![\"\"](\"https://www.googleapis.com/download/storage/v1/b/kaggle-forum-message-attachments/o/inbox%2F1314380%2F0c386d088aa94d741efb664e950915c7%2FScreenshot%202023-08-04%20at%201.43.26%20PM.png?generation=1691443737535603&alt=media\")
After evaluating all of your submissions, we are excited to announce the winners of the 2023 Kaggle AI Report Competition! We are truly grateful to everyone that shared an essay summarizing what the ML community has learned over the past 2 years.
\nWe will provide awards to the top scoring essays in each of the following categories:
\nIn @abireltaief''s essay \"Contemporary Large Language Models LLMs\", the author provides a fantastic overview of large language models (LLMs), covering LLM prompting, LLM fine-tuning, and LLM application development. A small but well-curated collection of recently published papers are reviewed, and the concepts within are explained in an accessible and approachable manner. Thanks for contributing such an informative essay!
\nHonorable mentions:
\nIn @radbear's essay \"Advances in AI Vision Models in the Last Two Years\", the author discusses recent advances in semantic segmentation, and the development of vision transformers and continual learning methods, amongst other topics. The essay is well-referenced and provides a wealth of resources for those that want to dig into these topics in more detail. Thanks @radbear!
\nHonorable mentions:
\nIn @rhysie's essay \"Learnings From the Typical Tabular Pipeline\", the author does a fantastic job of exploring the connections between the methods used in Kaggle competitions and the methods discussed in the academic literature. We really appreciated the deep-dive into Kaggle's monthly tabular playground series!
\n
Honorable mentions:
\nIn @iamleonie's essay \"Towards Green AI\", the author gives an overview of some AI research areas that are relevant to carbon footprints and the environment, and then explores the subset of Kaggle Competitions that specifically encourage submitting small and efficient models. Methods that can be used to reduce the carbon footprint of ML models are reviewed, and the essay leaves us feeling inspired to build more efficient ML systems. Thanks @iamleonie!
\nHonorable mentions:
\nIn @trushk's essay \"2023 Kaggle AI Report - Generative AI\", the author surveys the landscape of generative AI products that have been released over the past few years. These exciting tools allow ML practitioners and even novices to generate images, text, videos, or code, and this essay provides a great overview of the major advancements that occurred along the way.
\nHonorable mentions:
\nIn Team TISL's essay \"Exploring the landscape of AI Ethics\", the author introduces and explains a diverse set of concepts related to ethics and AI. Terms such as bias, fairness, equity, and explainability are introduced and explained, and pressing issues are discussed, such as climate change, copyright issues, and government regulation.
\n![\"\"](\"https://www.googleapis.com/download/storage/v1/b/kaggle-forum-message-attachments/o/inbox%2F1314380%2F683e981585d6f5d290d4f2db4b5a0015%2FScreenshot%202023-08-04%20at%202.58.25%20PM.png?generation=1691443712646125&alt=media\")
Honorable mentions:
\nIn @pluvias's essay, \"Kaggle AI Report Optimization\" the author explores recent advances in the fundamental area of optimization algorithms. The essay begins by giving an overview of more classical techniques, prior to moving on to highlight more recent advances such as MoMo and Sophia. Great work putting together this excellent resource @pluvia!
\nHonorable mentions:
\nThe Kaggle Team will reach out to the winners with instructions on how to collect their prizes.
", "messages": [ { "id": 2378929, "postDate": "2023-08-07T21:31:20.107Z", "content": "After evaluating all of your submissions, we are excited to announce the winners of the 2023 Kaggle AI Report Competition! We are truly grateful to everyone that shared an essay summarizing what the ML community has learned over the past 2 years.
\nWe will provide awards to the top scoring essays in each of the following categories:
\nIn @abireltaief''s essay \"Contemporary Large Language Models LLMs\", the author provides a fantastic overview of large language models (LLMs), covering LLM prompting, LLM fine-tuning, and LLM application development. A small but well-curated collection of recently published papers are reviewed, and the concepts within are explained in an accessible and approachable manner. Thanks for contributing such an informative essay!
\nHonorable mentions:
\nIn @radbear's essay \"Advances in AI Vision Models in the Last Two Years\", the author discusses recent advances in semantic segmentation, and the development of vision transformers and continual learning methods, amongst other topics. The essay is well-referenced and provides a wealth of resources for those that want to dig into these topics in more detail. Thanks @radbear!
\nHonorable mentions:
\nIn @rhysie's essay \"Learnings From the Typical Tabular Pipeline\", the author does a fantastic job of exploring the connections between the methods used in Kaggle competitions and the methods discussed in the academic literature. We really appreciated the deep-dive into Kaggle's monthly tabular playground series!
\n
Honorable mentions:
\nIn @iamleonie's essay \"Towards Green AI\", the author gives an overview of some AI research areas that are relevant to carbon footprints and the environment, and then explores the subset of Kaggle Competitions that specifically encourage submitting small and efficient models. Methods that can be used to reduce the carbon footprint of ML models are reviewed, and the essay leaves us feeling inspired to build more efficient ML systems. Thanks @iamleonie!
\nHonorable mentions:
\nIn @trushk's essay \"2023 Kaggle AI Report - Generative AI\", the author surveys the landscape of generative AI products that have been released over the past few years. These exciting tools allow ML practitioners and even novices to generate images, text, videos, or code, and this essay provides a great overview of the major advancements that occurred along the way.
\nHonorable mentions:
\nIn Team TISL's essay \"Exploring the landscape of AI Ethics\", the author introduces and explains a diverse set of concepts related to ethics and AI. Terms such as bias, fairness, equity, and explainability are introduced and explained, and pressing issues are discussed, such as climate change, copyright issues, and government regulation.
\n
Honorable mentions:
\nIn @pluvias's essay, \"Kaggle AI Report Optimization\" the author explores recent advances in the fundamental area of optimization algorithms. The essay begins by giving an overview of more classical techniques, prior to moving on to highlight more recent advances such as MoMo and Sophia. Great work putting together this excellent resource @pluvia!
\nHonorable mentions:
\nThe Kaggle Team will reach out to the winners with instructions on how to collect their prizes.
", "rawMarkdown": "# Congratulations everyone!\n\nAfter evaluating all of your submissions, we are excited to announce the winners of the [2023 Kaggle AI Report Competition](https://www.kaggle.com/competitions/2023-kaggle-ai-report/)! We are truly grateful to everyone that shared an essay summarizing what the ML community has learned over the past 2 years. \n\nWe will provide awards to the [top scoring](https://www.kaggle.com/competitions/2023-kaggle-ai-report/overview/evaluation) essays in each of the following categories: \n- Text data\n- Image and/or video data\n- Tabular and/or time series data\n- Kaggle Competitions\n- Generative AI\n- AI ethics\n- Other \n\n\n# Text data:\n\nIn @abireltaief''s essay \"[Contemporary Large Language Models LLMs](https://www.kaggle.com/code/abireltaief/contemporary-large-language-models-llms)\", the author provides a fantastic overview of large language models (LLMs), covering LLM prompting, LLM fine-tuning, and LLM application development. A small but well-curated collection of recently published papers are reviewed, and the concepts within are explained in an accessible and approachable manner. Thanks for contributing such an informative essay!\n\nHonorable mentions: \n - http://www.kaggle.com/code/mistylight/mini-giants-small-language-models\n - https://www.kaggle.com/code/flaussy/large-language-models-reasoning-ability\n \n# Image and/or video data:\n\nIn @radbear's essay \"[Advances in AI Vision Models in the Last Two Years](https://www.kaggle.com/code/radbear/advances-in-ai-vision-models-in-the-last-two-years)\", the author discusses recent advances in semantic segmentation, and the development of vision transformers and continual learning methods, amongst other topics. The essay is well-referenced and provides a wealth of resources for those that want to dig into these topics in more detail. Thanks @radbear!\n\nHonorable mentions: \n - https://www.kaggle.com/code/dokster/image-and-video-data-kaggle-ai-report\n - https://www.kaggle.com/omarrajaa/image-data-report\n\t\n# Tabular and/or time series data\n\n\nIn @rhysie's essay \"[Learnings From the Typical Tabular Pipeline](https://www.kaggle.com/code/rhysie/learnings-from-the-typical-tabular-pipeline)\", the author does a fantastic job of exploring the connections between the methods used in Kaggle competitions and the methods discussed in the academic literature. We really appreciated the deep-dive into Kaggle's monthly tabular playground series!\n\n\n\nHonorable mentions: \n - https://www.kaggle.com/code/kensverse/ai-report\n - https://www.kaggle.com/code/kobbiemanrique/2023-kaggle-ai-report-tabular-data\n\n# Kaggle Competitions\n\nIn @iamleonie's essay \"[Towards Green AI](https://www.kaggle.com/code/iamleonie/towards-green-ai/notebook)\", the author gives an overview of some AI research areas that are relevant to carbon footprints and the environment, and then explores the subset of Kaggle Competitions that specifically encourage submitting small and efficient models. Methods that can be used to reduce the carbon footprint of ML models are reviewed, and the essay leaves us feeling inspired to build more efficient ML systems. Thanks @iamleonie!\n\nHonorable mentions: \n - https://www.kaggle.com/code/thedrcat/how-to-win-a-kaggle-competition\n - https://www.kaggle.com/code/nghihuynh/kaggle-ai-report-medical-imaging-competitions\n\n\n\n# Generative AI\n\nIn @trushk's essay \"[2023 Kaggle AI Report - Generative AI](https://www.kaggle.com/code/trushk/2023-kaggle-ai-report-generative-ai)\", the author surveys the landscape of generative AI products that have been released over the past few years. These exciting tools allow ML practitioners and even novices to generate images, text, videos, or code, and this essay provides a great overview of the major advancements that occurred along the way.\n\nHonorable mentions: \n - https://www.kaggle.com/code/jay2333/understand-generate-and-transform-the-world\n - https://www.kaggle.com/code/pranavbelhekar/a-glimpse-into-the-realm-of-generative-ai/notebook\n\n# AI ethics\n\nIn Team TISL's essay \"[Exploring the landscape of AI Ethics](https://www.kaggle.com/code/patrikkenfack/exploring-the-landscape-of-ai-ethics)\", the author introduces and explains a diverse set of concepts related to ethics and AI. Terms such as bias, fairness, equity, and explainability are introduced and explained, and pressing issues are discussed, such as climate change, copyright issues, and government regulation. \n\n\n\nHonorable mentions: \n - https://www.kaggle.com/code/antongc/developments-in-ai-and-ethics-in-the-past-2-years\n - https://www.kaggle.com/code/shreyamishra610/ethical-ai-is-all-we-need\n\n\n\n# Other\n\nIn @pluvias's essay, \"[Kaggle AI Report Optimization](https://www.kaggle.com/code/pluvias/kaggle-ai-report-optimization/notebook)\" the author explores recent advances in the fundamental area of optimization algorithms. The essay begins by giving an overview of more classical techniques, prior to moving on to highlight more recent advances such as MoMo and Sophia. Great work putting together this excellent resource @pluvia! \n \nHonorable mentions: \n - https://www.kaggle.com/code/redberry/ai-models-for-biosciences-and-virus-data/\n - https://www.kaggle.com/code/diegoexe/kaggle-ai-report-medical-data/notebook\n - https://www.kaggle.com/code/hodajalali/graph-learning-and-complex-networks\n\n# Great work everyone!\nThe Kaggle Team will reach out to the winners with instructions on how to collect their prizes.\n\n", "votes": 51 }, { "id": 2379527, "postDate": "2023-08-08T07:45:32.797Z", "content": "Wow, thank you so much to the Kaggle team and the judges for selecting my entry as first place in the \"Kaggle Competitions\" category. I am especially honored because I thought there were some high-quality and impressive entries in this category.
", "rawMarkdown": "Wow, thank you so much to the Kaggle team and the judges for selecting my entry as first place in the \"Kaggle Competitions\" category. I am especially honored because I thought there were some high-quality and impressive entries in this category.", "votes": 6, "replies": [ { "id": 2379609, "postDate": "2023-08-08T08:25:47.510Z", "content": "Thanks for your humility:) . Your work is impressive too!: simple, highly enticing (efficient models, carbon footprint…), and original…very well deserved
", "rawMarkdown": "Thanks for your humility:) . Your work is impressive too!: simple, highly enticing (efficient models, carbon footprint...), and original...very well deserved\n", "votes": 1, "replies": [ { "id": 2379984, "postDate": "2023-08-08T11:27:03.897Z", "content": "Thank you so much for the kind words @abireltaief!
", "rawMarkdown": "Thank you so much for the kind words @abireltaief!" } ] }, { "id": 2379739, "postDate": "2023-08-08T09:10:10.977Z", "content": "Congratulations for the winning the essay competition in \"Kaggle Competitions\" category.
\nI recently published a notebook on \" Game Theory applications for ML competitions\"
\nYour prize winning NB will be useful reference.
\nhttps://www.kaggle.com/code/crsuthikshnkumar/gametheory-appln-ml-competitions-crsk
Congratulations, Leonie! A well-deserved winning essay in the \"Kaggle Competitions\" category!
", "rawMarkdown": "Congratulations, Leonie! A well-deserved winning essay in the \"Kaggle Competitions\" category!" } ] }, { "id": 2378937, "postDate": "2023-08-07T21:41:20.470Z", "content": "Congrats all the winners! @paultimothymooney can the teams access to their final evaluation rubric scores? The final LB was unexpected for us actually, so we would be very happy to know our weaknesses.
", "rawMarkdown": "Congrats all the winners! @paultimothymooney can the teams access to their final evaluation rubric scores? The final LB was unexpected for us actually, so we would be very happy to know our weaknesses.", "votes": 6, "replies": [ { "id": 2379042, "postDate": "2023-08-08T00:36:01.527Z", "content": "I second the request by @nlztrk for the final rubric @paultimothymooney I’m honestly baffled by the results.
", "rawMarkdown": "I second the request by @nlztrk for the final rubric @paultimothymooney I’m honestly baffled by the results.", "votes": 2 } ] }, { "id": 2381497, "postDate": "2023-08-09T08:18:50.147Z", "content": "I'm delighted to have won 1st place in the \"Text-data\" category. In addition, I'd like to extend my congratulations to the 6 winners in the other categories, as well as to everyone who received an honorable mention and all the competitors.
\nAdditionally, I'd like to congratulate the 7 GrandMasters who reviewed all of the entries as part of the judging panel for their professionalism, as well as the Kaggle team: Paul and the staff who worked behind the scenes.
\nI really feel like a double winner ? : yes, not only because I'm the winner in the category \"Text data\" but also because it's in lifting others that we can really feel it (A team to whom I gave meticulous, highly constructive feedback received an honorable mention: perhaps time did not allow them to include everything, but the final version has changed a lot and is really good.): Thank you, Kaggle, for including the feature of constructive peer-feedback!
\nOther submissions that are neither winners nor honorable mentions are nevertheless quite informative to read. The same is true in real life: you can learn from everyone, so I want to express my gratitude for all competitors and the whole incredible Kaggle community.
\nFurthermore, everyone has the ability and the potential to win at some point, so keep trying, inspiring the community, and taking part in these fantastic competitions. Best
Wow this was a pleasant surprise! I was able to get the first place in the \"Generative AI\" category which had a lot of very good entries. Thank you Paul and the Kaggle Team for hosting this competition which resulted in many high quality reports. Hope to see more of these competitions in the future! Also thanks to the KGM team that spent time reading and evaluating everyones report !
", "rawMarkdown": "Wow this was a pleasant surprise! I was able to get the first place in the \"Generative AI\" category which had a lot of very good entries. Thank you Paul and the Kaggle Team for hosting this competition which resulted in many high quality reports. Hope to see more of these competitions in the future! Also thanks to the KGM team that spent time reading and evaluating everyones report ! ", "votes": 2 }, { "id": 2379097, "postDate": "2023-08-08T02:05:49.350Z", "content": "Congratulations to the winners!
\nBut I'm extremely perplexed and dismayed by the results; there were numerous notebooks that were far more helpful, received more votes from the audience, and were forked many times by the participants themselves. And the audience only upvotes a notebook if they find it relevant and useful. So, what is the point of upvotes if they are ignored? If this is the case, the competitions should not include an upvote mechanism!
I have also noticed that the majority of essays that received high scores did not include the following prompt, which was highlighted below by @paultimothymooney.
\n\n\nUltimately the goal is to respond to one of the 7x prompts in a way that will score highest against the evaluation rubric, and the prompt begins with \"what the ML community has learned over the past 2 years of working with…\", which is a fairly broad mandate.
\n
Additionally, I've seen that few of the high-scoring individuals and teams have notebooks that cover multiple subjects, which is unfair.
\n\n\nEach team should write a single essay focusing on a single topic.
\n
And it would have been much better if the scores on the eligibility checklist had been included with the results so that the participants could observe where they performed poorly and work to improve for the upcoming competitions.
\nI wish the grading panel had double-checked these details before naming the winners.
Completely agree @sanjushasuresh, I'm honestly confused at some of the winning solutions compared to ones that I know were actual submissions and were of much higher quality. Especially in the 'Other' category. The winning essay mostly goes over gradient descent methods most of which are much older than the last two years. The writing is not particularly impressive either when compared to other entries. I don't mean to criticize the winner, but I just don't consider some of these $10,000 essays. 🤷♂️
", "rawMarkdown": "Completely agree @sanjushasuresh, I'm honestly confused at some of the winning solutions compared to ones that I know were actual submissions and were of much higher quality. Especially in the 'Other' category. The winning essay mostly goes over gradient descent methods most of which are much older than the last two years. The writing is not particularly impressive either when compared to other entries. I don't mean to criticize the winner, but I just don't consider some of these $10,000 essays. 🤷♂️", "votes": 2, "replies": [ { "id": 2379436, "postDate": "2023-08-08T06:51:21.680Z", "rawMarkdown": "", "isDeleted": true }, { "id": 2379728, "postDate": "2023-08-08T09:06:30.133Z", "content": "A lawyer, really? By the way, you have the right to protest and express yourself, just not in that way. You don't need to use force to get your point through or your voice heard, just as you don't have to speak loudly for people to hear you. I don't normally get involved in such kinds of conversations, but occasionally I do.
", "rawMarkdown": "A lawyer, really? By the way, you have the right to protest and express yourself, just not in that way. You don't need to use force to get your point through or your voice heard, just as you don't have to speak loudly for people to hear you. I don't normally get involved in such kinds of conversations, but occasionally I do." }, { "id": 2379859, "postDate": "2023-08-08T10:20:47.930Z", "content": "I agree with @jonbown I was going through the article just now and it's weird that the competition which explicitly mention we have to focus on recent advancements in last two years chose the winning article where major part of the article is based on the advancements made much prior in time. No doubt, the article is helpful and educative but I wonder what about this statement then?
\n\n\nThe goal of this competition is to tap into the diverse expertise of the Kaggle community to centralize and summarize the rapid advancements in AI from the past two years.
\n
There are so many notebooks in every category that actually went in-depth of the last 2 years advancements and are ranked much much lower.
", "rawMarkdown": "I agree with @jonbown I was going through the article just now and it's weird that the competition which explicitly mention we have to focus on recent advancements in last two years chose the winning article where major part of the article is based on the advancements made much prior in time. No doubt, the article is helpful and educative but I wonder what about this statement then? \n>The goal of this competition is to tap into the diverse expertise of the Kaggle community to centralize and summarize the rapid advancements in AI from the past two years. \n\nThere are so many notebooks in every category that actually went in-depth of the last 2 years advancements and are ranked much much lower.", "votes": 7 }, { "id": 2379980, "postDate": "2023-08-08T11:25:19.413Z", "rawMarkdown": "", "isDeleted": true }, { "id": 2380031, "postDate": "2023-08-08T11:47:01.160Z", "content": "Not to take anything from the winners - I know everyone worked hard. But I also completely disagree with the evaluation and the final ranking. I believe there are better works than the top 7, and we need full disclosure on the points, and the reasoning behind each.
", "rawMarkdown": "Not to take anything from the winners - I know everyone worked hard. But I also completely disagree with the evaluation and the final ranking. I believe there are better works than the top 7, and we need full disclosure on the points, and the reasoning behind each." } ] } ] }, { "id": 2388524, "postDate": "2023-08-13T12:55:48.750Z", "content": "Thank you for organizing this competition and congratulations to the winners !
\nI'm a bit disappointed with the results, I was expecting to at least have the points detailed in an evaluation grid.
", "rawMarkdown": "Thank you for organizing this competition and congratulations to the winners !\n\nI'm a bit disappointed with the results, I was expecting to at least have the points detailed in an evaluation grid.", "votes": 2 }, { "id": 2379019, "postDate": "2023-08-08T00:12:14.650Z", "content": "Congratulations for the winners. Thanks for sharing the info on top performing teams.
", "rawMarkdown": "Congratulations for the winners. Thanks for sharing the info on top performing teams." }, { "id": 2380993, "postDate": "2023-08-08T22:58:05.813Z", "content": "Congratulations to all the winners. The effort and merit is evident in all 7 essays.
\nAbout the complaints, we all knew what we where signing up for when we entered a subjective writing competition--where public relations efforts and individual considerations from the Kaggle team will be weighed. However, I still support the proposal for releasing the rubric for each essay for the learning value and experience it will provide.
\nI'm guessing the transparency will also alleviate the perceptions of unfairness that a lot of people are voicing here and on LinkedIn.
", "rawMarkdown": "Congratulations to all the winners. The effort and merit is evident in all 7 essays. \n\nAbout the complaints, we all knew what we where signing up for when we entered a subjective writing competition--where public relations efforts and individual considerations from the Kaggle team will be weighed. However, I still support the proposal for releasing the rubric for each essay for the learning value and experience it will provide. \n\nI'm guessing the transparency will also alleviate the perceptions of unfairness that a lot of people are voicing here and on LinkedIn.", "votes": 1 }, { "id": 2380639, "postDate": "2023-08-08T17:28:06.677Z", "content": "Thanks to all the participants and congratulations to the winners of Kaggle AI report 2023! Now we have a lot of essays which make us to think about the main destinations of AI development. I'm sure that all the participants were working hard during preparing the essays. And I'm sure that no less hard work was made by 7 Grand Masters during evaluating the essays. Trying to look at the evaluation process from the 7 GM perspective I understand that it was really hard to choose which essay's composition was more accurate & concise, and discussion was more thorough & detailed. I trust to the best 7 Kaggle thinkers in their grading. The requirement for scoring transparency sounds fair (similar to the requirements of transparency in AI decisions :)) ) but I don't think that it'll change the leaderboard. In other hand it can help to improve the essays writing skills in the future.
\nSpecial thanks to the Kaggle team for the opportunity to reflect publicly on one of the most hot topics of our days.
Is it possible to know category-wise rankings and score?
", "rawMarkdown": "Is it possible to know category-wise rankings and score?", "votes": 1 }, { "id": 2379423, "postDate": "2023-08-08T06:45:04.837Z", "content": "Congratulations to the winning teams ! I'm so happy to participate in this competition. I'm glad if you could provide our scores seperately according to the rubric because that'll be really helpful to improve more our writings at least through an email.
\nThank you!
", "rawMarkdown": "Congratulations to the winning teams ! I'm so happy to participate in this competition. I'm glad if you could provide our scores seperately according to the rubric because that'll be really helpful to improve more our writings at least through an email. \n\nThank you!", "votes": 1 }, { "id": 2378996, "postDate": "2023-08-07T23:27:44.807Z", "content": "Congratulations to all the winners and kudos to everyone who participated! What an exciting adventure this was. Hope Kaggle plans on hosting many more essay competitions in the future.
", "rawMarkdown": "Congratulations to all the winners and kudos to everyone who participated! What an exciting adventure this was. Hope Kaggle plans on hosting many more essay competitions in the future.", "votes": 1 }, { "id": 2387992, "postDate": "2023-08-13T05:29:36.890Z", "content": "🔥🔥🔥🔥🔥🔥🔥🔥
", "rawMarkdown": "🔥🔥🔥🔥🔥🔥🔥🔥", "votes": -1 }, { "id": 2381611, "postDate": "2023-08-09T09:31:54.997Z", "content": "Great announcement @paultimothymooney , congratulations to all winners. Best of luck for all
", "rawMarkdown": "Great announcement @paultimothymooney , congratulations to all winners. Best of luck for all ", "votes": -1 }, { "id": 2379179, "postDate": "2023-08-08T04:10:22.407Z", "content": "Congrats for all the winners and efforts taken by all the participants ! 💫
", "rawMarkdown": "Congrats for all the winners and efforts taken by all the participants ! 💫", "votes": -1 }, { "id": 3076720, "postDate": "2024-12-20T08:03:40.733Z", "content": "Congrats Leonie !!
", "rawMarkdown": "Congrats Leonie !!\n" }, { "id": 2380656, "postDate": "2023-08-08T17:35:56.650Z", "rawMarkdown": "", "votes": -1, "isDeleted": true }, { "id": 2379992, "postDate": "2023-08-08T11:30:32.900Z", "rawMarkdown": "", "isDeleted": true }, { "id": 2379415, "postDate": "2023-08-08T06:41:06.773Z", "rawMarkdown": "", "isDeleted": true } ], "comments": [ { "id": 2379527, "author_name": "Leonie", "author_url": "", "post_date": "2023-08-08T07:45:32.797000", "content": "Wow, thank you so much to the Kaggle team and the judges for selecting my entry as first place in the \"Kaggle Competitions\" category. I am especially honored because I thought there were some high-quality and impressive entries in this category.
", "votes": 6, "replies": [ { "id": 2379609, "author_name": "", "author_url": "", "post_date": "2023-08-08T08:25:47.510000", "content": "Thanks for your humility:) . Your work is impressive too!: simple, highly enticing (efficient models, carbon footprint…), and original…very well deserved
", "votes": 1, "replies": [ { "id": 2379984, "author_name": "Leonie", "author_url": "", "post_date": "2023-08-08T11:27:03.897000", "content": "Thank you so much for the kind words @abireltaief!
", "votes": 0, "replies": [] } ] }, { "id": 2379739, "author_name": "C R Suthikshn Kumar", "author_url": "", "post_date": "2023-08-08T09:10:10.977000", "content": "Congratulations for the winning the essay competition in \"Kaggle Competitions\" category.
\nI recently published a notebook on \" Game Theory applications for ML competitions\"
\nYour prize winning NB will be useful reference.
\nhttps://www.kaggle.com/code/crsuthikshnkumar/gametheory-appln-ml-competitions-crsk
Congratulations, Leonie! A well-deserved winning essay in the \"Kaggle Competitions\" category!
", "votes": 0, "replies": [] } ] }, { "id": 2378937, "author_name": "Anil Ozturk", "author_url": "", "post_date": "2023-08-07T21:41:20.470000", "content": "Congrats all the winners! @paultimothymooney can the teams access to their final evaluation rubric scores? The final LB was unexpected for us actually, so we would be very happy to know our weaknesses.
", "votes": 6, "replies": [ { "id": 2379042, "author_name": "Jonathan Bown", "author_url": "", "post_date": "2023-08-08T00:36:01.527000", "content": "I second the request by @nlztrk for the final rubric @paultimothymooney I’m honestly baffled by the results.
", "votes": 2, "replies": [] } ] }, { "id": 2381497, "author_name": "", "author_url": "", "post_date": "2023-08-09T08:18:50.147000", "content": "I'm delighted to have won 1st place in the \"Text-data\" category. In addition, I'd like to extend my congratulations to the 6 winners in the other categories, as well as to everyone who received an honorable mention and all the competitors.
\nAdditionally, I'd like to congratulate the 7 GrandMasters who reviewed all of the entries as part of the judging panel for their professionalism, as well as the Kaggle team: Paul and the staff who worked behind the scenes.
\nI really feel like a double winner ? : yes, not only because I'm the winner in the category \"Text data\" but also because it's in lifting others that we can really feel it (A team to whom I gave meticulous, highly constructive feedback received an honorable mention: perhaps time did not allow them to include everything, but the final version has changed a lot and is really good.): Thank you, Kaggle, for including the feature of constructive peer-feedback!
\nOther submissions that are neither winners nor honorable mentions are nevertheless quite informative to read. The same is true in real life: you can learn from everyone, so I want to express my gratitude for all competitors and the whole incredible Kaggle community.
\nFurthermore, everyone has the ability and the potential to win at some point, so keep trying, inspiring the community, and taking part in these fantastic competitions. Best
Wow this was a pleasant surprise! I was able to get the first place in the \"Generative AI\" category which had a lot of very good entries. Thank you Paul and the Kaggle Team for hosting this competition which resulted in many high quality reports. Hope to see more of these competitions in the future! Also thanks to the KGM team that spent time reading and evaluating everyones report !
", "votes": 2, "replies": [] }, { "id": 2379097, "author_name": "Sanjusha Suresh", "author_url": "", "post_date": "2023-08-08T02:05:49.350000", "content": "Congratulations to the winners!
\nBut I'm extremely perplexed and dismayed by the results; there were numerous notebooks that were far more helpful, received more votes from the audience, and were forked many times by the participants themselves. And the audience only upvotes a notebook if they find it relevant and useful. So, what is the point of upvotes if they are ignored? If this is the case, the competitions should not include an upvote mechanism!
I have also noticed that the majority of essays that received high scores did not include the following prompt, which was highlighted below by @paultimothymooney.
\n\n\nUltimately the goal is to respond to one of the 7x prompts in a way that will score highest against the evaluation rubric, and the prompt begins with \"what the ML community has learned over the past 2 years of working with…\", which is a fairly broad mandate.
\n
Additionally, I've seen that few of the high-scoring individuals and teams have notebooks that cover multiple subjects, which is unfair.
\n\n\nEach team should write a single essay focusing on a single topic.
\n
And it would have been much better if the scores on the eligibility checklist had been included with the results so that the participants could observe where they performed poorly and work to improve for the upcoming competitions.
\nI wish the grading panel had double-checked these details before naming the winners.
Completely agree @sanjushasuresh, I'm honestly confused at some of the winning solutions compared to ones that I know were actual submissions and were of much higher quality. Especially in the 'Other' category. The winning essay mostly goes over gradient descent methods most of which are much older than the last two years. The writing is not particularly impressive either when compared to other entries. I don't mean to criticize the winner, but I just don't consider some of these $10,000 essays. 🤷♂️
", "votes": 2, "replies": [ { "id": 2379436, "author_name": "", "author_url": "", "post_date": "2023-08-08T06:51:21.680000", "content": "", "votes": 0, "replies": [] }, { "id": 2379728, "author_name": "", "author_url": "", "post_date": "2023-08-08T09:06:30.133000", "content": "A lawyer, really? By the way, you have the right to protest and express yourself, just not in that way. You don't need to use force to get your point through or your voice heard, just as you don't have to speak loudly for people to hear you. I don't normally get involved in such kinds of conversations, but occasionally I do.
", "votes": 0, "replies": [] }, { "id": 2379859, "author_name": "AC", "author_url": "", "post_date": "2023-08-08T10:20:47.930000", "content": "I agree with @jonbown I was going through the article just now and it's weird that the competition which explicitly mention we have to focus on recent advancements in last two years chose the winning article where major part of the article is based on the advancements made much prior in time. No doubt, the article is helpful and educative but I wonder what about this statement then?
\n\n\nThe goal of this competition is to tap into the diverse expertise of the Kaggle community to centralize and summarize the rapid advancements in AI from the past two years.
\n
There are so many notebooks in every category that actually went in-depth of the last 2 years advancements and are ranked much much lower.
", "votes": 7, "replies": [] }, { "id": 2379980, "author_name": "", "author_url": "", "post_date": "2023-08-08T11:25:19.413000", "content": "", "votes": 0, "replies": [] }, { "id": 2380031, "author_name": "Xhoni Shollaj", "author_url": "", "post_date": "2023-08-08T11:47:01.160000", "content": "Not to take anything from the winners - I know everyone worked hard. But I also completely disagree with the evaluation and the final ranking. I believe there are better works than the top 7, and we need full disclosure on the points, and the reasoning behind each.
", "votes": 0, "replies": [] } ] } ] }, { "id": 2388524, "author_name": "Jo", "author_url": "", "post_date": "2023-08-13T12:55:48.750000", "content": "Thank you for organizing this competition and congratulations to the winners !
\nI'm a bit disappointed with the results, I was expecting to at least have the points detailed in an evaluation grid.
", "votes": 2, "replies": [] }, { "id": 2379019, "author_name": "C R Suthikshn Kumar", "author_url": "", "post_date": "2023-08-08T00:12:14.650000", "content": "Congratulations for the winners. Thanks for sharing the info on top performing teams.
", "votes": 0, "replies": [] }, { "id": 2380993, "author_name": "Adrian Dip", "author_url": "", "post_date": "2023-08-08T22:58:05.813000", "content": "Congratulations to all the winners. The effort and merit is evident in all 7 essays.
\nAbout the complaints, we all knew what we where signing up for when we entered a subjective writing competition--where public relations efforts and individual considerations from the Kaggle team will be weighed. However, I still support the proposal for releasing the rubric for each essay for the learning value and experience it will provide.
\nI'm guessing the transparency will also alleviate the perceptions of unfairness that a lot of people are voicing here and on LinkedIn.
", "votes": 1, "replies": [] }, { "id": 2380639, "author_name": "Veronika Sarikaya", "author_url": "", "post_date": "2023-08-08T17:28:06.677000", "content": "Thanks to all the participants and congratulations to the winners of Kaggle AI report 2023! Now we have a lot of essays which make us to think about the main destinations of AI development. I'm sure that all the participants were working hard during preparing the essays. And I'm sure that no less hard work was made by 7 Grand Masters during evaluating the essays. Trying to look at the evaluation process from the 7 GM perspective I understand that it was really hard to choose which essay's composition was more accurate & concise, and discussion was more thorough & detailed. I trust to the best 7 Kaggle thinkers in their grading. The requirement for scoring transparency sounds fair (similar to the requirements of transparency in AI decisions :)) ) but I don't think that it'll change the leaderboard. In other hand it can help to improve the essays writing skills in the future.
\nSpecial thanks to the Kaggle team for the opportunity to reflect publicly on one of the most hot topics of our days.
Is it possible to know category-wise rankings and score?
", "votes": 1, "replies": [] }, { "id": 2379423, "author_name": "Roshini Fernando", "author_url": "", "post_date": "2023-08-08T06:45:04.837000", "content": "Congratulations to the winning teams ! I'm so happy to participate in this competition. I'm glad if you could provide our scores seperately according to the rubric because that'll be really helpful to improve more our writings at least through an email.
\nThank you!
", "votes": 1, "replies": [] }, { "id": 2378996, "author_name": "Gouri Prakash", "author_url": "", "post_date": "2023-08-07T23:27:44.807000", "content": "Congratulations to all the winners and kudos to everyone who participated! What an exciting adventure this was. Hope Kaggle plans on hosting many more essay competitions in the future.
", "votes": 1, "replies": [] }, { "id": 2387992, "author_name": "Sajjad Amjad", "author_url": "", "post_date": "2023-08-13T05:29:36.890000", "content": "🔥🔥🔥🔥🔥🔥🔥🔥
", "votes": -1, "replies": [] }, { "id": 2381611, "author_name": "Tariq Mahmood", "author_url": "", "post_date": "2023-08-09T09:31:54.997000", "content": "Great announcement @paultimothymooney , congratulations to all winners. Best of luck for all
", "votes": -1, "replies": [] }, { "id": 2379179, "author_name": "Abdul khadar N", "author_url": "", "post_date": "2023-08-08T04:10:22.407000", "content": "Congrats for all the winners and efforts taken by all the participants ! 💫
", "votes": -1, "replies": [] }, { "id": 3076720, "author_name": "Arya Gupta", "author_url": "", "post_date": "2024-12-20T08:03:40.733000", "content": "Congrats Leonie !!
", "votes": 0, "replies": [] }, { "id": 2380656, "author_name": "", "author_url": "", "post_date": "2023-08-08T17:35:56.650000", "content": "", "votes": -1, "replies": [] }, { "id": 2379992, "author_name": "", "author_url": "", "post_date": "2023-08-08T11:30:32.900000", "content": "", "votes": 0, "replies": [] }, { "id": 2379415, "author_name": "", "author_url": "", "post_date": "2023-08-08T06:41:06.773000", "content": "", "votes": 0, "replies": [] } ], "raw_markdown_by_id": { "2378929": "# Congratulations everyone!\n\nAfter evaluating all of your submissions, we are excited to announce the winners of the [2023 Kaggle AI Report Competition](https://www.kaggle.com/competitions/2023-kaggle-ai-report/)! We are truly grateful to everyone that shared an essay summarizing what the ML community has learned over the past 2 years. \n\nWe will provide awards to the [top scoring](https://www.kaggle.com/competitions/2023-kaggle-ai-report/overview/evaluation) essays in each of the following categories: \n- Text data\n- Image and/or video data\n- Tabular and/or time series data\n- Kaggle Competitions\n- Generative AI\n- AI ethics\n- Other \n\n\n# Text data:\n\nIn @abireltaief''s essay \"[Contemporary Large Language Models LLMs](https://www.kaggle.com/code/abireltaief/contemporary-large-language-models-llms)\", the author provides a fantastic overview of large language models (LLMs), covering LLM prompting, LLM fine-tuning, and LLM application development. A small but well-curated collection of recently published papers are reviewed, and the concepts within are explained in an accessible and approachable manner. Thanks for contributing such an informative essay!\n\nHonorable mentions: \n - http://www.kaggle.com/code/mistylight/mini-giants-small-language-models\n - https://www.kaggle.com/code/flaussy/large-language-models-reasoning-ability\n \n# Image and/or video data:\n\nIn @radbear's essay \"[Advances in AI Vision Models in the Last Two Years](https://www.kaggle.com/code/radbear/advances-in-ai-vision-models-in-the-last-two-years)\", the author discusses recent advances in semantic segmentation, and the development of vision transformers and continual learning methods, amongst other topics. The essay is well-referenced and provides a wealth of resources for those that want to dig into these topics in more detail. Thanks @radbear!\n\nHonorable mentions: \n - https://www.kaggle.com/code/dokster/image-and-video-data-kaggle-ai-report\n - https://www.kaggle.com/omarrajaa/image-data-report\n\t\n# Tabular and/or time series data\n\n\nIn @rhysie's essay \"[Learnings From the Typical Tabular Pipeline](https://www.kaggle.com/code/rhysie/learnings-from-the-typical-tabular-pipeline)\", the author does a fantastic job of exploring the connections between the methods used in Kaggle competitions and the methods discussed in the academic literature. We really appreciated the deep-dive into Kaggle's monthly tabular playground series!\n\n\n\nHonorable mentions: \n - https://www.kaggle.com/code/kensverse/ai-report\n - https://www.kaggle.com/code/kobbiemanrique/2023-kaggle-ai-report-tabular-data\n\n# Kaggle Competitions\n\nIn @iamleonie's essay \"[Towards Green AI](https://www.kaggle.com/code/iamleonie/towards-green-ai/notebook)\", the author gives an overview of some AI research areas that are relevant to carbon footprints and the environment, and then explores the subset of Kaggle Competitions that specifically encourage submitting small and efficient models. Methods that can be used to reduce the carbon footprint of ML models are reviewed, and the essay leaves us feeling inspired to build more efficient ML systems. Thanks @iamleonie!\n\nHonorable mentions: \n - https://www.kaggle.com/code/thedrcat/how-to-win-a-kaggle-competition\n - https://www.kaggle.com/code/nghihuynh/kaggle-ai-report-medical-imaging-competitions\n\n\n\n# Generative AI\n\nIn @trushk's essay \"[2023 Kaggle AI Report - Generative AI](https://www.kaggle.com/code/trushk/2023-kaggle-ai-report-generative-ai)\", the author surveys the landscape of generative AI products that have been released over the past few years. These exciting tools allow ML practitioners and even novices to generate images, text, videos, or code, and this essay provides a great overview of the major advancements that occurred along the way.\n\nHonorable mentions: \n - https://www.kaggle.com/code/jay2333/understand-generate-and-transform-the-world\n - https://www.kaggle.com/code/pranavbelhekar/a-glimpse-into-the-realm-of-generative-ai/notebook\n\n# AI ethics\n\nIn Team TISL's essay \"[Exploring the landscape of AI Ethics](https://www.kaggle.com/code/patrikkenfack/exploring-the-landscape-of-ai-ethics)\", the author introduces and explains a diverse set of concepts related to ethics and AI. Terms such as bias, fairness, equity, and explainability are introduced and explained, and pressing issues are discussed, such as climate change, copyright issues, and government regulation. \n\n\n\nHonorable mentions: \n - https://www.kaggle.com/code/antongc/developments-in-ai-and-ethics-in-the-past-2-years\n - https://www.kaggle.com/code/shreyamishra610/ethical-ai-is-all-we-need\n\n\n\n# Other\n\nIn @pluvias's essay, \"[Kaggle AI Report Optimization](https://www.kaggle.com/code/pluvias/kaggle-ai-report-optimization/notebook)\" the author explores recent advances in the fundamental area of optimization algorithms. The essay begins by giving an overview of more classical techniques, prior to moving on to highlight more recent advances such as MoMo and Sophia. Great work putting together this excellent resource @pluvia! \n \nHonorable mentions: \n - https://www.kaggle.com/code/redberry/ai-models-for-biosciences-and-virus-data/\n - https://www.kaggle.com/code/diegoexe/kaggle-ai-report-medical-data/notebook\n - https://www.kaggle.com/code/hodajalali/graph-learning-and-complex-networks\n\n# Great work everyone!\nThe Kaggle Team will reach out to the winners with instructions on how to collect their prizes.\n\n", "2379527": "Wow, thank you so much to the Kaggle team and the judges for selecting my entry as first place in the \"Kaggle Competitions\" category. I am especially honored because I thought there were some high-quality and impressive entries in this category.", "2378937": "Congrats all the winners! @paultimothymooney can the teams access to their final evaluation rubric scores? The final LB was unexpected for us actually, so we would be very happy to know our weaknesses.", "2381497": "I'm delighted to have won 1st place in the \"Text-data\" category. In addition, I'd like to extend my congratulations to the **6 winners in the other categories**, as well as to everyone who **received an honorable mention** and **all the competitors**. \nAdditionally, I'd like to congratulate the 7 GrandMasters who reviewed all of the entries as part of the judging panel for their professionalism, as well as the Kaggle team: Paul and the staff who worked behind the scenes.\nI really feel like a **double winner** ? : yes, not only because I'm the winner in the category \"Text data\" but also because it's in lifting others that we can really feel it (A team to whom I gave meticulous, highly constructive feedback received an honorable mention: perhaps time did not allow them to include everything, but the final version has changed a lot and is really good.): Thank you, Kaggle, for including the feature of constructive peer-feedback! \n**Other submissions** that are neither winners nor honorable mentions are nevertheless quite informative to read. The same is true in real life: you can learn from everyone, so I want to express my gratitude for all competitors and the whole incredible Kaggle community.\nFurthermore, **everyone** has the ability and the potential to win at some point, so keep trying, inspiring the community, and taking part in these fantastic competitions. Best", "2380500": "Wow this was a pleasant surprise! I was able to get the first place in the \"Generative AI\" category which had a lot of very good entries. Thank you Paul and the Kaggle Team for hosting this competition which resulted in many high quality reports. Hope to see more of these competitions in the future! Also thanks to the KGM team that spent time reading and evaluating everyones report ! ", "2379097": "Congratulations to the winners!\nBut I'm extremely perplexed and dismayed by the results; there were numerous notebooks that were far more helpful, received more votes from the audience, and were forked many times by the participants themselves. And the audience only upvotes a notebook if they find it relevant and useful. So, what is the point of upvotes if they are ignored? If this is the case, the competitions should not include an upvote mechanism!\n\nI have also noticed that the majority of essays that received high scores did not include the following prompt, which was highlighted **[below](https://www.kaggle.com/competitions/2023-kaggle-ai-report/discussion/400084#2334533)** by @paultimothymooney.\n\n>Ultimately the goal is to respond to one of the 7x prompts in a way that will score highest against the evaluation rubric, and the prompt begins with \"**what the ML community has learned over the past 2 years of working with…**\", which is a fairly broad mandate.\n\nAdditionally, I've seen that few of the high-scoring individuals and teams have notebooks that cover multiple subjects, which is unfair. \n\n>**Each team should write a single essay focusing on a single topic.**\n\nAnd it would have been much better if the scores on the **[eligibility checklist](https://www.kaggle.com/competitions/2023-kaggle-ai-report/overview/evaluation)** had been included with the results so that the participants could observe where they performed poorly and work to improve for the upcoming competitions.\nI wish the grading panel had double-checked these details before naming the winners. ", "2388524": "Thank you for organizing this competition and congratulations to the winners !\n\nI'm a bit disappointed with the results, I was expecting to at least have the points detailed in an evaluation grid.", "2379019": "Congratulations for the winners. Thanks for sharing the info on top performing teams.", "2380993": "Congratulations to all the winners. The effort and merit is evident in all 7 essays. \n\nAbout the complaints, we all knew what we where signing up for when we entered a subjective writing competition--where public relations efforts and individual considerations from the Kaggle team will be weighed. However, I still support the proposal for releasing the rubric for each essay for the learning value and experience it will provide. \n\nI'm guessing the transparency will also alleviate the perceptions of unfairness that a lot of people are voicing here and on LinkedIn.", "2380639": "Thanks to all the participants and congratulations to the winners of Kaggle AI report 2023! Now we have a lot of essays which make us to think about the main destinations of AI development. I'm sure that all the participants were working hard during preparing the essays. And I'm sure that no less hard work was made by 7 Grand Masters during evaluating the essays. Trying to look at the evaluation process from the 7 GM perspective I understand that it was really hard to choose which essay's composition was more accurate & concise, and discussion was more thorough & detailed. I trust to the best 7 Kaggle thinkers in their grading. The requirement for scoring transparency sounds fair (similar to the requirements of transparency in AI decisions :)) ) but I don't think that it'll change the leaderboard. In other hand it can help to improve the essays writing skills in the future.\nSpecial thanks to the Kaggle team for the opportunity to reflect publicly on one of the most hot topics of our days.", "2379486": "Is it possible to know category-wise rankings and score?", "2379423": "Congratulations to the winning teams ! I'm so happy to participate in this competition. I'm glad if you could provide our scores seperately according to the rubric because that'll be really helpful to improve more our writings at least through an email. \n\nThank you!", "2378996": "Congratulations to all the winners and kudos to everyone who participated! What an exciting adventure this was. Hope Kaggle plans on hosting many more essay competitions in the future.", "2387992": "🔥🔥🔥🔥🔥🔥🔥🔥", "2381611": "Great announcement @paultimothymooney , congratulations to all winners. Best of luck for all ", "2379179": "Congrats for all the winners and efforts taken by all the participants ! 💫", "3076720": "Congrats Leonie !!\n", "2380656": "", "2379992": "", "2379415": "" } }