{"ID":2887833,"CreatedAt":"2026-06-01T04:54:23.091178241Z","UpdatedAt":"2026-06-01T04:54:23.091178241Z","DeletedAt":null,"paper_url":"https://arxiv.org/abs/2508.00384","arxiv_id":"2508.00384","title":"On Learning Closed-Loop Probabilistic Multi-Agent Simulator","abstract":"The rapid iteration of autonomous vehicle (AV) deployments leads to increasing needs for building realistic and scalable multi-agent traffic simulators for efficient evaluation. Recent advances in this area focus on closed-loop simulators that enable generating diverse and interactive scenarios. This paper introduces Neural Interactive Agents (NIVA), a probabilistic framework for multi-agent simulation driven by a hierarchical Bayesian model that enables closed-loop, observation-conditioned simulation through autoregressive sampling from a latent, finite mixture of Gaussian distributions. We demonstrate how NIVA unifies preexisting sequence-to-sequence trajectory prediction models and emerging closed-loop simulation models trained on Next-token Prediction (NTP) from a Bayesian inference perspective. Experiments on the Waymo Open Motion Dataset demonstrate that NIVA attains competitive performance compared to the existing method while providing embellishing control over intentions and driving styles.","short_abstract":"The rapid iteration of autonomous vehicle (AV) deployments leads to increasing needs for building realistic and scalable multi-agent traffic simulators for efficient evaluation. Recent advances in this area focus on closed-loop simulators that enable generating diverse and interactive scenarios. This paper introduces N...","url_abs":"https://arxiv.org/abs/2508.00384","url_pdf":"https://arxiv.org/pdf/2508.00384v1","authors":"[\"Juanwu Lu\",\"Rohit Gupta\",\"Ahmadreza Moradipari\",\"Kyungtae Han\",\"Ruqi Zhang\",\"Ziran Wang\"]","published":"2025-08-01T07:25:59Z","proceeding":"cs.RO","tasks":"[\"cs.RO\"]","methods":"[]","has_code":false}