VL-MemKnG hybridizes a spatio-temporal knowledge graph with segment-level memory to improve evidence retrieval for question answering on long egocentric navigation trajectories, raising Top-1 accuracy from 58% to 67% on the new WalkieKnowledgeT+ benchmark.
Findings of the Association for Computational Linguistics: ACL 2025 , pages =
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.RO 1years
2026 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
VL-MemKnG: Hybrid Memory with a Spatio-Temporal Knowledge Graph for Question Answering over Long Egocentric Navigation Trajectories
VL-MemKnG hybridizes a spatio-temporal knowledge graph with segment-level memory to improve evidence retrieval for question answering on long egocentric navigation trajectories, raising Top-1 accuracy from 58% to 67% on the new WalkieKnowledgeT+ benchmark.