Pu Wang

Spatial control in co-speech gesture generation can now be fine-tuned in real-time, overcoming the limitations of traditional models that freeze prior motion chunks.

Loop-Mamba not only restores old photos more effectively but also introduces a novel degradation-aware framework that outperforms state-of-the-art methods by leveraging persistent memory and semantic guidance.

By combining VLM-guided segmentation with spectral analysis, this approach offers a more interpretable and accurate alternative to traditional classifiers for veterinary diagnosis, sidestepping reliance on large labeled datasets.

Style transfer can now capture the essence of artistic abstraction, not just surface-level appearance, by explicitly reinterpreting image structure.

Achieve state-of-the-art collaborative perception accuracy, communication efficiency, and resilience to packet loss by directly communicating quantized point cloud indices instead of intermediate features.