Research map / Reinforcement learning / Visual and representation RL
VILTA: A VLM-in-the-Loop Adversary for Enhancing Driving Policy Robustness
Qimao Chen, Fang Li, Shaoqing Xu, Zhiyi Lai, Zixun Xie, Yuechen Luo, Shengyin Jiang, Hanbing Li, Long Chen, Bing Wang, Yi Zhang, Zhi-Xin Yang
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 |
|---|---|---|---|
| train with VLM-in-the-loop trajectory adversary (VILTA) instead of VLM-RL | RL driving policy · CARLA Challenging (3 towns total) | route completion: 2.10 -> 2.38 | Improved |
| train with VILTA instead of CAT | RL driving policy · CARLA Challenging (3 towns total) | crash rate: 1.70 -> 1.03 | Improved |
| alternate challenging and normal scenarios (VILTA) | RL driving policy · CARLA Normal (3 towns total) | route completion: 2.54 -> 2.70 | Improved |
| use VILTA instead of VLM-RL on normal scenario speed | RL driving policy · CARLA Normal | average speed: 68.65 -> 65.13 | Hurt |
| add trajectory post-processing (B-spline, sigmoid fusion, LQR) (ablation) | VILTA · CARLA Challenging | route completion: 0.73 -> 0.77 | Improved |
| add following reward (ablation) | VILTA · CARLA Challenging | route completion: 0.74 -> 0.77 | Improved |
| use Vision-Language-Editing instead of direct VLM generation (ablation) | VILTA · CARLA Challenging | crash rate: 0.53 -> 0.50 | Improved |
| raise challenging scenario frequency to 1 per 8 normal (from 1 per 2) (ablation) | VILTA · CARLA Challenging | route completion: 0.77 -> 0.87 | Improved |
| reduce challenging scenarios to 1 per 16 normal (ablation) | VILTA · CARLA Challenging | route completion: 0.87 -> 0.85 | Hurt |
Closest papers by meaning
- Drive-R1: Bridging Reasoning and Planning in VLMs for Autonomous Driving with Reinforcement Learning (AAAI 2026 · similarity 0.89)
- CoC-VLA: Delving into Adversarial Domain Transfer for Explainable Autonomous Driving via Chain-of-Causality Visual-Language-Action Model (NeurIPS 2025 · similarity 0.89)
- From Representational Complementarity to Dual Systems: Synergizing VLM and Vision-Only Backbones for End-to-End Driving (NeurIPS 2026 · similarity 0.88)
- Impromptu VLA: Open Weights and Open Data for Driving Vision-Language-Action Models (NeurIPS 2025 · similarity 0.88)
- DriveHierarchy: A Benchmark for Diagnosing VLM Driving Capabilities from Open-Loop Understanding to Closed-Loop Execution (NeurIPS 2026 · similarity 0.88)
- DriveAgent-R1: Advancing VLM-based Autonomous Driving with Active Perception and Hybrid Thinking (ICLR 2026 · similarity 0.88)
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.
