提出两种跨阶段一致性机制，提升自动驾驶视觉问答中规划与感知的逻辑连贯性。
Cross-Stage Coherence in Hierarchical Driving VQA: Explicit Baselines and Learned Gated Context Projectors
- 设计显式提示传递和隐式门控投影两种方法，分别实现零训练基线与高效参数微调。
- 隐式方法使规划阶段语义矛盾降低34%，跨阶段蕴含度提升50%，同时提高语言质量。
- 适合关注自动驾驶多阶段推理一致性的研究者，尤其关注模型可解释性与领域适配。
自动驾驶场景下的图结构视觉问答（GVQA）将推理划分为感知、预测和规划三个有序阶段，要求规划决策与模型自身感知保持一致。本文在DriveLM-nuScenes数据集上对比两种互补的跨阶段上下文传递机制。显式方法在领域自适应的40亿参数视觉语言模型（Mini-InternVL2-4B-DA-DriveLM）上评估三种基于提示的条件策略，无需额外训练，使NLI矛盾降低42.6%，建立强零训练基线。隐式方法引入门控上下文投影器，从一阶段提取隐藏状态向量，经归一化与门控后注入下一阶段输入嵌入；该投影器与阶段专用QLoRA适配器联合训练于通用80亿参数模型（InternVL3-8B-Instruct），仅更新约0.5%参数。隐式方法实现规划阶段NLI矛盾显著降低34%（置信区间95%，p<0.05），跨阶段蕴含度提升50%，采用多语言NLI分类器评估混合语言输出。规划语言质量提升（CIDEr +30.3%），但因缺乏驾驶领域预训练，词汇重叠与结构一致性下降。由于两方法使用不同基础模型，视为互补案例：显式方法提供表面一致性零训练基线，隐式门控投影则带来显著规划阶段语义增益，提示领域适配是实现全谱提升的关键下一步。
原文摘要 · Abstract (English)
Graph Visual Question Answering (GVQA) for autonomous driving organizes reasoning into ordered stages, namely Perception, Prediction, and Planning, where planning decisions should remain consistent with the model's own perception. We present a comparative study of cross-stage context passing on DriveLM-nuScenes using two complementary mechanisms. The explicit variant evaluates three prompt-based conditioning strategies on a domain-adapted 4B VLM (Mini-InternVL2-4B-DA-DriveLM) without additional training, reducing NLI contradiction by up to 42.6% and establishing a strong zero-training baseline. The implicit variant introduces gated context projectors, which extract a hidden-state vector from one stage and inject a normalized, gated projection into the next stage's input embeddings. These projectors are jointly trained with stage-specific QLoRA adapters on a general-purpose 8B VLM (InternVL3-8B-Instruct) while updating only approximately 0.5% of parameters. The implicit variant achieves a statistically significant 34% reduction in planning-stage NLI contradiction (bootstrap 95% CIs, p < 0.05) and increases cross-stage entailment by 50%, evaluated with a multilingual NLI classifier to account for mixed-language outputs. Planning language quality also improves (CIDEr +30.3%), but lexical overlap and structural consistency degrade due to the absence of driving-domain pretraining. Since the two variants use different base models, we present them as complementary case studies: explicit context passing provides a strong training-free baseline for surface consistency, while implicit gated projection delivers significant planning-stage semantic gains, suggesting domain adaptation as a plausible next ingredient for full-spectrum improvement.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。