Jian Gao

A two-agent alternative in which employer-side and candidate-side agents represent these roles, exchange evidence, and update their judgments before deciding who advances, which shapes who reaches human review and how reliably that access recurs.

Task-level natural-language priors can transform low-resource LLM training, boosting performance even with minimal data.

Unbounded interaction horizons and real-time responsiveness redefine the possibilities for immersive AI-driven environments.

Feature extraction from point clouds just got a whole lot better: RWoDSN outperforms existing methods by simultaneously accounting for spatial distribution, topological properties, and geometric characteristics.

Real-time 3D scene reconstruction from streaming video is now possible with a feed-forward transformer that outperforms traditional SLAM methods.

Medical MLLMs, despite their size and training data, stumble on basic image classification due to four key failure modes, revealing a disconnect between hype and clinical readiness.