Research map / Vision-language and multimodal / Multimodal LLMs
VLDrive: Vision-Augmented Lightweight MLLMs for Efficient Language-grounded Autonomous Driving
Ruifei Zhang, Wei Zhang, Xiao Tan, Sibei Yang, Xiang Wan, Xiaonan Luo, Guanbin Li
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 changed | Where | Measured | Effect |
|---|---|---|---|
| add CCDP (ablation) | VLDrive (LLaMA*) · LangAuto (25% data) | DS: 35.5 -> 39.8 | Improved |
| add MEFA (ablation) | VLDrive (LLaMA*) · LangAuto (25% data) | DS: 35.9 -> 39.8 | Improved |
| add DDIA (ablation) | VLDrive (LLaMA*) · LangAuto (25% data) | DS: 36.3 -> 39.8 | Improved |
| add all three modules to the baseline (ablation) | VLDrive (LLaMA*) · LangAuto (25% data) | DS: 30.0 -> 39.8 | Improved |
| use CCDP instead of structural pooling (ablation) | VLDrive · LangAuto (25% data) | DS: 33.6 -> 39.8 | Improved |
| add token reconstruction to dynamic pruning (ablation) | VLDrive · LangAuto (25% data) | DS: 36.3 -> 39.8 | Improved |
| raise token retention ratio from 5% to 30% (ablation) | VLDrive · LangAuto (25% data) | DS: 33.4 -> 39.8 | Improved |
| raise token retention ratio from 30% to 50% (ablation) | VLDrive · LangAuto (25% data) | DS: 39.8 -> 39.8 | No clear effect |
| memory bank of 10 frames instead of 5 (ablation) | VLDrive · LangAuto (25% data) | DS: 37.1 -> 39.8 | Improved |
| increase memory bank from 10 to 20 frames (ablation) | VLDrive · LangAuto (25% data) | DS: 39.8 -> 38.3 | Hurt |
| replace 7B LLaVA-v1.5 LLM with lightweight 1.3B TinyLLaMA plus vision enhancements | VLDrive · LangAuto | DS: 36.2 -> 43.8 | Improved |
| vanilla lightweight TinyLLaMA in LMDrive versus 7B LLaVA-v1.5 | LMDrive · LangAuto | DS: 36.2 -> 25.2 | Hurt |
Closest papers by meaning
- Prune2Drive: A Plug-and-Play Framework for Accelerating Vision-Language Models in Autonomous Driving (CVPR 2026 · similarity 0.90)
- MAPLM: A Real-World Large-Scale Vision-Language Benchmark for Map and Traffic Scene Understanding (CVPR 2024 · similarity 0.90)
- DriveGPT4-V2: Harnessing Large Language Model Capabilities for Enhanced Closed-Loop Autonomous Driving (CVPR 2025 · similarity 0.90)
- CoC-VLA: Delving into Adversarial Domain Transfer for Explainable Autonomous Driving via Chain-of-Causality Visual-Language-Action Model (NeurIPS 2025 · similarity 0.90)
- Less Language, More Latents: Annotation-Efficient VLAs for Driving (NeurIPS 2026 · similarity 0.90)
- VGGDrive: Empowering Vision-Language Models with Cross-View Geometric Grounding for Autonomous Driving (CVPR 2026 · similarity 0.90)
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.
