Profiles2021-2026
Hongbin Liu
Duke University
11Papers
0Last author
14Unique co-authors
4First author
0Solo
First and last author come from byline position, which some venues assign alphabetically.
Recent papers11
- Leave My Images Alone: Preventing Multi-Modal Large Language Models from Analyzing Images via Visual Prompt InjectionAuthor 2 of 4ACL 2026
- Tracing Back the Malicious Clients in Poisoning Attacks to Federated LearningAuthor 3 of 4NeurIPS 2025 · 8 citations
- Visual Hallucinations of Multi-modal Large Language ModelsAuthor 2 of 4ACL Findings 2024 · 30 citations
- Data Poisoning Based Backdoor Attacks to Contrastive LearningAuthor 2 of 4CVPR 2024 · 12 citations
- Mudjacking: Patching Backdoor Vulnerabilities in Foundation ModelsAuthor 1 of 3USENIX Security 2024 · 5 citations
- PointCert: Point Cloud Classification with Deterministic Certified Robustness GuaranteesAuthor 3 of 4CVPR 2023
- Almost Tight L0-norm Certified Robustness of Top-k Predictions against Adversarial PerturbationsAuthor 4 of 5ICLR 2022 · 26 citations
- StolenEncoder: Stealing Pre-trained Encoders in Self-supervised LearningAuthor 3 of 4CCS 2022 · 23 citations
- PoisonedEncoder: Poisoning the Unlabeled Pre-training Data in Contrastive LearningAuthor 1 of 2USENIX Security 2022
- EncoderMI: Membership Inference against Pre-trained Encoders in Contrastive LearningAuthor 1 of 5CCS 2021 · 63 citations
- PointGuard: Provably Robust 3D Point Cloud ClassificationAuthor 1 of 2CVPR 2021
