# Aligning Multimodal Representations through an Information Bottleneck

arXiv: https://arxiv.org/abs/2506.04870 · HTML: https://mlsanity.com/paper/2506.04870/

## Key facts

- Cited by 13 papers in the map (data through 14 Oct 2026).
- Builds on 50 of its 75 references that are in the map.
- First posted on arXiv 5 Jun 2025.

## Builds on (most cited first, 30 of 50)

- [Deep Residual Learning for Image Recognition](https://mlsanity.com/paper/1512.03385/) (2015)
- [Learning Transferable Visual Models From Natural Language Supervision](https://mlsanity.com/paper/2103.00020/) (2021)
- [An Image is Worth 16x16 Words: Transformers for Image Recognition at Scale](https://mlsanity.com/paper/2010.11929/) (2020)
- [Microsoft COCO: Common Objects in Context](https://mlsanity.com/paper/1405.0312/) (2014)
- [A Simple Framework for Contrastive Learning of Visual Representations](https://mlsanity.com/paper/2002.05709/) (2020)
- [Representation Learning with Contrastive Predictive Coding](https://mlsanity.com/paper/1807.03748/) (2018)
- [Scaling Laws for Neural Language Models](https://mlsanity.com/paper/2001.08361/) (2020)
- [Hierarchical Text-Conditional Image Generation with CLIP Latents](https://mlsanity.com/paper/2204.06125/) (2022)
- [BLIP-2: Bootstrapping Language-Image Pre-training with Frozen Image Encoders and Large Language Models](https://mlsanity.com/paper/2301.12597/) (2023)
- [BLIP: Bootstrapping Language-Image Pre-training for Unified Vision-Language Understanding and Generation](https://mlsanity.com/paper/2201.12086/) (2022)
- [Scaling Up Visual and Vision-Language Representation Learning With Noisy Text Supervision](https://mlsanity.com/paper/2102.05918/) (2021)
- [FaceNet: A Unified Embedding for Face Recognition and Clustering](https://mlsanity.com/paper/1503.03832/) (2015)
- [Unsupervised Feature Learning via Non-Parametric Instance-level Discrimination](https://mlsanity.com/paper/1805.01978/) (2018)
- [Similarity of Neural Network Representations Revisited](https://mlsanity.com/paper/1905.00414/) (2019)
- [Align before Fuse: Vision and Language Representation Learning with Momentum Distillation](https://mlsanity.com/paper/2107.07651/) (2021)
- [Understanding Contrastive Representation Learning through Alignment and Uniformity on the Hypersphere](https://mlsanity.com/paper/2005.10242/) (2020)
- [Contrastive Multiview Coding](https://mlsanity.com/paper/1906.05849/) (2019)
- [Learning deep representations by mutual information estimation and maximization](https://mlsanity.com/paper/1808.06670/) (2018)
- [ImageBind: One Embedding Space To Bind Them All](https://mlsanity.com/paper/2305.05665/) (2023)
- [Deep Visual-Semantic Alignments for Generating Image Descriptions](https://mlsanity.com/paper/1412.2306/) (2014)
- [Deep Variational Information Bottleneck](https://mlsanity.com/paper/1612.00410/) (2016)
- [What Makes for Good Views for Contrastive Learning?](https://mlsanity.com/paper/2005.10243/) (2020)
- [Deep Learning and the Information Bottleneck Principle](https://mlsanity.com/paper/1503.02406/) (2015)
- [Learning Representations by Maximizing Mutual Information Across Views](https://mlsanity.com/paper/1906.00910/) (2019)
- [EVA: Exploring the Limits of Masked Visual Representation Learning at Scale](https://mlsanity.com/paper/2211.07636/) (2022)
- [Do Vision Transformers See Like Convolutional Neural Networks?](https://mlsanity.com/paper/2108.08810/) (2021)
- [Data-Efficient Image Recognition with Contrastive Predictive Coding](https://mlsanity.com/paper/1905.09272/) (2019)
- [VideoCLIP: Contrastive Pre-training for Zero-shot Video-Text Understanding](https://mlsanity.com/paper/2109.14084/) (2021)
- [Mind the Gap: Understanding the Modality Gap in Multi-modal Contrastive Representation Learning](https://mlsanity.com/paper/2203.02053/) (2022)
- [Opening the Black Box of Deep Neural Networks via Information](https://mlsanity.com/paper/1703.00810/) (2017)

## Built on by (most cited first, 13 of 13)

- [ECHO: Toward Contextual Seq2Seq Paradigms in Large EEG Models](https://mlsanity.com/paper/2509.22556/) (2025)
- [LaME: Learning to Think in Latent Space for Multimodal Embedding via Information Bottleneck](https://mlsanity.com/paper/2606.13061/) (2026)
- [CrossJEPA: Cross-Modal Joint-Embedding Predictive Architecture for Efficient 3D Representation Learning from 2D Images](https://mlsanity.com/paper/2511.18424/) (2025)
- [A Comprehensive Information-Decomposition Analysis of Large Vision-Language Models](https://mlsanity.com/paper/2603.29676/) (2026)
- [Bridging Functional and Representational Similarity via Usable Information](https://mlsanity.com/paper/2601.21568/) (2026)
- [The Describe-Then-Generate Bottleneck: How VLM Descriptions Alter Image Generation Outcomes](https://mlsanity.com/paper/2509.18179/) (2025)
- [FDRMFL: Multimodal Federated Feature Extraction Model Based on Information Maximization and Contrastive Learning](https://mlsanity.com/paper/2512.02076/) (2025)
- [Toward Structural Multimodal Representations: Specialization, Selection, and Sparsification via Mixture-of-Experts](https://mlsanity.com/paper/2605.03348/) (2026)
- [Learning Emergent Modular Representations in Multi-modality Medical Vision Foundation Models](https://mlsanity.com/paper/2605.21861/) (2026)
- [Information-theoretic Multimodal Representation Learning for Electrocardiogram Signals](https://mlsanity.com/paper/2605.27583/) (2026)
- [OVA-IB: One vs All Information Bottleneck for Multi-Modal Alignment](https://mlsanity.com/paper/2605.29900/) (2026)
- [SGWIB:Sliced Gromov-Wasserstein Information Bottleneck for Video Highlight Detection](https://mlsanity.com/paper/2609.13966/) (2026)
- [Architecture-Dependent Fusion Pathways in MLLMs](https://mlsanity.com/paper/2610.03289/) (2026)

## Abstract (arXiv)

Contrastive losses have been extensively used as a tool for multimodal representation learning. However, it has been empirically observed that their use is not effective to learn an aligned representation space. In this paper, we argue that this phenomenon is caused by the presence of modality-specific information in the representation space. Although some of the most widely used contrastive losses maximize the mutual information between representations of both modalities, they are not designed to remove the modality-specific information. We give a theoretical description of this problem through the lens of the Information Bottleneck Principle. We also empirically analyze how different hyperparameters affect the emergence of this phenomenon in a controlled experimental setup. Finally, we propose a regularization term in the loss function that is derived by means of a variational approximation and aims to increase the representational alignment. We analyze in a set of controlled experiments and real-world applications the advantages of including this regularization term.