arXiv 2602.161442026-02cs.CLcs.LG

缺失-by-设计：可撤销多模态情感分析的可验证模态删除

Missing-by-Design: Certifiable Modality Deletion for Revocable Multimodal Sentiment Analysis

University of Macau（澳门大学）；Zhejiang University（浙江大学）；Fudan University（复旦大学）；Shanghai AI Laboratory（上海人工智能实验室）；Juntendo University（立命馆大学）；Tsinghua University（清华大学）；University of Chinese Academy of Sciences（中国科学院大学）

Rong Fu, Ziming Wang, Chunlei Meng, Jiekai Wu, Kangan Qian, Hao Zhang, Simon Fong

备注：: 21 pages, 6 figures. In the previous version, Juntendo University was erroneously listed as the affiliation; we must clarify that this paper has absolutely no relation to Juntendo University. Therefore, we have replaced this affiliation in the new version

阅读 PDF

arXiv 页面 ↗HTML 原文 ↗

AI总结：

本文提出MBD框架，通过结构化表示学习和可验证参数修改流程，实现可撤销多模态情感分析中的模态删除，实验表明其在不完整输入下具有强预测性能，并实现隐私与效用的平衡。

摘要：

随着多模态系统越来越多地处理敏感个人数据，能够选择性撤销特定数据模态的能力已成为隐私合规和用户自主性的重要要求。我们提出了缺失-by-设计（MBD），一种统一的可撤销多模态情感分析框架，结合了结构化表示学习和可验证的参数修改流程。可撤销性对于隐私敏感应用至关重要，其中用户或监管机构可能要求删除特定模态的信息。MBD学习属性感知的嵌入，并利用生成器进行重建以恢复缺失通道，同时保持任务相关信号。对于删除请求，框架采用显著性驱动的候选选择和校准的高斯更新，生成可验证的模态删除证书。在基准数据集上的实验表明，MBD在不完整输入下实现了强预测性能，并提供了实用的隐私-效用权衡，使精确反向学习成为全重训练的高效替代方案。

英文摘要：

As multimodal systems increasingly process sensitive personal data, the ability to selectively revoke specific data modalities has become a critical requirement for privacy compliance and user autonomy. We present Missing-by-Design (MBD), a unified framework for revocable multimodal sentiment analysis that combines structured representation learning with a certifiable parameter-modification pipeline. Revocability is critical in privacy-sensitive applications where users or regulators may request removal of modality-specific information. MBD learns property-aware embeddings and employs generator-based reconstruction to recover missing channels while preserving task-relevant signals. For deletion requests, the framework applies saliency-driven candidate selection and a calibrated Gaussian update to produce a machine-verifiable Modality Deletion Certificate. Experiments on benchmark datasets show that MBD achieves strong predictive performance under incomplete inputs and delivers a practical privacy-utility trade-off, positioning surgical unlearning as an efficient alternative to full retraining.

机构、总结、中文摘要均由AI分析而来，可能存在错误，请以论文正文为准。