paper-with-me

Papers

Profit: Benchmarking Personalization and Robustness Trade-off in Federated Prompt Tuning

2023-10-06 · Liam Collins, Shanshan Wu, Sewoong Oh, Khe Chai Sim

In many applications of federated learning (FL), clients desire models that are personalized using their local data, yet are also robust in the sense that they retain general global knowledge. However, the presence of data heterogeneity across clients induces a fundamental trade-off between personalization (i.e., adaptation to a local distribution) and robustness (i.e., not forgetting previously learned general knowledge). It is critical to understand how to navigate this personalization vs robustness trade-off when designing federated systems, which are increasingly moving towards a paradigm of fine-tuning large foundation models. Due to limited computational and communication capabilities in most federated settings, this foundation model fine-tuning must be done using parameter-efficient fine-tuning (PEFT) approaches. While some recent work has studied federated approaches to PEFT, the personalization vs robustness trade-off of federated PEFT has been largely unexplored. In this work, we take a step towards bridging this gap by benchmarking fundamental FL algorithms -- FedAvg and FedSGD plus personalization (via client local fine-tuning) -- applied to one of the most ubiquitous PEFT approaches to large language models (LLMs) -- prompt tuning -- in a multitude of hyperparameter settings under varying levels of data heterogeneity. Our results show that federated-trained prompts can be surprisingly robust when using a small learning rate with many local epochs for personalization, especially when using an adaptive optimizer as the client optimizer during federated training. We also demonstrate that simple approaches such as adding regularization and interpolating two prompts are effective in improving the personalization vs robustness trade-off in computation-limited settings with few local updates allowed for personalization.

📄 PDF Abstract BibTeX arXiv:2310.04627

Code (0)

등록된 구현이 없습니다.

Tasks

BenchmarkingFederated LearningGeneral KnowledgeNavigateparameter-efficient fine-tuning

Similar Papers 제목 키워드 기반

Motley: Benchmarking Heterogeneity and Personalization in Federated Learning

2022-06-18 · Shanshan Wu, Tian Li, Zachary Charles, Yu Xiao 외

Personalized federated learning considers learning models unique to each client in a heterogeneous network. The resulting client-specific models have been purported to improve metrics such as accuracy, fairness, and robu…

BenchmarkingFairnessFederated LearningMeta-Learning+3

Understanding the Statistical Accuracy-Communication Trade-off in Personalized Federated Learning with Minimax Guarantees

2024-10-11 · Xin Yu, Zelin He, Ying Sun, Lingzhou Xue 외

Personalized federated learning (PFL) offers a flexible framework for aggregating information across distributed clients with heterogeneous data. This work considers a personalized federated learning setting that simulta…

Federated LearningPersonalized Federated Learning

Federated Evaluation and Tuning for On-Device Personalization: System Design & Applications

2021-02-16 · Matthias Paulik, Matt Seigel, Henry Mason, Dominic Telaar 외

We describe the design of our federated task processing system. Originally, the system was created to support two specific federated tasks: evaluation and tuning of on-device ML systems, primarily for the purpose of pers…

Federated Learning

Certified Robustness in Federated Learning

2022-06-06 · Motasem Alfarra, Juan C. Pérez, Egor Shulgin, Peter Richtárik 외

Federated learning has recently gained significant attention and popularity due to its effectiveness in training machine learning models on distributed data privately. However, as in the single-node supervised learning s…

Federated Learning

Ditto: Fair and Robust Federated Learning Through Personalization

2020-12-08 · Tian Li, Shengyuan Hu, Ahmad Beirami, Virginia Smith

Fairness and robustness are two important concerns for federated learning systems. In this work, we identify that robustness to data and model poisoning attacks and fairness, measured as the uniformity of performance acr…

FairnessFederated LearningModel PoisoningMulti-Task Learning+1