Haokai Ma
2 papers · Latest:
Cryptography & Security
SnapGuard: Lightweight Prompt Injection Detection for Screenshot-Based Web Agents
SnapGuard is a lightweight, multimodal method to detect prompt injection in screenshot-based web agents, outperforming large VLMs in speed and efficiency.
2604.25562
Machine LearningLess Approximates More: Harmonizing Performance and Confidence Faithfulness via Hybrid Post-Training for High-Stakes Tasks
HyTuning improves LLM accuracy and confidence faithfulness for high-stakes tasks by adaptively combining reasoning distillation and internal feedback.
2604.08454
📬 Weekly AI Paper Digest
Get the top 10 AI/ML arXiv papers from the week — summarized, scored, and delivered to your inbox every Monday.