用预训练嵌入评估联邦学习前的系统表现，提前预判效果
Task2vec Readiness: Diagnostics for Federated Learning from Pre-Training Embeddings
- 基于任务嵌入计算客户端一致性、分散度等无监督指标
- 指标与最终性能相关性超0.9，跨数据集和客户端数稳定
- 适合想选优质客户端或优化联邦策略的研究者
联邦学习性能高度依赖客户端间异质性，但实践者缺乏训练前预测系统行为的方法。本文提出基于Task2Vec嵌入的就绪度指标，量化联邦系统在训练前的对齐程度，并与最终性能显著相关。通过计算客户端嵌入的凝聚性、分散度和密度等无监督指标，在CIFAR-10、FEMNIST、PathMNIST、BloodMNIST等多个数据集上，于10至20个客户端、狄利克雷异质性参数α∈{0.05,…,5.0}范围内测试，结果表明部分就绪度指标与最终性能之间的皮尔逊和斯皮尔曼相关系数普遍超过0.9，验证其作为联邦学习结果可靠代理的有效性。该方法为异构联邦学习提供了一种可解释、训练前的诊断工具，可用于预测与指导客户端选择。
原文摘要 · Abstract (English)
Federated learning (FL) performance is highly sensitive to heterogeneity across clients, yet practitioners lack reliable methods to anticipate how a federation will behave before training. We propose readiness indices, derived from Task2Vec embeddings, that quantifies the alignment of a federation prior to training and correlates with its eventual performance. Our approach computes unsupervised metrics -- such as cohesion, dispersion, and density -- directly from client embeddings. We evaluate these indices across diverse datasets (CIFAR-10, FEMNIST, PathMNIST, BloodMNIST) and client counts (10--20), under Dirichlet heterogeneity levels spanning $α\in \{0.05,\dots,5.0\}$ and FedAVG aggregation strategy. Correlation analyses show consistent and significant Pearson and Spearman coefficients between some of the Task2Vec-based readiness and final performance, with values often exceeding 0.9 across dataset$\times$client configurations, validating this approach as a robust proxy for FL outcomes. These findings establish Task2Vec-based readiness as a principled, pre-training diagnostic for FL that may offer both predictive insight and actionable guidance for client selection in heterogeneous federations.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。