Summary of To Generate or to Retrieve? on the Effectiveness Of Artificial Contexts For Medical Open-domain Question Answering, by Giacomo Frisoni et al.
To Generate or to Retrieve? On the Effectiveness of Artificial Contexts for Medical Open-Domain Question Answering
by Giacomo Frisoni, Alessio Cocchieri, Alex Presepi, Gianluca Moro, Zaiqiao Meng
First submitted to arxiv on: 4 Mar 2024
Categories
- Main: Computation and Language (cs.CL)
- Secondary: Artificial Intelligence (cs.AI)
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 A novel approach to medical open-domain question answering is presented, which decouples knowledge from model parameters to enable training on low-resource hardware. The proposed “generate-then-read” framework, MedGENIE, outperforms existing retrieve-then-read methods on multiple-choice question answering tasks in medicine, including MedQA-USMLE, MedMCQA, and MMLU. This is achieved by constructing artificial contexts through prompting, leveraging domain-specific large language models. Experimental results show that generated passages are more effective than retrieved ones in attaining higher accuracy, using up to 706 times fewer parameters than a zero-shot closed-book baseline. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary Medical experts want computers to be able to answer medical questions better. To do this, they need access to lots of information about medicine. A new way to help them is by letting computers create their own information based on what’s already known, rather than just looking it up from a big library. This helps computers make more accurate answers. The new method, called MedGENIE, does better than other methods at answering medical questions and uses much less computer power. |
Keywords
» Artificial intelligence » Prompting » Question answering » Zero shot
