Summary of Understanding Adam Optimizer Via Online Learning Of Updates: Adam Is Ftrl in Disguise, by Kwangjun Ahn et al.
Understanding Adam Optimizer via Online Learning of Updates: Adam is FTRL in Disguise
by Kwangjun Ahn, Zhiyu Zhang, Yunbum Kook, Yan Dai
First submitted to arxiv on: 2 Feb 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 paper provides new insights into the Adam optimizer’s algorithmic components, challenging existing analyses that show convergence rates comparable to non-adaptive algorithms like SGD. By applying an online learning framework, the authors demonstrate that Adam corresponds to a principled approach called Follow-the-Regularized-Leader (FTRL). This framework allows for the design of good optimizers by designing effective online learners. The study highlights the importance of Adam’s algorithmic components and sheds light on their benefits from an online learning perspective. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary The paper is about understanding how the Adam optimizer works better. It looks at how Adam compares to other algorithms that don’t adapt as well, like SGD. The authors found a new way to think about Adam by using an “online learning” approach. This helped them see why Adam’s special components are important and how they make it work better. |
Keywords
* Artificial intelligence * Online learning
