Ke Chen

The paper identifies that the convenience of host-acting agents leads to semantic under-specification in user goals, which forces the agent to generate potentially risky execution plans.

The paper proposes an infrastructure, clawgang and meowtrade, to transform private, non-transferable agent memories into verifiable, tradable economic commodities.

TRACE proposes a novel method to mitigate catastrophic forgetting in continual LLM fine-tuning by identifying and isolating a small, task-specific subset of essential parameters for each task.

The paper introduces Med-HEAL, a comprehensive framework and dataset for systematically identifying and mitigating hallucinations in medical LLMs, demonstrating that a self-critique pipeline significantly improves model accuracy.

MOSS-Audio is a unified audio-language model designed for comprehensive understanding of speech, environmental sounds, and music, achieving strong performance across various audio-grounded tasks.