# Contribution-Aware Bandwidth Allocation for Multimodal Split Learning

arXiv: https://arxiv.org/abs/2609.01406 · HTML: https://mlsanity.com/paper/2609.01406/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Builds on 23 of its 33 references that are in the map.
- First posted on arXiv 1 Sep 2026.

## Builds on (most cited first, 23 of 23)

- [PyTorch: An Imperative Style, High-Performance Deep Learning Library](https://mlsanity.com/paper/1912.01703/) (2019)
- [Communication-Efficient Learning of Deep Networks from Decentralized Data](https://mlsanity.com/paper/1602.05629/) (2016)
- [Split learning for health: Distributed deep learning without sharing raw patient data](https://mlsanity.com/paper/1812.00564/) (2018)
- [SplitFed: When Federated Learning Meets Split Learning](https://mlsanity.com/paper/2004.12088/) (2020)
- [Balanced Multimodal Learning via On-the-fly Gradient Modulation](https://mlsanity.com/paper/2203.15332/) (2022)
- [BottleNet++: An End-to-End Approach for Feature Compression in Device-Edge Co-Inference Systems](https://mlsanity.com/paper/1910.14315/) (2019)
- [Split Learning in 6G Edge Networks](https://mlsanity.com/paper/2306.12194/) (2023)
- [MBQ: Modality-Balanced Quantization for Large Vision-Language Models](https://mlsanity.com/paper/2412.19509/) (2024)
- [Enhancing multimodal cooperation via sample-level modality valuation](https://mlsanity.com/paper/2309.06255/) (2023)
- [MM-SHAP: A Performance-agnostic Metric for Measuring Multimodal Contributions in Vision and Language Models & Tasks](https://mlsanity.com/paper/2212.08158/) (2022)
- [A Federated Learning Framework for Healthcare IoT devices](https://mlsanity.com/paper/2005.05083/) (2020)
- [Reducing Communication for Split Learning by Randomized Top-k Sparsification](https://mlsanity.com/paper/2305.18469/) (2023)
- [Communication-Efficient Split Learning via Adaptive Feature-Wise Compression](https://mlsanity.com/paper/2307.10805/) (2023)
- [Improving Multimodal Learning via Imbalanced Learning](https://mlsanity.com/paper/2507.10203/) (2025)
- [SHAPE: An Unified Approach to Evaluate the Contribution and Cooperation of Individual Modalities](https://mlsanity.com/paper/2205.00302/) (2022)
- [Communication-Efficient Multimodal Federated Learning: Joint Modality and Client Selection](https://mlsanity.com/paper/2401.16685/) (2024)
- [Distributed LLMs and Multimodal Large Language Models: A Survey on Advances, Challenges, and Future Directions](https://mlsanity.com/paper/2503.16585/) (2025)
- [C3-SL: Circular Convolution-Based Batch-Wise Compression for Communication-Efficient Split Learning](https://mlsanity.com/paper/2207.12397/) (2022)
- [Split Federated Learning for 6G Enabled-Networks: Requirements, Challenges and Future Directions](https://mlsanity.com/paper/2309.09086/) (2023)
- [Adaptive Compression-Aware Split Learning and Inference for Enhanced Network Efficiency](https://mlsanity.com/paper/2311.05739/) (2023)
- [SL-ACC: A Communication-Efficient Split Learning Framework with Adaptive Channel-wise Compression](https://mlsanity.com/paper/2508.12984/) (2025)
- [Contribution-Guided Asymmetric Learning for Robust Multimodal Fusion under Imbalance and Noise](https://mlsanity.com/paper/2510.26289/) (2025)
- [AutoEncoder-Compressed Parallel Split Learning for Pre-trained Model Fine-Tuning](https://mlsanity.com/paper/2607.17913/) (2026)

## Abstract (arXiv)

Multimodal models are increasingly the default option for perception at the network edge, yet they are trained almost entirely in the datacenter, because a client holding several sensor streams cannot host an encoder per modality. Split Learning makes such training feasible by keeping only the first layers on the device, at the cost of an uplink that must carry smashed activations for every modality at every step. Existing compression schemes give each modality the same keep-ratio, so the shared budget is divided in proportion to smashed-activation dimension, a quantity unrelated to how much each modality contributes to the fused prediction. We make that division an explicit decision and call it inter-modality allocation: under a fixed uplink budget, every policy transmits the same expected payload and differs only in how that payload is split across modalities. Our allocator, ModalShare, sets each modality's keep-ratio from a Shapley contribution score that the server computes over coalitions of activations it has already received. Measuring this score adds no uplink traffic and no client-side computation, and needs no prior knowledge of which stream is which. ModalShare improves accuracy over equal keep-ratios by 15.4 and 12.4 percentage points on CREMA-D and MVSA at matched payload in 5x compression, with strong performance across three compressors, three datasets, and four budgets. We show that existing compressors underperform in multimodal settings, with ModalShare recovering what gains are left behind.