Cross-Stage Coherence in Hierarchical Driving VQA: A Consistency Protocol and Parameter-Efficient Gated Context Routing

arXiv:2604.22560 · cs.CV, cs.AI · Submitted 2026-04-24 · Read on arXiv

cs.CV, cs.AI

Submitted: 2026-04-24

Updated: 2026-10-13

Terminology

• Consistency — 3× in this paper · explained in 3 episodes
• Adaptation — 1× in this paper · explained in 1 episode
• Domain Adaptation — 1× in this paper · explained in 5 episodes
• Embeddings — 1× in this paper · explained in 2 episodes

Sources

• Cosmos-Reason1: From Physical Common Sense To Embodied Reasoning
• QLoRA: Efficient Finetuning of Quantized LLMs
• VERDI: VLM-Embedded Reasoning for Autonomous Driving
• Mini-InternVL: A Flexible-Transfer Pocket Multimodal Model with 5% Parameters and 90% Performance
• VLAD: A VLM-Augmented Autonomous Driving Framework with Hierarchical Planning and Interpretable Decision Process
• Training Large Language Models to Reason in a Continuous Latent Space
• LoRA: Low-Rank Adaptation of Large Language Models
• Precise Drive with VLM: First Prize Solution for PRCV 2024 Drive LM challenge
• DriveTransformer: Unified Transformer for Scalable End-to-End Autonomous Driving
• AlphaDrive: Unleashing the Power of VLMs in Autonomous Driving via Reinforcement Learning and Reasoning
• Building Efficient Universal Classifiers with Natural Language Inference
• ALN-P3: Unified Language Alignment for Perception, Prediction, and Planning in Autonomous Driving
• Multi-Modal Adapter for Vision-Language Models
• CODI: Compressing Chain-of-Thought into Continuous Space via Self-Distillation
• HiP-AD: Hierarchical and Multi-Granularity Planning with Deformable Attention for Autonomous Driving in a Single Decoder
• DriveVLM: The Convergence of Autonomous Driving and Large Vision-Language Models
• DriveCoT: Integrating Chain-of-Thought Reasoning with End-to-End Driving
• Are VLMs Ready for Autonomous Driving? An Empirical Study from the Reliability, Data, and Metric Perspectives
• OpenDriveVLA: Towards End-to-end Autonomous Driving with Large Vision Language Action Model
• InternVL3: Exploring Advanced Training and Test-Time Recipes for Open-Source Multimodal Models

Related papers

• Loss Knows Best: Detecting Annotation Errors in Videos via Loss Trajectories
• AnchorWeave: World-Consistent Video Generation with Retrieved Local Spatial Memories
• Benchmarking the Robustness of Foundation Models for Mammography under Domain Shift
• MambaX-Net: Dual-Input Mamba-Enhanced Cross-Attention Network for Longitudinal MRI Segmentation
• TeleOCR: Navigating Document Parsing Across Digital and Camera-Captured Documents
• A Survey on Efficient Vision-Language-Action Models