Multi-Task Reinforcement Learning with Soft Modularization

Benchmark Model Rank Results
meta-learning-on-mt50SoftModule#1Average Success Rate: 60.0%