自举多视图学习以应对测试时噪声对应
Bootstrapping Multi-view Learning for Test-time Noisy Correspondence
针对推理阶段跨视图异步采样导致的错配，BML 用可控噪声增强构造训练集并训练轻量估计器输出可靠性权重来调制融合，可直接迁移到多传感器/多模态融合（如 RGB-深度、多相机 BEV）在部署时出现错配的鲁棒融合环节。
原文依据 BML performs in-place TNC bootstrapping to construct a controllable noise-augmented training set that simulates realistic correspondence distortion
摘要译文
由模型翻译 · 术语统一 · 请对照原文核对关键结论多视图学习通过融合互补视图来提升感知能力，但真实部署中常常遭遇测试时噪声对应（TNC）问题——由异步采样、瞬时网络拥塞或其他扰动引起的跨视图错配。此类错配会引入语义不一致，并显著降低性能。现有补救方法通常从干净且对齐良好的训练数据中估计各视图的可靠性，再外推到推理阶段的噪声融合，从而造成训练-测试任务鸿沟，并降低对 TNC 的鲁棒性。为弥合这一鸿沟，我们提出自举多视图学习（BML）——一个即插即用的框架，显式地学习在 TNC 下进行融合。具体而言，BML 执行原位 TNC 自举，构建可控的噪声增强训练集以模拟真实的对应失真，从而无需外部数据即可消除任务鸿沟。与以往以无监督方式建模可靠性的不确定性方法不同，BML 提出一种揭示监督范式，其中轻量估计器联合建模视图内预测不确定性（视图质量）与视图间预测差异（对应一致性），在任务目标与自举监督的共同引导下产生校准的可靠性权重。一旦部署，这些可靠性权重直接调制融合，抑制受损视图同时保留有信息量的视图。在涵盖多种噪声比例的 11 个基准上，BML 一致优于最优基线，并保持对 TNC 的鲁棒性。
英文原文摘要 1698 字符
Multi-view learning fuses complementary views to improve perception, but real-world deployments often suffer from Test-time Noisy Correspondence (TNC) -- cross-view misalignment caused by asynchronous sampling, transient network congestion, or other disturbances. Such misalignment introduces semantic inconsistency and significantly degrades performance. Existing remedies typically estimate view-specific reliability from clean, well-aligned training data and then extrapolate to noisy fusion at inference, resulting in a train-test task gap and reduced robustness against TNC. To bridge this gap, we propose \underline \textcolor red B ootstrapping \underline \textcolor red M ulti-view \underline \textcolor red L earning (BML) -- a plug-and-play framework that explicitly learns to fuse under TNC. Specifically, BML performs in-place TNC bootstrapping to construct a controllable noise-augmented training set that simulates realistic correspondence distortion, thereby eliminating the task gap without external data. Unlike prior uncertainty-based approaches that model reliability in an unsupervised manner, BML presents a reveal-supervised paradigm, wherein a lightweight estimator jointly models intra-view predictive uncertainty (view quality) and inter-view prediction discrepancy (correspondence consistency) to produce calibrated reliability weights guided by both task objectives and bootstrapped supervision. Once deployed, these reliability weights directly modulate fusion, suppressing corrupted views while preserving informative ones. Across 11 benchmarks spanning diverse noise ratios, BML consistently outperforms state-of-the-art baselines and maintains robustness against TNC.