Suorong Yang

Δ-AttnMask: Attention-Guided Masked Hidden States for Efficient Data Selection and Augmentation

Exposing Hallucinations To Suppress Them: VLMs Representation Editing With Generative Anchors

DONOD: Efficient and Generalizable Instruction Fine-Tuning for LLMs via Model-Intrinsic Data Selection

On-the-Fly Data Augmentation via Gradient-Guided and Sample-Aware Influence Estimation

A CLIP-Powered Framework for Robust and Generalizable Data Selection

Neural-Driven Image Editing

When Dynamic Data Selection Meets Data Augmentation: Achieving Enhanced Training Acceleration