Research Areas
In the DI²Lab, we focus on exploring the intrinsic mechanisms of Large Language Models(LLMs) as well as deep models, while also paying close attention to the cross-fertilization of AI with downstream tasks.
Multi-modal Perception Fusion
Multimodal fusion sensing of vision, radar, text, and speech for ground water unmanned vehicles, human walking head and ECG signs, drones and robots, etc.
LLM Mechanism and Trustworthy AI
Research and enhancement of model credibility in safety-sensitive scenarios such as medical care and autonomous driving based on causal mechanism research.
Machine Consciousness
Constructing Artificial Consciousness, Exploring the Origin of Consciousness, and Scientifically Assessing the Benefits and Risks of Machine Consciousness.
Application Domains
In addition to the above research directions, DI²Lab also focuses on the use of advanced technologies to solve real-world problems.
