Hello Kagglers, Astrophysicists,
\n\nNew PLAsTiCC paper is submitted to arXiv > astro-ph.\nhttps://arxiv.org/abs/1903.11756
\n\nThank you,
\n\nSho NAKAMURA
", "messages": [ { "id": "502759", "postDate": "03/29/2019 03:20:53", "content": "Hello Kagglers, Astrophysicists,
\n\nNew PLAsTiCC paper is submitted to arXiv > astro-ph.\nhttps://arxiv.org/abs/1903.11756
\n\nThank you,
\n\nSho NAKAMURA
", "rawMarkdown": "Hello Kagglers, Astrophysicists, \n\nNew PLAsTiCC paper is submitted to arXiv > astro-ph.\n[https://arxiv.org/abs/1903.11756](https://arxiv.org/abs/1903.11756)\n\nThank you, \n\nSho NAKAMURA", "votes": null }, { "id": "503172", "postDate": "03/29/2019 15:01:20", "content": "Dear PLAsTiCC Kagglers,
\n\nAs mentioned above, we have posted a scientific article about the simulation and models used in PLASTICC:\n https://arxiv.org/abs/1903.11756. \nThe article has also been submitted to a peer-reviewed journal:\n https://www.astrosociety.org/publication/pasp
\n\nWe hope that at least some of the information in our \"sim+model\" article will useful and interesting, and also answer some of your questions about the data. Feel free to post questions, or even ask where in the article a particular topic is addressed. Note that your comments & questions might be addressed in the next version of the article when we address comments from the anonymous journal referee.
\n\nFinally, I really enjoyed this whole experience of working with so many talented astronomers to develop the models and the data set. It was also really fun to watch how quickly the Kaggle community figured out our data set and develop some really cool classification methods. At the start of this project, I was concerned that astronomers' domain knowledge would give them too much of an advantage. However, the Kagglers were amazingly diligent and determined, and congratulations on pushing Kyle right up to the last hours of the competition. And FYI, I had never met or communicated with Kyle before or during the competition.
\n\nBest wishes to all on your future challenges.\n-Rick
", "rawMarkdown": "Dear PLAsTiCC Kagglers,\n\nAs mentioned above, we have posted a scientific article about the simulation and models used in PLASTICC:\n https://arxiv.org/abs/1903.11756. \nThe article has also been submitted to a peer-reviewed journal:\n https://www.astrosociety.org/publication/pasp\n\nWe hope that at least some of the information in our \"sim+model\" article will useful and interesting, and also answer some of your questions about the data. Feel free to post questions, or even ask where in the article a particular topic is addressed. Note that your comments & questions might be addressed in the next version of the article when we address comments from the anonymous journal referee. \n\nFinally, I really enjoyed this whole experience of working with so many talented astronomers to develop the models and the data set. It was also really fun to watch how quickly the Kaggle community figured out our data set and develop some really cool classification methods. At the start of this project, I was concerned that astronomers' domain knowledge would give them too much of an advantage. However, the Kagglers were amazingly diligent and determined, and congratulations on pushing Kyle right up to the last hours of the competition. And FYI, I had never met or communicated with Kyle before or during the competition.\n\nBest wishes to all on your future challenges.\n-Rick", "votes": null }, { "id": "503173", "postDate": "03/29/2019 15:03:29", "content": "Thank you for sharing this, @nakasho! This paper describes how the training and test sets were simulated from the underlying models. We're also working on a paper describing how we validated the models to make sure they reflected reality and the steps we took to prevent leakage and ensure the simulations did not give any unanticipated clues to the truth.
\n\nPlease let us know what questions you have about the models/simulations, and we will be happy to discuss it.
\n\nBest wishes,\n-Gautham\nOn behalf of the PLAsTiCC team
", "rawMarkdown": "Thank you for sharing this, @nakasho! This paper describes how the training and test sets were simulated from the underlying models. We're also working on a paper describing how we validated the models to make sure they reflected reality and the steps we took to prevent leakage and ensure the simulations did not give any unanticipated clues to the truth. \n\nPlease let us know what questions you have about the models/simulations, and we will be happy to discuss it.\n\nBest wishes,\n-Gautham\nOn behalf of the PLAsTiCC team", "votes": null } ], "comments": [ { "id": 503172, "author_name": "richardkessler", "author_url": "", "post_date": "03/29/2019 15:01:20", "content": "Dear PLAsTiCC Kagglers,
\n\nAs mentioned above, we have posted a scientific article about the simulation and models used in PLASTICC:\n https://arxiv.org/abs/1903.11756. \nThe article has also been submitted to a peer-reviewed journal:\n https://www.astrosociety.org/publication/pasp
\n\nWe hope that at least some of the information in our \"sim+model\" article will useful and interesting, and also answer some of your questions about the data. Feel free to post questions, or even ask where in the article a particular topic is addressed. Note that your comments & questions might be addressed in the next version of the article when we address comments from the anonymous journal referee.
\n\nFinally, I really enjoyed this whole experience of working with so many talented astronomers to develop the models and the data set. It was also really fun to watch how quickly the Kaggle community figured out our data set and develop some really cool classification methods. At the start of this project, I was concerned that astronomers' domain knowledge would give them too much of an advantage. However, the Kagglers were amazingly diligent and determined, and congratulations on pushing Kyle right up to the last hours of the competition. And FYI, I had never met or communicated with Kyle before or during the competition.
\n\nBest wishes to all on your future challenges.\n-Rick
", "votes": null, "replies": [] }, { "id": 503173, "author_name": "gsnarayan", "author_url": "", "post_date": "03/29/2019 15:03:29", "content": "Thank you for sharing this, @nakasho! This paper describes how the training and test sets were simulated from the underlying models. We're also working on a paper describing how we validated the models to make sure they reflected reality and the steps we took to prevent leakage and ensure the simulations did not give any unanticipated clues to the truth.
\n\nPlease let us know what questions you have about the models/simulations, and we will be happy to discuss it.
\n\nBest wishes,\n-Gautham\nOn behalf of the PLAsTiCC team
", "votes": null, "replies": [] } ], "raw_markdown_by_id": { "502759": "Hello Kagglers, Astrophysicists, \n\nNew PLAsTiCC paper is submitted to arXiv > astro-ph.\n[https://arxiv.org/abs/1903.11756](https://arxiv.org/abs/1903.11756)\n\nThank you, \n\nSho NAKAMURA", "503172": "Dear PLAsTiCC Kagglers,\n\nAs mentioned above, we have posted a scientific article about the simulation and models used in PLASTICC:\n https://arxiv.org/abs/1903.11756. \nThe article has also been submitted to a peer-reviewed journal:\n https://www.astrosociety.org/publication/pasp\n\nWe hope that at least some of the information in our \"sim+model\" article will useful and interesting, and also answer some of your questions about the data. Feel free to post questions, or even ask where in the article a particular topic is addressed. Note that your comments & questions might be addressed in the next version of the article when we address comments from the anonymous journal referee. \n\nFinally, I really enjoyed this whole experience of working with so many talented astronomers to develop the models and the data set. It was also really fun to watch how quickly the Kaggle community figured out our data set and develop some really cool classification methods. At the start of this project, I was concerned that astronomers' domain knowledge would give them too much of an advantage. However, the Kagglers were amazingly diligent and determined, and congratulations on pushing Kyle right up to the last hours of the competition. And FYI, I had never met or communicated with Kyle before or during the competition.\n\nBest wishes to all on your future challenges.\n-Rick", "503173": "Thank you for sharing this, @nakasho! This paper describes how the training and test sets were simulated from the underlying models. We're also working on a paper describing how we validated the models to make sure they reflected reality and the steps we took to prevent leakage and ensure the simulations did not give any unanticipated clues to the truth. \n\nPlease let us know what questions you have about the models/simulations, and we will be happy to discuss it.\n\nBest wishes,\n-Gautham\nOn behalf of the PLAsTiCC team" }, "source": "meta" }