Track papers across labs, topics, and trends.
Search papers, labs, and topics across Lattice.
This paper addresses the challenge of learning discriminative visual representations in Federated Learning (FL) by introducing FedSAP, a framework that stabilizes representation learning through a deterministic alignment curriculum and a geometry-driven proxy separation loss. These mechanisms mitigate the issues of gradient pressure and poorly organized embedding spaces that arise from early training rounds, particularly under non-IID conditions. Experimental results demonstrate that FedSAP achieves up to 4 percentage points improvement in recognition performance over existing prototype-based methods, especially in high heterogeneity scenarios.
FedSAP's innovative alignment curriculum and proxy separation loss lead to significantly improved recognition performance in Federated Learning, particularly when data is highly heterogeneous.
Learning discriminative visual representations from distributed, heterogeneous data is a fundamental challenge in Federated Learning (FL). Prototype-based methods address statistical heterogeneity by sharing class-level representations across clients but create a distance-dependent gradient pressure that is particularly severe during early training rounds: alignment pressure applied to immature global prototypes, aggregated from noisy local representations, generates large gradients that suppress the emergence of local discriminative structure. The result is a poorly organized embedding space and degraded recognition performance, particularly under severe non-IID conditions. We propose FedSAP, a framework that stabilises federated representation learning through two complementary mechanisms: a deterministic alignment curriculum that delays global alignment until local representations become stable and a geometry-driven proxy separation loss that enforces inter-class structure on the unit hypersphere using the existing prototype bank without introducing additional parameters or communication overhead. Together, these mechanisms produce compact, well-separated class clusters without altering the underlying communication protocol between federation's participants. Experiments across three benchmarks and varying degrees of heterogeneity show gains of up to 4 percentage points over the prototype-based baselines evaluated, with improvements most pronounced under high heterogeneity. The representational nature of our framework further enables a straightforward extension to semi-supervised settings, where unlabelled data is incorporated with minimal modification, underscoring the generality of scheduled alignment as a design principle.