AI Paper Atlas

Research map / Vision-language and multimodal / Robot action and navigation

LASAR: Towards Spatio-temporal Reasoning with Latent Cognitive Map

Jinzhou Tang, Sidi Liu, Waikit Xiu, Weixing Chen, Keze Wang

CVPR 2026 · Vision-language and multimodal › Robot action and navigation · approach: Embodied multimodal LLM agents · 0 citations (OpenAlex)

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
add ST-CRL loss (ablation)LASAR (IL+QA) · MindCraft-TestQA-Acc: 60.6 -> 63.5Improved
add semantic atlas and episodic discriminability losses (ablation)LASAR with ST-CRL · MindCraft-TestQA-Acc: 63.5 -> 65.3Improved
add geometric feature fusion (VGGT) (ablation)LASAR · MindCraft-TestGCA: 62.1 -> 70.4Improved
add geometric feature fusion (VGGT) (ablation)LASAR · MindCraft-TestCMC: 66.3 -> 75.8Improved
add Spatial Semantic Memory (latent cognitive map) (ablation)LASAR · MindCraft-TestCMC: 58.2 -> 75.8Improved
add Spatial Semantic Memory (latent cognitive map) (ablation)LASAR · MindCraft-TestSR@WA: 45.7 -> 35.2Improved
add Spatial Semantic Memory (latent cognitive map) (ablation)LASAR · MindCraft-TestQA-Acc: 62.1 -> 65.3Improved
add ST-CRL and auxiliary losses to IL+QA trainingLASAR (IL+QA) · VSI-Bench (zero-shot)Avg: 44.8 -> 48.9Improved
add MindCraft QA training to imitation-onlyLASAR (IL) · VSI-Bench (zero-shot)Avg: 37.8 -> 44.8Improved
add QA loss to imitation-only trainingLASAR (IL) · R2R val-unseenSPL: 52.7 -> 47.4Hurt
add ST-CRL and auxiliary losses to IL+QA trainingLASAR (IL+QA) · R2R val-unseenSPL: 47.4 -> 53.9Improved
use LASAR instead of NaVILAnavigation on R2R · R2R val-unseenSR: 54.8 -> 57.0Improved

Closest papers by meaning

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.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy