For many years, our team has been focusing on AI for life science research, forming a research system with theoretical algorithms, intelligent systems, and AI applications in the scientific field as the core. In recent years, we have conducted in-depth research on multimodal artificial intelligence algorithms and system development, cross-domain autonomous evolutionary intelligence and applications, and published a number of high-level SCI papers and artificial intelligence conference papers including Nature, Nature Commun., IEEE TMI, AAAI, etc. as the first and correspondence (including parallel).

In the research of AI theoretical algorithms, we focus on multimodal rapid evolutionary intelligence and propose multiple network architectures including spectral neural network (WaveNet) for reconstructing information propagation matrix, behavioral perception relationship network (BARN) for multi-information monitoring, and multi-channel joint monitoring and classification network (JDC).

In the development of AI systems, we have developed an animal behavior recognition system for drug safety evaluation, which has been tried in many companies. In the bionic intelligent system, we have developed a bionic octopus hand, which can be used for physical quantity measurement in dark environments and flexible operation of human-computer interaction scenes.

In the application of life sciences, the proposed large model of behavior recognition is applied to the long-term behavior recognition of macaques, which promotes the discovery of the driving force of spinal cord aging. The proposed large-scale model of microscopic image processing is applied to the discovery of DNA-RNA-PROTEIN genetic variations, promoting the revelation of the mechanisms of human diseases.