Hey Kagglers,
\n\nThis competition has closed! We hope you were able to learn a lot, and use your machine learning skills to help the scientific community. We had over 47,800 submissions from 2700 teams (3300 competitors) from 84 countries! For 582 users (including 8 individuals in our top 10!), this was their first competition. We also had 2 new Grandmasters and 11 new Masters at the conclusion of this competition.
\n\nWe always love being able to host competitions that further the scientific community, and we've already heard from the competition sponsor that that the solutions were very advanced and interesting. Based on what we've seen in the leaderboard so far, it certainly makes for an interesting answer to the question \"Will this be won by a quantum chemistry specialist?\"
\n\nThank you all for your hard work in this competition and congratulations to our winners and to those who gained a new ranking! And a continued congratulation to the many of you may reach a new tier because of this competition, as well as far more of you for which this was your first competitions.
\n\nWe're also continually pleased to see how Kaggle Competitions serve as a great medium and vehicle through which the greater data science community can learn and develop their machine learning abilities and skills. Beginners and experts can come together to start, grow, and succeed. Further, in this competition, many can use their machine learning skills to help a great cause like this one.
\n\nWe also would like to give special thanks to Lars from CHAMPS and the University of Bristol for allowing us to run this competition. They've been a pleasure to work with, and we hope to work with them again in the future.
\n\nWe've cleaned the leaderboard and disqualified some teams that have violated the rules. If you think you were removed by mistake, or believe you have evidence that suggests another team cheated, please contact compliance. Please fill in all the fields honestly.
\n\nThe top potential winning teams will be contacted via email shortly to provide their winning solutions for sponsor review. Should we need to move down the leaderboard for any reason, we will do so and continue to make contact with subsequent teams.
\n\nYou should see your points and medals awarded soon, however it may take a few hours to propagate through.
\n\nIn the meantime, I've created this thread to be a curated list of your work. As you post your solutions, top kernels, and general discussion posts, I'll update this thread to keep track. If you write a paper, thesis, or present at a conference, and would like to share your work - please let us know so we can share with the Kaggle community! Additionally, if you performed any data manipulation, or maybe used a new technique in this competition that you'd like to share to further the industry, please let us know and we'll post it here! That way, should you be looking for takeaways in the future, or if you've just now stumbled across the competition, you have a place to see what surfaced from the work performed.
\n\nHappy Modeling!\nKaggle Team
\n\nTop Solutions\n- 1st Place Solution\n- 2nd Place Solution\n- 3rd Place Solution\n- 4th Place Solution\n- 5th Place Solution\n- 6th Place Solution\n- 7th Place Solution\n- 8th Place Solution\n- 10th Place Solution\n- 12th Place Solution\n- 13th Place Solution\n- 14th Place Solution\n- 15th Place Solution\n- 19th Place Solution\n- 20th Place Solution\n- 22nd Place Solution\n- 24th Place Solution\n- 37th Place Solution\n- 42nd Place Solution\n- 43rd Place Solution\n- 57th Place Solution\n- 62nd Place Solution
\n\nTop Notebooks\n- Molecular Properties EDA and Models\n- Brute Force Feature Engineering\n- Using Meta-Features to Improve Models\n- Keras Neural Net\n- How To: Easy Visualization of Molecules\n- Data Table, Simple Features\n- Distance is all you need\n- EachType\n- Simple EDA LightGBM Autotuning with HyperOpt\n- Ultra Fast Distance Matrix Computation\n- Just Speed Up Calculate Distance from Benchmark\n- MPNN\n- Predicting Mulliken Chanrges with ACSF Descriptors\n- SchNet Starter Kit\n- Dataset with Number of Bonds between Atoms\n- Bond Calculation\n- Validation, Feature Selection, Interpretation, etc\n- Exploring Molecular Properties Data\n- Atomic Distance Benchmark
\n\nHigh Scoring Kernels\n- -2.87894\n- -2.878\n- -2.609\n- -2.31496\n- -2.24207
\n\nOther Top Discussions\n- Which Graph CNN is Best\n- If not using Kernels, what are your Hardware specs?\n- Rectified Adam - cool new optimizer\n- How to obtain bonding data
", "messages": [ { "id": 612744, "postDate": "2019-08-30T01:17:12.617Z", "content": "Hey Kagglers,
\n\nThis competition has closed! We hope you were able to learn a lot, and use your machine learning skills to help the scientific community. We had over 47,800 submissions from 2700 teams (3300 competitors) from 84 countries! For 582 users (including 8 individuals in our top 10!), this was their first competition. We also had 2 new Grandmasters and 11 new Masters at the conclusion of this competition.
\n\nWe always love being able to host competitions that further the scientific community, and we've already heard from the competition sponsor that that the solutions were very advanced and interesting. Based on what we've seen in the leaderboard so far, it certainly makes for an interesting answer to the question \"Will this be won by a quantum chemistry specialist?\"
\n\nThank you all for your hard work in this competition and congratulations to our winners and to those who gained a new ranking! And a continued congratulation to the many of you may reach a new tier because of this competition, as well as far more of you for which this was your first competitions.
\n\nWe're also continually pleased to see how Kaggle Competitions serve as a great medium and vehicle through which the greater data science community can learn and develop their machine learning abilities and skills. Beginners and experts can come together to start, grow, and succeed. Further, in this competition, many can use their machine learning skills to help a great cause like this one.
\n\nWe also would like to give special thanks to Lars from CHAMPS and the University of Bristol for allowing us to run this competition. They've been a pleasure to work with, and we hope to work with them again in the future.
\n\nWe've cleaned the leaderboard and disqualified some teams that have violated the rules. If you think you were removed by mistake, or believe you have evidence that suggests another team cheated, please contact compliance. Please fill in all the fields honestly.
\n\nThe top potential winning teams will be contacted via email shortly to provide their winning solutions for sponsor review. Should we need to move down the leaderboard for any reason, we will do so and continue to make contact with subsequent teams.
\n\nYou should see your points and medals awarded soon, however it may take a few hours to propagate through.
\n\nIn the meantime, I've created this thread to be a curated list of your work. As you post your solutions, top kernels, and general discussion posts, I'll update this thread to keep track. If you write a paper, thesis, or present at a conference, and would like to share your work - please let us know so we can share with the Kaggle community! Additionally, if you performed any data manipulation, or maybe used a new technique in this competition that you'd like to share to further the industry, please let us know and we'll post it here! That way, should you be looking for takeaways in the future, or if you've just now stumbled across the competition, you have a place to see what surfaced from the work performed.
\n\nHappy Modeling!\nKaggle Team
\n\nTop Solutions\n- 1st Place Solution\n- 2nd Place Solution\n- 3rd Place Solution\n- 4th Place Solution\n- 5th Place Solution\n- 6th Place Solution\n- 7th Place Solution\n- 8th Place Solution\n- 10th Place Solution\n- 12th Place Solution\n- 13th Place Solution\n- 14th Place Solution\n- 15th Place Solution\n- 19th Place Solution\n- 20th Place Solution\n- 22nd Place Solution\n- 24th Place Solution\n- 37th Place Solution\n- 42nd Place Solution\n- 43rd Place Solution\n- 57th Place Solution\n- 62nd Place Solution
\n\nTop Notebooks\n- Molecular Properties EDA and Models\n- Brute Force Feature Engineering\n- Using Meta-Features to Improve Models\n- Keras Neural Net\n- How To: Easy Visualization of Molecules\n- Data Table, Simple Features\n- Distance is all you need\n- EachType\n- Simple EDA LightGBM Autotuning with HyperOpt\n- Ultra Fast Distance Matrix Computation\n- Just Speed Up Calculate Distance from Benchmark\n- MPNN\n- Predicting Mulliken Chanrges with ACSF Descriptors\n- SchNet Starter Kit\n- Dataset with Number of Bonds between Atoms\n- Bond Calculation\n- Validation, Feature Selection, Interpretation, etc\n- Exploring Molecular Properties Data\n- Atomic Distance Benchmark
\n\nHigh Scoring Kernels\n- -2.87894\n- -2.878\n- -2.609\n- -2.31496\n- -2.24207
\n\nOther Top Discussions\n- Which Graph CNN is Best\n- If not using Kernels, what are your Hardware specs?\n- Rectified Adam - cool new optimizer\n- How to obtain bonding data
", "rawMarkdown": "Hey Kagglers,\n\nThis competition has closed! We hope you were able to learn a lot, and use your machine learning skills to help the scientific community. We had over 47,800 submissions from 2700 teams (3300 competitors) from 84 countries! For 582 users (including 8 individuals in our top 10!), this was their first competition. We also had 2 new Grandmasters and 11 new Masters at the conclusion of this competition.\n\nWe always love being able to host competitions that further the scientific community, and we've already heard from the competition sponsor that that the solutions were very advanced and interesting. Based on what we've seen in the leaderboard so far, it certainly makes for an interesting answer to the question \"Will this be won by a quantum chemistry specialist?\" \n\nThank you all for your hard work in this competition and congratulations to our winners and to those who gained a new ranking! And a continued congratulation to the many of you may reach a new tier because of this competition, as well as far more of you for which this was your first competitions.\n\nWe're also continually pleased to see how Kaggle Competitions serve as a great medium and vehicle through which the greater data science community can learn and develop their machine learning abilities and skills. Beginners and experts can come together to start, grow, and succeed. Further, in this competition, many can use their machine learning skills to help a great cause like this one.\n\nWe also would like to give special thanks to Lars from CHAMPS and the University of Bristol for allowing us to run this competition. They've been a pleasure to work with, and we hope to work with them again in the future.\n\nWe've cleaned the leaderboard and disqualified some teams that have violated the rules. If you think you were removed by mistake, or believe you have evidence that suggests another team cheated, please contact [compliance](https://www.kaggle.com/compliance). Please fill in all the fields honestly.\n\nThe top potential winning teams will be contacted via email shortly to provide their winning solutions for sponsor review. Should we need to move down the leaderboard for any reason, we will do so and continue to make contact with subsequent teams.\n\nYou should see your points and medals awarded soon, however it may take a few hours to propagate through.\n\nIn the meantime, I've created this thread to be a curated list of your work. As you post your solutions, top kernels, and general discussion posts, I'll update this thread to keep track. If you write a paper, thesis, or present at a conference, and would like to share your work - please let us know so we can share with the Kaggle community! Additionally, if you performed any data manipulation, or maybe used a new technique in this competition that you'd like to share to further the industry, please let us know and we'll post it here! That way, should you be looking for takeaways in the future, or if you've just now stumbled across the competition, you have a place to see what surfaced from the work performed.\n\nHappy Modeling!\nKaggle Team\n\nTop Solutions\n- [1st Place Solution](https://www.kaggle.com/c/champs-scalar-coupling/discussion/106575)\n- [2nd Place Solution](https://www.kaggle.com/c/champs-scalar-coupling/discussion/106468)\n- [3rd Place Solution](https://www.kaggle.com/c/champs-scalar-coupling/discussion/106572)\n- [4th Place Solution](https://www.kaggle.com/c/champs-scalar-coupling/discussion/106534)\n- [5th Place Solution](https://www.kaggle.com/c/champs-scalar-coupling/discussion/106864)\n- [6th Place Solution](https://www.kaggle.com/c/champs-scalar-coupling/discussion/106407)\n- [7th Place Solution](https://www.kaggle.com/c/champs-scalar-coupling/discussion/106421)\n- [8th Place Solution](https://www.kaggle.com/c/champs-scalar-coupling/discussion/106347)\n- [10th Place Solution](https://www.kaggle.com/c/champs-scalar-coupling/discussion/106271)\n- [12th Place Solution](https://www.kaggle.com/c/champs-scalar-coupling/discussion/106275)\n- [13th Place Solution](https://www.kaggle.com/c/champs-scalar-coupling/discussion/106377)\n- [14th Place Solution](https://www.kaggle.com/c/champs-scalar-coupling/discussion/106298)\n- [15th Place Solution](https://www.kaggle.com/c/champs-scalar-coupling/discussion/106288)\n- [19th Place Solution](https://www.kaggle.com/c/champs-scalar-coupling/discussion/106905)\n- [20th Place Solution](https://www.kaggle.com/c/champs-scalar-coupling/discussion/106379)\n- [22nd Place Solution](https://www.kaggle.com/c/champs-scalar-coupling/discussion/106424)\n- [24th Place Solution](https://www.kaggle.com/c/champs-scalar-coupling/discussion/106404)\n- [37th Place Solution](https://www.kaggle.com/c/champs-scalar-coupling/discussion/106384)\n- [42nd Place Solution](https://www.kaggle.com/c/champs-scalar-coupling/discussion/106263)\n- [43rd Place Solution](https://www.kaggle.com/c/champs-scalar-coupling/discussion/106280)\n- [57th Place Solution](https://www.kaggle.com/c/champs-scalar-coupling/discussion/106359)\n- [62nd Place Solution](https://www.kaggle.com/c/champs-scalar-coupling/discussion/106310)\n\n\nTop Notebooks\n- [Molecular Properties EDA and Models](https://www.kaggle.com/artgor/molecular-properties-eda-and-models)\n- [Brute Force Feature Engineering](https://www.kaggle.com/artgor/brute-force-feature-engineering)\n- [Using Meta-Features to Improve Models](https://www.kaggle.com/artgor/using-meta-features-to-improve-model)\n- [Keras Neural Net](https://www.kaggle.com/todnewman/keras-neural-net-for-champs)\n- [How To: Easy Visualization of Molecules](https://www.kaggle.com/borisdee/how-to-easy-visualization-of-molecules)\n- [Data Table, Simple Features](https://www.kaggle.com/titericz/giba-r-data-table-simple-features-1-17-lb)\n- [Distance is all you need](https://www.kaggle.com/criskiev/distance-is-all-you-need-lb-1-481)\n- [EachType](https://www.kaggle.com/adrianoavelar/eachtype)\n- [Simple EDA LightGBM Autotuning with HyperOpt](https://www.kaggle.com/kabure/simple-eda-lightgbm-autotuning-w-hyperopt)\n- [Ultra Fast Distance Matrix Computation](https://www.kaggle.com/cpmpml/ultra-fast-distance-matrix-computation)\n- [Just Speed Up Calculate Distance from Benchmark](https://www.kaggle.com/seriousran/just-speed-up-calculate-distance-from-benchmark)\n- [MPNN](https://www.kaggle.com/fnands/1-mpnn)\n- [Predicting Mulliken Chanrges with ACSF Descriptors](https://www.kaggle.com/borisdee/predicting-mulliken-charges-with-acsf-descriptors)\n- [SchNet Starter Kit](https://www.kaggle.com/toshik/schnet-starter-kit)\n- [Dataset with Number of Bonds between Atoms](https://www.kaggle.com/asauve/dataset-with-number-of-bonds-between-atoms)\n- [Bond Calculation](https://www.kaggle.com/adrianoavelar/bond-calculaltion-lb-0-82)\n- [Validation, Feature Selection, Interpretation, etc](https://www.kaggle.com/artgor/validation-feature-selection-interpretation-etc)\n- [Exploring Molecular Properties Data](https://www.kaggle.com/robikscube/exploring-molecular-properties-data)\n- [Atomic Distance Benchmark](https://www.kaggle.com/inversion/atomic-distance-benchmark)\n\nHigh Scoring Kernels\n- [-2.87894](https://www.kaggle.com/chechir/champs-ensemble)\n- [-2.878](https://www.kaggle.com/senkin13/15th-stacking)\n- [-2.609](https://www.kaggle.com/cedben/squid)\n- [-2.31496](https://www.kaggle.com/fnands/mpnn-skip-2)\n- [-2.24207](https://www.kaggle.com/jesucristo/single-lgbm-2-242-top54)\n\nOther Top Discussions\n- [Which Graph CNN is Best](https://www.kaggle.com/c/champs-scalar-coupling/discussion/93972)\n- [If not using Kernels, what are your Hardware specs?](https://www.kaggle.com/c/champs-scalar-coupling/discussion/95294)\n- [Rectified Adam - cool new optimizer](https://www.kaggle.com/c/champs-scalar-coupling/discussion/104361)\n- [How to obtain bonding data](https://www.kaggle.com/c/champs-scalar-coupling/discussion/93928)\n\n", "votes": 65 }, { "id": 612770, "postDate": "2019-08-30T01:55:01.237Z", "content": "This is an amazing display of the incredible skill and creativity we have in Kaggle community! I'm personally humbled as well as exceedingly impressed with the solutions that have been developed for this competition.
", "rawMarkdown": "This is an amazing display of the incredible skill and creativity we have in Kaggle community! I'm personally humbled as well as exceedingly impressed with the solutions that have been developed for this competition.", "votes": 7, "replies": [ { "id": 612782, "postDate": "2019-08-30T02:11:35.213Z", "content": "Thanks @inversion, Great job on this amazing competition. Wish I had more time. Was in the midst of my company acquisition. Love the incredible works of the top winners which always bring in the next level of AI innovations. Keep this type of competitions coming....
", "rawMarkdown": "Thanks @inversion, Great job on this amazing competition. Wish I had more time. Was in the midst of my company acquisition. Love the incredible works of the top winners which always bring in the next level of AI innovations. Keep this type of competitions coming....", "votes": 2 } ] }, { "id": 613367, "postDate": "2019-08-30T12:59:21.383Z", "content": "This was an amazing competition. I was interested in it from the very beginning and it proved to be really cool. It was organized very well, for which I want to thank admins and hosts.
\n\nThe problem was unique, and so were the approaches. I'm glad that my kernels were useful to people.
\n\nI was able to take part in a wonderful team and learned a lot. And, of course, I became Competition Master.
\n\nThanks for this competition!
", "rawMarkdown": "This was an amazing competition. I was interested in it from the very beginning and it proved to be really cool. It was organized very well, for which I want to thank admins and hosts.\n\nThe problem was unique, and so were the approaches. I'm glad that my kernels were useful to people.\n\nI was able to take part in a wonderful team and learned a lot. And, of course, I became Competition Master.\n\nThanks for this competition!", "votes": 5, "replies": [ { "id": 627096, "postDate": "2019-09-15T12:05:58.057Z", "content": "Great job!
", "rawMarkdown": "Great job! " } ] }, { "id": 613138, "postDate": "2019-08-30T08:54:20.703Z", "content": "Thanks for the detailed write-up, @addisonhoward !\nIt was a first competition into which I really put A LOT of time and efforts. It was really good experience!\nBig thanks to @inversion , who handled issues during this competition just brilliantly.\nAnd of course thanks to the best team ever @yassinealouini @harshit92 @muhakabartay @felipemello , it was fun!
", "rawMarkdown": "Thanks for the detailed write-up, @addisonhoward !\nIt was a first competition into which I really put A LOT of time and efforts. It was really good experience!\nBig thanks to @inversion , who handled issues during this competition just brilliantly.\nAnd of course thanks to the best team ever @yassinealouini @harshit92 @muhakabartay @felipemello , it was fun!", "votes": 5 }, { "id": 613340, "postDate": "2019-08-30T12:33:29.623Z", "content": "I was a skeptic at the beginning. But it ended up a very enjoyable learning experience for me. Thank you for organizing it! Let's do one in Astronomy next.
", "rawMarkdown": "I was a skeptic at the beginning. But it ended up a very enjoyable learning experience for me. Thank you for organizing it! Let's do one in Astronomy next.", "votes": 6, "replies": [ { "id": 613343, "postDate": "2019-08-30T12:37:33.883Z", "content": "Astronomy would be SOOOOOO cool!
", "rawMarkdown": "Astronomy would be SOOOOOO cool!", "votes": 2 }, { "id": 613536, "postDate": "2019-08-30T15:36:51.627Z", "content": "You may have a look at this one for astronomy https://www.kaggle.com/c/PLAsTiCC-2018 even if it's already ended
", "rawMarkdown": "You may have a look at this one for astronomy https://www.kaggle.com/c/PLAsTiCC-2018 even if it's already ended", "votes": 1 }, { "id": 613541, "postDate": "2019-08-30T15:44:03.973Z", "content": "I hope they will do another one with real data when the telescope will take its first lights (2020-2021)
", "rawMarkdown": "I hope they will do another one with real data when the telescope will take its first lights (2020-2021)" }, { "id": 613548, "postDate": "2019-08-30T15:52:48.710Z", "content": "and I won't even imagine what we could do when the cherenkov telescope array will be build https://en.wikipedia.org/wiki/Cherenkov_Telescope_Array It will be an observatory so the data will be public!!!
", "rawMarkdown": "and I won't even imagine what we could do when the cherenkov telescope array will be build https://en.wikipedia.org/wiki/Cherenkov_Telescope_Array It will be an observatory so the data will be public!!!" } ] }, { "id": 674084, "postDate": "2019-11-15T21:53:38.773Z", "content": "Great initiative to put the solutions online. We can learn from it so much
", "rawMarkdown": "Great initiative to put the solutions online. We can learn from it so much", "votes": 1 }, { "id": 619876, "postDate": "2019-09-06T17:24:00.857Z", "content": "@inversion @addisonhoward Thank you for hosting this amazing competition. You mentioned about \"open-source implementation of the solution\" in the competition overview. Any update on that of the price wining solutions? I am really curious to see the source code.
", "rawMarkdown": "@inversion @addisonhoward Thank you for hosting this amazing competition. You mentioned about \"open-source implementation of the solution\" in the competition overview. Any update on that of the price wining solutions? I am really curious to see the source code.", "votes": 1 }, { "id": 615741, "postDate": "2019-09-02T10:15:26.350Z", "content": "5th place solution : https://www.kaggle.com/c/champs-scalar-coupling/discussion/106864#latest-615484
", "rawMarkdown": "5th place solution : https://www.kaggle.com/c/champs-scalar-coupling/discussion/106864#latest-615484", "votes": 1 }, { "id": 615525, "postDate": "2019-09-02T03:52:55.917Z", "content": "19th place solution is here.\n- 19th place solution
https://www.kaggle.com/c/champs-scalar-coupling/discussion/106905#latest-615483
This was my first competition in chemistry. Went to my school days to understand the different bonds and how can it be correlated. It was indeed a great competition to work on. Learnt lot of things in this competition. Thanks to organizers to conducting these kind of competitions. Of course, congrats to winners for their amazing work.
", "rawMarkdown": "This was my first competition in chemistry. Went to my school days to understand the different bonds and how can it be correlated. It was indeed a great competition to work on. Learnt lot of things in this competition. Thanks to organizers to conducting these kind of competitions. Of course, congrats to winners for their amazing work. ", "votes": 2, "replies": [ { "id": 627094, "postDate": "2019-09-15T12:04:45.373Z", "content": "@manojprabhaakr congratulations on your work!
", "rawMarkdown": "@manojprabhaakr congratulations on your work! ", "votes": 1 }, { "id": 627899, "postDate": "2019-09-16T14:50:14.040Z", "content": "Thanks @shrutikaggle :).
", "rawMarkdown": "Thanks @shrutikaggle :). ", "votes": 1 } ] }, { "id": 613535, "postDate": "2019-08-30T15:36:24.157Z", "content": "Thank all organizers and other kagglers for this amazing competition. It was very fun and tough, we've learnt a lot from this. We will post our writeup solution very soon. Sorry for the delay!
", "rawMarkdown": "Thank all organizers and other kagglers for this amazing competition. It was very fun and tough, we've learnt a lot from this. We will post our writeup solution very soon. Sorry for the delay! ", "votes": 2 }, { "id": 612783, "postDate": "2019-08-30T02:12:29.417Z", "content": "Thanks so much for the recap @inversion and @addisonhoward + companionship throughout the competition
", "rawMarkdown": "Thanks so much for the recap @inversion and @addisonhoward + companionship throughout the competition", "votes": 2 }, { "id": 615938, "postDate": "2019-09-02T14:19:46.960Z", "content": "71% evaluation means what about outliers? why?\nI would be happy to get the test set SCC values.. posted accessible to all.\nPlease let me know the link if available.. (privately available to gold medalists and award winners)\nWe put some efforts, time, computational resources (even if it is laptop to compute 25m entries), but did not get any email from organizers.. irresective of winning/losing..\nI did only one submission based on calculation done on 25m entries \nmy resutls was 0.8 R2 .. (for entire dataset for buidling models PLS etc) and for the diverse molecular dataset, without any filters..\nif DSO, PSO, SD and FC not given which is critical for SCC .. getting SCC from XYZ without calculating energy , forces, distance is quite surprising.. and magnetic properties calculation depends on these factors (from my Pauling's book)
\n\nI am happy to see the results declared in hurry.. would be happy to see the research papers coming out of prediction results (71% of total). if the predicted is directly coorelated with test set values originally available to organizers only, then the relationship with the submitted prediction (y'=m'x+c') can be built and that would correct prediction.
\n\ntop 99% is very interesting.. the SCC does not belong to one group of compounds (it has 6 types of compounds/SCCs) building 6 models and identify the suitable model based on structural evaluation for the accuracy is not correct as it kills the sprit of competition that should not use extensive computational resources or time. Time and computational complexity of final test prediction should be a criteria for true success\nit is my personal view.\nkarthi
", "rawMarkdown": "71% evaluation means what about outliers? why?\nI would be happy to get the test set SCC values.. posted accessible to all.\nPlease let me know the link if available.. (privately available to gold medalists and award winners)\nWe put some efforts, time, computational resources (even if it is laptop to compute 25m entries), but did not get any email from organizers.. irresective of winning/losing..\nI did only one submission based on calculation done on 25m entries \nmy resutls was 0.8 R2 .. (for entire dataset for buidling models PLS etc) and for the diverse molecular dataset, without any filters..\nif DSO, PSO, SD and FC not given which is critical for SCC .. getting SCC from XYZ without calculating energy , forces, distance is quite surprising.. and magnetic properties calculation depends on these factors (from my Pauling's book)\n\nI am happy to see the results declared in hurry.. would be happy to see the research papers coming out of prediction results (71% of total). if the predicted is directly coorelated with test set values originally available to organizers only, then the relationship with the submitted prediction (y'=m'x+c') can be built and that would correct prediction.\n\ntop 99% is very interesting.. the SCC does not belong to one group of compounds (it has 6 types of compounds/SCCs) building 6 models and identify the suitable model based on structural evaluation for the accuracy is not correct as it kills the sprit of competition that should not use extensive computational resources or time. Time and computational complexity of final test prediction should be a criteria for true success\nit is my personal view.\nkarthi\n", "votes": -1 }, { "id": 618899, "postDate": "2019-09-05T16:00:53.247Z", "content": "I enjoyed participating in and learning a lot from this competition. Thanks!
", "rawMarkdown": "I enjoyed participating in and learning a lot from this competition. Thanks!" }, { "id": 613361, "postDate": "2019-08-30T12:54:06.673Z", "rawMarkdown": "", "isDeleted": true }, { "id": 612754, "postDate": "2019-08-30T01:25:58.483Z", "rawMarkdown": "", "votes": 3, "isDeleted": true }, { "id": 616460, "postDate": "2019-09-03T06:35:06.757Z", "content": "Thanks a lot for sharing.
", "rawMarkdown": "Thanks a lot for sharing." } ], "comments": [ { "id": 612770, "author_name": "inversion", "author_url": "", "post_date": "2019-08-30T01:55:01.237000", "content": "This is an amazing display of the incredible skill and creativity we have in Kaggle community! I'm personally humbled as well as exceedingly impressed with the solutions that have been developed for this competition.
", "votes": 7, "replies": [ { "id": 612782, "author_name": "Patrick Chan", "author_url": "", "post_date": "2019-08-30T02:11:35.213000", "content": "Thanks @inversion, Great job on this amazing competition. Wish I had more time. Was in the midst of my company acquisition. Love the incredible works of the top winners which always bring in the next level of AI innovations. Keep this type of competitions coming....
", "votes": 2, "replies": [] } ] }, { "id": 613367, "author_name": "Andrey Lukyanenko", "author_url": "", "post_date": "2019-08-30T12:59:21.383000", "content": "This was an amazing competition. I was interested in it from the very beginning and it proved to be really cool. It was organized very well, for which I want to thank admins and hosts.
\n\nThe problem was unique, and so were the approaches. I'm glad that my kernels were useful to people.
\n\nI was able to take part in a wonderful team and learned a lot. And, of course, I became Competition Master.
\n\nThanks for this competition!
", "votes": 5, "replies": [ { "id": 627096, "author_name": "Pythondragon", "author_url": "", "post_date": "2019-09-15T12:05:58.057000", "content": "Great job!
", "votes": 0, "replies": [] } ] }, { "id": 613138, "author_name": "Stanislav Blinov", "author_url": "", "post_date": "2019-08-30T08:54:20.703000", "content": "Thanks for the detailed write-up, @addisonhoward !\nIt was a first competition into which I really put A LOT of time and efforts. It was really good experience!\nBig thanks to @inversion , who handled issues during this competition just brilliantly.\nAnd of course thanks to the best team ever @yassinealouini @harshit92 @muhakabartay @felipemello , it was fun!
", "votes": 5, "replies": [] }, { "id": 613340, "author_name": "Mycroft Holmes", "author_url": "", "post_date": "2019-08-30T12:33:29.623000", "content": "I was a skeptic at the beginning. But it ended up a very enjoyable learning experience for me. Thank you for organizing it! Let's do one in Astronomy next.
", "votes": 6, "replies": [ { "id": 613343, "author_name": "Stanislav Blinov", "author_url": "", "post_date": "2019-08-30T12:37:33.883000", "content": "Astronomy would be SOOOOOO cool!
", "votes": 2, "replies": [] }, { "id": 613536, "author_name": "Raphaël CC", "author_url": "", "post_date": "2019-08-30T15:36:51.627000", "content": "You may have a look at this one for astronomy https://www.kaggle.com/c/PLAsTiCC-2018 even if it's already ended
", "votes": 1, "replies": [] }, { "id": 613541, "author_name": "Raphaël CC", "author_url": "", "post_date": "2019-08-30T15:44:03.973000", "content": "I hope they will do another one with real data when the telescope will take its first lights (2020-2021)
", "votes": 0, "replies": [] }, { "id": 613548, "author_name": "Raphaël CC", "author_url": "", "post_date": "2019-08-30T15:52:48.710000", "content": "and I won't even imagine what we could do when the cherenkov telescope array will be build https://en.wikipedia.org/wiki/Cherenkov_Telescope_Array It will be an observatory so the data will be public!!!
", "votes": 0, "replies": [] } ] }, { "id": 674084, "author_name": "ck", "author_url": "", "post_date": "2019-11-15T21:53:38.773000", "content": "Great initiative to put the solutions online. We can learn from it so much
", "votes": 1, "replies": [] }, { "id": 619876, "author_name": "Zhenlan", "author_url": "", "post_date": "2019-09-06T17:24:00.857000", "content": "@inversion @addisonhoward Thank you for hosting this amazing competition. You mentioned about \"open-source implementation of the solution\" in the competition overview. Any update on that of the price wining solutions? I am really curious to see the source code.
", "votes": 1, "replies": [] }, { "id": 615741, "author_name": "redshellspy", "author_url": "", "post_date": "2019-09-02T10:15:26.350000", "content": "5th place solution : https://www.kaggle.com/c/champs-scalar-coupling/discussion/106864#latest-615484
", "votes": 1, "replies": [] }, { "id": 615525, "author_name": "toshi_k", "author_url": "", "post_date": "2019-09-02T03:52:55.917000", "content": "19th place solution is here.\n- 19th place solution
https://www.kaggle.com/c/champs-scalar-coupling/discussion/106905#latest-615483
This was my first competition in chemistry. Went to my school days to understand the different bonds and how can it be correlated. It was indeed a great competition to work on. Learnt lot of things in this competition. Thanks to organizers to conducting these kind of competitions. Of course, congrats to winners for their amazing work.
", "votes": 2, "replies": [ { "id": 627094, "author_name": "Pythondragon", "author_url": "", "post_date": "2019-09-15T12:04:45.373000", "content": "@manojprabhaakr congratulations on your work!
", "votes": 1, "replies": [] }, { "id": 627899, "author_name": "Manoj Prabhakar", "author_url": "", "post_date": "2019-09-16T14:50:14.040000", "content": "Thanks @shrutikaggle :).
", "votes": 1, "replies": [] } ] }, { "id": 613535, "author_name": "Tu Nguyen", "author_url": "", "post_date": "2019-08-30T15:36:24.157000", "content": "Thank all organizers and other kagglers for this amazing competition. It was very fun and tough, we've learnt a lot from this. We will post our writeup solution very soon. Sorry for the delay!
", "votes": 2, "replies": [] }, { "id": 612783, "author_name": "Michel ML", "author_url": "", "post_date": "2019-08-30T02:12:29.417000", "content": "Thanks so much for the recap @inversion and @addisonhoward + companionship throughout the competition
", "votes": 2, "replies": [] }, { "id": 615938, "author_name": "m karthikeyan", "author_url": "", "post_date": "2019-09-02T14:19:46.960000", "content": "71% evaluation means what about outliers? why?\nI would be happy to get the test set SCC values.. posted accessible to all.\nPlease let me know the link if available.. (privately available to gold medalists and award winners)\nWe put some efforts, time, computational resources (even if it is laptop to compute 25m entries), but did not get any email from organizers.. irresective of winning/losing..\nI did only one submission based on calculation done on 25m entries \nmy resutls was 0.8 R2 .. (for entire dataset for buidling models PLS etc) and for the diverse molecular dataset, without any filters..\nif DSO, PSO, SD and FC not given which is critical for SCC .. getting SCC from XYZ without calculating energy , forces, distance is quite surprising.. and magnetic properties calculation depends on these factors (from my Pauling's book)
\n\nI am happy to see the results declared in hurry.. would be happy to see the research papers coming out of prediction results (71% of total). if the predicted is directly coorelated with test set values originally available to organizers only, then the relationship with the submitted prediction (y'=m'x+c') can be built and that would correct prediction.
\n\ntop 99% is very interesting.. the SCC does not belong to one group of compounds (it has 6 types of compounds/SCCs) building 6 models and identify the suitable model based on structural evaluation for the accuracy is not correct as it kills the sprit of competition that should not use extensive computational resources or time. Time and computational complexity of final test prediction should be a criteria for true success\nit is my personal view.\nkarthi
", "votes": -1, "replies": [] }, { "id": 618899, "author_name": "Mahtab Noor Shaan", "author_url": "", "post_date": "2019-09-05T16:00:53.247000", "content": "I enjoyed participating in and learning a lot from this competition. Thanks!
", "votes": 0, "replies": [] }, { "id": 613361, "author_name": "", "author_url": "", "post_date": "2019-08-30T12:54:06.673000", "content": "", "votes": 0, "replies": [] }, { "id": 612754, "author_name": "", "author_url": "", "post_date": "2019-08-30T01:25:58.483000", "content": "", "votes": 3, "replies": [] }, { "id": 616460, "author_name": "Rashidul H", "author_url": "", "post_date": "2019-09-03T06:35:06.757000", "content": "Thanks a lot for sharing.
", "votes": 0, "replies": [] } ], "raw_markdown_by_id": { "612744": "Hey Kagglers,\n\nThis competition has closed! We hope you were able to learn a lot, and use your machine learning skills to help the scientific community. We had over 47,800 submissions from 2700 teams (3300 competitors) from 84 countries! For 582 users (including 8 individuals in our top 10!), this was their first competition. We also had 2 new Grandmasters and 11 new Masters at the conclusion of this competition.\n\nWe always love being able to host competitions that further the scientific community, and we've already heard from the competition sponsor that that the solutions were very advanced and interesting. Based on what we've seen in the leaderboard so far, it certainly makes for an interesting answer to the question \"Will this be won by a quantum chemistry specialist?\" \n\nThank you all for your hard work in this competition and congratulations to our winners and to those who gained a new ranking! And a continued congratulation to the many of you may reach a new tier because of this competition, as well as far more of you for which this was your first competitions.\n\nWe're also continually pleased to see how Kaggle Competitions serve as a great medium and vehicle through which the greater data science community can learn and develop their machine learning abilities and skills. Beginners and experts can come together to start, grow, and succeed. Further, in this competition, many can use their machine learning skills to help a great cause like this one.\n\nWe also would like to give special thanks to Lars from CHAMPS and the University of Bristol for allowing us to run this competition. They've been a pleasure to work with, and we hope to work with them again in the future.\n\nWe've cleaned the leaderboard and disqualified some teams that have violated the rules. If you think you were removed by mistake, or believe you have evidence that suggests another team cheated, please contact [compliance](https://www.kaggle.com/compliance). Please fill in all the fields honestly.\n\nThe top potential winning teams will be contacted via email shortly to provide their winning solutions for sponsor review. Should we need to move down the leaderboard for any reason, we will do so and continue to make contact with subsequent teams.\n\nYou should see your points and medals awarded soon, however it may take a few hours to propagate through.\n\nIn the meantime, I've created this thread to be a curated list of your work. As you post your solutions, top kernels, and general discussion posts, I'll update this thread to keep track. If you write a paper, thesis, or present at a conference, and would like to share your work - please let us know so we can share with the Kaggle community! Additionally, if you performed any data manipulation, or maybe used a new technique in this competition that you'd like to share to further the industry, please let us know and we'll post it here! That way, should you be looking for takeaways in the future, or if you've just now stumbled across the competition, you have a place to see what surfaced from the work performed.\n\nHappy Modeling!\nKaggle Team\n\nTop Solutions\n- [1st Place Solution](https://www.kaggle.com/c/champs-scalar-coupling/discussion/106575)\n- [2nd Place Solution](https://www.kaggle.com/c/champs-scalar-coupling/discussion/106468)\n- [3rd Place Solution](https://www.kaggle.com/c/champs-scalar-coupling/discussion/106572)\n- [4th Place Solution](https://www.kaggle.com/c/champs-scalar-coupling/discussion/106534)\n- [5th Place Solution](https://www.kaggle.com/c/champs-scalar-coupling/discussion/106864)\n- [6th Place Solution](https://www.kaggle.com/c/champs-scalar-coupling/discussion/106407)\n- [7th Place Solution](https://www.kaggle.com/c/champs-scalar-coupling/discussion/106421)\n- [8th Place Solution](https://www.kaggle.com/c/champs-scalar-coupling/discussion/106347)\n- [10th Place Solution](https://www.kaggle.com/c/champs-scalar-coupling/discussion/106271)\n- [12th Place Solution](https://www.kaggle.com/c/champs-scalar-coupling/discussion/106275)\n- [13th Place Solution](https://www.kaggle.com/c/champs-scalar-coupling/discussion/106377)\n- [14th Place Solution](https://www.kaggle.com/c/champs-scalar-coupling/discussion/106298)\n- [15th Place Solution](https://www.kaggle.com/c/champs-scalar-coupling/discussion/106288)\n- [19th Place Solution](https://www.kaggle.com/c/champs-scalar-coupling/discussion/106905)\n- [20th Place Solution](https://www.kaggle.com/c/champs-scalar-coupling/discussion/106379)\n- [22nd Place Solution](https://www.kaggle.com/c/champs-scalar-coupling/discussion/106424)\n- [24th Place Solution](https://www.kaggle.com/c/champs-scalar-coupling/discussion/106404)\n- [37th Place Solution](https://www.kaggle.com/c/champs-scalar-coupling/discussion/106384)\n- [42nd Place Solution](https://www.kaggle.com/c/champs-scalar-coupling/discussion/106263)\n- [43rd Place Solution](https://www.kaggle.com/c/champs-scalar-coupling/discussion/106280)\n- [57th Place Solution](https://www.kaggle.com/c/champs-scalar-coupling/discussion/106359)\n- [62nd Place Solution](https://www.kaggle.com/c/champs-scalar-coupling/discussion/106310)\n\n\nTop Notebooks\n- [Molecular Properties EDA and Models](https://www.kaggle.com/artgor/molecular-properties-eda-and-models)\n- [Brute Force Feature Engineering](https://www.kaggle.com/artgor/brute-force-feature-engineering)\n- [Using Meta-Features to Improve Models](https://www.kaggle.com/artgor/using-meta-features-to-improve-model)\n- [Keras Neural Net](https://www.kaggle.com/todnewman/keras-neural-net-for-champs)\n- [How To: Easy Visualization of Molecules](https://www.kaggle.com/borisdee/how-to-easy-visualization-of-molecules)\n- [Data Table, Simple Features](https://www.kaggle.com/titericz/giba-r-data-table-simple-features-1-17-lb)\n- [Distance is all you need](https://www.kaggle.com/criskiev/distance-is-all-you-need-lb-1-481)\n- [EachType](https://www.kaggle.com/adrianoavelar/eachtype)\n- [Simple EDA LightGBM Autotuning with HyperOpt](https://www.kaggle.com/kabure/simple-eda-lightgbm-autotuning-w-hyperopt)\n- [Ultra Fast Distance Matrix Computation](https://www.kaggle.com/cpmpml/ultra-fast-distance-matrix-computation)\n- [Just Speed Up Calculate Distance from Benchmark](https://www.kaggle.com/seriousran/just-speed-up-calculate-distance-from-benchmark)\n- [MPNN](https://www.kaggle.com/fnands/1-mpnn)\n- [Predicting Mulliken Chanrges with ACSF Descriptors](https://www.kaggle.com/borisdee/predicting-mulliken-charges-with-acsf-descriptors)\n- [SchNet Starter Kit](https://www.kaggle.com/toshik/schnet-starter-kit)\n- [Dataset with Number of Bonds between Atoms](https://www.kaggle.com/asauve/dataset-with-number-of-bonds-between-atoms)\n- [Bond Calculation](https://www.kaggle.com/adrianoavelar/bond-calculaltion-lb-0-82)\n- [Validation, Feature Selection, Interpretation, etc](https://www.kaggle.com/artgor/validation-feature-selection-interpretation-etc)\n- [Exploring Molecular Properties Data](https://www.kaggle.com/robikscube/exploring-molecular-properties-data)\n- [Atomic Distance Benchmark](https://www.kaggle.com/inversion/atomic-distance-benchmark)\n\nHigh Scoring Kernels\n- [-2.87894](https://www.kaggle.com/chechir/champs-ensemble)\n- [-2.878](https://www.kaggle.com/senkin13/15th-stacking)\n- [-2.609](https://www.kaggle.com/cedben/squid)\n- [-2.31496](https://www.kaggle.com/fnands/mpnn-skip-2)\n- [-2.24207](https://www.kaggle.com/jesucristo/single-lgbm-2-242-top54)\n\nOther Top Discussions\n- [Which Graph CNN is Best](https://www.kaggle.com/c/champs-scalar-coupling/discussion/93972)\n- [If not using Kernels, what are your Hardware specs?](https://www.kaggle.com/c/champs-scalar-coupling/discussion/95294)\n- [Rectified Adam - cool new optimizer](https://www.kaggle.com/c/champs-scalar-coupling/discussion/104361)\n- [How to obtain bonding data](https://www.kaggle.com/c/champs-scalar-coupling/discussion/93928)\n\n", "612770": "This is an amazing display of the incredible skill and creativity we have in Kaggle community! I'm personally humbled as well as exceedingly impressed with the solutions that have been developed for this competition.", "613367": "This was an amazing competition. I was interested in it from the very beginning and it proved to be really cool. It was organized very well, for which I want to thank admins and hosts.\n\nThe problem was unique, and so were the approaches. I'm glad that my kernels were useful to people.\n\nI was able to take part in a wonderful team and learned a lot. And, of course, I became Competition Master.\n\nThanks for this competition!", "613138": "Thanks for the detailed write-up, @addisonhoward !\nIt was a first competition into which I really put A LOT of time and efforts. It was really good experience!\nBig thanks to @inversion , who handled issues during this competition just brilliantly.\nAnd of course thanks to the best team ever @yassinealouini @harshit92 @muhakabartay @felipemello , it was fun!", "613340": "I was a skeptic at the beginning. But it ended up a very enjoyable learning experience for me. Thank you for organizing it! Let's do one in Astronomy next.", "674084": "Great initiative to put the solutions online. We can learn from it so much", "619876": "@inversion @addisonhoward Thank you for hosting this amazing competition. You mentioned about \"open-source implementation of the solution\" in the competition overview. Any update on that of the price wining solutions? I am really curious to see the source code.", "615741": "5th place solution : https://www.kaggle.com/c/champs-scalar-coupling/discussion/106864#latest-615484", "615525": "19th place solution is here.\n- 19th place solution