Summary of Context Clues: Evaluating Long Context Models For Clinical Prediction Tasks on Ehrs, by Michael Wornow et al.
Context Clues: Evaluating Long Context Models for Clinical Prediction Tasks on EHRs
by Michael Wornow, Suhana Bedi, Miguel Angel Fuentes Hernandez, Ethan Steinberg, Jason Alan Fries, Christopher Re, Sanmi Koyejo, Nigam H. Shah
First submitted to arxiv on: 9 Dec 2024
Categories
- Main: Machine Learning (cs.LG)
- Secondary: Artificial Intelligence (cs.AI); Computational Engineering, Finance, and Science (cs.CE)
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 paper presents a systematic evaluation of the effect of context length on modeling Electronic Health Records (EHRs) using Foundation Models (FMs). Existing FMs trained on EHRs have achieved state-of-the-art results, but their context windows are limited to <1k tokens. The authors apply recent advancements in subquadratic long-context architectures, such as Mamba, to EHR data and demonstrate that longer context models improve predictive performance. They also evaluate the robustness of these models across three properties of EHR data: copy-forwarded diagnoses, irregular time intervals between events, and natural increase in disease complexity over time. The results show that longer context models are more robust to extreme levels of these properties. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary The paper explores how long-context architectures can be used to model Electronic Health Records (EHRs). EHRs contain important information for healthcare, but existing models have limitations due to their short context windows. By using a new type of architecture called Mamba, the authors are able to create longer context models that perform better at predicting certain tasks. They also test how well these models work when dealing with different types of data in EHRs. |
Keywords
» Artificial intelligence » Context length
