Yi Ren

HUGS achieves a remarkable balance between grasp success and diversity, synthesizing 3.2 million grasps that can adaptively handle objects from screws to large boxes.

On-device retrieval is nearly solved, but ensuring safe and helpful responses from generative models remains a critical challenge in medical AI applications.

A groundbreaking benchmark reveals that existing medical QA systems overlook critical maternal and reproductive health queries, paving the way for more effective AI-driven healthcare solutions.

Aggregating moral judgments can obscure critical individual perspectives, but modeling these differences reveals deeper insights into moral classification.

Avatar V achieves unprecedented avatar video fidelity by directly conditioning on full video references, capturing dynamic behaviors that previous methods missed.

Achieve near-perfect structured information extraction from historical documents with a surprisingly simple framework that auto-accepts the vast majority of labels via cross-model consensus.

Injecting optical flow into VLMs lets them spot subtle video transitions that other methods miss, opening the door to more robust video understanding.

Ditch the static image: this method generates realistic talking avatars by learning from *videos* of the subject in completely different scenes.

Coordinating human input with autonomous orientation control enables stable teleoperation of multiple objects at high accelerations, even in complex scenarios.