Mingyu Wang

Ocean University of China

Author Statistics

Papers

Citation

H-Index

i-10 index

Research Trends

Author Order

Document Type

Co-Authors

Yunlong Li

Applied Science Private University

Jianping Li

Ocean University of China

Ying Hu

Central South University

Chu Zhang

Shenzhen University

Xianfen Diao

Shenzhen University

Yiyi Zhang

Guangxi University

Luo Yang-xia

Xi'an University of Finance and Economics

Hongxin Lin

Fujian Normal University

Jian Hu

Nanjing University of Science and Technology

Bingsheng Huang

Shenzhen University

Cooperative Institutions

Chinese Academy of Sciences

Zhejiang University

University of Chinese Academy of Sciences

Institute of Optics and Electronics, Chinese Academy of Sciences

Shenzhen University

Chongqing University

University of Science and Technology Beijing

Shenzhen University Health Science Center

Xi'an University of Finance and Economics

Xinjiang University

Author Statistics

Papers

Citation

H-Index

i-10 index

Research Field

Numerical research on chaos control

Mingyu Wang

In this paper, the numerical simulation method is applied to investigate the control for a given chaotic system. In order to obtain a better understanding of the relationship between the Lyapunov exponents and evolution of chaos, a stable Duffing system is examined. By way of adding a pulsing perturbation of the resistance coefficient, the pulsing control method is applied to control the chaotic motion of Duffing system. This control method is equivalent to impulse control, so the pulse intensity, positive pulse duration, pulse period and other control parameters on the results of chaos control are discussed in detail.

Duffing equation

Control of chaos

Impulse control

10.1109/icise.2010.5688647

Cite

Citations (0)

A method for testing the feasibility of 3D map design based on multi-agent AI-driven

2021 2nd International Conference on Big Data Economy and Information Management (BDEIM) (2021)

Yang Bai Zhuxin Xue Yiyi Zhang Mingyu Wang Yingchao Ren

As one of the most cutting-edge technologies of graphics and multimedia, the core of the three-dimensional map is the spatial layout design of the virtual world. The three-dimensional digital map has significant spatial design significance. It uses a method consistent with geographic theory to integrate the real and the fiction and attach it to various applications such as factory layout, workshop layout, production line planning, and so on. Therefore, the three-dimensional digital map as the spatial background of production and operation to carry various resources is an important part of digital design. Among them, ensuring that the virtual avatar representing a real person can move in an open and feasible space and reach the expected endpoint is an important content of three-dimensional map path planning. To solve the problem of 3D map design level testing, this paper proposes an automated testing method driven by ai multi-agents, which simulates three typical human behaviors and conducts automatic, fast, and comprehensive testing of 3D maps to evaluate the level of 3D map design.

Avatar

Digital mapping

Factory (object-oriented programming)

Interactive Design

10.1109/bdeim55082.2021.00094

Cite

Citations (0)

Self-Consistency Training with Hierarchical Temporal Aggregation for Sound Event Detection

2022 Asia-Pacific Signal and Information Processing Association Annual Summit and Conference (APSIPA ASC) (2022)

Yunlong Li Xiujuan Zhu Mingyu Wang Ying Hu

In this paper, we propose a sound event detection (SED) method based on the self-consistency training (SCT) strategy and a hierarchical temporal aggregation (HTA) module, named SCT-HTA. This method adopts Mean Teacher (MT) semi-supervised learning method, exploiting a dual-branch convolutional recurrent neural network (CRNN) structure including the main branch and auxiliary branch. We adopt an SCT strategy to apply the self-consistency regularization in addition to the MT loss to maintain the consistency between the outputs of the auxiliary and main branches. Furthermore, an HTA module is designed to aggregate the information at different temporal resolutions. We also explored three aggregators to be applied in the HTA module and four kinds of combinations of pooling methods in the localization modules of two branches. Experimental results demonstrate that our proposed SCT-HTA method outperforms the four compared methods. The results show that the max pooling aggregator has a better ability to highlight the location of sound events. And the "linear softmax + attention" combination of the pooling method achieves the best performance.

10.23919/apsipaasc55919.2022.9980285

Cite

Citations (2)

Exploring the potential of contemporary deep learning methods in purifying polluted information

Machine Learning Science and Technology (2024)

Mingyu Wang Jianping Li

Abstract Accurately predicting chaotic dynamical systems is a crucial task in various fields, and recent advancements have leveraged deep learning for this purpose. However, in the era of big data, the inevitable challenge of data contamination caused by invalid information from other interfering systems becomes increasingly prominent and complicates accurate predictions. Although contemporary deep learning methods have shown their potential, very few studies have focused on developing algorithms specifically designed to address the problem posed by such data contamination. Thus, exploring the ability of contemporary deep learning methods to purify polluted information fills an important gap in the current body of research. This study explores the performance and stability of several modern deep learning techniques for predicting chaotic systems using datasets polluted by invalid information. Our findings reveal that while most of the state-of-the-art deep learning methods exhibit reduced and unstable predictive performance owing to such contamination, the dynamical system deep learning (DSDL) method stands out, remaining unaffected by any interference. This breakthrough illustrates DSDL's unique ability to purify invalid data and uncover the inherent rules of chaotic systems. As we move forward, DSDL paves the way for a more reliable and interpretable model, ensuring that we can confidently predict chaotic systems with precision even in the most challenging environments.

10.1088/2632-2153/ad8983

Cite

Citations (0)

Interpretable predictions of chaotic dynamical systems using dynamical system deep learning

Scientific Reports (2024)

Mingyu Wang Jianping Li

Abstract Making accurate predictions of chaotic dynamical systems is an essential but challenging task with many practical applications in various disciplines. However, the current dynamical methods can only provide short-term precise predictions, while prevailing deep learning techniques with better performances always suffer from model complexity and interpretability. Here, we propose a new dynamic-based deep learning method, namely the dynamical system deep learning (DSDL), to achieve interpretable long-term precise predictions by the combination of nonlinear dynamics theory and deep learning methods. As validated by four chaotic dynamical systems with different complexities, the DSDL framework significantly outperforms other dynamical and deep learning methods. Furthermore, the DSDL also reduces the model complexity and realizes the model transparency to make it more interpretable. We firmly believe that the DSDL framework is a promising and effective method for comprehending and predicting chaotic dynamical systems.

Interpretability

Dynamical system (definition)

Nonlinear dynamical systems

10.1038/s41598-024-53169-y

Cite

Citations (1)

Identifying Software Theft Based on Classification of Multi-Attribute Features

Journal of Software (2014)

Ye Cai Guo Mingyu Wang Luo Yang-xia

Due to the low performance caused by the traditional embedded watermark and the shortages about low accuracy and weak anti-aggressive of single-attribute birthmark in checking obfuscated software theft, a software identification scheme is proposed which is based on classification of multi-dimensional features. After disassembly analysis and static analysis on protecting software and its resisting semantics-preserving transformations, the algorithm extracts features from many dimensions, which combines the statistic and semantic features to reflect the behavior characteristic of the software, analyzing and detecting theft based on similarities of software instead of traditional ways depending on a trusted third party or alone-similarity threshold. Through giving the formal description about the algorithm, depicting the algorithm realization, after comparisons and analysis from the qualitative and quantitative, theoretical and experimental aspects, the results show that the algorithm contributes to the resistance to attacks, as well as the robustness and credibility, and has advantages compared with similar methods.

10.4304/jsw.9.6.1401-1411

Cite

Citations (2)

Low-carbon Electricity Consumption Mechanism Considering the Interaction and Coordination of Source and Charge

Journal of Physics Conference Series (2023)

Hezhong Xu Mingfeng Xu Chengxuan Hui Zeming Lv Zhiping Liu

Abstract With the increasing shortage of traditional energy and environmental pollution, the “double carbon” goal was put forward. The construction of a new power system is an important way to achieve this goal, and how to maximize the consumption of new energy under the interaction and coordination of source and charge is the focus of this paper. Through discussing the new development trend of the present situation in the construction of electric power systems and the source charge interaction, the theoretical foundation is established to promote low-carbon electricity given the source of the collaborative optimization model. Then, the power grid company’s profit is maximized by considering the source charge interactive coordination. Finally, the process is optimized to establish the objective function and variables of the scope of the restrictions, so as to ensure the improvement and construction of the low-carbon electricity consumption mechanism.

10.1088/1742-6596/2427/1/012008

Cite

Citations (0)

Design and implementation of a modified high performance and low power CIC interpolation filter

Xiaopeng Liu Han Yan Guo Liang Mingyu Wang Lu Liao

This paper presents a modified cascaded integral comb (CIC) interpolation filter in order to improve filter characteristics and reduce power consumption at the same time. The modified CIC interpolation filter is a two-stage multiplier-less CIC-based interpolator. The first stage is a cascaded CIC filter whereas the second stage is a cascaded CIC filter and a second-order compensator. In an effort to reduce power consumption, the poly-phase decomposition and no-recursive algorithm is used when the modified filter is implemented. Simulation and synthesis results indicate that the stop-band attenuation is up to 137.7 dB and the pass-band drop is only 0.0003 dB with the filter interpolation factor 16. Working at 50 MHz clock frequency, the filter can reduce the power consumption of 16.78%. This new interpolator is implemented on Altera Cyclone III EP3C10E144C8 FPGA.

Voltage-controlled filter

Interpolation

All-pass filter

Butterworth filter

10.1109/edssc.2011.6117685

Cite

Citations (4)

Direction-of-Arrival Estimation of Non-equidistant Linear Array Based on Vectors

International Conference on Electric Information and Control Engineering (2012)

Mingyu Wang Youkui Zhang

Direction of arrival estimation of non-equidistant linear array in the sound field based on vectors is studied in this paper. For getting higher resolution of an array with only few sesors in the signal processing, sound pressure and velocity are collected, a similar uniform linear array of equally spaced linear array with two-dimensional vector sensors is used, MUSIC high-resolution method is used in received-signal processing. At last, the simulation results demonstrate that the method not only get higher resolution of a single target, but also has a basic resolutin of multiple targets.

Equidistant

Direction of arrival

Sensor array

Array Processing

SIGNAL (programming language)

10.1109/iceice.2012.158

Cite

Citations (0)

Class-Balancing Deep Active Learning with Auto-Feature Mixing and Minority Push-Pull Sampling

Lecture notes in computer science (2024)

Hongxin Lin Chu Zhang Mingyu Wang Bin Huang Jingjing Shao

Feature (linguistics)

10.1007/978-3-031-72390-2_2

Cite

Citations (0)