# Countering Multi-modal Representation Collapse through Rank-targeted Fusion

arXiv: https://arxiv.org/abs/2511.06450 · HTML: https://mlsanity.com/paper/2511.06450/

## Key facts

- Cited by 1 papers in the map (data through 14 Oct 2026).
- Builds on 53 of its 79 references that are in the map.
- First posted on arXiv 9 Nov 2025.

## Builds on (most cited first, 30 of 53)

- [Learning Transferable Visual Models From Natural Language Supervision](https://mlsanity.com/paper/2103.00020/) (2021)
- [SGDR: Stochastic Gradient Descent with Warm Restarts](https://mlsanity.com/paper/1608.03983/) (2016)
- [InternVL: Scaling up Vision Foundation Models and Aligning for Generic Visual-Linguistic Tasks](https://mlsanity.com/paper/2312.14238/) (2023)
- [Ego4D: Around the World in 3,000 Hours of Egocentric Video](https://mlsanity.com/paper/2110.07058/) (2021)
- [ViLT: Vision-and-Language Transformer Without Convolution or Region Supervision](https://mlsanity.com/paper/2102.03334/) (2021)
- [ImageBind: One Embedding Space To Bind Them All](https://mlsanity.com/paper/2305.05665/) (2023)
- [VICReg: Variance-Invariance-Covariance Regularization for Self-Supervised Learning](https://mlsanity.com/paper/2105.04906/) (2021)
- [Multimodal Machine Learning: A Survey and Taxonomy](https://mlsanity.com/paper/1705.09406/) (2017)
- [NTU RGB+D: A Large Scale Dataset for 3D Human Activity Analysis](https://mlsanity.com/paper/1604.02808/) (2016)
- [Rescaling Egocentric Vision](https://mlsanity.com/paper/2006.13256/) (2020)
- [Temporal Convolutional Networks for Action Segmentation and Detection](https://mlsanity.com/paper/1611.05267/) (2016)
- [Understanding Dimensional Collapse in Contrastive Self-supervised Learning](https://mlsanity.com/paper/2110.09348/) (2021)
- [MS-TCN: Multi-Stage Temporal Convolutional Network for Action Segmentation](https://mlsanity.com/paper/1903.01945/) (2019)
- [Deep Modular Co-Attention Networks for Visual Question Answering](https://mlsanity.com/paper/1906.10770/) (2019)
- [Pixel-BERT: Aligning Image Pixels with Text by Deep Multi-Modal Transformers](https://mlsanity.com/paper/2004.00849/) (2020)
- [Multimodal Deep Learning](https://mlsanity.com/paper/2301.04856/) (2023)
- [Revisiting Skeleton-based Action Recognition](https://mlsanity.com/paper/2104.13586/) (2021)
- [UniT: Multimodal Multitask Learning with a Unified Transformer](https://mlsanity.com/paper/2102.10772/) (2021)
- [ASFormer: Transformer for Action Segmentation](https://mlsanity.com/paper/2110.08568/) (2021)
- [RankMe: Assessing the downstream performance of pretrained self-supervised representations by their rank](https://mlsanity.com/paper/2210.02885/) (2022)
- [When will you do what? - Anticipating Temporal Occurrences of Activities](https://mlsanity.com/paper/1804.00892/) (2018)
- [What Would You Expect? Anticipating Egocentric Actions with Rolling-Unrolling LSTMs and Modality Attention](https://mlsanity.com/paper/1905.09035/) (2019)
- [Alleviating Over-segmentation Errors by Detecting Action Boundaries](https://mlsanity.com/paper/2007.06866/) (2020)
- [Diffusion Action Segmentation](https://mlsanity.com/paper/2303.17959/) (2023)
- [Self-Supervised Model Adaptation for Multimodal Semantic Segmentation](https://mlsanity.com/paper/1808.03833/) (2018)
- [Future Transformer for Long-term Action Anticipation](https://mlsanity.com/paper/2205.14022/) (2022)
- [Unified Fully and Timestamp Supervised Temporal Action Segmentation via Sequence to Sequence Translation](https://mlsanity.com/paper/2209.00638/) (2022)
- [Anticipative Feature Fusion Transformer for Multi-Modal Action Anticipation](https://mlsanity.com/paper/2210.12649/) (2022)
- [Long-Term Anticipation of Activities with Cycle Consistency](https://mlsanity.com/paper/2009.01142/) (2020)
- [Anticipating human actions by correlating past with the future with Jaccard similarity measures](https://mlsanity.com/paper/2105.12414/) (2021)

## Built on by (most cited first, 1 of 1)

- [Information Router for Mitigating Modality Dominance in Vision-Language Models](https://mlsanity.com/paper/2604.16264/) (2026)

## Abstract (arXiv)

Multi-modal fusion methods often suffer from two types of representation collapse: feature collapse where individual dimensions lose their discriminative power (as measured by eigenspectra), and modality collapse where one dominant modality overwhelms the other. Applications like human action anticipation that require fusing multifarious sensor data are hindered by both feature and modality collapse. However, existing methods attempt to counter feature collapse and modality collapse separately. This is because there is no unifying framework that efficiently addresses feature and modality collapse in conjunction. In this paper, we posit the utility of effective rank as an informative measure that can be utilized to quantify and counter both the representation collapses. We propose textitRank-enhancing Token Fuser, a theoretically grounded fusion framework that selectively blends less informative features from one modality with complementary features from another modality. We show that our method increases the effective rank of the fused representation. To address modality collapse, we evaluate modality combinations that mutually increase each others' effective rank. We show that depth maintains representational balance when fused with RGB, avoiding modality collapse. We validate our method on action anticipation, where we present textttR3D, a depth-informed fusion framework. Extensive experiments on NTURGBD, UTKinect, and DARai demonstrate that our approach significantly outperforms prior state-of-the-art methods by up to 3.74%. Our code is available at: hrefhttps://github.com/olivesgatech/R3Dhttps://github.com/olivesgatech/R3D.