Summary of On the Limitations Of Markovian Rewards to Express Multi-objective, Risk-sensitive, and Modal Tasks, by Joar Skalse and Alessandro Abate
On the Limitations of Markovian Rewards to Express Multi-Objective, Risk-Sensitive, and Modal Tasks
by Joar Skalse, Alessandro Abate
First submitted to arxiv on: 26 Jan 2024
Categories
- Main: Artificial Intelligence (cs.AI)
- Secondary: Machine Learning (cs.LG)
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 paper investigates the limitations of scalar, Markovian reward functions in Reinforcement Learning (RL), focusing on three classes: multi-objective RL, risk-sensitive RL, and modal RL. Researchers derive necessary and sufficient conditions for expressing each class using a scalar, Markovian reward. Surprisingly, they find that most instances in these classes cannot be expressed with standard rewards. The study contributes to a deeper understanding of what reward functions can and cannot express. Additionally, the authors highlight the importance of modal problems, which have been overlooked in RL literature until now. They also outline potential solutions for some of these problems using bespoke RL algorithms. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary This paper looks at how well certain types of rewards work in Reinforcement Learning (RL). It’s like trying to describe a complicated problem with just one number. The researchers tested three kinds of rewards on different tasks and found that most of the time, they’re not enough. They showed what kind of problems these rewards can solve and which ones they can’t. This helps us understand what we can expect from these rewards. The paper also introduces a new type of problem called modal problems, which haven’t been studied much before. |
Keywords
* Artificial intelligence * Reinforcement learning
