Summary of Bandit-feedback Online Multiclass Classification: Variants and Tradeoffs, by Yuval Filmus et al.
Bandit-Feedback Online Multiclass Classification: Variants and Tradeoffs
by Yuval Filmus, Steve Hanneke, Idan Mehalel, Shay Moran
First submitted to arxiv on: 12 Feb 2024
Categories
- Main: Machine Learning (cs.LG)
- Secondary: Machine Learning (stat.ML)
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 In this paper, researchers investigate the trade-offs between bandit feedback and full information in multiclass classification tasks within an online setting. They examine how an adaptive adversary can amplify losses compared to an oblivious one, as well as the impact of randomized learners on reducing loss compared to deterministic ones. The study focuses on the mistake bound model, providing nearly tight answers to these questions. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary Machine learning researchers are trying to figure out what happens when we use bandit feedback instead of getting all the information in a multiclass classification task that’s happening online. They’re looking at how an attacker who can adapt to our moves can make things worse, and also seeing if using random choices helps us avoid mistakes. |
Keywords
* Artificial intelligence * Classification * Machine learning
