 Prof. Lei Cheng, Zhejiang University
Personal profile: Lei Cheng is currently ZJU Young Professor with the College of Information Science and Electronic Engineering, Zhejiang University, Hangzhou, China. He received the B.Eng. degree from Zhejiang University in 2013, and the Ph.D. degree from The University of Hong Kong in 2018. He was a Research Scientist in Shenzhen Research Institute of Big Data, The Chinese University of Hong Kong, Shenzhen, from 2018 to 2021.
He is the author of the book “Bayesian Tensor Decomposition for Signal Processing and Machine Learning”, Springer, 2023. He was a Tutorial Speaker in ICASSP 2023 and Invited Speaker in ASA 2024. He is now Associate Editor for Elsevier Signal Processing and Young Editor for Acta Acustica.
Presentation title: Towards environment-aware underwater acoustics: Ocean sound speed field reconstruction beyond tensor neural network
Abstract: Obtaining accurate ocean sound speed fields (SSFs) across a three-dimensional (3D) geographic region is the stepping stone towards environment-aware underwater acoustics. However, the scarcity of measurements due to the high cost of underwater sensors, combined with the high dimensionality of complex 3D SSF, makes the reconstruction problem highly ill-conditioned, thus demanding advanced models and methods. Our recent work has analyzed the reconstruction error and identified one promising way: finding a representation model that is both concise and expressive. Following this path, we proposed a tensor neural network (TNN) model, which leverages the conciseness of tensor models and the expressive power of deep learning.
To further distill knowledge, we drew inspiration from the recent accomplishments of large-scale machine learning models such as DALLE-2 and SORA. We explored effective strategies for leveraging pre-trained deep models trained on natural images, as well as training deep generative models using our established 3D SSF datasets. Experimental results compellingly demonstrate the superior performance of our approach compared to state-of-the-art methods. Our findings have been published in five papers in the Journal of the Acoustical Society of America (JASA) over the past two years.
|