Stochastic Prediction of Multi-Agent Interactions from Partial Observations

2019/02/25 by Sun, Chen, Karlsson, Per, Wu, Jiajun +2
#Computer Vision and Pattern Recognition (cs.CV) #FOS: Computer and information sciences #Machine Learning (cs.LG) #Machine Learning (stat.ML)

paper · doi:10.48550/arxiv.1902.09641

Abstract

We present a method that learns to integrate temporal information, from a learned dynamics model, with ambiguous visual information, from a learned vision model, in the context of interacting agents. Our method is based on a graph-structured variational recurrent neural network (Graph-VRNN), which is trained end-to-end to infer the current state of the (partially observed) world, as well as to forecast future states. We show that our method outperforms various baselines on two sports datasets, one based on real basketball trajectories, and one generated by a soccer game engine.

Related

• Time-series Imputation of Temporally-occluded Multiagent Trajectories
• A Graph Attention Based Approach for Trajectory Prediction in Multi-agent Sports Games
• Neural Relational Inference for Interacting Systems
• Learning Interacting Dynamical Systems with Latent Gaussian Process ODEs
• STEMFold: Stochastic Temporal Manifold for Multi-Agent Interactions in the Presence of Hidden Agents