Summary of Label Smoothing Improves Machine Unlearning, by Zonglin Di et al.
Label Smoothing Improves Machine Unlearning
by Zonglin Di, Zhaowei Zhu, Jinghan Jia, Jiancheng Liu, Zafar Takhirov, Bo Jiang, Yuanshun Yao, Sijia Liu, Yang Liu
First submitted to arxiv on: 11 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 paper proposes a new machine unlearning (MU) approach called UGradSL, which eliminates previously learned data from a model while minimizing computation cost and maintaining performance. Inspired by label smoothing and differential privacy, UGradSL uses an inverse process of label smoothing to achieve better MU results. Theoretical analyses show that properly introducing label smoothing improves MU performance. Experimental results on six datasets demonstrate the effectiveness and robustness of UGradSL, achieving up to 66% improvement in unlearning accuracy with minimal additional computation. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary Machine unlearning (MU) is a way to remove old information from a computer model. The goal is to balance how much computing power is needed with how well the model works after unlearning. A new method called UGradSL helps achieve this balance by using “smoothed labels”. This means that the old information is gradually erased, not suddenly removed. The paper shows that UGradSL is a simple and effective way to unlearn information from a model. |
