Summary of Vit-mul: a Baseline Study on Recent Machine Unlearning Methods Applied to Vision Transformers, by Ikhyun Cho et al.
ViT-MUL: A Baseline Study on Recent Machine Unlearning Methods Applied to Vision Transformers
by Ikhyun Cho, Changyeon Park, Julia Hockenmaier
First submitted to arxiv on: 7 Feb 2024
Categories
- Main: Computer Vision and Pattern Recognition (cs.CV)
- Secondary: 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 A novel machine learning technique called Machine Unlearning (MUL) aims to erase specific training data points from a trained model. Although recent work on MUL has focused on ResNet-based models, Vision Transformers (ViT) have become the dominant architecture, making it essential to study MUL tailored to ViTs. This paper presents comprehensive experiments using recent MUL algorithms and datasets, with ablation studies and findings that could provide valuable insights and inspire further research. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary Machine unlearning is a new way to erase specific training data points from a trained model. Right now, most research on this topic focuses on ResNet-based models, but Vision Transformers are the most popular type of model, so it’s important to study MUL with these models too. This paper does some experiments and studies to help us understand how well MUL works with ViTs. |
Keywords
* Artificial intelligence * Machine learning * Resnet * Vit
