Representation Disentanglement

Overview

Representation disentanglement aims to make different properties—such as content and style, identity and expression, or task-relevant and domain-specific signals—independently accessible in a model’s features. Common approaches include variational objectives, orthogonality and independence constraints, adversarial or contrastive learning, sparse autoencoders, and architectures with separate feature streams or subspaces. The methods are used in settings such as controllable image and speech generation, domain generalization, privacy and fairness, and mechanistic interpretability, with evaluation through probes, leakage tests, reconstruction quality, and downstream performance. A central challenge is that factors may be correlated or not uniquely identifiable, while suppressing unwanted information can also reduce useful task information or reconstruction fidelity.

Latest papers

CardsList

104 papers6 papers in the last four weeks, down 25% on the four weeks before. 0.0% of all new papers.

  1. Location Matters: Layer-wise Prompt Conditioning for Voice--Style Disentanglement in Text-to-Speech

    Nov 18, 2026David Lindevelt, Suzan Verberne, Joost BroekensTTS SynthesisExpressive Speech Synthesis

  2. When Does Label-Free Deconfounding of Document Embeddings Work?

    Nov 14, 2026Anna Kiepura, Yingqiang Gao, Elliott Ash +1Representation DisentanglementCausal Inference with Unobserved Confounding

  3. WaveletLoRA: Frequency-Aware Content-Style Decomposition for Personalized Image Generation

    Nov 11, 2026Peiyao Wang, Jiahui Sun, Weining Wang +1Image GenerationSubject-Driven Image Generation

  4. Exploring Post-Hoc Disentanglement of Audio Foundation Models for Music Similarity

    Nov 10, 2026Lucas de Souza Brandão, Gabriel da Costa Merlin, Diego Furtado SilvaRepresentation DisentanglementDisentangled Representation Learning

  5. Disentangling Channel Semantics in Vision Transformers via Token Decorrelation and Composition-Aware Modulation

    Nov 7, 2026Daeun Kim, Hyesong Choi, Hyejin Park +1Vision TransformerVisual Representation Learning

  6. Compact Motion Embeddings for Feed-forward Unified Human Animation

    Nov 5, 2026Seonwook Park, Hyunsoo Cha, Amrita Mazumdar +5Human Avatar AnimationMotion Representation Learning

  7. Isolating a Language Model's Judgment with a Label-Free Subspace

    Oct 15, 2026Chaemin Jang, Donggyu Lee, Jeesoo Kim +2LLM InterpretabilityRepresentation Disentanglement

  8. DeTac: Dual-branch Decoupling of Texture and Force Signals in Vision-Based Tactile Sensors

    Oct 12, 2026Zhirui Zhang, Yunkai Tang, Jiahang Li +6Vision-Based Tactile SensingRepresentation Disentanglement

  9. DisFace3DNet: Explainable Facial Attractiveness Prediction via 3D Component Disentanglement

    Oct 8, 2026Fenggui Rao, Yan Luximon, Jie ZhangInterpretable MLRepresentation Disentanglement

  10. Representation Disentanglement for Fair Chest X-Ray Diagnosis

    Oct 5, 2026Yujie Sun, Ruizhe Li, Xiaowu SunChest X-Ray ClassificationRepresentation Disentanglement

  11. On Color Alignment in VAE Latent Spaces and Its Applications

    Oct 5, 2026Julian D. Santamaria, Kai Wang, Jesús Malo +2Variational AutoencodersImage Generation

  12. Revisiting Cross-Reconstruction for Generalizable Deepfake Detection

    Oct 1, 2026Bingjian Yang, Shilei Zhao, Zheng WangImage Forgery DetectionDigital Image Forensics

  13. Reason in Style: Discovering and Controlling Style in Language Models

    Sep 30, 2026Ioana Marinescu, Eric Karl Oermann, Kyunghyun ChoUnsupervised LearningLanguage Model Steering

  14. SafeStyle: Calibrated Style Residual Injection for Controllable Style-Leakage Trade-off in Diffusion Stylization

    Sep 18, 2026Zhangping Yang, Min Li, Song Yan +4Image Style TransferDiffusion-Based Image Editing

  15. Do speech foundation models really learn words?

    Sep 9, 2026Robin Huo, Ewan DunbarRepresentation LearningSpeech Foundation Models

  16. High-probability guarantees for linear accessibility in feature superposition

    Sep 9, 2026Enrico VompaFeature SuperpositionNeural Representation Geometry

  17. Reachability Is Not Generalization: Understanding Verb--Noun Decomposition in Assembly Action Recognition

    Sep 4, 2026Changyi Li, Yu XiaoCompositional GeneralizationHuman Activity Recognition

  18. Entangled Representations Amplify Collateral Damage in Unlearning

    Sep 2, 2026Evžen Wybitul, Tim G. J. Rudner, Christian Schroeder de WittMachine UnlearningRepresentation Disentanglement

  19. A Deeper Analysis of Block-Sparse Featurizers

    Aug 27, 2026Alexandru-Iulius Jerpelea, Amith AnanthramSparse AutoencodersRepresentation Disentanglement

  20. Defending Wearable VLMs Against Private Attribute Inference

    Aug 26, 2026Zhimin Li, Pan Wang, Jingxian Chen +4Privacy Protection in VLMsPrivacy-Utility Tradeoff

  21. LightAIR: Lightweight Action Inversion and Riemannian Rectification for Text-based Person Anomaly Search

    Aug 10, 2026Yulun Zhang, Zixu Li, Zhiwei Chen +6Text-Based Person SearchRepresentation Disentanglement

  22. Support Operation Factorization: Compositional Readout of Frozen Vision Encoders under Controlled Interventions

    Aug 6, 2026Zhongyao Wang, Wanli Ouyang, Taoyong Cui +1Visual Representation LearningRepresentation Disentanglement

  23. MIDAS: Mutual Information Disentanglement with Uncertainty-Aware Fusion for Incomplete Multimodal Sentiment Analysis

    Aug 6, 2026Yuhua Wen, Yingying Zhou, Qifei Li +4Missing-Modality LearningMultimodal Disentangled Representation Learning

  24. QuerySplat: Decoupling Geometry and Appearance Representations in 3DGS Prediction

    Aug 2, 2026Yinglong Li, Donghui Shen, Xiaoyu Zhang +53D Gaussian SplattingRepresentation Disentanglement

  25. What to Remove, What to Preserve: Dual-Ambiguity Rectification for All-in-One Image Restoration

    Jul 30, 2026Cencen Liu, Wen Yin, Dongyang Zhang +6Representation LearningRepresentation Disentanglement

  26. Controlling Embedding Spaces with Text-Conditioned Transformations

    Jul 24, 2026Joseph Fioresi, Fabian Caba Heilbron, Pankaj Nathani +2Representation DisentanglementMultimodal Embedding