Xinmei Tian

9

Papers

13

Total Citations

Papers (9)

Robust Training of Federated Models with Extremely Label Deficiency

A Similarity Paradigm Through Textual Regularization Without Forgetting

Detecting Generated Images by Fitting Natural Image Distributions

NeurIPS 2025arXiv

Epistemic Uncertainty for Generated Image Detection

NeurIPS 2025arXiv

Enhanced Motion-Text Alignment for Image-to-Video Transfer Learning

Sheared Backpropagation for Fine-tuning Foundation Models

Towards Theoretical Understandings of Self-Consuming Generative Models

Interpreting and Improving Large Language Models in Arithmetic Calculation

From Yes-Men to Truth-Tellers: Addressing Sycophancy in Large Language Models with Pinpoint Tuning