# Federated learning for unpaired multimodal data through a homogeneous transformer model

arXiv: https://arxiv.org/abs/2601.17986 · HTML: https://mlsanity.com/paper/2601.17986/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Builds on 18 of its 30 references that are in the map.
- First posted on arXiv 25 Jan 2026.

## Builds on (most cited first, 18 of 18)

- [LoRA: Low-Rank Adaptation of Large Language Models](https://mlsanity.com/paper/2106.09685/) (2021)
- [LLaMA: Open and Efficient Foundation Language Models](https://mlsanity.com/paper/2302.13971/) (2023)
- [Communication-Efficient Learning of Deep Networks from Decentralized Data](https://mlsanity.com/paper/1602.05629/) (2016)
- [Similarity of Neural Network Representations Revisited](https://mlsanity.com/paper/1905.00414/) (2019)
- [DINOv3](https://mlsanity.com/paper/2508.10104/) (2025)
- [Vision-Language Models for Vision Tasks: A Survey](https://mlsanity.com/paper/2304.00685/) (2023)
- [DoRA: Weight-Decomposed Low-Rank Adaptation](https://mlsanity.com/paper/2402.09353/) (2024)
- [The Platonic Representation Hypothesis](https://mlsanity.com/paper/2405.07987/) (2024)
- [Towards Building the Federated GPT: Federated Instruction Tuning](https://mlsanity.com/paper/2305.05644/) (2023)
- [FLoRA: Federated Fine-Tuning Large Language Models with Heterogeneous Low-Rank Adaptations](https://mlsanity.com/paper/2409.05976/) (2024)
- [FedPara: Low-Rank Hadamard Product for Communication-Efficient Federated Learning](https://mlsanity.com/paper/2108.06098/) (2021)
- [Multimodal Federated Learning via Contrastive Representation Ensemble](https://mlsanity.com/paper/2302.08888/) (2023)
- [Harnessing the Universal Geometry of Embeddings](https://mlsanity.com/paper/2505.12540/) (2025)
- [FedMultimodal: A Benchmark For Multimodal Federated Learning](https://mlsanity.com/paper/2306.09486/) (2023)
- [Do Vision and Language Encoders Represent the World Similarly?](https://mlsanity.com/paper/2401.05224/) (2024)
- [Universally Converging Representations of Matter Across Scientific Foundation Models](https://mlsanity.com/paper/2512.03750/) (2025)
- [Multimodal Federated Learning in Healthcare: a Review](https://mlsanity.com/paper/2310.09650/) (2023)
- [Towards Multi-modal Transformers in Federated Learning](https://mlsanity.com/paper/2404.12467/) (2024)

## Abstract (arXiv)

Training of multimodal foundation models is currently restricted to centralized data centers containing massive, aligned datasets (e.g., image-text pairs). However, in realistic federated environments, data is often unpaired and fragmented across disjoint nodes; one node may hold sensor data, while another holds textual logs. These datasets are strictly private and share no common samples. Current federated learning (FL) methods fail in this regime, as they assume local clients possess aligned pairs or require sharing raw feature embeddings, which violates data sovereignty. We propose a novel framework to train a global multimodal transformer across decentralized nodes with disjoint modalities. We introduce a small public anchor set to align disjoint private manifolds. Using Gram matrices calculated from these public anchors, we enforce semantic alignment across modalities through centered kernel alignment without ever transmitting private samples, offering a mathematically superior privacy guarantee compared to prototype sharing. Further, we introduce a subspace-stabilized fine-tuning method to handle FL with huge transformer models. We strictly decouple domain-specific magnitude shifts from semantic direction, ensuring that nodes with varying sensor characteristics align geometrically to the global consensus. Lastly, we propose precision weighted averaging, where efficiently obtained uncertainty estimates are used to downweight uncertain nodes. This paper establishes the mathematical backbone for federated unpaired foundation models, enabling a global model to learn a unified representation of the world from fragmented, disjoint, and private data silos without requiring centralized storage or paired samples.