Prism: Spectral Parameter Sharing for Multi-Agent Reinforcement Learning
Not on the map. Why?
Abstract (arXiv)
Parameter sharing is a key strategy in multi-agent reinforcement learning (MARL) for improving scalability, yet conventional fully shared architectures often collapse into homogeneous behaviors. Recent methods introduce diversity through clustering, pruning, or masking, but typically compromise resource efficiency. We propose Prism, a parameter sharing framework that induces inter-agent diversity by representing shared networks in the spectral domain via singular value decomposition (SVD). All agents share the singular vector directions while learning distinct spectral masks on singular values. This mechanism encourages inter-agent diversity and preserves scalability. Extensive experiments on both homogeneous (LBF, SMACv2) and heterogeneous (MaMuJoCo) benchmarks show that Prism achieves competitive performance with superior resource efficiency.
Key facts. Cited by 1 paper in the map (data through 7 Oct 2026). Builds on 17 of its 22 references that are in the map. First posted on arXiv 6 Feb 2026.
Builds on
17 of 22 references.
References · 17
- 2013Estimating or Propagating Gradients Through Stochastic Neurons for Conditional Computationcited by 2,031
- 2017Multi-Agent Actor-Critic for Mixed Cooperative-Competitive Environmentscited by 1,519
- 2018QMIX: Monotonic Value Function Factorisation for Deep Multi-Agent Reinforcement Learningcited by 697
- 2020FACMAC: Factored Multi-Agent Centralised Policy Gradientscited by 178
- 2024Deep Reinforcement Learning for Robotics: A Survey of Real-World Successescited by 153
- 2022SMACv2: An Improved Benchmark for Cooperative Multi-Agent Reinforcement Learningcited by 118
- 2020Shared Experience Actor-Critic for Multi-Agent Reinforcement Learningcited by 114
- 2021Scaling Multi-Agent Reinforcement Learning with Selective Parameter Sharingcited by 85
- 2021Initialization and Regularization of Factorized Neural Layerscited by 58
- 2020Learning Low-rank Deep Neural Networks via Singular Vector Orthogonality Regularization and Singular Value Sparsificationcited by 52
- 2019Reducing Overestimation Bias in Multi-Agent Domains Using Double Centralized Criticscited by 46
- 2015A Game-Theoretic Model and Best-Response Learning Method for Ad Hoc Coordination in Multiagent Systemscited by 41
5 more
- 2022Revisiting Some Common Practices in Cooperative Multi-Agent Reinforcement Learningcited by 32
- 2022Transferring Multi-Agent Reinforcement Learning Policies for Autonomous Driving using Sim-to-Realcited by 16
- 2023Parameter Sharing with Network Pruning for Scalable Multi-Agent Deep Reinforcement Learningcited by 13
- 2024HyperMARL: Adaptive Hypernetworks for Multi-Agent RLcited by 10
- 2023Adaptive parameter sharing for multi-agent reinforcement learningcited by 6
Built on by
Citing papers per year: 2026: 1.
1 paper.