O-MAPL trains cooperative MARL agents end-to-end from trajectory preferences by maximizing a Bradley-Terry likelihood in soft Q-space with linear value factorization and local weighted behavior cloning.
Title resolution pending
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.LG 1years
2025 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
O-MAPL: Offline Multi-agent Preference Learning
O-MAPL trains cooperative MARL agents end-to-end from trajectory preferences by maximizing a Bradley-Terry likelihood in soft Q-space with linear value factorization and local weighted behavior cloning.