Summary of Inferring Preferences From Demonstrations in Multi-objective Reinforcement Learning, by Junlin Lu et al.
Inferring Preferences from Demonstrations in Multi-objective Reinforcement Learning
by Junlin Lu, Patrick Mannion, Karl Mason
First submitted to arxiv on: 30 Sep 2024
Categories
- Main: Artificial Intelligence (cs.AI)
- 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 This research proposes a dynamic weight-based preference inference (DWPI) algorithm that can infer the preferences of agents acting in multi-objective decision-making problems from demonstrations. The algorithm is designed to handle complex decision-making scenarios where multiple objectives are involved, and human or agent decision-makers’ preferences are unknown. In these situations, the DWPI algorithm uses demonstrated behaviors to infer preferences, which is crucial for various applications such as traffic management, item gathering, and treasure hunting (Deep Sea Treasure). The proposed algorithm outperforms existing preference inference algorithms in terms of time efficiency and inference accuracy on three multi-objective Markov decision processes. Additionally, the algorithm can maintain its performance even when inferring preferences from sub-optimal demonstrations. Furthermore, DWPI does not require any interaction with the user during inference, making it a valuable tool for real-world applications. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary This research solves a big problem in decision-making where multiple goals are involved and we don’t know what people or agents want to achieve. The solution is an algorithm that can figure out people’s preferences by looking at how they behave when given different options. This algorithm works well even when the person making the decision doesn’t make the best choice. It also doesn’t need any extra help from humans, which makes it useful for real-life situations like traffic control or finding treasures. |
Keywords
» Artificial intelligence » Inference
