Identifiable Shared Component Analysis of Unpaired Multimodal Mixtures
Not on the map. Why?
Abstract (arXiv)
A core task in multi-modal learning is to integrate information from multiple feature spaces (e.g., text and audio), offering modality-invariant essential representations of data. Recent research showed that, classical tools such as it canonical correlation analysis (CCA) provably identify the shared components up to minor ambiguities, when samples in each modality are generated from a linear mixture of shared and private components. Such identifiability results were obtained under the condition that the cross-modality samples are aligned/paired according to their shared information. This work takes a step further, investigating shared component identifiability from multi-modal linear mixtures where cross-modality samples are unaligned. A distribution divergence minimization-based loss is proposed, under which a suite of sufficient conditions ensuring identifiability of the shared components are derived. Our conditions are based on cross-modality distribution discrepancy characterization and density-preserving transform removal, which are much milder than existing studies relying on independent component analysis. More relaxed conditions are also provided via adding reasonable structural constraints, motivated by available side information in various applications. The identifiability claims are thoroughly validated using synthetic and real-world data.
Key facts. Cited by 3 papers in the map (data through 7 Oct 2026). Builds on 47 of its 77 references that are in the map. First posted on arXiv 28 Sep 2024.
Builds on
47 of 77 references.
References · 47
- 2015Deep Residual Learning for Image Recognitioncited by 51,809
- 2014Adam: A Method for Stochastic Optimizationcited by 36,960
- 2021Learning Transferable Visual Models From Natural Language Supervisioncited by 33,456
- 2020A Simple Framework for Contrastive Learning of Visual Representationscited by 11,979
- 2022Generative Adversarial Networkscited by 5,603
- 2016Matching Networks for One Shot Learningcited by 2,246
- 2014Unsupervised Domain Adaptation by Backpropagationcited by 1,906
- 2015Learning Transferable Features with Deep Adaptation Networkscited by 1,323
- 2017Adversarial Discriminative Domain Adaptationcited by 1,309
- 2017Deep Hashing Network for Unsupervised Domain Adaptationcited by 1,181
- 2019StarGAN v2: Diverse Image Synthesis for Multiple Domainscited by 1,095
- 2017StarGAN: Unified Generative Adversarial Networks for Multi-Domain Image-to-Image Translationcited by 1,036
35 more
- 2018Multimodal Unsupervised Image-to-Image Translationcited by 795
- 2015Unsupervised Deep Embedding for Clustering Analysiscited by 763
- 2016Deep Transfer Learning with Joint Adaptation Networkscited by 609
- 2017Word Translation Without Parallel Datacited by 480
- 2019Bridging Theory and Algorithm for Domain Adaptationcited by 332
- 2016FastText.zip: Compressing text classification modelscited by 331
- 2019Meta-Dataset: A Dataset of Datasets for Learning to Learn from Few Examplescited by 312
- 2021Self-Supervised Learning with Data Augmentations Provably Isolates Content from Stylecited by 308
- 2017Unsupervised Machine Translation Using Monolingual Corpora Onlycited by 272
- 2017Deep Clustering via Joint Convolutional Autoencoder Embedding and Relative Entropy Minimizationcited by 113
- 2018Unsupervised Alignment of Embeddings with Wasserstein Procrustescited by 96
- 2019The Incomplete Rosetta Stone Problem: Identifiability Results for Multi-View Nonlinear ICAcited by 86
- 2022A Closer Look at Smoothness in Domain Adversarial Trainingcited by 85
- 2018Nonnegative Matrix Factorization for Signal and Data Analytics: Identifiability, Algorithms, and Applicationscited by 55
- 2015Complete Dictionary Recovery over the Sphere I: Overview and the Geometric Picturecited by 49
- 2016Deep Variational Canonical Correlation Analysiscited by 44
- 2021Understanding Latent Correlation-Based Multiview Learning and Self-Supervision: An Identifiability Perspectivecited by 37
- 2014The Why and How of Nonnegative Matrix Factorizationcited by 34
- 2023Free Lunch for Domain Adversarial Training: Environment Label Smoothingcited by 33
- 2023Unpaired Multi-Domain Causal Representation Learningcited by 32
- 2023Unbalancedness in Neural Monge Maps Improves Unpaired Domain Translationcited by 32
- 2016Robust Volume Minimization-Based Matrix Factorization for Remote Sensing and Document Clusteringcited by 28
- 2017Estimating Mixed Memberships with Sharp Eigenvector Deviationscited by 22
- 2023Self-Supervised Disentanglement by Leveraging Structure in Data Augmentationscited by 16
- 2016Anchor-Free Correlated Topic Modeling: Identifiability and Algorithmcited by 15
- 2018MAGAN: Aligning Biological Manifoldscited by 15
- 2022Polytopic Matrix Factorization: Determinant Maximization Based Criterion and Identifiabilitycited by 14
- 2020Generalized Canonical Correlation Analysis: A Subspace Intersection Approachcited by 11
- 2013Independent Vector Analysis: Identification Conditions and Performance Boundscited by 7
- 2020Kernel of CycleGAN as a Principle homogeneous spacecited by 6
- 2021Shared Independent Component Analysis for Multi-Subject Neuroimagingcited by 6
- 2023Revisiting Deep Generalized Canonical Correlation Analysiscited by 6
- 2024Towards Identifiable Unsupervised Domain Translation: A Diversified Distribution Matching Approachcited by 5
- 2017The Role of Minimal Complexity Functions in Unsupervised Learning of Semantic Mappingscited by 4
- 2020Cell-Edge Detection via Selective Cooperation and Generalized Canonical Correlationcited by 3
Built on by
Citing papers per year: 2025: 1; 2026: 2.
3 papers.