Summary of Are We Making Progress in Unlearning? Findings From the First Neurips Unlearning Competition, by Eleni Triantafillou et al.
Are we making progress in unlearning? Findings from the first NeurIPS unlearning competition
by Eleni Triantafillou, Peter Kairouz, Fabian Pedregosa, Jamie Hayes, Meghdad Kurmanji, Kairan Zhao, Vincent Dumoulin, Julio Jacques Junior, Ioannis Mitliagkas, Jun Wan, Lisheng Sun Hosoya, Sergio Escalera, Gintare Karolina Dziugaite, Peter Triantafillou, Isabelle Guyon
First submitted to arxiv on: 13 Jun 2024
Categories
- Main: Machine Learning (cs.LG)
- Secondary: None
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 NeurIPS unlearning competition aimed to spur innovation in novel algorithms and robust evaluation methods. With nearly 1,200 teams participating, it achieved great success. This paper analyzes top solutions, discussing benchmarking unlearning as a research problem. A formal notion of unlearning was used, incorporating model utility for holistic evaluation. The effectiveness of different instantiations under various compute costs is analyzed, revealing implications for standardizing evaluation. Top-performing methods surpass existing algorithms under our framework. Trade-offs between generalizability and strengths/weaknesses are examined, paving the way for advancing benchmarking and algorithm development in unlearning. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary Unlearning competition: Over 1,200 teams from around the world participated in a competition that aimed to develop new algorithms and discuss ways to evaluate them. This paper looks at the top solutions and talks about how to measure “forgetting” or losing information. It also compares different methods for evaluating this process. |
