Before anything else, our team would like to say thank you to the competition organizers, EU-OPENSCREEN ERIC and SLAS, for hosting this interesting and challenging competition. Also, We would also like to thank all of the researchers who were involved in collecting and curating this high-quality compound solubility dataset. In this article, we would like to provide and explain in detail the solution that we employed to tackle this problem.
\nIn short, our solution is LM/GNN ensemble. To reach the point of the final solution, we have done various experiments and discovered many interesting outcomes. The key contribution to the performance of our solution is the utilization of the Pre-train Language Model for molecule representation. Attached, we will describe the detail and some stories behind the development of our solution.
\nhttps://storage.googleapis.com/kaggle-forum-message-attachments/2124821/18648/olab-solution.pdf
", "comments": [ { "id": 2240904, "author_name": "mipypf", "author_url": "", "post_date": "2023-05-01T01:39:22.063000", "content": "Thanks for sharing solution. I found it very interesting. I also saw your talk at the recent NVIDIA GTC 2023.
\nThe solution was very nice and informative, so I asked you a question. (I was so caught up in your talk at NVIDIA GTC 2023 that I missed asking you this question…) I would appreciate it if you could reply to my question when you are free.
", "votes": 0, "replies": [] } ], "comment_count": 1, "messages": [ { "id": 2124821, "postDate": "2023-02-01T08:31:18.523Z", "content": "Before anything else, our team would like to say thank you to the competition organizers, EU-OPENSCREEN ERIC and SLAS, for hosting this interesting and challenging competition. Also, We would also like to thank all of the researchers who were involved in collecting and curating this high-quality compound solubility dataset. In this article, we would like to provide and explain in detail the solution that we employed to tackle this problem.
\nIn short, our solution is LM/GNN ensemble. To reach the point of the final solution, we have done various experiments and discovered many interesting outcomes. The key contribution to the performance of our solution is the utilization of the Pre-train Language Model for molecule representation. Attached, we will describe the detail and some stories behind the development of our solution.
\nhttps://storage.googleapis.com/kaggle-forum-message-attachments/2124821/18648/olab-solution.pdf
", "rawMarkdown": "#Introduction\nBefore anything else, our team would like to say thank you to the competition organizers, EU-OPENSCREEN ERIC and SLAS, for hosting this interesting and challenging competition. Also, We would also like to thank all of the researchers who were involved in collecting and curating this high-quality compound solubility dataset. In this article, we would like to provide and explain in detail the solution that we employed to tackle this problem.\n\n#Overview\nIn short, our solution is LM/GNN ensemble. To reach the point of the final solution, we have done various experiments and discovered many interesting outcomes. The key contribution to the performance of our solution is the utilization of the Pre-train Language Model for molecule representation. Attached, we will describe the detail and some stories behind the development of our solution.\n\n[https://storage.googleapis.com/kaggle-forum-message-attachments/2124821/18648/olab-solution.pdf](https://storage.googleapis.com/kaggle-forum-message-attachments/2124821/18648/olab-solution.pdf)", "votes": 12 }, { "id": 2240904, "postDate": "2023-05-01T01:39:22.063Z", "content": "Thanks for sharing solution. I found it very interesting. I also saw your talk at the recent NVIDIA GTC 2023.
\nThe solution was very nice and informative, so I asked you a question. (I was so caught up in your talk at NVIDIA GTC 2023 that I missed asking you this question…) I would appreciate it if you could reply to my question when you are free.
", "rawMarkdown": "Thanks for sharing solution. I found it very interesting. I also saw your talk at the recent NVIDIA GTC 2023.\n\nThe solution was very nice and informative, so I asked you a question. (I was so caught up in your talk at NVIDIA GTC 2023 that I missed asking you this question...) I would appreciate it if you could reply to my question when you are free." } ], "complete": true, "source": "live", "fetched_at": "2026-10-19T16:45:28Z", "competition_id": 38158, "competition_slug": "euos-slas", "team_id": 9203074, "solution_url": "/competitions/euos-slas/writeups/olab-olab-solution" }