Yang Zhang

Yang Zhang, Xinran Li, Jianing Ye et al.

World models have recently attracted growing interest in Multi-Agent Reinforcement Learning (MARL) due to their ability to improve sample efficiency for policy learning. However, accurately modeling environments in MARL is challenging due to the exponentially large joint action space and highly uncertain dynamics inherent in multi-agent systems. To address this, we reduce modeling complexity by shifting from jointly modeling the entire state-action transition dynamics to focusing on the state space alone at each timestep through sequential agent modeling. Specifically, our approach enables the model to progressively resolve uncertainty while capturing the structured dependencies among agents, providing a more accurate representation of how agents influence the state. Interestingly, this sequential revelation of agents' actions in a multi-agent system aligns with the reverse process in diffusion models--a class of powerful generative models known for their expressiveness and training stability compared to autoregressive or latent variable models. Leveraging this insight, we develop a flexible and robust world model for MARL using diffusion models. Our method, Diffusion-Inspired Multi-Agent world model (DIMA), achieves state-of-the-art performance across multiple multi-agent control benchmarks, significantly outperforming prior world models in terms of final return and sample efficiency, including MAMuJoCo and Bi-DexHands. DIMA establishes a new paradigm for constructing multi-agent world models, advancing the frontier of MARL research. Codes are open-sourced at https://github.com/breez3young/DIMA.

Jun Pang, Yang Zhang

The popularity of online social networks (OSNs) makes the protection of users' private information an important but scientifically challenging problem. In the literature, relationship-based access control schemes have been proposed to address this problem. However, with the dynamic developments of OSNs, we identify new access control requirements which cannot be fully captured by the current schemes. In this paper, we focus on public information in OSNs and treat it as a new dimension which users can use to regulate access to their resources. We define a new OSN model containing users and their relationships as well as public information. Based on this model, we introduce a variant of hybrid logic for formulating access control policies. We exploit a type of category information and relationship hierarchy to further extend our logic for its usage in practice. In the end, we propose a few solutions to address the problem of information reliability in OSNs, and formally model collaborative access control in our access control scheme.