[NeurIPS 2026] Linguistic Trajectory Encoding (LTE): object-centric spatiotemporal memory for embodied agents. CPU and real-video demos with SAM3, ViPE, DINOv2, and Qwen3-VL. Spatial Memory Benchmark (SMB). arXiv:2609.04802.
video-understanding long-term-memory embodied-ai ego4d spatiotemporal-memory trajectory-compression neurips-2026 linguistic-trajectory-encoding object-centric-memory spatial-memory-benchmark egolife natural-language-retrieval
-
Updated
Sep 26, 2026 - Python