Summary of Adapting Open-source Large Language Models For Cost-effective, Expert-level Clinical Note Generation with On-policy Reinforcement Learning, by Hanyin Wang et al.
Adapting Open-Source Large Language Models for Cost-Effective, Expert-Level Clinical Note Generation with On-Policy Reinforcement Learning
by Hanyin Wang, Chufan Gao, Bolun Liu, Qiping Xu, Guleid Hussein, Mohamad El Labban, Kingsley Iheasirim, Hariprasad Korsapati, Chuck Outcalt, Jimeng Sun
First submitted to arxiv on: 25 Apr 2024
Categories
- Main: Computation and Language (cs.CL)
- Secondary: Artificial Intelligence (cs.AI); Machine Learning (cs.LG)
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 study presents a domain- and task-specific adaptation process for the open-source LLaMA-2 model, enabling it to generate high-quality clinical notes from outpatient patient-doctor dialogues. The process incorporates continued pre-training, supervised fine-tuning, and reinforcement learning from both AI and human feedback. A new approach, DistillDirect, is introduced for on-policy reinforcement learning with Gemini 1.0 Pro as the teacher model. The resulting model, LLaMA-Clinic, generates clinical notes comparable in quality to those authored by physicians. In a blinded physician reader study, 90.4% of individual evaluations rated the notes generated by LLaMA-Clinic as “acceptable” or higher across all three criteria: real-world readiness, completeness, and accuracy. The study also highlights key considerations for future clinical note-generation tasks, emphasizing the importance of pre-defining a best-practice note format. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary This study is about making computers that can help doctors write better notes. The computer models are trained on conversations between patients and doctors to make them sound more like real doctor’s notes. The new approach helps the computer model learn from both its own mistakes and feedback from humans. The result is a computer that can write notes just as well as a human doctor. Doctors even liked the computer’s notes better than their own! This study shows that computers can be really helpful in writing doctor’s notes, but it also warns that we need to make sure the computers are following the same rules and guidelines as doctors. |
Keywords
» Artificial intelligence » Fine tuning » Gemini » Llama » Reinforcement learning » Supervised » Teacher model
