Multi-View Information Bottleneck Without Variational Approximation

作者：Qi Zhang, Shujian Yu, Jingmin Xin, Badong Chen · 发表于：ICASSP 2022 - 2022 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP) · 年份：2022 · DOI：10.1109/icassp43922.2022.9747614 · 被引用次数：10 · 研究领域：Stochastic Gradient Optimization Techniques、Neural Networks and Applications、Machine Learning and Algorithms

By "intelligently" fuse the complementary information across different views, multi-view learning is able to improve the performance of classification task. In this work, we extend the information bottleneck principle to supervised multi-view learning scenario and use the recently proposed matrix-based Rényi’s α-order entropy functional to optimize the resulting objective directly, without the necessity of variational approximation or adversarial training. Empirical results in both synthetic and real-world datasets suggest that our method enjoys improved robustness to noise and redundant information in each view, especially given limited training samples. Code is available at https://github.com/archy666/MEIB.