Learning Robust Representations via Multi-View Information Bottleneck

By Marco Federici, Anjan Dutta, Patrick Forr'e, Nate Kushman, Zeynep Akata. International Conference on Learning Representations, 2020

345 citations recorded by the source corpus.

The information bottleneck method provides an information-theoretic view of representation learning. The original formulation, however, can only be applied in the supervised setting where task-specific labels are available at learning time. We extend this method to the unsupervised setting, by taking advantage of multi-view data, which provides two views of the same underlying entity. A theoretical analysis leads to the definition of a new multi-view model which produces state-of-the-art results on two standard multi-view datasets, Sketchy and MIR-Flickr. We also extend our theory to the single-view setting by taking advantage of standard data augmentation techniques, empirically showing better generalization capabilities when compared to traditional unsupervised approaches.

Subfields: Artificial Intelligence, Computer Vision and Pattern Recognition

View Paper: https://www.semanticscholar.org/paper/4bf832104a47c380eb4413b20a9d5bf06649684f

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Artificial Intelligence

In Computer Vision and Pattern Recognition

Citations in the corpus

Deep multi-view learning methods: A review (2021, methodology): “For the deep multi-view object recognition methods, we compare the classification accuracy (Accuracy) [49] of the following 660 approaches: multi-view CNN (MVCNN) [49], Kd-tree based network (KD-Net) [51], multi-view harmonized bilinear network (MHBN) [52], group-view CNN (GVCNN) [50] and dynamic ro”
What Makes Multimodal Learning Better than Single (Provably) (2021, background): “Besides, previous multi-view analysis [37, 1, 45, 13] typically assumes that each view alone is sufficient to predict the target accurately, which may not hold in our multi-modal setting.”
Dual Contrastive Prediction for Incomplete Multi-View Representation Learning (2022, methodology): “used assumption of multi-view data [42] as below:”
Semi-Supervised and Unsupervised Deep Visual Learning: A Survey (2022, background): “The multi-view information bottleneck model [140] extends the original information bottleneck principle to unsupervised learning and trains an encoder to retain all the relevant information for predicting the label while minimizing the excess information in the representation.”
Dual Temperature Helps Contrastive Learning Without Many Negative Samples: Towards Understanding and Simplifying MoCo (2022, background): “The core of unsupervised learning is to learn augmentation-invariant representation, for which CL is at the core of its development [17,34,41,42,50].”
From Canonical Correlation Analysis to Self-supervised Graph Neural Networks (2021, methodology): “The most relevant one names Self-supervised Information Bottleneck: Definition 2.”
Self-supervised Learning from a Multi-view Perspective (2020, methodology): “, 2018) and the Multiview Information Bottleneck method (Federici et al., 2020).”
SimCVD: Simple Contrastive Voxel-Wise Representation Distillation for Semi-Supervised Medical Image Segmentation (2021, background): “Recently, contrastive learning, a type of self-supervised learning, has received a lot of interests [46], [49], [50], [51], [26], [52], [24].”