Yucheng Wang

CogEvol slashes learning artifact generation time from minutes to seconds while ensuring high reliability and quality.

Confidence miscalibration in medical AI can be mitigated, leading to both higher diagnostic accuracy and improved trust in clinical applications.

Living-Harness enables agents to learn from past failures dynamically, leading to substantial performance improvements in interactive tasks.

Kimi K3's innovative architecture achieves a 2.5x scaling efficiency improvement, enabling robust performance across diverse long-horizon tasks.

MLLMs can mislead medical professionals by misaligning confidence with accuracy, but a new calibration method cuts Expected Calibration Error by 40%.

Over-reliance on agentic decomposition can actually *hurt* audio understanding when a strong audio frontend already provides sufficient information, highlighting the importance of conditional evidence acquisition.

Forget painstakingly collecting real-world defect data: high-fidelity synthetic anomalies, automatically generated from product designs using an MLLM, can dramatically improve 3D anomaly detection.

Forget fixed residual connections: Attention Residuals let each layer selectively attend to previous layers, boosting performance and gradient flow in deep LLMs.

MLLMs are often overconfident, but a new confidence-driven training and test-time scaling approach can boost accuracy by 8.8% across benchmarks.

G-STAR tackles long-form, multi-speaker ASR by giving Speech-LLMs time-aware speaker tracking, enabling robust identity linking across chunks.