Professor Lei Cheng,Zhejiang University

Accurate acquisition of multidimensional underwater environmental information is a prerequisite for reliable acoustic detection and communication. However, in large-scale deep-sea areas, the environmental data obtained by existing observation networks are sparse and noisy, and advanced multidimensional environmental field reconstruction and inference technologies are urgently needed. To address the challenges of insufficient utilization of multidimensional information in existing methods, this report will take the multidimensional extension of the matrix model - tensor model as the cornerstone, and introduce the collaborative modeling and algorithm design research of cutting-edge artificial intelligence models such as neural networks, Plug-and-Play, and diffusion models. The theoretical inadequacies of existing work will be pointed out, and future important research directions will be explored. In addition, this report will introduce the modeling of an integer programming problem for collaborative detection of underwater vehicles based on the environment information and explore future optimization directions.