{"ID":5551955,"CreatedAt":"2026-07-02T01:54:51.863792489Z","UpdatedAt":"2026-07-04T05:21:04.038526805Z","DeletedAt":null,"paper_url":"https://arxiv.org/abs/2607.00457","arxiv_id":"2607.00457","title":"Multi-scale Mixture of World Models for Embodied Agents in Evolving Environments","abstract":"Embodied agents operating in the real world require multi-scale reasoning and knowledge adaptation as conditions change. We identify two challenges in applying Mixture of Experts (MoE) to this setting: routing lacks an explicit notion of scale, preventing targeted updates at specific scales, and a uniform update policy cannot accommodate the different rates at which knowledge at each scale becomes outdated. We present MuSix, a framework that addresses both challenges through scale-aware world model mixture and evolution. A two-stage routing mechanism grounds scale selection in experiential distance, a measure of situational novelty inspired by Construal Level Theory: a meta-router first maps this quantity to a weight over continuous scale space, then per-scale base routers select world models within the identified scale. For adaptation, scale-dependent forgetting rates allow low-scale knowledge to refresh rapidly while high-scale abstractions persist, and gated inter-scale transfer maintains coherence across the hierarchy. Experiments on EmbodiedBench and HAZARD show that MuSix improves over state-of-the-art baselines on multi-scale reasoning and dynamic adaptation.","short_abstract":"Embodied agents operating in the real world require multi-scale reasoning and knowledge adaptation as conditions change. We identify two challenges in applying Mixture of Experts (MoE) to this setting: routing lacks an explicit notion of scale, preventing targeted updates at specific scales, and a uniform update policy...","url_abs":"https://arxiv.org/abs/2607.00457","url_pdf":"https://arxiv.org/pdf/2607.00457v1","authors":"[\"Jinwoo Jang\",\"Daniel J. Rho\",\"Sihyung Yoon\",\"Hyunsuk Cho\",\"Honguk Woo\"]","published":"2026-07-01T05:23:56Z","proceeding":"cs.AI","tasks":"[\"cs.AI\"]","methods":"[\"Mixture of Experts\"]","has_code":false}