analiza aiqbench
MemoryVLN addresses critical bottlenecks in embodied navigation by introducing a memory architecture that compresses long trajectories while preserving spatial grounding through 2D, 3D, and pose tokens. This approach effectively mitigates context-length limits and unstable alignment issues inherent to large vision-language models in long-horizon tasks. The method demonstrates significant performance gains on standard benchmarks by enabling reliable reuse of past spatial experiences without redundant observations.
MemoryVLN: From Ego-Aware Memory Representation to Cross-Context Memory Reuse in Vision-Language Navigation
arXiv:2610.18478v1 Announce Type: cross Abstract: Large vision-language models (LVLMs) have shown strong potential for embodied vision-language navigation (VLN), yet their performance degrades on long-horizon trajectories due to context-length limits, redundant historical observations, and unstable spatial alignment across time. These issues make it difficult to both compress long navigation histories and reliably reuse past spatial experience. To address these limitations, we propose \textbf{MemoryVLN}, a unified memory-driven navigation system built around memory construction and management. During memory construction, \emph{Long Sequence Compression} employs forgetting-curve-inspired temporal sampling to preserve dense recent evidence and sparse yet representative distant observations under a limited token budget. \emph{Ego-Aware Memory Representation} further integrates 2D semantic tokens, 3D geometric tokens, and 6D agent poses, allowing each memory to encode both what was observed and where it was observed. During memory management, \emph{Hierarchical Memory Bank} maintains in-episode and context-level memories through cross-context warm-start and confidence-controlled write-back. \emph{Memory Enhancement} then retrieves the Top-$K$ relevant memories according to the current navigation state and fuses them with the current visual tokens. Together, these mechanisms transform raw navigation history into compact, spatially grounded, selectively retrievable, and reusable navigation experience. Extensive experiments on R2R-CE and RxR-CE demonstrate consistent improvements over the state-of-the-art baselines by a substantial margin (i.e., 2\% and 6.5\% in successful rate).
Abstract:Large vision-language models (LVLMs) have shown strong potential for embodied vision-language navigation (VLN), yet their performance degrades on long-horizon trajectories due to context-length limits, redundant historical observations, and unstable spatial alignment across time. These issues make it difficult to both compress long navigation histories and reliably reuse past spatial experience. To address these limitations, we propose \textbf{MemoryVLN}, a unified memory-driven navigation system built around memory construction and management. During memory construction, \emph{Long Sequence Compression} employs forgetting-curve-inspired temporal sampling to preserve dense recent evidence and sparse yet representative distant observations under a limited token budget. \emph{Ego-Aware Memory Representation} further integrates 2D semantic tokens, 3D geometric tokens, and 6D agent poses, allowing each memory to encode both what was observed and where it was observed. During memory management, \emph{Hierarchical Memory Bank} maintains in-episode and context-level memories through cross-context warm-start and confidence-controlled write-back. \emph{Memory Enhancement} then retrieves the Top-$K$ relevant memories according to the current navigation state and fuses them with the current visual tokens. Together, these mechanisms transform raw navigation history into compact, spatially grounded, selectively retrievable, and reusable navigation experience. Extensive experiments on R2R-CE and RxR-CE demonstrate consistent improvements over the state-of-the-art baselines by a substantial margin (i.e., 2\% and 6.5\% in successful rate).
Robotics (cs.RO); Artificial Intelligence (cs.AI)
arXiv:2610.18478 [cs.RO]
(or arXiv:2610.18478v1 [cs.RO] for this version)
https://doi.org/10.48550/arXiv.2610.18478 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website.
Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them.
Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs.
Text integral sub licența deschisă a publicației · 299 cuvinte
Subiecte conexe
MIND: Lightweight and Effective Memory Injection Defense for LLM Agents via Intent-Aware Information Bottleneck
arXiv · 15 oct. 2026
Enhancing Long-Video VLM Embeddings with Query-Aware Streaming Latent Reasoning
arXiv · 7 oct. 2026
Beyond RAG: Task-aware knowledge compression for enterprise AI on AWS
AWS · 27 iul. 2026
Benchmarking Kimi-K3 Across vLLM, SGLang, and ATOM on MI350X
AMD · 22 sept. 2026