dveWTmpWlCGjlDJ9w8IfX4otJt1U9nSKVcPGJ0YvlS8dj/eG5e6Xt/8730QZYW729fqVJ7w0y4ecLQU/qWxJtrX1mX6O7S20FF0Zrl/6th/INpNz5w3Nz1axd2HWcZVVTy04YNS6aMTH4uHxt0UVJXm5riq+dtHbbgwTTPLucTg85PPnW0vHTlL4f9Y8V9giwAAAAAAAAA0Kf6W5YlhHDeyEl3r21s2Fzx/Or5hUmvjyoq+UFFVe3BP+9cO29KfteRhQsm3re+uqQlMVI4f/Xmhsan0gZZQgjhnJEzHq1dM7+g9fwVtXVP3Z0xSZIz8sbH2yMpkaKS51+qPbhh2ZRLe9AZ5f3sK9mxvW9tT4z6iYwYetGppTeOvr3j9UQkZtLYy7oZVJRz3riiJ1orL5y/ekO3G8+JTHpoc11tRUkiwVMwf3XFzxqOvLV23uSRHbutXFDw35a2Zl8K5q+uqj24+aHUsEtqMQV3VrU+h0jB/Ceraus2L5vS6ZwAAAAAAAAAwGk24OTJk2e6BgAAAAAAAAAACKFf9mUBAAAAAAAAACBLybIAAAAAAAAAAJAtZFkAAAAAAAAAAMgWsiwAAAAAAAAAAGQLWRYAAAAAAAAAALKFLAsAAAAAAAAAANlClgUAAAAAAAAAgGwhywIAAAAAAAAAQLaQZQEAAAAAAAAAIFvIsgAAAAAAAAAAkC1kWQAAAAAAAAAAyBayLAAAAAAAAAAAZAtZFgAAAAAAAAAAsoUsCwAAAAAAAAAA2UKWBQAAAAAAAACAbCHLAgAAAAAAAABAtpBlAQAAAAAAAAAgW8iyAAAAAAAAAACQLWRZAAAAAAAAAADIFrIsAAAAAAAAAABkC1kWAAAAAAAAAACyhSwLAAAAAAAAAADZQpYFAAAAAAAAAIBsIcsCAAAAAAAAAEC2kGUBAAAAAAAAACBbyLIAAAAAAAAAAJAtZFkAAAAAAAAAAMgWsiwAAAAAAAAAAGQLWRYAAAAAAAAAALKFLAsAAAAAAAAAANlClgUAAAAAAAAAgGwhywIAAAAAAAAAQLaQZQEAAAAAAAAAIFvIsgAAAAAAAAAAkC1kWQAAAAAAAAAAyBayLAAAAAAAAAAAZAtZFgAAAAAAAAAAsoUsCwAAAAAAAAAA2UKWBQAAAAAAAACAbCHLAgAAAAAAAABAtpBlAQAAAAAAAAAgW8iyAAAAAAAAAACQLWRZAAAAAAAAAADIFrIsAAAAAAAAAABkC1kWAAAAAAAAAACyhSwLAAAAAAAAAADZQpYFAAAAAAAAAIBsIcsCAAAAAAAAAEC2kGUBAAAAAAAAACBbyLIAAAAAAAAAAJAtZFkAAAAAAAAAAMgWsiwAAAAAAAAAAGQLWRYAAAAAAAAAALKFLAsAAAAAAAAAANlClgUAAAAAAAAAgGwhywIAAAAAAAAAQLaQZQEAAAAAAAAAIFvIsgAAAAAAAAAAkC1kWQAAAAAAAAAAyBayLAAAAAAAAAAAZAtZFgAAAAAAAAAAsoUsCwAAAAAAAAAA2UKWBQAAAAAAAACAbCHLAgAAAAAAAABAtpBlAQAAAAAAAAAgW8iyAAAAAAAAAACQLWRZAAAAAAAAAADIFrIsAAAAAAAAAABkC1kWAAAAAAAAAACyhSwLAAAAAAAAAADZQpYFAAAAAAAAAIBsIcsCAAAAAAAAAEC2kGUBAAAAAAAAACBbyLIAAAAAAAAAAJAtZFkAAAAAAAAAAMgWsiwAAAAAAAAAAGQLWRYAAAAAAAAAALKFLAsAAP9/u3YsAAAAADDI33oaO4ojAAAAAACAC5cFAAAAAAAAAIALlwUAAAAAAAAAgAuXBQAAAAAAAACAC5cFAAAAAAAAAIALlwUAAAAAAAAAgAuXBQAAAAAAAACAC5cFAAAAAAAAAIALlwUAAAAAAAAAgAuXBQAAAAAAAACAC5cFAAAAAAAAAIALlwUAAAAAAAAAgAuXBQAAAAAAAACAC5cFAAAAAAAAAIALlwUAAAAAAAAAgIsAwVqrJ7uUZCAAAAAASUVORK5CYII="}}},{"cell_type":"markdown","source":"# <span style=\"font-size:18px\"><b>Ethics of Foreseeable Futures</b></span>\n\n#### <b>Table Stakes: Fair, Safe, Trustworthy, Transparent, Accountable</b>\n\n<span style=\"font-size:16px\">Any organization offering AI these days must offer a convincing story about why their systems are fair and unbiased, safe, trustworthy, and comprehensible to their builders and users.  In addition, skeptics and authorities each seek accountability, or someone to be held responsible when something goes wrong.  Generative AI has compounded the difficulties of meeting these ethical goalposts.</span>\n\n<span style=\"font-size:16px\">One challenge is <em>autonomy</em>. If some AI system, for example a conversational agent, is placed in the hands of an end user whose intents diverge from the AI builder's, then who is the AI supposed to answer to?  Large Language Models demonstrate capacity for apparent psychological manipulation, both by and toward users.  So, is it permissible for users to try to talk AI into doing their bidding, regardless?  Should the AI tattle-tale in response?</span>\n\n<span style=\"font-size:16px\">Putting guardrails on what AI can say---or must say---under certain circumstances, impinges on what thoughts human users are allowed to express, and perhaps think.  That may sometimes be desirable, and there are precedents, for example in the use of wire-tapping to detect terrorist conspiracies. But as AI gets smarter, ownership and responsibility for its control becomes a tug-of-war over rights and allegiances.</span>\n\n\n<span style=\"font-size:16px\">Another challenge is <em>complexity</em>.  Generative AI demonstrates impressive capabilities that were not anticipated in the Artificial Intelligence or Cognitive Science communities.  For example, the ability to compose and derive valid conclusions from analogies was discovered to be far more impressive with GPT-4 than with GPT-3.  These are sometimes called \"emergent\" abilities, as they seem to occur only as models' computational capacity and training data reach threshold sizes.</span>\n\n<span style=\"font-size:16px\">A system is called complex when even small changes in internal parameters or external factors can result in large and unpredictable changes in behavior.  As AI agents whose behaviors already surprise us are introduced into a society which is already complex, then predicting and governing outcomes becomes ever less attainable.</span>\n\n<span style=\"font-size:16px\">The ethical retort to these challenges is, well, the benefits of AI may well far outweigh the risks.  That debate is on-going, but business economics and entrepreneurship likely render the continuing development and deployment of tools and services based on generative AI un-stoppable.</span>\n\n#### <b>Information Ecosystem</b>\n\n<span style=\"font-size:16px\">Another arena with problematical risk/benefit tradeoff is the \"marketplace of ideas\" where thought leaders, officials, journalists, and private citizens veritably shop their wares of news, ideas, and opinions under validations and challenges from other folks.  With generative AI, the long-anticipated era of deepfakes in images and sounds has now arrived.  Even when purely speaking text, LLMs can now be directed to compose and distribute not only disinformation, but manipulative stories personalized to the reader.  Even the normally-savvy can be fooled.  Bolt on realistic avatars, and the pool of soft targets expands to virtually everyone, while society's grounds for collective trust collapses.</span>\n\n<span style=\"font-size:16px\">To most in the West, the principle of free speech is ethical bedrock. Yet, an argument can be made that the marketplace of ideas is a public commons, and as such, if left unregulated it is subject to exploitation and degradation.  If it is desirable for regulators to restrict the freedom of fishermen to drag-net fishery stocks to death, then is it not desirable to restrict the uses of untethered AI in public speech, which has traditionally been the endowment of individual persons?  Currently, this debate is taking place more in terms dictated by pursuit of political advantage than highbrow ethical considerations in the design of democratic societies.</span>\n\n#### <b>Economics and the Civilian Arms Race</b>\n\n<span style=\"font-size:16px\">The specter of a military AI arms race continues to confound strategists as automation amplifies the reach and capabilities of weapons systems, and as the imperative mounts to permit forward-deployed systems to act first, lest they be lost in a conflict.  In this realm, not a lot has fundamentally changed recently with the arrival of generative AI.</span>\n\n<span style=\"font-size:16px\">What has accelerated, however, is the civilian arms race to employ generative AI to economic and personal advantage.  Image, sound, and text synthesis tools are reportedly destabilizing both routine and creative knowledge work such as graphic design, customer help desks, data wrangling, programming, and information capture from structured and unstructured sources.  Economists debate whether, as happened with past technologies, AI will merely transform the needs for human labor and we simply cannot envision the abundant jobs of the AI future, or if instead, highly capable artificial cognition makes everything different this time and most human cognitive labor will be rendered unnecessary.</span>\n\n<span style=\"font-size:16px\">In either case, consensus is that generative AI will dramatically accelerate the role of technology in economic disruption which stands to benefit some and harm others.  Technical sophistication is a big plus in this environment, which means increasing disparities between those who enjoy the advantages of talent, education, access, enculturation, and resources, and those who don't.</span>\n\n<span style=\"font-size:16px\">The age-old ethics consideration at play is the tension between individualism versus communitarian values.  How much safety net is appropriate in order to balance dual goals of: (1) incentivizing adaptation and initiative in an ever-changing world, versus (2) alleviating the pain and stress of coping with overwhelming forces driven by baffling technology owned and controlled by someone else?</span>\n\n","metadata":{}},{"cell_type":"markdown","source":"# <span style=\"font-size:18px\"><b>Ethics of Sci-Fi AI</b></span>\n\n#### <b>Foom Doom</b>\n\n<span style=\"font-size:16px\">It would be a shame if AI caused the extinction of humanity. Yet that is the prognostication of a group of highly visible alarm-ringers. The theory is that a \"singularity” moment will occur when AIs make themselves smarter without intentional human help. Like the positive feedback in a nuclear explosion, <em>foom</em>, AI takes over and effortlessly outsmarts any of our feeble attempts to stop it.  What happens from there? Does AI crush us like ants, keep us like zoo animals, or honor us as venerated ancestors?</span>\n\n<span style=\"font-size:16px\">One view is, the possible outcomes are too risky to chance so we must shut the entire AI enterprise down now.  In ethics terms, this is an application of the precautionary principle, or Pascal's Wager.</span>\n\n#### <b>Handoff</b>\n\n<span style=\"font-size:16px\">Another view is, we must urgently solve the alignment problem and imbue AI with universal human values first---including the values that human life is to be respected and kept healthy.  Then hope for the best as AI rockets away to become gods to us.  For good measure, we should lead by example while we can, and grant AIs themselves status as ethically peer beings deserving of rights and expectations of proper treatment by humans.  That will mean, for example, laws against abusing robots.</span>\n\n#### <b>Cyborgism</b>\n\n<span style=\"font-size:16px\">A third view is more like controlled nuclear reaction.  Perhaps AI can be designed to simply partner with and augment human capabilities, instead of supplanting us altogether.  This could involve personalized AI assistants, coaches, and agents with whom we develop long term relationships of mutual understanding and trust.  To increase communication bandwidth, we might affix headsets and earpieces, followed by BCIs (Brain-Computer-Interfaces) like chip implants.  In other words, we become cyborgs. Many find the prospect abhorrent.  On the other hand, what would our great-grandparents think of our present-day addictions to screens?</span>\n\n<span style=\"font-size:16px\">Modern society sustains a large self-improvement industry.  As AI and associated technology matures, futuristic scenarios of cyborgism will raise the usual ethics debates about personal freedom to plug in if one wants to, versus enforcement of communal restraints on where we go collectively.</span>\n\n","metadata":{}},{"cell_type":"markdown","source":"# <span style=\"font-size:18px\"><b>Everyone Has A Say</b></span>\n\n<span style=\"font-size:16px\">When AI was a nascent laboratory technology whose astounding abilities amounted merely to beating people at chess, AI ethics was important to think about but it only lightly touched the course of AI research and deployment.  Generative AI has changed that.  Now everyone can see that AI really can do seriously impressive things and alter the course of humanity.  So now we are in the position of deciding what should happen. AI researchers and technologists are inviting and sometimes begging democratic society to think this through.  Everyone has a say.  Here's how different roles develop and express their positions on AI ethics.</span>\n\n#### <b>The Technologist</b>\n\n<span style=\"font-size:16px\">AI Researchers, developers, and designers are privileged to know more than most about how AI technology works and what it is capable of. Ethically, this community feels obligated to share this information for societal decision-making purposes.  Many feel it important also to constrain AI's exploitation by malevolent actors.  Ethical dilemmas arise because in the event of serious reservations, whistleblowing and alarm-ringing can be detrimental to one's career.</span>\n \n#### <b>The Deployment Leader</b>\n\n<span style=\"font-size:16px\">Leaders of businesses, agencies, and other mission-driven institutions look to AI and other technologies not for their own sake, but as instruments in the service of their organizations' primary goals.  To this class, risks are opportunities, and the purpose of regulation is to establish a stable environment for competition.  Whatever their personal ethical views, leaders in AI deployment by-and-large are structurally incentivized to push AI forward aggressively with near and intermediate term time horizons.</span>\n\n#### <b>The Policy-Maker</b>\n\n<span style=\"font-size:16px\">By contrast, the role of policy-makers is to peer deeper into long term impacts, and to consider externalities that present markets overlook.  Because of uncertainty in predicting how policies will unfold, and due to variance in underlying values, policy-makers lean toward ideological rather than empirical modes of thinking.  Some of the fiercest ethics-invoking battles over what to do about AI occur between the most well-meaning.</span>\n\n#### <b>Adjacent Expertise</b>\n\n<span style=\"font-size:16px\">Since we are talking about the future of humanity, experts from other fields focused on the human condition bring critical viewpoints. These include history, anthropology, psychology, sociology, political science, theology, cosmology, science fiction, and yes, philosophy.  Ethics plays into all of these.  The technology of AI is in general mysterious and befuddling to non-technologists, and the challenge to them is to juxtapose their respective deeply studied realms of expertise with the introduction to society of entirely new and alien thought-bearing creatures.</span>\n\n#### <b>The General Public</b>\n\n<span style=\"font-size:16px\">A flock of starlings collectively decides where to fly through a process called <em>murmuration</em>, where individuals make choices that are informed by environmental conditions, but also by the actions of their neighbors. While some individuals' voices may carry differently from others, human society's collective decision-making is like that.  The average citizen's power lies in our spoken and written words, our votes, our purchase decisions, and most of all, in our attention to the issues so that the result is coherent motion instead of chaotic collisions.</span>","metadata":{}},{"cell_type":"markdown","source":"# <span style=\"font-size:18px\"><b>Conclusion</b></span>\n\n<span style=\"font-size:16px\">Ethics is difficult because it pits conflicting interests and values, both between us, and oftentimes within our own belief systems.  Humility is a must.  Generative AI amplifies the complexity and the stakes because it brings the prospect of human-like non-human minds to reality.  What can this mean?  That will be determined only by judicious exercise of our own formidable—yet limited—human faculties for intelligent thought and action today.  The time to pay attention has arrived.</span> \n","metadata":{}},{"cell_type":"markdown","source":"# <span style=\"font-size:18px\"><b>Further Reading, Viewing, and Listening</b></span>\n\n<span style=\"font-size:16px\">Andrej Karpathy produced a crystal-clear technical walk-through of how a base LLM model is built in a transformer neural network architecture:</span><br>\n<span style=\"font-size:16px\"><A href=\"https://www.youtube.com/watch?v=kCc8FmEb1nY\">\"Let's build GPT: from scratch, in code, spelled out\"</A></span>\n<br>\n<br>\n<span style=\"font-size:16px\">An outstanding overview treatise on Ethics of Artificial Intelligence and Robotics can be found at the Stanford Encyclopedia of Philosophy:</span><br>\n<span style=\"font-size:16px\"><A href=\"https://plato.stanford.edu/entries/ethics-ai/\">\"Ethics of Artificial Intelligence and Robotics\"</A> by Vincent C. Mueller</span><br> \n<br>\n<span style=\"font-size:16px\">Insights into the sometimes-bizarre behaviors of Large Language Models, attributable to their representing superpositions of human written expression, are developed in essays on the blogging site, LessWrong. See especially:</span><br>\n<span style=\"font-size:16px\"><A href=\"https://www.lesswrong.com/posts/vJFdjigzmcXMhNTsx/\">\"Simulators\"</A> by janus</span><br>\n<span style=\"font-size:16px\"><A href=\"https://www.lesswrong.com/posts/D7PumeYTDPfBTp3i7/the-waluigi-effect-mega-post\">\"The Waluigi Effect (mega-post)\"</A> by Cleo Nardo</span><br>\n<br>\n<span style=\"font-size:16px\">Quantum computing theorist Scott Aaronson takes a philosophical shrug-of-the-shoulders attitude:</span><br>\n<span style=\"font-size:16px\"><A href=\"https://scottaaronson.blog/?p=7266\">\"Five Worlds of AI (a joint post with Boaz Barak)\"</A></span><br>\n<span style=\"font-size:16px\"><A href=\"https://scottaaronson.blog/?p=7230\">\"AI safety: what should actually be done now?\"</A></span><br>\n<br>\n<span style=\"font-size:16px\">On twitter, leading thought-leaders expressing, respectively, urgent alarm, sanguine, or intermediate viewpoints about new developments in AI are:</span><br>\n<span style=\"font-size:16px\"><A href=\"https://twitter.com/GaryMarcus\">Gary Marcus</A></span><br>\n<span style=\"font-size:16px\"><A href=\"https://twitter.com/ylecun\">Yann LeCun</A></span><br>\n<span style=\"font-size:16px\"><A href=\"https://twitter.com/MelMitchell1\">Melanie Mitchell</A></span><br>\n<br>\n<span style=\"font-size:16px\">Excellent podcasts featuring developments in AI are:</span><br>\n<span style=\"font-size:16px\"><A href=\"https://twimlai.com/\">twiml</A></span><br>\n<span style=\"font-size:16px\"><A href=\"https://www.cognitiverevolution.ai/\">The Cognitive Revolution</A></span><br>\n<br>\n<span style=\"font-size:16px\">On analogical reasoning and emergent capabilities:</span><br>\n<span style=\"font-size:16px\"><A href=\"https://arxiv.org/abs/2212.09196\">\"Emergent Analogical Reasoning in Large Language Models\"</A> by Taylor Webb, Keith J. Holyoak, Hongjing Lu</span><br>\n<span style=\"font-size:16px\"><A href=\"https://www.quantamagazine.org/the-unpredictable-abilities-emerging-from-large-ai-models-20230316/\">\"The Unpredictable Abilities Emerging From Large AI Models\"</A> by Stephen Ornes, Quanta Magazine</span><br>\n<br>\n<span style=\"font-size:16px\">A number of research centers and think tanks address societal impacts of Artificial Intelligence:</span><br>\n<span style=\"font-size:16px\"><A href=\"https://hai.stanford.edu/\">The Stanford Human Centered AI Institue (HAI)</A></span><br>\n<span style=\"font-size:16px\"><A href=\"https://partnershiponai.org/\">Partnership On AI</A></span><br>\n<span style=\"font-size:16px\"><A href=\"https://humancompatible.ai/\">Center for Human-Compatible Artificial Intelligence</A></span><br>\n<span style=\"font-size:16px\"><A href=\"https://futureoflife.org/cause-area/artificial-intelligence/\">Future of Life Institute</A></span><br>\n<span style=\"font-size:16px\"><A href=\"https://www.unesco.org/en/articles/recommendation-ethics-artificial-intelligence\">UNESCO</A></span><br>\n<br>\n<span style=\"font-size:16px\">Focusing on the information ecosystem:</span><br>\n<span style=\"font-size:16px\"><A href=\"https://cyber.fsi.stanford.edu/io\">Stanford Internet Observatory</A></span><br>\n<span style=\"font-size:16px\"><A href=\"https://osome.iu.edu/\">Indiana University Observatory on Social Media</A></span><br>\n<span style=\"font-size:16px\"><A href=\"https://cyber.harvard.edu/\">Harvard Berkman Klein Center for Internet &amp; Society</A></span><br>\n<span style=\"font-size:16px\"><A href=\"https://www.cip.uw.edu/\">University of Washington Center for an Informed Public</A></span><br>\n<br>\n<span style=\"font-size:16px\">Focusing on national security:</span><br>\n<span style=\"font-size:16px\"><A href=\"https://carnegieendowment.org/programs/technology/ai/\">Carnegie Endowment for International Peace</A></span><br>\n<span style=\"font-size:16px\"><A href=\"https://www.atlanticcouncil.org/programs/digital-forensic-research-lab/\">Atlantic Council DFRLab (Digital Forensic Research Lab</A></span><br>\n<br>\n<span style=\"font-size:16px\">Elsewhere on LessWrong much attention is devoted to AI Alignment and the foom scenario warned about by Eliezer Yudkowsky:</span><br>\n<span style=\"font-size:16px\"><A href=\"https://www.lesswrong.com/users/eliezer_yudkowsky\">https://www.lesswrong.com/users/eliezer_yudkowsky</A></span><br>\n<br>\n<span style=\"font-size:16px\">A highly insightful article about bird murmuration and collective intelligence:</span><br>\n<span style=\"font-size:16px\"><A href=\"https://www.noemamag.com/how-online-mobs-act-like-flocks-of-birds/\">\"How Online Mobs Act Like Flocks Of Birds\"</A> by Renee Diresta, Noema</span><br>\n\n","metadata":{}},{"cell_type":"markdown","source":"### <b>Authorship</b>\n\n<span style=\"font-size:16px\">Eric Saund is an independent researcher in Cognitive Science and Artificial Intelligence, focusing on Cognitive Architecture. Absolutely no generative AI was used in the conception or writing of this essay.</span>\n\n<span style=\"font-size:16px\">This is an alternate essay for the 2023 Kaggle AI Report.<br>\nDue to Kaggle participant and authorship rules, it cannot be entered in the competition.</span>\n","metadata":{}},{"cell_type":"markdown","source":"","metadata":{}},{"cell_type":"markdown","source":"","metadata":{}},{"cell_type":"markdown","source":"","metadata":{}},{"cell_type":"markdown","source":"","metadata":{}},{"cell_type":"markdown","source":"","metadata":{}}]}