Cover Image

Perturb and Recover: Fine-tuning for Effective Backdoor Removal from CLIP

Fuente: arXiv

Saved in:

Bibliographic Details
Main Authors:	Singh, Naman Deep, Croce, Francesco, Hein, Matthias
Format:	Preprint
Published:	2024
Subjects:	Machine Learning Cryptography and Security Computer Vision and Pattern Recognition
Online Access:	Acceder al recurso
Tags:	Add Tag No Tags, Be the first to tag this record!

Similar Items

Robust CLIP: Unsupervised Adversarial Fine-Tuning of Vision Embeddings for Robust Large Vision-Language Models
by: Schlarmann, Christian, et al.
Published: (2024)

Adversarially Robust CLIP Models Can Induce Better (Robust) Perceptual Metrics
by: Croce, Francesco, et al.
Published: (2025)

Advancing Compositional Awareness in CLIP with Efficient Fine-Tuning
by: Peleg, Amit, et al.
Published: (2025)

ToxicTextCLIP: Text-Based Poisoning and Backdoor Attacks on CLIP Pre-training
by: Yao, Xin, et al.
Published: (2025)

BackWeak: Backdooring Knowledge Distillation Simply with Weak Triggers and Fine-tuning
by: Wang, Shanmin, et al.
Published: (2025)

Wicked Oddities: Selectively Poisoning for Effective Clean-Label Backdoor Attacks
by: Nguyen, Quang H., et al.
Published: (2024)

Recovering the Pre-Fine-Tuning Weights of Generative Models
by: Horwitz, Eliahu, et al.
Published: (2024)

Towards Reliable Evaluation and Fast Training of Robust Semantic Segmentation Models
by: Croce, Francesco, et al.
Published: (2023)

CLIP-Inspector: Model-Level Backdoor Detection for Prompt-Tuned CLIP via OOD Trigger Inversion
by: Jindal, Akshit, et al.
Published: (2026)

Under-confidence Backdoors Are Resilient and Stealthy Backdoors
by: Peng, Minlong, et al.
Published: (2022)

Beyond Full Poisoning: Effective Availability Attacks with Partial Perturbation
by: Zhe, Yu, et al.
Published: (2024)

Backdoor Federated Learning by Poisoning Backdoor-Critical Layers
by: Zhuang, Haomin, et al.
Published: (2023)

Backdoor Cleaning without External Guidance in MLLM Fine-tuning
by: Rong, Xuankun, et al.
Published: (2025)

Universal Backdoor Attacks
by: Schneider, Benjamin, et al.
Published: (2023)

Unveiling and Mitigating Backdoor Vulnerabilities based on Unlearning Weight Changes and Backdoor Activeness
by: Lin, Weilin, et al.
Published: (2024)

How to Backdoor Consistency Models?
by: Wang, Chengen, et al.
Published: (2024)

MedBlindTuner: Towards Privacy-preserving Fine-tuning on Biomedical Images with Transformers and Fully Homomorphic Encryption
by: Panzade, Prajwal, et al.
Published: (2024)

Hiding-in-Plain-Sight (HiPS) Attack on CLIP for Targetted Object Removal from Images
by: Daw, Arka, et al.
Published: (2024)

Does CLIP Know My Face?
by: Hintersdorf, Dominik, et al.
Published: (2022)

Memory Backdoor Attacks on Neural Networks
by: Luzon, Eden, et al.
Published: (2024)

Invisible Backdoor Attacks on Diffusion Models
by: Li, Sen, et al.
Published: (2024)

Backdoor Attack with Sparse and Invisible Trigger
by: Gao, Yinghua, et al.
Published: (2023)

IU: Imperceptible Universal Backdoor Attack
by: Lin, Hsin, et al.
Published: (2026)

Better Safe than Sorry: Pre-training CLIP against Targeted Data Poisoning and Backdoor Attacks
by: Yang, Wenhan, et al.
Published: (2023)

Backdooring CLIP through Concept Confusion
by: Hu, Lijie, et al.
Published: (2025)

Generating Potent Poisons and Backdoors from Scratch with Guided Diffusion
by: Souri, Hossein, et al.
Published: (2024)

Differentially Private Bias-Term Fine-tuning of Foundation Models
by: Bu, Zhiqi, et al.
Published: (2022)

Mudjacking: Patching Backdoor Vulnerabilities in Foundation Models
by: Liu, Hongbin, et al.
Published: (2024)

Beating Backdoor Attack at Its Own Game
by: Liu, Min, et al.
Published: (2023)

Removing the Trigger, Not the Backdoor: Alternative Triggers and Latent Backdoors
by: Abad, Gorka, et al.
Published: (2026)

Unsupervised Backdoor Detection and Mitigation for Spiking Neural Networks
by: Li, Jiachen, et al.
Published: (2025)

Forensics Adapter: Unleashing CLIP for Generalizable Face Forgery Detection
by: Cui, Xinjie, et al.
Published: (2024)

DisDet: Exploring Detectability of Backdoor Attack on Diffusion Models
by: Sui, Yang, et al.
Published: (2024)

Identifying Physically Realizable Triggers for Backdoored Face Recognition Networks
by: Raj, Ankita, et al.
Published: (2025)

VillanDiffusion: A Unified Backdoor Attack Framework for Diffusion Models
by: Chou, Sheng-Yen, et al.
Published: (2023)

CorruptEncoder: Data Poisoning based Backdoor Attacks to Contrastive Learning
by: Zhang, Jinghuai, et al.
Published: (2022)

PointBA: Towards Backdoor Attacks in 3D Point Cloud
by: Li, Xinke, et al.
Published: (2021)

Bad-PFL: Exploring Backdoor Attacks against Personalized Federated Learning
by: Fan, Mingyuan, et al.
Published: (2025)

Adversarially Guided Stateful Defense Against Backdoor Attacks in Federated Deep Learning
by: Ali, Hassan, et al.
Published: (2024)

UFID: A Unified Framework for Input-level Backdoor Detection on Diffusion Models
by: Guan, Zihan, et al.
Published: (2024)