# TallyTrain: Communication-Efficient Federated Distillation

arXiv: https://arxiv.org/abs/2607.00173 · HTML: https://mlsanity.com/paper/2607.00173/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Builds on 24 of its 29 references that are in the map.
- First posted on arXiv 30 Jun 2026.

## Builds on (most cited first, 24 of 24)

- [Distilling the Knowledge in a Neural Network](https://mlsanity.com/paper/1503.02531/) (2015)
- [Communication-Efficient Learning of Deep Networks from Decentralized Data](https://mlsanity.com/paper/1602.05629/) (2016)
- [Pointer Sentinel Mixture Models](https://mlsanity.com/paper/1609.07843/) (2016)
- [Advances and Open Problems in Federated Learning](https://mlsanity.com/paper/1912.04977/) (2019)
- [Federated Optimization in Heterogeneous Networks](https://mlsanity.com/paper/1812.06127/) (2018)
- [Membership Inference Attacks against Machine Learning Models](https://mlsanity.com/paper/1610.05820/) (2016)
- [Accurate, Large Minibatch SGD: Training ImageNet in 1 Hour](https://mlsanity.com/paper/1706.02677/) (2017)
- [Byzantine-Robust Distributed Learning: Towards Optimal Statistical Rates](https://mlsanity.com/paper/1803.01498/) (2018)
- [Ensemble Distillation for Robust Model Fusion in Federated Learning](https://mlsanity.com/paper/2006.07242/) (2020)
- [Local SGD Converges Fast and Communicates Little](https://mlsanity.com/paper/1805.09767/) (2018)
- [Tighter Theory for Local SGD on Identical and Heterogeneous Data](https://mlsanity.com/paper/1909.04746/) (2019)
- [Communication-Efficient On-Device Machine Learning: Federated Distillation and Augmentation under Non-IID Private Data](https://mlsanity.com/paper/1811.11479/) (2018)
- [FedKD: Communication Efficient Federated Learning via Knowledge Distillation](https://mlsanity.com/paper/2108.13323/) (2021)
- [DiLoCo: Distributed Low-Communication Training of Language Models](https://mlsanity.com/paper/2311.08105/) (2023)
- [Cronus: Robust and Heterogeneous Collaborative Learning with Black-Box Knowledge Transfer](https://mlsanity.com/paper/1912.11279/) (2019)
- [Overcoming Data and Model Heterogeneities in Decentralized Federated Learning via Synthetic Anchors](https://mlsanity.com/paper/2405.11525/) (2024)
- [Decentralized Sporadic Federated Learning: A Unified Algorithmic Framework with Convergence Guarantees](https://mlsanity.com/paper/2402.03448/) (2024)
- [FedSPD: A Soft-clustering Approach for Personalized Decentralized Federated Learning](https://mlsanity.com/paper/2410.18862/) (2024)
- [DFML: Decentralized Federated Mutual Learning](https://mlsanity.com/paper/2402.01863/) (2024)
- [NTK-DFL: Enhancing Decentralized Federated Learning in Heterogeneous Settings via Neural Tangent Kernel](https://mlsanity.com/paper/2410.01922/) (2024)
- [Mitigating Privacy-Utility Trade-off in Decentralized Federated Learning via f-Differential Privacy](https://mlsanity.com/paper/2510.19934/) (2025)
- [Competitive Advantage Attacks to Decentralized Federated Learning](https://mlsanity.com/paper/2310.13862/) (2023)
- [On the Fast Adaptation of Delayed Clients in Decentralized Federated Learning: A Centroid-Aligned Distillation Approach](https://mlsanity.com/paper/2508.02993/) (2025)
- [Decentralized Federated Learning by Partial Message Exchange](https://mlsanity.com/paper/2603.01730/) (2026)

## Abstract (arXiv)

Federated learning is bandwidth-bound on two orthogonal axes: model size, which limits how often parameter-averaging methods can afford to merge, and class count, which makes per-probe soft-label distillation prohibitive at large vocabularies. Both ceilings tighten as modern systems scale. We collapse the class-count axis to lceil log2 C rceil bits per probe by transmitting only each peer's argmax class index, where C is the number of output classes. The resulting protocol, TallyTrain, is not merely compressed: under non-IID training it can be preferable to soft-label distillation, because under-trained peers are confidently wrong and majority voting filters this noise where soft-label averaging amplifies it. Across standard benchmarks, TallyTrain matches or beats soft-label distillation at up to three orders of magnitude less communication. We also relax the model-size axis: we compose the cheap hard-label consensus with sparse parameter merges to obtain a bandwidth-bridge variant, which Pareto-dominates every tested operating point of the standard FedAvg, FedProx and FedDF baselines.