Continual Reinforcement Learning on Meta-World MT5 (final) benchmark leaderboard · SOTA2 Research

Continual Reinforcement Learning on Meta-World MT5 (final)

0.166AFSRP&C

Evaluation Results

Method	Links
P&C2026.06		0.166	0.034	0.312	0.16
Nash-MTL2026.06		0.18	0.02	0.985	0.295
EWC2026.06		0.21	0.021	0.719	0.229
A-GEM2026.06		0.229	0.023	0.949	0.404
CLEAR2026.06		0.229	0.003	0.949	0.347
PCGrad2026.06		0.237	0.016	0.046	0.288
CompoNet2026.06		0.257	0.048	0.012	0.477
DER2026.06		0.323	0.106	0.836	0.214
R&D2026.06		0.353	0.123	0.7	0.504
Sleeping Robots2026.06		0.578	0.245	0.28	0.767