Xiaoyu Xu

Not every missing modality needs to be repaired for optimal sentiment analysis, and SIEVE learns to make this decision dynamically at the sample level.

CanonicalGS achieves a remarkable 2.5 dB improvement in novel view synthesis quality by stabilizing scene representations against noisy input data.

Robots can now recover from falls and seamlessly transition to walking on diverse terrains without prior terrain knowledge, thanks to a novel decoupling of training phases and surface types.