Salvaging Federated Learning by Local Adaptation

联合学习 计算机科学 差别隐私 适应(眼睛) 激励 任务(项目管理) 质量(理念) 方案(数学) 差速器(机械装置) 人工智能 机器学习 数据挖掘 心理学 数学分析 哲学 数学 管理 认识论 神经科学 工程类 经济 微观经济学 航空航天工程
作者
Changyuan Yu,Eugene Bagdasaryan,Vitaly Shmatikov
出处
期刊:Cornell University - arXiv 被引量:102
标识
DOI:10.48550/arxiv.2002.04758
摘要

Federated learning (FL) is a heavily promoted approach for training ML models on sensitive data, e.g., text typed by users on their smartphones. FL is expressly designed for training on data that are unbalanced and non-iid across the participants. To ensure privacy and integrity of the fedeated model, latest FL approaches use differential privacy or robust aggregation. We look at FL from the \emph{local} viewpoint of an individual participant and ask: (1) do participants have an incentive to participate in FL? (2) how can participants \emph{individually} improve the quality of their local models, without re-designing the FL framework and/or involving other participants? First, we show that on standard tasks such as next-word prediction, many participants gain no benefit from FL because the federated model is less accurate on their data than the models they can train locally on their own. Second, we show that differential privacy and robust aggregation make this problem worse by further destroying the accuracy of the federated model for many participants. Then, we evaluate three techniques for local adaptation of federated models: fine-tuning, multi-task learning, and knowledge distillation. We analyze where each is applicable and demonstrate that all participants benefit from local adaptation. Participants whose local models are poor obtain big accuracy improvements over conventional FL. Participants whose local models are better than the federated model\textemdash and who have no incentive to participate in FL today\textemdash improve less, but sufficiently to make the adapted federated model better than their local models.
最长约 10秒,即可获得该文献文件

科研通智能强力驱动
Strongly Powered by AbleSci AI
更新
PDF的下载单位、IP信息已删除 (2025-6-4)

科研通是完全免费的文献互助平台,具备全网最快的应助速度,最高的求助完成率。 对每一个文献求助,科研通都将尽心尽力,给求助人一个满意的交代。
实时播报
溪夕er完成签到,获得积分10
刚刚
SSS水鱼发布了新的文献求助10
1秒前
ACE发布了新的文献求助10
1秒前
2秒前
splemeth完成签到,获得积分10
3秒前
桐桐应助liars采纳,获得10
4秒前
yishan101完成签到,获得积分10
5秒前
6秒前
可靠觅珍应助高大的依秋采纳,获得30
8秒前
8秒前
9秒前
澡雪发布了新的文献求助10
9秒前
大模型应助YXC采纳,获得10
10秒前
10秒前
蟪蛄鸪完成签到 ,获得积分20
11秒前
11秒前
Bryan应助科研通管家采纳,获得10
13秒前
科研通AI5应助科研通管家采纳,获得10
13秒前
FashionBoy应助科研通管家采纳,获得10
14秒前
饱满南松发布了新的文献求助10
14秒前
乐乐应助科研通管家采纳,获得10
14秒前
丘比特应助科研通管家采纳,获得10
14秒前
领导范儿应助科研通管家采纳,获得10
14秒前
852应助科研通管家采纳,获得10
14秒前
Owen应助科研通管家采纳,获得30
14秒前
Bryan应助科研通管家采纳,获得10
14秒前
Akim应助科研通管家采纳,获得10
14秒前
乐乐应助科研通管家采纳,获得10
14秒前
花花发布了新的文献求助10
16秒前
多金完成签到,获得积分10
16秒前
闪电侠完成签到 ,获得积分10
17秒前
18秒前
多金发布了新的文献求助10
19秒前
19秒前
dodo应助王者归来采纳,获得200
19秒前
精明的沅应助甜美的一笑采纳,获得10
20秒前
23秒前
Rondab应助HIT_C采纳,获得10
24秒前
卢建军应助虚幻初之采纳,获得20
25秒前
LL关闭了LL文献求助
25秒前
高分求助中
A new approach to the extrapolation of accelerated life test data 1000
Picture Books with Same-sex Parented Families: Unintentional Censorship 700
ACSM’s Guidelines for Exercise Testing and Prescription, 12th edition 500
Nucleophilic substitution in azasydnone-modified dinitroanisoles 500
不知道标题是什么 500
Indomethacinのヒトにおける経皮吸収 400
Phylogenetic study of the order Polydesmida (Myriapoda: Diplopoda) 370
热门求助领域 (近24小时)
化学 材料科学 医学 生物 工程类 有机化学 生物化学 物理 内科学 纳米技术 计算机科学 化学工程 复合材料 遗传学 基因 物理化学 催化作用 冶金 细胞生物学 免疫学
热门帖子
关注 科研通微信公众号,转发送积分 3975516
求助须知:如何正确求助?哪些是违规求助? 3519930
关于积分的说明 11200130
捐赠科研通 3256278
什么是DOI,文献DOI怎么找? 1798183
邀请新用户注册赠送积分活动 877425
科研通“疑难数据库(出版商)”最低求助积分说明 806320