Research map / Multi-agent systems and games / LLM agents
ClarifyVC: Clarifying Ambiguous Commands in Vehicle Control with a Hybrid Data Augmentation Pipeline
Hange Zhou, Zhonglin Jiang, yingjie cui, Mingzhe Zhang, Xiaotang Wang, Hengwei Dai, Qiyao Yu, Yong Chen, Yongqi Zhang
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.
Introduces ClarifyVC, a framework with a hybrid data-augmentation pipeline, reference models, and an evaluation protocol for resolving ambiguous natural-language vehicle-control commands.
Problem
Natural language interfaces for vehicle control must handle vague commands, evolving dialogue context, and strict protocol constraints.
Approach
A unified framework pairing an agent-orchestrated hybrid data-augmentation pipeline that generates diverse, ambiguity-rich dialogues with reference models trained on that data and a dedicated evaluation protocol.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| supervised fine-tune on ClarifyVC-Data vs zero-shot | Qwen2.5-7B · ClarifyVC-Data single-round | accuracy: 74.3 -> 89.0 | Improved |
| supervised fine-tune on ClarifyVC-Data vs zero-shot | LLaMA3-70B · ClarifyVC-Data single-round | accuracy: 81.2 -> 94.1 | Improved |
| add 4 in-context examples | small models (Qwen2.5-0.5B) · ClarifyVC-Data single-round | accuracy: 59.2 -> 64.5 | Improved |
| add 4 in-context examples | LLaMA3-70B · ClarifyVC-Data single-round | accuracy: 81.2 -> 80.1 | Hurt |
| use 7B SFT model instead of larger backbones | Qwen2.5 · ClarifyVC-Data | accuracy: 89.0 vs 95.8 (72B) | No clear effect |
| scale model size under SFT | Qwen2.5 0.5B to 72B · ClarifyVC-Data single-round | accuracy: 75.1 -> 95.8 | Improved |
| add fuzz injection (FIA) stage (ablation) | data pipeline · ambiguity diversity | ambiguity diversity: 0.68 -> 0.85 | Improved |
| add adversarial generation (AGA) stage (ablation) | data pipeline · ambiguity diversity | ambiguity diversity: 0.71 -> 0.85 | Improved |
| run AGA before FIA rather than reversed (ablation) | data pipeline · ambiguity diversity | ambiguity diversity: 0.79 -> 0.85 | Improved |
| add multi-turn evolution (MEA) stage (ablation) | data pipeline · avg score | avg score drop when removed: -4.6% | Improved |
| constrain FIA with drift control and expert-curated pairs (ablation) | ambiguity injection · intent preservation | intent preservation: 89-92% -> 98.9% | Improved |
| use ClarifyVC-Data in place of existing datasets as a benchmark | zero-shot LLMs · benchmark difficulty | intent accuracy: 92.5% -> 70.5% (GPT-4, APIGen vs ClarifyVC-Data) | Hurt |
Experimental setup
Benchmarks: ClarifyVC-Eval
Closest papers by meaning
- ChatScene: Knowledge-Enabled Safety-Critical Scenario Generation for Autonomous Vehicles (CVPR 2024 · similarity 0.85)
- Uncertainty-Aware Clarification in LLM Agents with Information Gain (ICML 2026 · similarity 0.84)
- Hybrid-Driving: An Autonomous Driving Decision Framework Integrating Large Language Models, Knowledge Graphs and Driving Rules (AAAI 2025 · similarity 0.83)
- LMDrive: Closed-Loop End-to-End Driving with Large Language Models (CVPR 2024 · similarity 0.83)
- AutoVLA: A Vision-Language-Action Model for End-to-End Autonomous Driving with Adaptive Reasoning and Reinforcement Fine-Tuning (NeurIPS 2025 · similarity 0.83)
- WOMD-Reasoning: A Large-Scale Dataset for Interaction Reasoning in Driving (ICML 2025 · similarity 0.83)
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.
