Research map / Probabilistic and scientific modeling / Dynamical systems
Learning Hierarchical World Models with Adaptive Temporal Abstractions from Discrete Latent Dynamics
Read the paperCheck your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's abstract only. Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Proposes THICK, a method that learns a hierarchical world model for model-based RL by using discrete latent dynamics with adaptive temporal abstractions.
Problem
Flat world models in model-based RL struggle to reason across multiple time scales, while hierarchical world models are less commonly used.
Approach
Learns a world model hierarchy via discrete latent dynamics (Temporal Hierarchies from Invariant Context Kernels), with lower and higher levels operating at adaptive temporal abstractions.
Closest papers by meaning
- Simplifying Latent Dynamics with Softly State-Invariant World Models (NeurIPS 2024 · similarity 0.90)
- Beyond Prediction: Steering VLM Agents with Retrospective World Modeling (NeurIPS 2026 · similarity 0.88)
- Dynamics-Aligned Latent Imagination in Contextual World Models for Zero-Shot Generalization (NeurIPS 2025 · similarity 0.88)
- Emergent Hierarchical Reasoning in LLMs through Reinforcement Learning (ICLR 2026 · similarity 0.88)
- HaM-World: Soft-Hamiltonian World Models with Selective Memory for Planning (NeurIPS 2026 · similarity 0.88)
- Modeling Unseen Environments with Language-guided Composable Causal Components in Reinforcement Learning (ICLR 2025 · similarity 0.88)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.