Di Lin

Anchoring positional embeddings in video generation can drastically enhance the quality of looping videos, achieving unprecedented temporal coherence and visual fidelity.

Explanation-guided learning can significantly boost the robustness of medical NER, achieving consistent performance gains across multiple models.

IViT achieves 93.80% accuracy in skin disease detection while reducing feature redundancy by 29.5%, striking a crucial balance between interpretability and performance.

LLMs can be culturally insensitive even when they possess relevant cultural knowledge, revealing a disconnect between knowledge and safety alignment.