Summary of Approximate Nullspace Augmented Finetuning For Robust Vision Transformers, by Haoyang Liu et al.
Approximate Nullspace Augmented Finetuning for Robust Vision Transformers
by Haoyang Liu, Aditya Singh, Yijiang Li, Haohan Wang
First submitted to arxiv on: 15 Mar 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 The proposed approach enhances the robustness of vision transformers (ViTs) by exploiting the concept of nullspace from linear algebra. The paper shows that many pretrained ViTs have a non-trivial nullspace due to the patch embedding layer, allowing for resilience to input variations. By synthesizing approximate nullspace elements using an optimization strategy, the model demonstrates robustness to both adversarial and natural image perturbations after fine-tuning with synthesized noise. This work has implications for the real-world deployment of ViTs. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary Deep learning models, like vision transformers (ViTs), need to be super strong in real-life situations. In this study, scientists found a way to make ViTs more robust by using an idea from math called nullspace. They showed that many pre-trained ViTs have a special property that helps them ignore certain types of noise. Then, they figured out how to create fake versions of these noise patterns and used them to train the model. As a result, the model became much better at handling unexpected changes in images. |
Keywords
* Artificial intelligence * Deep learning * Embedding * Fine tuning * Optimization
