Profiles2022-2026
Xingjun Ma
Fudan University
29Papers
6Last author
37Unique co-authors
0First author
0Solo
First and last author come from byline position, which some venues assign alphabetically.
Recent papers29
- SIDE: Surrogate Conditional Data Extraction from Diffusion ModelsAuthor 4 of 4AAAI 2026 · 9 citations
- Toward Universal and Transferable Jailbreak Attacks on Vision-Language ModelsAuthor 4 of 7ICLR 2026 · 4 citations
- AgenticEval: Toward Agentic and Self-Evolving Safety Evaluation of Large Language ModelsAuthor 7 of 8ACL Findings 2026 · 1 citation
- BackdoorAgent: A Unified Framework for Backdoor Attacks on LLM-based AgentsAuthor 8 of 9ACL Findings 2026 · 1 citation
- Deliberative Searcher: Improving LLM Reliability via Reinforcement Learning with ConstraintsAuthor 4 of 5ACL 2026 · 1 citation
- PaperAsk: A Benchmark for Reliability Evaluation of LLMs in Paper Search and ReadingAuthor 5 of 5WWW 2026 · 1 citation
- FakeWorld 1.0: An Omni-modal Benchmark for Fake Media and ContentAuthor 12 of 13ICML 2026
- Just Ask: Curious Code Agents Reveal System Prompts in Frontier LLMsAuthor 5 of 8ICML 2026
- Perturbation Effects on Robustness and Individual FairnessAuthor 4 of 7KDD 2026
- Towards Context-Invariant Safety Alignment for Large Language ModelsAuthor 6 of 7ICML 2026
- JailBound: Jailbreaking Internal Safety Boundaries of Vision-Language ModelsAuthor 7 of 8NeurIPS 2025 · 14 citations
- SAMA: Towards Multi-Turn Referential Grounded Video Chat with Large Language ModelsAuthor 5 of 6NeurIPS 2025 · 9 citations
- HoneypotNet: Backdoor Attacks Against Model ExtractionAuthor 5 of 5AAAI 2025 · 4 citations
- Ideator: Jailbreaking and Benchmarking Large Vision-Language Models Using ThemselvesAuthor 8 of 9ICCV 2025 · 2 citations
- BlueSuffix: Reinforced Blue Teaming for Vision-Language Models Against Jailbreak AttacksAuthor 5 of 6ICLR 2025
- Detecting Backdoor Samples in Contrastive Language Image PretrainingAuthor 4 of 5ICLR 2025
- FedAPT: Federated Adversarial Prompt Tuning for Vision-Language ModelsAuthor 3 of 4ACM MM 2025
- Optimizing Cross-Client Domain Coverage for Federated Instruction Tuning of Large Language ModelsAuthor 3 of 6EMNLP Findings 2025
- StolenLoRA: Exploring LoRA Extraction Attacks via Synthetic DataAuthor 4 of 4ICCV 2025
- TAPT: Test-Time Adversarial Prompt Tuning for Robust Inference in Vision-Language ModelsAuthor 5 of 5CVPR 2025
- Towards Million-Scale Adversarial Robustness Evaluation With Stronger Individual AttacksAuthor 5 of 5CVPR 2025
- X-Transfer Attacks: Towards Super Transferable Adversarial Attacks on CLIPAuthor 4 of 5ICML 2025
- White-box Multimodal Jailbreaks Against Large Vision-Language ModelsAuthor 2 of 6ACM MM 2024 · 22 citations
- AdvQDet: Detecting Query-Based Adversarial Attacks with Adversarial Contrastive Prompt TuningAuthor 3 of 6ACM MM 2024 · 6 citations
- Reconstructive Neuron Pruning for Backdoor DefenseAuthor 3 of 7ICML 2023 · 86 citations
- Distilling Cognitive Backdoor Patterns within an ImageAuthor 2 of 4ICLR 2023 · 7 citations
- Transferable Unlearnable ExamplesAuthor 4 of 6ICLR 2023 · 6 citations
- Fine-mixing: Mitigating Backdoors in Fine-tuned Language ModelsAuthor 3 of 5EMNLP Findings 2022 · 16 citations
- Backdoor Attacks on Crowd CountingAuthor 3 of 10ACM MM 2022 · 11 citations
