paper-with-me

홈 › Papers

TallyTrain: Communication-Efficient Federated Distillation

2026-06-30 · Radhakrishna Achanta, Will Reed 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 $\lceil \log_2 C \rceil$ bits per probe by transmitting only each peer's $\arg\max$ 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.

📄 PDF Abstract BibTeX arXiv:2607.00173

Code (0)

등록된 구현이 없습니다.

Tasks

Federated Learning

Similar Papers 제목 키워드 기반

Improving Communication Efficiency of Federated Distillation via Accumulating Local Updates

2023-12-07 · Zhiyuan Wu, Sheng Sun, Yuwei Wang, Min Liu 외

As an emerging federated learning paradigm, federated distillation enables communication-efficient model training by transmitting only small-scale knowledge during the learning process. To further improve the communicati…

Federated Learning

Communication-Efficient Federated Distillation

2020-12-01 · Felix Sattler, Arturo Marban, Roman Rischke, Wojciech Samek

Communication constraints are one of the major challenges preventing the wide-spread adoption of Federated Learning systems. Recently, Federated Distillation (FD), a new algorithmic paradigm for Federated Learning with f…

Federated Learningimage-classificationImage ClassificationLanguage Modeling+2

FedKD: Communication Efficient Federated Learning via Knowledge Distillation

2021-08-30 · Chuhan Wu, Fangzhao Wu, Lingjuan Lyu, Yongfeng Huang 외

Federated learning is widely used to learn intelligent models from decentralized data. In federated learning, clients need to communicate their local model updates in each iteration of model learning. However, model upda…

Federated LearningKnowledge Distillation

Wireless Federated Distillation for Distributed Edge Learning with Heterogeneous Data

2019-07-05 · Jin-Hyun Ahn, Osvaldo Simeone, Joonhyuk Kang

Cooperative training methods for distributed machine learning typically assume noiseless and ideal communication channels. This work studies some of the opportunities and challenges arising from the presence of wireless …

BIG-bench Machine LearningFederated Learning

Communication-Aware Knowledge Distillation for Federated LLM Fine-Tuning over Wireless Networks

2025-09-01 · Xinlu Zhang, Na Yan, Yang Su, Yansha Deng 외 arxiv

Federated learning (FL) for large language models (LLMs) offers a privacy-preserving scheme, enabling clients to collaboratively fine-tune locally deployed LLMs or smaller language models (SLMs) without exchanging raw da…

Knowledge DistillationFederated Learning