Zhanxing Zhu

Class-imbalance bias in long-tailed semi-supervised learning can be systematically mitigated with a novel dynamic pruning approach that reallocates gradient budgets for better generalization.

RLVR's reasoning boost isn't just about more data – it's a Goldilocks problem: too easy and LLMs skip the reasoning, too hard and they break down, but just right and they become reasoning powerhouses.

Accurately simulating 3D flexible deformations just got easier: MAVEN leverages higher-dimensional mesh elements to outperform existing GNNs, especially when dealing with large deformations and prolonged contacts.