Summary of Value Function Interference and Greedy Action Selection in Value-based Multi-objective Reinforcement Learning, by Peter Vamplew et al.
Value function interference and greedy action selection in value-based multi-objective reinforcement learning
by Peter Vamplew, Cameron Foale, Richard Dazeley
First submitted to arxiv on: 9 Feb 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 A novel approach to multi-objective reinforcement learning (MORL) tackles the challenge of optimizing for multiple conflicting objectives. The authors modify widely-used scalar RL methods like Q-learning to handle vector-valued rewards and develop a scalarisation or ordering operator that reflects user utility. However, they show that this can lead to value-function interference in stochastic environments, causing convergence to sub-optimal policies. A simple example illustrates the problem’s persistence in deterministic settings. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary Reinforcement learning is like training an AI to make good decisions. When we have multiple goals, it gets tricky! This paper shows how to handle these “multi-objective” problems. They take popular methods and adapt them to work with many goals at once. But they also find a problem: when the AI tries to balance different goals, it can get stuck in bad habits. This happens more often when things are random and unpredictable. |
Keywords
* Artificial intelligence * Reinforcement learning
