mlsanityData to 14 Oct 2026

SecureSplit: Mitigating Backdoor Attacks in Split Learning

arXiv 2601.14054 ↗First posted 20 Jan 2026cs.CR cs.DC cs.LGcited by 0

Not on the map. Why?

Abstract (arXiv)

Split Learning (SL) offers a framework for collaborative model training that respects data privacy by allowing participants to share the same dataset while maintaining distinct feature sets. However, SL is susceptible to backdoor attacks, in which malicious clients subtly alter their embeddings to insert hidden triggers that compromise the final trained model. To address this vulnerability, we introduce SecureSplit, a defense mechanism tailored to SL. SecureSplit applies a dimensionality transformation strategy to accentuate subtle differences between benign and poisoned embeddings, facilitating their separation. With this enhanced distinction, we develop an adaptive filtering approach that uses a majority-based voting scheme to remove contaminated embeddings while preserving clean ones. Rigorous experiments across four datasets (CIFAR-10, MNIST, CINIC-10, and ImageNette), five backdoor attack scenarios, and seven alternative defenses confirm the effectiveness of SecureSplit under various challenging conditions.

Key facts. No citing papers in the map yet (data through 14 Oct 2026). Builds on 22 of its 48 references that are in the map. First posted on arXiv 20 Jan 2026.

Builds on

22 of 48 references.

References · 22

  1. 2016Communication-Efficient Learning of Deep Networks from Decentralized Datacited by 7,669
  2. 2018UMAP: Uniform Manifold Approximation and Projection for Dimension Reductioncited by 2,898
  3. 2016Deep Learning with Differential Privacycited by 2,728
  4. 2019Local Model Poisoning Attacks to Byzantine-Robust Federated Learningcited by 493
  5. 2018Split learning for health: Distributed deep learning without sharing raw patient datacited by 395
  6. 2020FLTrust: Byzantine-robust Federated Learning via Trust Bootstrappingcited by 346
  7. 2020SplitFed: When Federated Learning Meets Split Learningcited by 310
  8. 2021Adversarial Neuron Pruning Purifies Backdoored Deep Modelscited by 235
  9. 2020fastai: A Layered API for Deep Learningcited by 183
  10. 2019Detailed comparison of communication efficiency of split learning and federated learningcited by 77
  11. 2022FederatedScope: A Flexible Federated Learning Platform for Heterogeneitycited by 53
  12. 2024Byzantine-Robust Decentralized Federated Learningcited by 48
10 more
  1. 2022BlindFL: Vertical Federated Machine Learning without Peeking into Your Datacited by 36
  2. 2022AFLGuard: Byzantine-robust Asynchronous Federated Learningcited by 29
  3. 2020Asymmetrical Vertical Federated Learningcited by 26
  4. 2025Do We Really Need to Design New Byzantine-robust Aggregation Rules?cited by 15
  5. 2025SafeSplit: A Novel Defense Against Client-Side Backdoor Attacks in Split Learning (Full Version)cited by 7
  6. 2025Byzantine-Robust Federated Learning over Ring-All-Reduce Distributed Computingcited by 6
  7. 2025Poisoning Attacks and Defenses to Federated Unlearningcited by 5
  8. 2024Securing Distributed Network Digital Twin Systems Against Model Poisoning Attackscited by 4
  9. 2024VFLIP: A Backdoor Defense for Vertical Federated Learning via Identification and Purificationcited by 4
  10. 2025Toward Malicious Clients Detection in Federated Learningcited by 4

Built on by

Not cited yet.

Similar, not cited

Close in wording, no citation either way

  1. 2023On Feasibility of Server-side Backdoor Attacks on Split Learning
  2. 2026ProtoGuard-SL: Prototype Consistency Based Backdoor Defense for Vertical Split Learning ↗
  3. 2024Dullahan: Stealthy Backdoor Attack against Without-Label-Sharing Split Learning
  4. 2023Backdoor Defense via Adaptively Splitting Poisoned Dataset
  5. 2026Client-Cooperative Split Learning ↗
  6. 2024Unlearning Backdoor Attacks through Gradient-Based Model Pruning
  7. 2023Shared Adversarial Unlearning: Backdoor Mitigation by Unlearning Shared Adversarial Examples
  8. 2025BadFU: Backdoor Federated Learning through Adversarial Machine Unlearning