Summary of Are We Making Much Progress? Revisiting Chemical Reaction Yield Prediction From An Imbalanced Regression Perspective, by Yihong Ma et al.
Are we making much progress? Revisiting chemical reaction yield prediction from an imbalanced regression perspective
by Yihong Ma, Xiaobao Huang, Bozhao Nan, Nuno Moniz, Xiangliang Zhang, Olaf Wiest, Nitesh V. Chawla
First submitted to arxiv on: 6 Feb 2024
Categories
- Main: Machine Learning (cs.LG)
- Secondary: Chemical Physics (physics.chem-ph)
GrooveSquid.com Paper Summaries
GrooveSquid.com’s goal is to make artificial intelligence research accessible by summarizing AI papers in simpler terms. Each summary below covers the same AI paper, written at different levels of difficulty. The medium difficulty and low difficulty versions are original summaries written by GrooveSquid.com, while the high difficulty version is the paper’s original abstract. Feel free to learn from the version that suits you best!
| Summary difficulty | Written by | Summary |
|---|---|---|
| High | Paper authors | High Difficulty Summary Read the original abstract here |
| Medium | GrooveSquid.com (original content) | Medium Difficulty Summary The proposed paper tackles the challenge of accurately predicting chemical reaction yields, particularly for high-yield reactions which are crucial in synthesis planning. Recent advancements have improved overall yield predictions, but there remains a significant gap in predicting high-yield reactions. The authors argue that this performance gap stems from the imbalance in real-world data distribution towards low-yield reactions. Existing methods treat different yield ranges equally, assuming a balanced training distribution. This paper reframes reaction yield prediction as an imbalanced regression problem and demonstrates the effectiveness of simple cost-sensitive re-weighting methods on three real-world datasets. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary This research helps chemists predict how well chemical reactions will work before doing experiments in the lab. Chemists want to make sure they’re using the best reaction for their needs, but it’s hard to get accurate results without trying many different reactions first. Scientists have made progress predicting yields, but there’s still a problem with predicting high-yield reactions. This is because most real-world data shows low-yield reactions, making it harder to predict well-performing reactions. The researchers suggest that we need to change how we approach this problem and use new methods that take into account the imbalance in the data. |
Keywords
* Artificial intelligence * Regression
