List your company

Multi-task Reinforcement Learning on Meta-World MT10 Conditioned v1

71.8Average Success RateSoft Modularization (Shallow)
25.20837.30449.461.496Mar 30, 2020

Evaluation results

MethodLinks
Soft Modularization (Shallow)Network Depth=Shallow2020.0371.8
Soft Modularization (Deep)Network Depth=Deep2020.0368.4
MT-MH-SACImplementation=Internal implementation baseline2020.0367.4
MT-SACImplementation=Internal implementation baseline2020.0342.6
Mix-ExpertImplementation=Internal implementation baseline2020.0340
Hard RoutingImplementation=Internal implementation baseline2020.0327