Summary of Understanding the Gains From Repeated Self-distillation, by Divyansh Pareek et al.
Understanding the Gains from Repeated Self-Distillation
by Divyansh Pareek, Simon S. Du, Sewoong Oh
First submitted to arxiv on: 5 Jul 2024
Categories
- Main: Machine Learning (cs.LG)
- Secondary: Machine Learning (stat.ML)
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 an investigation into the effectiveness of self-distillation, a technique that improves performance by having student models with the same architecture as teacher models. The authors aim to understand how much gain can be achieved by applying multiple steps of self-distillation, using linear regression as a canonical task. Their analysis shows that multi-step self-distillation can significantly reduce excess risk, with potential gains up to 47% on certain datasets. They also analyze the impact of input dimensionality on this improvement. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary Self-Distillation is a special type of knowledge distillation where the student model has the same architecture as the teacher model. Despite using the same architecture and training data, self-distillation has been empirically observed to improve performance. The paper investigates how much gain can be achieved by applying multiple steps of self-distillation, focusing on linear regression tasks from the UCI repository. |
Keywords
» Artificial intelligence » Distillation » Knowledge distillation » Linear regression » Student model » Teacher model
