Summary of Fast Training Of Large Kernel Models with Delayed Projections, by Amirhesam Abedsoltan et al.
Fast training of large kernel models with delayed projections
by Amirhesam Abedsoltan, Siyuan Ma, Parthe Pandit, Mikhail Belkin
First submitted to arxiv on: 25 Nov 2024
Categories
- Main: Machine Learning (stat.ML)
- 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 A novel methodology for building kernel machines that can scale efficiently with both data size and model size is proposed in this paper. The algorithm, called EigenPro4, introduces delayed projections to Preconditioned Stochastic Gradient Descent (PSGD), enabling the training of larger models than previously feasible. This breakthrough in kernel-based learning has significant implications for scaling classical kernel machines to handle large datasets and model sizes. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary This new approach is designed to overcome the limitations of traditional kernel machines, which have historically struggled with scaling. By introducing delayed projections to PSGD, EigenPro4 enables the training of larger models while maintaining comparable or better classification accuracy. The algorithm’s effectiveness is demonstrated across multiple datasets, resulting in significant speed-ups over existing methods. |
Keywords
» Artificial intelligence » Classification » Stochastic gradient descent
