AI Paper Atlas

Research map / Vision-language and multimodal / Multimodal LLMs

Reason2Drive: Towards Interpretable and Chain-based Reasoning for Autonomous Driving

Ming Nie, Renyuan Peng, Chunwei Wang, Xinyue Cai, Jianhua Han, Hang Xu, Li Zhang

ECCV 2024 · Vision-language and multimodal › Multimodal LLMs · approach: Chain-of-thought multimodal reasoning · 97 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 reasoning-chain tasks to instruction tuning data (ablation)Reason2Drive model · Reason2Drive valADRScore: 0.282 -> 0.351Improved
add perception task data on top of reasoning (ablation)Reason2Drive model · Reason2Drive valADRScore: 0.351 -> 0.407Improved
add prediction task data on top of perception and reasoning (ablation)Reason2Drive model · Reason2Drive valADRScore: 0.407 -> 0.463Improved
use multi-frame video features instead of single-frame image features (ablation)Reason2Drive model · Reason2Drive valADRScore-S: 0.379 -> 0.394Improved
add region-level features from the prior tokenizer (ablation)Reason2Drive model · Reason2Drive valADRScore-S: 0.394 -> 0.418Improved
add positional embedding to perceptual priors (ablation)Reason2Drive model · Reason2Drive valADRScore-S: 0.418 -> 0.432Improved
pre-train the instructed vision decoder (ablation)Reason2Drive model · Reason2Drive valADRScore-S: 0.361 -> 0.396Improved
add text embedding input to the vision decoder (ablation)Reason2Drive model · Reason2Drive valADRScore-S: 0.396 -> 0.425Improved
use ViT-G/14 EVA-CLIP encoder instead of ViT-L/14 (ablation)Reason2Drive model · Reason2Drive valADRScore-S: 0.397 -> 0.432Improved
pre-train on Reason2Drive before fine-tuning for control signalsInstructBLIP · nuScenes control signalsSpeed RMSE: 3.743 -> 3.086Improved
replace textual perception priors and LLM-only decoding with prior tokenizer and instructed vision decoderInstructBLIP Vicuna-7B · Reason2Drive valADRScore-S: 0.214 -> 0.432Improved
train on nuScenes only and test on unseen Waymo+ONCEOurs · Reason2Drive generalizationADRScore-S: 0.443 -> 0.385Mixed

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