Summary of Learning Adversarial Mdps with Stochastic Hard Constraints, by Francesco Emanuele Stradi et al.
Learning Adversarial MDPs with Stochastic Hard Constraints
by Francesco Emanuele Stradi, Matteo Castiglioni, Alberto Marchesi, Nicola Gatti
First submitted to arxiv on: 6 Mar 2024
Categories
- Main: Machine Learning (cs.LG)
- Secondary: None
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 investigates online learning in constrained Markov decision processes (CMDPs) with adversarial losses and stochastic hard constraints under bandit feedback. The authors propose three algorithms for tackling this problem: one for general CMDPs, another that assumes the existence of a policy satisfying the constraints, and a third that only requires the existence of a feasible policy. These algorithms achieve sublinear regret and cumulative positive constraints violation. The paper’s main contribution is the study of CMDPs involving both adversarial losses and hard constraints, which enables the adoption of these algorithms in a wider range of applications. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary Online learning in constrained Markov decision processes (CMDPs) means finding the best way to make decisions while following certain rules. This problem gets even harder when there are bad things happening and we don’t know what will happen next. The researchers came up with three ways to solve this problem, each one good at handling different situations. They showed that their methods work well in many different scenarios and can be used in a lot of real-world applications. |
Keywords
* Artificial intelligence * Online learning
