This is the pre-proceedings for the RLC 2026. You may expect minor changes.
Reinforcement Learning Journal, vol. 7, 2026, pp. TBD.
Will be presented at the Reinforcement Learning Conference (RLC), Montréal, Quebec, Canada, August 15–17, 2026.
Keywords: Multi-agent RL, Transfer Learning, Sample Efficiency, Observation Space Mismatch
Multi-agent reinforcement learning (MARL) addresses the problem of training multiple agents that pursue collaborative, competitive, or mixed objectives. Prior work has investigated transfer learning between source and target domains in MARL; however, the majority of existing approaches impose the constraint that the dimensionalities of the observation space and the global state space must be identical across domains. In this paper, we propose a method that explicitly accounts for and reconciles the dimensionality mismatch between the source and target domains. The proposed approach, ASALT, incorporates both observation-level and state-level adapters that map the target-domain observations and global states into a shared embedding space, thereby enabling more effective knowledge transfer across both actors and critics: observation-adapter transfer conveys individual skill, while state-adapter transfer conveys strategic/coordination knowledge captured only by the global state. Experimental results on multiple configurations in standard benchmark environments demonstrate that ASALT surpasses existing baselines in terms of sample efficiency and global return in cooperative settings, but its effectiveness depends on the degree of mismatch between source and target domains. Furthermore, our findings indicate that ASALT mitigates negative transfer, which is most pronounced in domains consisting of agents with differing observation and action spaces.
Anurag Akula, Satheesh K Perepu, Abhishek Sarkar, and Kaushik Dey. "ASALT: Adaptive State Alignment for Lateral Transfer in Multi-agent Reinforcement Learning." Reinforcement Learning Journal, vol. 7, 2026, pp. TBD.
BibTeX:@article{akula2026asalt,
title={ASALT: Adaptive State Alignment for Lateral Transfer in Multi-agent Reinforcement Learning},
author={Anurag Akula and Satheesh K Perepu and Abhishek Sarkar and Kaushik Dey},
journal={Reinforcement Learning Journal},
volume={7},
pages={},
year={2026}
}