Summary of Decentralized Multi-task Online Convex Optimization Under Random Link Failures, by Wenjing Yan and Xuanyu Cao
Decentralized Multi-Task Online Convex Optimization Under Random Link Failures
by Wenjing Yan, Xuanyu Cao
First submitted to arxiv on: 4 Jan 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 A decentralized optimization method for multi-task online convex optimization is presented in this paper, focusing on handling random link failures between agents due to network congestion or other factors. The authors develop a robust decentralized saddle-point algorithm that replaces missing neighbor decisions with their latest received values and bounds the accumulated deviation caused by these replacements. The algorithm achieves O(sqrt(T)) regret and O(T^(3/4)) constraint violations in the full information scenario, matching the performance bounds of algorithms with perfect communications. The approach is extended to the two-point bandit feedback scenario, where only partial cost function values are disclosed to agents. Numerical simulations confirm the effectiveness of the proposed algorithm. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary Decentralized optimization helps groups of computers work together efficiently. When these computers share information with each other, problems can occur due to network issues or hardware failures. This paper solves this problem by creating an algorithm that can adapt when some information is missing. The algorithm replaces missing data with what it knows, and then adjusts for the errors caused by these replacements. The result is a way for computers to work together more efficiently even when there are occasional problems. |
Keywords
* Artificial intelligence * Multi task * Optimization
