Extending Environments To Measure Self-Reflection In Reinforcement Learning
arXiv:2110.06890 · doi:10.2478/jagi-2022-0001
Abstract
We consider an extended notion of reinforcement learning in which the environment can simulate the agent and base its outputs on the agent's hypothetical behavior. Since good performance usually requires paying attention to whatever things the environment's outputs are based on, we argue that for an agent to achieve on-average good performance across many such extended environments, it is necessary for the agent to self-reflect. Thus weighted-average performance over the space of all suitably well-behaved extended environments could be considered a way of measuring how self-reflective an agent is. We give examples of extended environments and introduce a simple transformation which experimentally seems to increase some standard RL agents' performance in a certain type of extended environment.
arXiv comment24 pages, 2 figures, 1 table, 2 listings
References in corpus (9)
- The Arcade Learning Environment: An Evaluation Platform for General Agents
- Benchmarking Neural Network Robustness to Common Corruptions and Perturbations
- Provably Bounded-Optimal Agents
- Leveraging Procedural Generation to Benchmark Reinforcement Learning
- Gotta Learn Fast: A New Benchmark for Generalization in RL
- Bad Universal Priors and Notions of Optimality
- Risks from Learned Optimization in Advanced Machine Learning Systems
- Detecting Qualia in Natural and Artificial Agents
- Reward-Punishment Symmetric Universal Intelligence