Recently, the paper Multilayer Bootstrap Networks issued in the journal of Neural Networks in 2018, which was published by Professor Zhang Xiaolei of the School of Marine Science and Technology in NPU, won the best paper award and was also the unique one of this journal in 2020.

Fig. (1) Award Ceremony Held at the International Joint Conference of Neural Networks

This paper filled the long-term gap of the integrated learning method, one of the four categories of nonlinear methods of machine learning, in the unsupervised dimensionality reduction problem, broke through the traditional framework of deep neural network optimization relying on the back-propagation algorithm, realized the original innovation in the research of machine learning algorithm, effectively promoted the interpretability of artificial intelligence and deep learning in the neural network, developed the basic research such as small data-oriented depth model, as well as the related application technologies in the fields of data mining and multimedia information processing, which has an outstanding academic value and social significance. What’s more, this paper has received only support from the National Natural Science Foundation of China.

Fig. (2) Journal and Paper Cover

Neural Networks, the flagship journal of the International Neural Network Society, selects the only best paper award of the year from all the papers published in the past two years. After more than 100 Editorial Committee members from all over the world participated in the strict review process of nomination and voting, the best paper award was finally nominated by Professor Li Xuelong, deputy director of the academic committee of NPU, who is also the Editorial Board Member of the journal.

Fig. (3) Structure Diagram of Multi-layer Bootstrap Network

Fig. (4) Example of Results of Multi-layer Bootstrap Network Applied to Data Dimensionality Reduction

Fig. (5) Geometric Principle Analysis of Multilayer Bootstrap Network

Meanwhile, as one of the most important awards of International Neural Network Society, the winners of each year are international famous scholars, such as, the winner in 2017 is Professor Jürgen Schmidhuber, the founder of modern artificial intelligence and deep learning as well as the father of the world-famous neural network LSTM; Professor Stephen Grossberg, the founder of computational neuroscience, the first president of the International Neural Network Society and the first editor in chief of Neural Networks, is the winner of 2019.

Source: School of Marine Science and Technology