Hong-Yu Liu

Few-step diffusion models can achieve state-of-the-art 4-NFE generation without auxiliary score models or adversarial training, matching or exceeding multi-step teachers.

A set's "thickness" precisely dictates its observability for the $|x|$ potential Schrödinger equation across all positive times.

Interactive AI worlds can now speak and react in real-time with spatially accurate sound, bridging the sensory gap in generative environments.

OPSD-V shows that incorporating real video context during training can dramatically enhance the performance of autoregressive video generators, leading to superior visual quality and motion fidelity.

RCT-AD achieves a 61.5 nuScenes Detection Score by intelligently filtering unreliable sensor data, making autonomous driving safer in challenging urban environments.

LLMs can now effectively analyze deep learning frameworks for bugs without the need for costly runtime execution, revealing 31 previously undetected issues in PyTorch.

Open-source LongCat-Video-Avatar 1.5 leapfrogs closed-source competitors in audio-driven video generation by prioritizing practical engineering over architectural novelty, delivering commercial-grade quality and speed.

LLMs can achieve expert-level accuracy in specialized industrial domains by grounding them in traceable, queryable knowledge graphs built from fragmented scientific literature.

Today's best AI agents can only solve 55% of real-world academic tasks that university students find challenging, revealing a significant gap between current AI capabilities and the demands of academic workflows.

Autoregressive Transformers can now generate high-fidelity, animatable 4D Gaussian avatars from single portrait images, offering a new paradigm for controllable avatar creation.