Xing Xie

This article examines challenges and solutions related to noisy labels in training data.

A new method enhances security of Vision Transformers against adversarial attacks.

A new framework improves learning from incomplete data labels.

A new approach to integrate diverse cultural insights into language models.

Introducing a fresh approach to assess large language models effectively.

KIEval offers interactive evaluation to address data contamination in language models.

A look at the importance of aligning AI systems with human values.

Examining how noise in pre-training data impacts model performance.

A system to improve cultural representation in AI through simulated dialogues.

A new method for assessing LLMs aligns with human values.

TrendSim simulates social media trends to study misinformation effects.