CARL: A Benchmark for Contextual and Adaptive Reinforcement Learning
arXiv:2110.02102
Abstract
While Reinforcement Learning has made great strides towards solving ever more complicated tasks, many algorithms are still brittle to even slight changes in their environment. This is a limiting factor for real-world applications of RL. Although the research community continuously aims at improving both robustness and generalization of RL algorithms, unfortunately it still lacks an open-source set of well-defined benchmark problems based on a consistent theoretical framework, which allows comparing different approaches in a fair, reliable and reproducibleway. To fill this gap, we propose CARL, a collection of well-known RL environments extended to contextual RL problems to study generalization. We show the urgent need of such benchmarks by demonstrating that even simple toy environments become challenging for commonly used approaches if different contextual instances of this task have to be considered. Furthermore, CARL allows us to provide first evidence that disentangling representation learning of the states from the policy learning with the context facilitates better generalization. By providing variations of diverse benchmarks from classic control, physical simulations, games and a real-world application of RNA design, CARL will allow the community to derive many more such insights on a solid empirical foundation.
References in corpus (19)
- Continuous control with deep reinforcement learning
- Model-Agnostic Meta-Learning for Fast Adaptation of Deep Networks
- Robust Adversarial Reinforcement Learning
- Meta-World: A Benchmark and Evaluation for Multi-Task and Meta Reinforcement Learning
- Agent57: Outperforming the Atari Human Benchmark
- Leveraging Procedural Generation to Benchmark Reinforcement Learning
- Provably efficient RL with Rich Observations via Latent State Decoding
- Robust Reinforcement Learning on State Observations with Learned Optimal Adversary
- The NetHack Learning Environment
- Environment Probing Interaction Policies
- Multi-Task Reinforcement Learning with Context-based Representations
- Evolving Reinforcement Learning Algorithms
- MiniHack the Planet: A Sandbox for Open-Ended Reinforcement Learning Research
- Robots Learn Increasingly Complex Tasks with Intrinsic Motivation and Automatic Curriculum Learning
- Probabilistic Active Meta-Learning
- TOAD-GAN: Coherent Style Level Generation from a Single Example
- Self-Paced Context Evaluation for Contextual Reinforcement Learning
- A New Representation of Successor Features for Transfer across Dissimilar Environments
- Learning Task Informed Abstractions