Liwei Chen

REAL’s innovative graph-based memory management achieves a 22.72% improvement in long-term memory performance for LLMs by preserving fact evolution and enhancing retrieval accuracy.

Discovering stability and receptivity in complex systems no longer requires known equations, thanks to a new neural operator framework that learns directly from data.

Dramatically improve ASR for low-resource languages by cleverly leveraging readily available subtitles in higher-resource languages.