Summary of Towards Stable and Storage-efficient Dataset Distillation: Matching Convexified Trajectory, by Wenliang Zhong and Haoyu Tang and Qinghai Zheng and Mingzhu Xu and Yupeng Hu and Liqiang Nie
Towards Stable and Storage-efficient Dataset Distillation: Matching Convexified Trajectory
by Wenliang Zhong, Haoyu Tang, Qinghai Zheng, Mingzhu Xu, Yupeng Hu, Liqiang Nie
First submitted to arxiv on: 28 Jun 2024
Categories
- Main: Machine Learning (cs.LG)
- Secondary: None
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, Matching Convexified Trajectory (MCT), aims to improve upon existing Dataset Distillation methods by introducing a novel technique that leverages insights from Neural Tangent Kernel methods. By creating a convex combination of expert trajectories, MCT guides the student network to converge rapidly and stably, enabling continuous sampling during distillation and thorough learning. The method addresses three limitations of traditional Matching Training Trajectories (MTT): instability, low convergence speed, and high storage consumption. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary The paper proposes a new approach to dataset distillation called Matching Convexified Trajectory (MCT). MCT is designed to improve upon existing methods by providing better guidance for the student trajectory. It does this by creating a convex combination of expert trajectories using insights from Neural Tangent Kernel methods. This allows the student network to learn more quickly and thoroughly, while also being easier to store. |
Keywords
» Artificial intelligence » Distillation
