# SSL-Cleanse: Trojan Detection and Mitigation in Self-Supervised Learning

arXiv: https://arxiv.org/abs/2303.09079 · HTML: https://mlsanity.com/paper/2303.09079/

## Key facts

- Cited by 11 papers in the map (data through 14 Oct 2026).
- Placed in Backdoor and data poisoning attacks, one of 12 directions in LLM safety and security.
- Builds on 28 of its 51 references that are in the map.
- First posted on arXiv 16 Mar 2023.

## Placed in

- LLM safety and security / [Backdoor and data poisoning attacks](https://mlsanity.com/safety/backdoor-and-data-poisoning-attacks/)

## Builds on (most cited first, 28 of 28)

- [Deep Residual Learning for Image Recognition](https://mlsanity.com/paper/1512.03385/) (2015)
- [A Simple Framework for Contrastive Learning of Visual Representations](https://mlsanity.com/paper/2002.05709/) (2020)
- [Momentum Contrast for Unsupervised Visual Representation Learning](https://mlsanity.com/paper/1911.05722/) (2019)
- [Bootstrap your own latent: A new approach to self-supervised Learning](https://mlsanity.com/paper/2006.07733/) (2020)
- [Exploring Simple Siamese Representation Learning](https://mlsanity.com/paper/2011.10566/) (2020)
- [ALBERT: A Lite BERT for Self-supervised Learning of Language Representations](https://mlsanity.com/paper/1909.11942/) (2019)
- [Improved Baselines with Momentum Contrastive Learning](https://mlsanity.com/paper/2003.04297/) (2020)
- [BadNets: Identifying Vulnerabilities in the Machine Learning Model Supply Chain](https://mlsanity.com/paper/1708.06733/) (2017)
- [Targeted Backdoor Attacks on Deep Learning Systems Using Data Poisoning](https://mlsanity.com/paper/1712.05526/) (2017)
- [Fine-Pruning: Defending Against Backdooring Attacks on Deep Neural Networks](https://mlsanity.com/paper/1805.12185/) (2018)
- [A Survey on Contrastive Self-supervised Learning](https://mlsanity.com/paper/2011.00362/) (2020)
- [Hidden Trigger Backdoor Attacks](https://mlsanity.com/paper/1910.00033/) (2019)
- [Graph Self-Supervised Learning: A Survey](https://mlsanity.com/paper/2103.00111/) (2021)
- [Blind Backdoors in Deep Learning Models](https://mlsanity.com/paper/2005.03823/) (2020)
- [Universal Litmus Patterns: Revealing Backdoor Attacks in CNNs](https://mlsanity.com/paper/1906.10842/) (2019)
- [BadEncoder: Backdoor Attacks to Pre-trained Encoders in Self-Supervised Learning](https://mlsanity.com/paper/2108.00352/) (2021)
- [Practical Detection of Trojan Neural Networks: Data-Limited and Data-Free Cases](https://mlsanity.com/paper/2007.15802/) (2020)
- [Backdoor Attacks on Self-Supervised Learning](https://mlsanity.com/paper/2105.10123/) (2021)
- [On Certifying Robustness against Backdoor Attacks via Randomized Smoothing](https://mlsanity.com/paper/2002.11750/) (2020)
- [Zero-Shot Self-Supervised Learning for MRI Reconstruction](https://mlsanity.com/paper/2102.07737/) (2021)
- [PoisonedEncoder: Poisoning the Unlabeled Pre-training Data in Contrastive Learning](https://mlsanity.com/paper/2205.06401/) (2022)
- [TrojViT: Trojan Insertion in Vision Transformers](https://mlsanity.com/paper/2208.13049/) (2022)
- [CorruptEncoder: Data Poisoning based Backdoor Attacks to Contrastive Learning](https://mlsanity.com/paper/2211.08229/) (2022)
- [ASSET: Robust Backdoor Data Detection Across a Multiplicity of Deep Learning Paradigms](https://mlsanity.com/paper/2302.11408/) (2023)
- [TrojText: Test-time Invisible Textual Trojan Insertion](https://mlsanity.com/paper/2303.02242/) (2023)
- [Cassandra: Detecting Trojaned Networks from Adversarial Perturbations](https://mlsanity.com/paper/2007.14433/) (2020)
- [ESTAS: Effective and Stable Trojan Attacks in Self-supervised Encoders with One Target Unlabelled Sample](https://mlsanity.com/paper/2211.10908/) (2022)
- [A Framework using Contrastive Learning for Classification with Noisy Labels](https://mlsanity.com/paper/2104.09563/) (2021)

## Built on by (most cited first, 11 of 11)

- [BadCLIP: Dual-Embedding Guided Backdoor Attack on Multimodal Contrastive Learning](https://mlsanity.com/paper/2311.12075/) (2023)
- [TrojLLM: A Black-box Trojan Prompt Attack on Large Language Models](https://mlsanity.com/paper/2306.06815/) (2023)
- [GhostEncoder: Stealthy Backdoor Attacks with Dynamic Triggers to Pre-trained Encoders in Self-supervised Learning](https://mlsanity.com/paper/2310.00626/) (2023)
- [DeDe: Detecting Backdoor Samples for SSL Encoders via Decoders](https://mlsanity.com/paper/2411.16154/) (2024)
- [Backdooring Self-Supervised Contrastive Learning by Noisy Alignment](https://mlsanity.com/paper/2508.14015/) (2025)
- [TrojFair: Trojan Fairness Attacks](https://mlsanity.com/paper/2312.10508/) (2023)
- [Circumventing Backdoor Space via Weight Symmetry](https://mlsanity.com/paper/2506.07467/) (2025)
- [Dataset Ownership Verification for Pre-trained Masked Models](https://mlsanity.com/paper/2507.12022/) (2025)
- [Backdooring CLIP through Concept Confusion](https://mlsanity.com/paper/2503.09095/) (2025)
- [REALM: A Unified Red-Teaming Benchmark for Physical-World VLMs](https://mlsanity.com/paper/2606.23892/) (2026)
- [The Platonic Defense: Backdoor Defense for Self-Supervised Encoders in the Era of Large Scale Pre-training](https://mlsanity.com/paper/2606.29451/) (2026)

## Abstract (arXiv)

Self-supervised learning (SSL) is a prevalent approach for encoding data representations. Using a pre-trained SSL image encoder and subsequently training a downstream classifier, impressive performance can be achieved on various tasks with very little labeled data. The growing adoption of SSL has led to an increase in security research on SSL encoders and associated Trojan attacks. Trojan attacks embedded in SSL encoders can operate covertly, spreading across multiple users and devices. The presence of backdoor behavior in Trojaned encoders can inadvertently be inherited by downstream classifiers, making it even more difficult to detect and mitigate the threat. Although current Trojan detection methods in supervised learning can potentially safeguard SSL downstream classifiers, identifying and addressing triggers in the SSL encoder before its widespread dissemination is a challenging task. This challenge arises because downstream tasks might be unknown, dataset labels may be unavailable, and the original unlabeled training dataset might be inaccessible during Trojan detection in SSL encoders. We introduce SSL-Cleanse as a solution to identify and mitigate backdoor threats in SSL encoders. We evaluated SSL-Cleanse on various datasets using 1200 encoders, achieving an average detection success rate of 82.2% on ImageNet-100. After mitigating backdoors, on average, backdoored encoders achieve 0.3% attack success rate without great accuracy loss, proving the effectiveness of SSL-Cleanse. The source code of SSL-Cleanse is available at https://github.com/UCF-ML-Research/SSL-Cleanse.