Research
Spatial Memory Intelligence: Endowing World Models with Understanding-Driven Long-Term Memory
arXiv:2610.02521v1 Announce Type: new Abstract: Long-video generation and world models have shown strong potential for interactive entertainment and embodied simulation by predicting future observatio
arXiv:2610.02521v1 Announce Type: new Abstract: Long-video generation and world models have shown strong potential for interactive entertainment and embodied simulation by predicting future observations conditioned on user actions and historical memory. However, as memory sequences grow longer and their structures become increasingly complex, managing long-range spatial context becomes increasingly challenging, calling for a more intelligent and systematic memory-management strategy. Building on the advancing spatial reasoning capabilities of multimodal large language models (MLLMs) and the broader vision of unified models, we propose Spatial Memory Intelligence (SMI), the first framework to systematically employ an understanding model for spatial-memory management in long-video world models. SMI introduces four coordinated atomic operations: spatial clustering, within-cluster sparsification, action-aware retrieval, and reliability-aware filtering. Extensive experiments across multiple baselines, benchmarks, and world-model backbones demonstrate the effectiveness and generalizability of SMI, achieving comprehensive improvements in memory sparsity, generation stability, and spatial consistency.
Related
- Honeycomb: Constant-Size Scene Memory Representation for Video World Models
- Compression and Retrieval: Implicit Memory Retrieval for Video World Models
- WorldCrafter: Consistent Video World Model with Implicit 3D-aware Memory
- Latent Spatial Memory for Video World Models
- World2VLM: Distilling World Model Imagination into VLMs for Dynamic Spatial Reasoning
Source: arXiv cs.CV | 2026-10-05