Report Title: Dynamic Analysis and Finite Time Control of Memristor Neural Networks

Presenter: Professor Wang Leimin, Doctoral Supervisor

Unit: China Address University (Wuhan)

Time: 14:00-15:00, Monday, May 26, 2025

Location: Room A321, School of Mathematics

Report Summary:

As the fourth fundamental circuit component following resistors, capacitors and inductors, the resistance value of the memristor can change with the amount of charge flowing through it and remain unchanged after power failure. This characteristic endows it with a natural memory function, providing the possibility of simulating the plasticity and memory function of neuronal synapses to construct neural network models that are closer to the characteristics of biological neurons. With the gradual maturation of memristor technology, memristor neural networks, as an important direction of next-generation neuromorphic computing, are receiving extensive attention from both the academic and industrial communities. As a new type of network model that integrates the unique memory characteristics of memristors, memristor neural networks not only inherit the powerful information processing capabilities of neural networks, but also exhibit more rich and complex dynamic behaviors due to the unique memory and nonlinear characteristics of memristors. This report focuses on the dynamic characteristics analysis of memristor neural networks, aiming to reveal how the characteristics of memristors affect the working mechanism and operation rules of neural networks, and provide a theoretical basis for network design, optimization and application. Meanwhile, developing effective intermittent control methods to achieve the stability and synchronization of the network within a limited time, solving the deficiencies of traditional control methods in convergence speed and accuracy, improving the performance of neural networks in applications, and also providing a brand-new perspective and tools for solving complex and nonlinear computing problems in the real world.

Biography of the Speaker:

Wang Leimin, male, graduated with a doctorate from the School of Artificial Intelligence and Automation of Huazhong University of Science and Technology in 2016. Currently, he is a professor and doctoral supervisor at the School of Automation of China University of Geosciences (Wuhan), a scholar of China University of Geosciences (Wuhan) - a young top-notch talent, a Member of the Chinese Association of Automation, a member of the Artificial Intelligence Society, and an IEEE Senior Member. Member of the TCCT Networked Control System and Stochastic System Control Group. In 2024, he/she won the Third Prize of the Natural Science Award of Hubei Province (ranked first), the Second Prize of the Natural Science Award of the Chinese Association of Automation (ranked third), and was nominated for the Excellent Doctoral Dissertation Award of the Chinese Association of Automation in 2017. From 2020 to 2024, he/she was consecutively included in the list of the world's top 2% of scientists released by Stanford University for five years. The current main research directions are the theory and application of time-delay neural networks, memristors and memristor systems, finite-time control of nonlinear systems, chaotic image encryption, etc. He has presided over more than ten projects including the National Natural Science Foundation of China's general project and youth Project, the Guangdong Provincial Natural Science Foundation of China's general Project, the Hubei Provincial Key Laboratory Fund, and defense and enterprise horizontal projects. He has successively participated in multiple national and provincial-level projects such as the National Natural Science Foundation of China's key project, general project, and Hubei Provincial Key Research and Development Project. Served as an editorial board member of SCI journals such as Mathematics, Scientific Reports, Frontiers in Physics and IEEE ACCESS, and as a column editor for multiple journals. More than 100 papers have been published in top international academic journals related to neural networks, artificial intelligence and automatic control such as IEEE TNNLS, IEEE TFS, IEEE TCYB, NNs, among which 80 are the first or corresponding papers. The papers have been cited more than 3,500 times on Google Scholar and have an H-index of 36. Eight papers have successively entered the top 1% of ESI highly cited papers, two have entered the top 0.1% of ESI hot papers, and one has been selected as an outstanding scientific and technological paper of Hubei Association for Science and Technology in 2023. Six national invention patents and multiple software Copyrights have been authorized.