Summary of Demoshapley: Valuation Of Demonstrations For In-context Learning, by Shan Xie et al.
DemoShapley: Valuation of Demonstrations for In-Context Learning
by Shan Xie, Man Luo, Chadly Daniel Stern, Mengnan Du, Lu Cheng
First submitted to arxiv on: 10 Oct 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 The proposed DemoShapley approach assesses the influence of individual demonstration instances on few-shot learning models leveraging in-context learning (ICL). By distinguishing between contributing and hindering demonstrations, DemoShapley enhances model performance in terms of accuracy and fairness. The methodology demonstrates versatility and effectiveness in optimizing ICL demonstration selection, even generalizing to domains distinct from those used for training. Additionally, DemoShapley aids in identifying noisy data within the demonstration set. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary Large language models are getting better at learning new tasks without needing a lot of practice. This is thanks to something called “in-context learning.” However, researchers have found that what works well depends on which examples or “demonstrations” are used during training. To solve this problem, scientists created an approach called DemoShapley. It helps decide which demonstrations are helpful and which might actually make things worse. This new method makes models work better, be fairer, and even learn from different types of data. Plus, it can spot bad examples in the demonstration set. |
Keywords
* Artificial intelligence * Few shot
