AI Paper Atlas

Research map / Vision-language and multimodal / Multimodal LLMs

Ask Less, See More: Communication-Conditioned Token Pruning for Vehicle-to-Vehicle Cooperative Autonomous Driving with Multimodal Large Language Models

Shiqi Sun, Yantao Lu, Bingkun Sun, Ning Liu, Bo Jiang, Ying Zhang, Jinchao Chen, Chenglie Du

ICML 2026 · Vision-language and multimodal › Multimodal LLMs · approach: Training-free token reduction for MLLMs

Read the paperCheck your idea against 137,217 papers

AI-extracted summary. Written by AI from the paper's full text. Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Proposes communication-conditioned token pruning to reduce the communication and inference cost of V2V cooperative driving with MLLMs, based only on the abstract's stated problem framing.

Problem

Existing vehicle-to-vehicle (V2V) MLLM frameworks share and fuse dense tokens, causing high communication and inference costs for language-based cooperative driving in occluded scenarios.

Approach

Conditions token pruning on communication so that only informative tokens are shared and fused, moving beyond conventional feature-sharing V2V perception; details beyond the abstract are unavailable.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
add spatial-redundancy term (lambda 1.2) to token scoring instead of lambda 0.0 (ablation)V2V-CCM · V2V-QA 70% pruningQ5 L2 error (m): 6.08 -> 5.21Improved
add spatial-redundancy term (lambda 1.2) instead of lambda 0.0 (ablation)V2V-CCM · V2V-QA 70% pruningQ5 collision rate (%): 7.54 -> 6.28Improved
increase lambda from 1.2 to 1.5 (ablation)V2V-CCM · V2V-QA 70% pruningQ5 L2 error (m): 5.21 -> 5.34Hurt
increase lambda from 0.5 to 1.0 (ablation)V2V-CCM · V2V-QA 70% pruningQ5 L2 error (m): 5.61 -> 5.43Improved
use communication-conditioned pruning instead of VisionZip at 70%V2V-LLM · V2V-QA V2V-split 70% pruningQ5 L2 error (m): 6.97 -> 5.21Improved
use communication-conditioned pruning instead of VisionZip at 70%V2V-GoT · V2V-GoT-QA 70% pruningQ9 L2 error (m): 5.48 -> 3.01Improved
use communication-conditioned pruning instead of Random at 30%V2V-LLM · V2V-QA V2V-split 30% pruningQ5 L2 error (m): 6.78 -> 5.05Improved
prune 30% of tokens with V2V-CCMV2V-GoT · V2V-GoT-QAQ9 L2 error (m): 2.62 -> 2.65No clear effect
prune 30% of tokens with V2V-CCMV2V-LLM · V2V-QA V2V-splitgrounding F1 (QGr): 40.7 -> 40.5No clear effect
prune 70% of tokens with V2V-CCMV2V-LLM · V2V-QA V2V-splitQ5 L2 error (m): 4.99 -> 5.21Hurt
prune 70% of tokens with V2V-CCMV2V-LLM · V2V-LLM efficiencycommunication bandwidth (MB): 0.4068 -> 0.1251Improved
prune 70% of tokens with V2V-CCMV2V-LLM · H200 GPUTTFT (ms): 45.5 -> 22.0Improved

Earlier methods this paper compares with

Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.

Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.

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