Summary of Small Scale Data-free Knowledge Distillation, by He Liu et al.
Small Scale Data-Free Knowledge Distillation
by He Liu, Yikai Wang, Huaping Liu, Fuchun Sun, Anbang Yao
First submitted to arxiv on: 12 Jun 2024
Categories
- Main: Computer Vision and Pattern Recognition (cs.CV)
- Secondary: Artificial Intelligence (cs.AI); 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 paper proposes a novel approach to data-free knowledge distillation, which enables the training of smaller student networks without accessing the original training data. This method, called Small Scale Data-free Knowledge Distillation (SSD-KD), improves upon existing methods by introducing a modulating function to balance synthetic samples and a priority sampling function to select proper samples. SSD-KD uses a dynamic replay buffer and reinforcement learning strategy to condition distillation training on an extremely small scale of synthetic samples, resulting in faster training efficiency while maintaining competitive model performance. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary The paper’s main contribution is the development of a new data-free knowledge distillation method that can perform well with only a fraction of the original training data. The method uses generative adversarial networks to synthesize synthetic samples and then trains a smaller student network on these samples. The authors show that their approach can achieve better performance than existing methods while reducing the amount of training data needed. |
Keywords
* Artificial intelligence * Distillation * Knowledge distillation * Reinforcement learning
