Di Wang
4 papers · Latest:
Machine Learning
Benign Overfitting in Adversarial Training for Vision Transformers
This paper provides the first theoretical analysis showing benign overfitting in adversarially trained Vision Transformers, leading to strong generalization.
2604.19724
Machine LearningUnderstanding and Improving Continuous Adversarial Training for LLMs via In-context Learning Theory
This paper theoretically explains Continuous Adversarial Training (CAT) for LLMs using in-context learning and proposes an improved regularization method.
2604.12817
Cryptography & SecurityCoLA: A Choice Leakage Attack Framework to Expose Privacy Risks in Subset Training
This paper introduces CoLA, a framework demonstrating that subset training can leak sensitive information about data selection, challenging privacy assumptions.
2604.12342
Artificial IntelligenceAnalysis of LLM Performance on AWS Bedrock: Receipt-item Categorisation Case Study
This paper evaluates AWS Bedrock LLMs for receipt-item categorization, finding Claude 3.7 Sonnet offers the best balance of accuracy and cost.
2604.01615
📬 Weekly AI Paper Digest
Get the top 10 AI/ML arXiv papers from the week — summarized, scored, and delivered to your inbox every Monday.