ProtoGuard-SL: Prototype Consistency Based Backdoor Defense for Vertical Split Learning

TL;DR

ProtoGuard-SL is a novel server-side defense for vertical split learning that detects backdoor attacks by analyzing class-conditional embedding consistency and filtering anomalous data.

Contribution

It introduces a prototype-based, conformal filtering approach to enhance robustness against poisoning attacks in split learning.

Findings

Achieves state-of-the-art backdoor detection performance.

Effectively identifies poisoned embeddings across multiple datasets.

Robust against various adaptive attack strategies.

Abstract

Vertical split learning (SL) enables collaborative model training across parties holding complementary features without sharing raw data, but recent work has shown that it is highly vulnerable to poisoning-based backdoor attacks operating on intermediate embeddings. By compromising malicious clients, adversaries can inject stealthy triggers that manipulate the server-side model while remaining difficult to detect, and existing defenses provide limited robustness against adaptive attacks. In this paper, we propose ProtoGuard-SL, a server-side defense that improves the robustness of split learning by exploiting class-conditional representation consistency in the embedding space. Our approach is motivated by the observation that benign embeddings within the same class exhibit stable semantic alignment, whereas poisoned embeddings inevitably disrupt this structure. ProtoGuard-SL adopts a two-stage framework that constructs robust class prototypes and transforms embeddings into a prototype-consistency representation, followed by a class-conditional, distribution-free conformal filtering strategy to identify and remove anomalous embeddings. Extensive experiments are conducted on three datasets, CIFAR-10, SVHN, and Bank Marketing, under three different attack settings demonstrate that our method achieves state-of-the-art performance.