Corpus ID: 220793224

Dynamic Relational Inference in Multi-Agent Trajectories

@article{Xiao2020DynamicRI,
  title={Dynamic Relational Inference in Multi-Agent Trajectories},
  author={Ruichao Xiao and R. Yu},
  journal={ArXiv},
  year={2020},
  volume={abs/2007.13524}
}
  • Ruichao Xiao, R. Yu
  • Published 2020
  • Computer Science, Mathematics
  • ArXiv
  • Inferring interactions from multi-agent trajectories has broad applications in physics, vision and robotics. Neural relational inference (NRI) is a deep generative model that can reason about relations in complex dynamics without supervision. In this paper, we take a careful look at this approach for relational inference in multi-agent trajectories. First, we discover that NRI can be fundamentally limited without sufficient long-term observations. Its ability to accurately infer interactions… CONTINUE READING

    Figures and Tables from this paper

    References

    SHOWING 1-10 OF 55 REFERENCES
    Neural Relational Inference for Interacting Systems
    • 223
    • Highly Influential
    • PDF
    Relational Forward Models for Multi-Agent Learning
    • 29
    • PDF
    Neural Relational Inference with Fast Modular Meta-learning
    • 14
    • PDF
    Theory of Minds: Understanding Behavior in Groups Through Inverse Planning
    • 24
    • PDF
    Interaction Networks for Learning about Objects, Relations and Physics
    • 578
    • PDF
    Learning Policy Representations in Multiagent Systems
    • 36
    • PDF
    Coordinated Multi-Agent Imitation Learning
    • 83
    • PDF
    Multi-Agent Actor-Critic for Mixed Cooperative-Competitive Environments
    • 893
    • PDF
    Learning to Simulate Complex Physics with Graph Networks
    • 49
    • PDF
    Flexible Neural Representation for Physics Prediction
    • 93
    • PDF