As @ambrosm writes in his post here: \"The most important result of every endeavor is what we learn from itβ¦\"
\nLet us organize webinars to share experience from the competition.
\nThe competition is important for the research community and deeper understanding of such tasks would be beneficial for biology and potentially medicine (in some feature).
Everybody is very welcome to participate - it does not matter your standing on the LB, what matters - your will to share the experience - please notify me if you have can make a webinar !
\n======================
\nπ @SBERLOGACOMPETE webinar on data science and bioinformatics:
\nπ¨βπ¬ Tianyu Liu (Yale University) \"Lessons we learned from the Kaggle competition: Open Problems in Single-cell Analysis\"
\nβοΈ 22 December, Thursday, 17.00 (Paris Time)
(The topics will include the overview of this problem including the difficulties and solutions, the models we used, and the functions we used to try but we did not observe score improvement based on public leaderboard. If time permits, I will also talk about how to use SHAP to explain our neural network or other model in this problem.
\nAbout the reporter: Tianyu Liu team got the silver medal and was one of the invited teams to make a talk at NearIPS workshop.
\nZoom link will be available at
\nhttps://t.me/sberlogabig shortly before the start.
And also will be shared here on Kaggle.
\n======================
\nPast webinars:
\nπ¨βπ¬ Andrey Shevtsov, Stas Borodykin \"Ideas of the top solutions of the Kaggle competition on single-cell data\".
\nβοΈ 25 November, Friday, 17.00 (Paris Time)
\nπ Presentation: https://t.me/sberlogacompete/3653
\nπΉ Video: https://youtu.be/F2X1bHGwDUE
\n(Sorry it is in Russian - we checked there were no English speaking participants on the lecture).
π¨βπ¬ Oleg Khudyakov \"Lessons learnt from the competition\".
\nβοΈ 17 November, Thursday, 17.00 (Paris Time)
\nπ Presentation: https://www.kaggle.com/code/kaggledummie007/lesson-how-to-overfit-on-public-lb
\nπ Notebooks: https://www.kaggle.com/code/kaggledummie007/cite-pca-feature-importance
\nπΉ Video: https://youtu.be/bEfpDf6q6-s
\n(Sorry it is in Russian - we checked there were no English speaking participants on the lecture).
As @ambrosm writes in his post here: \"The most important result of every endeavor is what we learn from itβ¦\"
\nLet us organize webinars to share experience from the competition.
\nThe competition is important for the research community and deeper understanding of such tasks would be beneficial for biology and potentially medicine (in some feature).
Everybody is very welcome to participate - it does not matter your standing on the LB, what matters - your will to share the experience - please notify me if you have can make a webinar !
\n======================
\nπ @SBERLOGACOMPETE webinar on data science and bioinformatics:
\nπ¨βπ¬ Tianyu Liu (Yale University) \"Lessons we learned from the Kaggle competition: Open Problems in Single-cell Analysis\"
\nβοΈ 22 December, Thursday, 17.00 (Paris Time)
(The topics will include the overview of this problem including the difficulties and solutions, the models we used, and the functions we used to try but we did not observe score improvement based on public leaderboard. If time permits, I will also talk about how to use SHAP to explain our neural network or other model in this problem.
\nAbout the reporter: Tianyu Liu team got the silver medal and was one of the invited teams to make a talk at NearIPS workshop.
\nZoom link will be available at
\nhttps://t.me/sberlogabig shortly before the start.
And also will be shared here on Kaggle.
\n======================
\nPast webinars:
\nπ¨βπ¬ Andrey Shevtsov, Stas Borodykin \"Ideas of the top solutions of the Kaggle competition on single-cell data\".
\nβοΈ 25 November, Friday, 17.00 (Paris Time)
\nπ Presentation: https://t.me/sberlogacompete/3653
\nπΉ Video: https://youtu.be/F2X1bHGwDUE
\n(Sorry it is in Russian - we checked there were no English speaking participants on the lecture).
π¨βπ¬ Oleg Khudyakov \"Lessons learnt from the competition\".
\nβοΈ 17 November, Thursday, 17.00 (Paris Time)
\nπ Presentation: https://www.kaggle.com/code/kaggledummie007/lesson-how-to-overfit-on-public-lb
\nπ Notebooks: https://www.kaggle.com/code/kaggledummie007/cite-pca-feature-importance
\nπΉ Video: https://youtu.be/bEfpDf6q6-s
\n(Sorry it is in Russian - we checked there were no English speaking participants on the lecture).
https://us02web.zoom.us/j/88379340212?pwd=M3JEbWk5L1hGNWIvMTNsTjdqU1p0QT09
", "rawMarkdown": "https://us02web.zoom.us/j/88379340212?pwd=M3JEbWk5L1hGNWIvMTNsTjdqU1p0QT09", "votes": 2 }, { "id": 2071275, "postDate": "2022-12-20T20:21:59.080Z", "content": "π @SBERLOGACOMPETE webinar on data science and bioinformatics:
\nπ¨βπ¬ Tianyu Liu (Yale University) \"Lessons we learned from the Kaggle competition: Open Problems in Single-cell Analysis\"
\nβοΈ 22 December, Thursday, 17.00 (Paris Time)
(The topics will include the overview of this problem including the difficulties and solutions, the models we used, and the functions we used to try but we did not observe score improvement based on public leaderboard. If time permits, I will also talk about how to use SHAP to explain our neural network or other model in this problem.
\nAbout the reporter: Tianyu Liu team got the silver medal and was one of the invited teams to make a talk at NearIPS workshop.
\nZoom link will be available at
\nhttps://t.me/sberlogabig shortly before the start.
And also will be shared here on Kaggle.
", "rawMarkdown": "π @SBERLOGACOMPETE webinar on data science and bioinformatics:\nπ¨βπ¬ Tianyu Liu (Yale University) \"Lessons we learned from the Kaggle competition: Open Problems in Single-cell Analysis\"\nβοΈ 22 December, Thursday, 17.00 (Paris Time)\n\n[Add to Google calendar](https://calendar.google.com/calendar/render?action=TEMPLATE&dates=20221222T160000Z%2F20221222T173000Z&details=Zoom%20link%20will%20be%20available%20at%20%0Ahttps%3A%2F%2Ft.me%2Fsberlogabig%20shortly%20before%20the%20start.&text=%20%40SBERLOGABIG%20%20Tianyu%20Liu%20%28Yale%20University%29%20%22%20Lessons%20we%20Learned%20from%20the%20Kaggle%20competition%3A%20%20Open%20Problems%20in%20Single-cell%20Analysis%22%20)\n\n(The topics will include the overview of this problem including the difficulties and solutions, the models we used, and the functions we used to try but we did not observe score improvement based on public leaderboard. If time permits, I will also talk about how to use SHAP to explain our neural network or other model in this problem.\n\nAbout the reporter: Tianyu Liu team got the silver medal and was one of the invited teams to make a talk at NearIPS workshop.\n\nZoom link will be available at\nhttps://t.me/sberlogabig shortly before the start.\n\nAnd also will be shared here on Kaggle.", "votes": 2 }, { "id": 2041994, "postDate": "2022-11-24T11:13:46.647Z", "content": "π¨βπ¬ Andrey Shevtsov, Stas Borodykin \"Ideas of the top solutions of the Kaggle competition on single-cell data\".
\nβοΈ 25 November, Friday, 17.00 (Paris Time)
\nπ Presentation: https://t.me/sberlogacompete/3653
\nπΉ Video: https://youtu.be/F2X1bHGwDUE
\n(Sorry it is in Russian - we checked there were no English speaking participants on the lecture).
\nRemarkable feature of Kaggle - publication of main ideas of the solutions after the end of the competition. These posts provide collection of beautiful insights how to improve your machine learning models. The present talk will be devoted to the survey of some of these ideas.
\nZoom link will be available at
\nhttps://t.me/sberlogabig shortly before the start.
\nVideo records of the talks: https://www.youtube.com/c/SciBerloga - subscribe our channel !
https://us02web.zoom.us/j/88379340212?pwd=M3JEbWk5L1hGNWIvMTNsTjdqU1p0QT09
", "votes": 2, "replies": [] }, { "id": 2071275, "author_name": "Alexander Chervov", "author_url": "", "post_date": "2022-12-20T20:21:59.080000", "content": "π @SBERLOGACOMPETE webinar on data science and bioinformatics:
\nπ¨βπ¬ Tianyu Liu (Yale University) \"Lessons we learned from the Kaggle competition: Open Problems in Single-cell Analysis\"
\nβοΈ 22 December, Thursday, 17.00 (Paris Time)
(The topics will include the overview of this problem including the difficulties and solutions, the models we used, and the functions we used to try but we did not observe score improvement based on public leaderboard. If time permits, I will also talk about how to use SHAP to explain our neural network or other model in this problem.
\nAbout the reporter: Tianyu Liu team got the silver medal and was one of the invited teams to make a talk at NearIPS workshop.
\nZoom link will be available at
\nhttps://t.me/sberlogabig shortly before the start.
And also will be shared here on Kaggle.
", "votes": 2, "replies": [] }, { "id": 2041994, "author_name": "Alexander Chervov", "author_url": "", "post_date": "2022-11-24T11:13:46.647000", "content": "π¨βπ¬ Andrey Shevtsov, Stas Borodykin \"Ideas of the top solutions of the Kaggle competition on single-cell data\".
\nβοΈ 25 November, Friday, 17.00 (Paris Time)
\nπ Presentation: https://t.me/sberlogacompete/3653
\nπΉ Video: https://youtu.be/F2X1bHGwDUE
\n(Sorry it is in Russian - we checked there were no English speaking participants on the lecture).
\nRemarkable feature of Kaggle - publication of main ideas of the solutions after the end of the competition. These posts provide collection of beautiful insights how to improve your machine learning models. The present talk will be devoted to the survey of some of these ideas.
\nZoom link will be available at
\nhttps://t.me/sberlogabig shortly before the start.
\nVideo records of the talks: https://www.youtube.com/c/SciBerloga - subscribe our channel !