Summary of Increasing Both Batch Size and Learning Rate Accelerates Stochastic Gradient Descent, by Hikaru Umeda et al.
Increasing Both Batch Size and Learning Rate Accelerates Stochastic Gradient Descent
by Hikaru Umeda, Hideaki Iiduka
First submitted to arxiv on: 13 Sep 2024
Categories
- Main: Machine Learning (cs.LG)
- Secondary: Optimization and Control (math.OC)
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 paper investigates the performance of mini-batch stochastic gradient descent (SGD) by analyzing four schedulers: constant batch size and decaying learning rate, increasing batch size and decaying learning rate, increasing batch size and increasing learning rate, and increasing batch size and warm-up decaying learning rate. The study reveals that using schedulers (i), (ii), or (iii) does not always minimize the expectation of the full gradient norm of the empirical loss, whereas scheduler (iv) accelerates mini-batch SGD and minimizes the full gradient norm faster than others. The results are supported by numerical analyses. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary The paper looks at how to make a type of machine learning algorithm called mini-batch stochastic gradient descent work better. It tries four different ways to do this, called schedulers, and finds that one of them makes the algorithm work faster and better than the others. This is important because it could help people use these algorithms more effectively. |
Keywords
» Artificial intelligence » Machine learning » Stochastic gradient descent
