Summary of Factual: a Novel Framework For Contrastive Learning Based Robust Sar Image Classification, by Xu Wang et al.
FACTUAL: A Novel Framework for Contrastive Learning Based Robust SAR Image Classification
by Xu Wang, Tian Ye, Rajgopal Kannan, Viktor Prasanna
First submitted to arxiv on: 4 Apr 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 Deep learning models for synthetic aperture radar (SAR) automatic target recognition (ATR) have shown improved performance but are vulnerable to adversarial attacks. Existing works improve robustness by training on adversarial samples, neglecting real-world feasibility of such attacks. This paper proposes FACTUAL, a novel Contrastive Learning framework for Adversarial Training and robust SAR classification. FACTUAL consists of two components: a novel perturbation scheme that incorporates realistic physical adversarial attacks (OTSA) to build a supervised adversarial pre-training network and a linear classifier cascaded after the encoder. By pre-training and fine-tuning on both clean and adversarial samples, this model achieves high prediction accuracy on both cases. It outperforms previous state-of-the-art methods with 99.7% accuracy on clean samples and 89.6% on perturbed samples. |
Low | GrooveSquid.com (original content) | Low Difficulty Summary This paper is about making computer models better at recognizing things in pictures taken from space using a special kind of radar. These models are good, but they can be tricked into making mistakes by adding fake information to the pictures. The people who came up with this new model, called FACTUAL, tried to make it more honest and accurate. They did this by training the model on both real and fake pictures, so it can learn to tell the difference. Their new model is really good at recognizing things in pictures taken from space, even when there’s fake information added. This is important because it could help us use radar technology to do things like track ships or find hidden buildings. |
Keywords
» Artificial intelligence » Classification » Deep learning » Encoder » Fine tuning » Supervised