The Center is dedicated to advancing the frontiers of language modeling and human-computer interaction, with breakthroughs in three major areas:

1. Speech Large Language Models (LLM)

The objective is to develop an innovative generation of end-to-end speech-language models. By decoupling knowledge from language, these models will enable low-resource, high-efficiency multilingual interaction and support lightweight deployment on terminal devices.

1. Speech Brain-Computer Interface (Speech BCI)

This research aims to construct a bidirectionally aligned brain-computer interaction foundation model. By integrating the deep semantic understanding of large language models with the neural signal decoding capabilities of brain-computer interfaces, it seeks to bridge the technological gap in interpreting high-level intents within brain-computer interactions.

1. Brain-inspired Auditory Intelligence

By leveraging the sparse perceptual mechanisms of the human auditory system, the goal is to construct an ultra-low-power, brain-inspired large auditory model based on event-driven spiking neural networks. This model is designed to enable intelligent processing and cognition of complex acoustic scenes.