Summary of A Versatile Influence Function For Data Attribution with Non-decomposable Loss, by Junwei Deng et al.
A Versatile Influence Function for Data Attribution with Non-Decomposable Loss
by Junwei Deng, Weijing Tang, Jiaqi W. Ma
First submitted to arxiv on: 2 Dec 2024
Categories
- Main: Machine Learning (cs.LG)
- Secondary: Machine Learning (stat.ML)
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 bridges the gap in influence function applications by extending beyond M-estimators and developing the Versatile Influence Function (VIF). VIF allows for straightforward application to machine learning models trained with non-decomposable losses. It leverages auto-differentiation, eliminating case-specific derivations. The authors demonstrate VIF’s effectiveness across Cox regression, node embedding, and listwise learning-to-rank tasks. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary The paper helps machines learn better by creating a new way to understand how individual pieces of data affect model predictions. This is called influence function and it usually only works with simple loss functions. But what about when the loss function depends on multiple data points? The authors found a way to make influence function work in these cases too, using something called Versatile Influence Function (VIF). They tested VIF on three different tasks and showed that it’s fast and accurate. |
Keywords
» Artificial intelligence » Embedding » Loss function » Machine learning » Regression
