Summary of An Unsupervised Approach to Achieve Supervised-level Explainability in Healthcare Records, by Joakim Edin et al.
An Unsupervised Approach to Achieve Supervised-Level Explainability in Healthcare Records
by Joakim Edin, Maria Maistro, Lars Maaløe, Lasse Borgholt, Jakob D. Havtorn, Tuukka Ruotsalo
First submitted to arxiv on: 13 Jun 2024
Categories
- Main: Machine Learning (cs.LG)
- Secondary: None
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 This paper proposes an approach to produce plausible and faithful explanations for language models used in electronic healthcare records without requiring human-annotated evidence spans. The authors demonstrate the effectiveness of this method on the automated medical coding task, showing that adversarial robustness training improves explanation plausibility and introducing AttInGrad, a new explanation method superior to previous ones. By combining both contributions in a fully unsupervised setup, the authors produce explanations of comparable quality or better than those achieved through supervised approaches. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary This study is about making computer models that help doctors with patient records more understandable. Right now, these models are like black boxes – they make decisions without explaining how or why. Doctors want to know what’s going on inside these models because it can affect patient care. The researchers found a way to make the models explain themselves without needing extra information from doctors. They tested this method on a task that involves coding medical records and showed that it works really well. |
Keywords
» Artificial intelligence » Supervised » Unsupervised
