Summary of Adaptive Teaching with Shared Classifier For Knowledge Distillation, by Jaeyeon Jang et al.
Adaptive Teaching with Shared Classifier for Knowledge Distillation
by Jaeyeon Jang, Young-Ik Kim, Jisu Lim, Hyeonseong Lee
First submitted to arxiv on: 12 Jun 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 This paper proposes a new knowledge distillation (KD) method called adaptive teaching with a shared classifier (ATSC). ATSC builds on recent discoveries that sharing the classifier of a teacher network can significantly boost the performance of a student network. The approach involves self-adjusting the teacher network to better align with the learning needs of the student network, and using a shared classifier to enhance the student’s performance. Experiments demonstrate the effectiveness of ATSC on CIFAR-100 and ImageNet datasets in single-teacher and multiteacher scenarios. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary This paper shares knowledge from a smart teacher to help students learn better. The teacher adjusts itself to fit the student’s needs, and they both use the same tool (classifier) to get smarter together. The researchers tested this idea on pictures and showed that it works really well, even with many teachers helping at once. |
Keywords
» Artificial intelligence » Knowledge distillation
