Search papers, labs, and topics across Lattice.
This paper investigates Federated Active Learning (FAL) in low-budget scenarios, revealing that homogeneous data requires more coordination to prevent redundant queries, while heterogeneous data fosters diversity. The authors introduce a novel FAL framework that employs federated representation learning to align client data in a shared embedding space, allowing for effective global query selection while maintaining local annotation. Remarkably, this approach outperforms existing FAL methods, even with significantly larger annotation budgets, highlighting the importance of centralized coordination in privacy-constrained environments.
In low-budget Federated Active Learning, homogeneous data demands more coordination than heterogeneous data, flipping the conventional wisdom on its head.
Federated Active Learning (FAL) addresses the dual challenges of data privacy and label scarcity, where the absence of a global data view introduces additional hurdles for coordinated query selection. We study cross-silo FAL in the low-budget regime, where annotation decisions are most critical. We characterize, both theoretically and empirically, a heterogeneity reversal: in low-budget settings, homogeneous (IID) data requires stronger coordination to avoid redundant queries, whereas heterogeneous data naturally promotes diversity; this trend reverses at higher budgets. Thus, in contrast to the standard federated learning (FL) narrative where heterogeneity is a primary challenge, we show that IID settings are more challenging for query selection in FAL. Motivated by these findings, we propose a new FAL framework that utilizes federated representation learning to align client data in a shared embedding space. This enables the server to perform globally coordinated active selection over optionally obfuscated client embeddings, while annotation remains local to each client. Although our framework operates in the more challenging low-budget regime, it achieves performance that surpasses existing FAL methods even when they are given substantially larger annotation budgets, demonstrating the value of centralized coordination under privacy constraints.
