Summary of Finite-time Convergence and Sample Complexity Of Actor-critic Multi-objective Reinforcement Learning, by Tianchen Zhou et al.
Finite-Time Convergence and Sample Complexity of Actor-Critic Multi-Objective Reinforcement Learning
by Tianchen Zhou, FNU Hairi, Haibo Yang, Jia Liu, Tian Tong, Fan Yang, Michinari Momma, Yan Gao
First submitted to arxiv on: 5 May 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 tackles the under-explored multi-objective reinforcement learning (MORL) problem by introducing an innovative actor-critic algorithm named MOAC. This algorithm iteratively makes trade-offs among conflicting reward signals to find a policy. The authors provide the first analysis of finite-time Pareto-stationary convergence and corresponding sample complexity in both discounted and average reward settings. The approach has two salient features: mitigating cumulative estimation bias using an optimal common gradient descent direction, and initializing policy gradients with samples from the environment. This enhances the practicality and robustness of MOAC. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary The paper introduces a new algorithm called MOAC that helps solve multi-objective reinforcement learning problems. These types of problems happen when we have multiple goals or rewards in a situation. The authors show how their algorithm can find a good solution by balancing these conflicting goals. They also prove that their algorithm works well and provide some guarantees about how fast it will work. |
Keywords
* Artificial intelligence * Gradient descent * Reinforcement learning
