Search papers, labs, and topics across Lattice.
Zhejiang University
3
0
7
0
Synthetic data that looks good can still tank your model's performance – Optimsyn uses influence functions to find the *actually* useful synthetic examples and optimize your generation rubrics.
Forget real-world video datasets: training VLMs on just 7.7K synthetic videos with temporal primitives beats 165K real-world examples, unlocking surprisingly effective transfer learning for video reasoning.
Multimodal models are often blind at birth: a new "Visual Attention Score" reveals they struggle to focus on visual inputs during cold-start, but a simple attention-guided fix can boost performance by 7%.
