- Việc làm
CVTạo CV online chuyên nghiệp, miễn phí
PortfolioTrình bày kỹ năng chuyên môn và dự án đã thực hiện một cách sinh động, trực quan.CVTạo CV online chuyên nghiệp, miễn phíTạo CVTạo CV online chuyên nghiệp, miễn phíMẫu CVTruy cập kho mẫu CV online ấn tượng, chuẩn quốc tế.CV mẫuTham khảo các CV mẫu đẹp đã được người dùng CakeResume ứng tuyển thành công.Cẩm nang nghề nghiệpTìm hiểu cách viết CV cho từng ngành nghề và vị trí.Trợ giúp viết CVTham khảo lời khuyên từ các chuyên gia tuyển dụng của CakeResume.- Kho tài nguyênCẩm nang ứng viênThế giới công sởVề đối tácCẩm nang doanh nghiệpVề CakeResumeBài viết nổi bật
- Tuyển dụng
- Tải app CakeResume
CakeResume Talent Search
Kế hoạch tuyển dụng đơn giản và hiệu quả nhất
Tìm kiếm 40k+ CV và chủ động liên hệ với ứng viên để đạt hiệu quả tuyển dụng cao hơn. Sự lựa chọn của hàng trăm công ty.
- Duyệt tất cả kết quả tìm kiếm
- Bắt đầu trò chuyện không giới hạn
- CV chỉ có thể truy cập bởi các công ty trả phí
- Xem địa chỉ email và số điện thoại của người dùng
Definition of Reputation Credits
Profile
I am Justin, who graduated from the department of Computer Science Engineering at National Sun Yat-sen University. I am friendly, optimistic, and willing to learn new knowledge.
As a software engineer, I am proficient in using Python, C/C++, and Java, and have an understanding of Git, which I have utilized for collaborative development projects with team members. Additionally, I have experience with Azure CI/CD, Docker, and Kubernetes (K8s), which has allowed me to proficiently manage and deploy applications to the cloud. These technologies has enabled me to streamline the software development process and enhance the overall quality of the projects.
I have served as the co-PI of a project under the Ministry of Science and Technology, honing my skills in coordination and teamwork. During my university studies, I also acted as a teaching assistant for courses in Artificial Intelligence, Algorithms, and Individual Study, helping instructors address students' inquiries.
My research focus is on neural network training algorithms to enhance the accuracy of deep learning models. I have proposed a novel optimization algorithm in my thesis that combines meta-heuristic algorithms and gradient-based optimization techniques, effectively improving the accuracy of deep learning models. The effectiveness of the proposed algorithm is demonstrated through experiments on various types of datasets and neural network models.
Work Experience
Engineer of MTIT, TSMC September 2021 - April 2022
#VB #ASP.NET #SQL #Azure
Develop and operate the full automation systems running in 200mm FABs.
Engage with FAB users to develop high value requirements and solutions to conquer the challenges about manufacturing.
Transform repeatable tasks into automation tools (CI/CD)
Skills
Software Engineer
- S.O.L.I.D
- Design Pattern
- MVC
Programming Language
- Python
- C/C++
- Java
Deep Learning
- Neural Network Optimization Algorithm
- Hyper-Parameter Tuning Algorithm
Optimization Algorithm
- Meta-heuristic Algorithm
- Gradient-based Algorithm
Publications
Thesis
An Effective Optimizer based on Global and Local Searched Experiences for Neural Network Training.
This thesis proposes a novel hybrid optimizer, GLAdam, which combines the benefits of meta-heuristic and gradient-based methods. GLAdam calculates the update direction by incorporating both global and local searched experiences, leading to an improved optimization process. The performance of GLAdam was evaluated through time series numerical forecasting and image classification experiments, demonstrating its effectiveness in training machine learning models.
Conference paper
ACM ICEA, “An Effective Optimizer based on Global and Local Searched Experiences for Short-term Electricity Consumption Forecasting”, Korea, 2020
This study presents a novel optimization algorithm, GLAdam, aimed at addressing the limitations of conventional gradient-based optimization methods. GLAdam incorporates a heuristic mechanism that leverages past search experiences, resulting in a more efficient exploration-exploitation trade-off during the optimization process. The results of experiments on time series numerical forecasting and image classification datasets show that GLAdam outperforms popular optimization algorithms such as Adagrad, RMSprop, and Adam, with an improvement in accuracy of 5.37% compared to the best performing algorithm.
ACM ICEA, “An Effective Multi-Swarm Algorithm for Optimizing Hyperparameters of DNN”, Korea, 2020
This study proposes an improved Multi-Swarm Particle Swarm Optimization (MSPSO) algorithm for optimizing hyperparameters of Deep Neural Networks (DNNs). The proposed algorithm outperforms traditional methods and was evaluated on Taipei passenger data, demonstrating improved accuracy in predicting the number of passengers for Taipei metro stations compared to other machine learning algorithms, DNN, and PSO with DNN.
Ministry of Science and Technology Program
A High-Efficiency Smart Grid Management System Combining Deep learning and Meta-heuristic Algorithms — 2020
- Using particle swarm optimization algorithm and search economic algorithm to improve the optimizer in deep learning to provide an accurate electric load forecasting model
- Using genetic algorithms to adaptively adjust the convolutional neural network structure and feature extraction of abnormal power consumption in smart grids
Towards Deep Learning for Next-Generation Automation: A Case Study of Intelligent Traffic Control Systems — 2021
- Using AutoML to predict traffic flow on plane roads and predict people flow in mass transit systems
- Using federated learning to control traffic lights at multiple intersections
- Road Travel Recommendation Using Reinforcement Learning
Profile
I am Justin, who graduated from the department of Computer Science Engineering at National Sun Yat-sen University. I am friendly, optimistic, and willing to learn new knowledge.
As a software engineer, I am proficient in using Python, C/C++, and Java, and have an understanding of Git, which I have utilized for collaborative development projects with team members. Additionally, I have experience with Azure CI/CD, Docker, and Kubernetes (K8s), which has allowed me to proficiently manage and deploy applications to the cloud. These technologies has enabled me to streamline the software development process and enhance the overall quality of the projects.
I have served as the co-PI of a project under the Ministry of Science and Technology, honing my skills in coordination and teamwork. During my university studies, I also acted as a teaching assistant for courses in Artificial Intelligence, Algorithms, and Individual Study, helping instructors address students' inquiries.
My research focus is on neural network training algorithms to enhance the accuracy of deep learning models. I have proposed a novel optimization algorithm in my thesis that combines meta-heuristic algorithms and gradient-based optimization techniques, effectively improving the accuracy of deep learning models. The effectiveness of the proposed algorithm is demonstrated through experiments on various types of datasets and neural network models.
Work Experience
Engineer of MTIT, TSMC September 2021 - April 2022
#VB #ASP.NET #SQL #Azure
Develop and operate the full automation systems running in 200mm FABs.
Engage with FAB users to develop high value requirements and solutions to conquer the challenges about manufacturing.
Transform repeatable tasks into automation tools (CI/CD)
Skills
Software Engineer
- S.O.L.I.D
- Design Pattern
- MVC
Programming Language
- Python
- C/C++
- Java
Deep Learning
- Neural Network Optimization Algorithm
- Hyper-Parameter Tuning Algorithm
Optimization Algorithm
- Meta-heuristic Algorithm
- Gradient-based Algorithm
Publications
Thesis
An Effective Optimizer based on Global and Local Searched Experiences for Neural Network Training.
This thesis proposes a novel hybrid optimizer, GLAdam, which combines the benefits of meta-heuristic and gradient-based methods. GLAdam calculates the update direction by incorporating both global and local searched experiences, leading to an improved optimization process. The performance of GLAdam was evaluated through time series numerical forecasting and image classification experiments, demonstrating its effectiveness in training machine learning models.
Conference paper
ACM ICEA, “An Effective Optimizer based on Global and Local Searched Experiences for Short-term Electricity Consumption Forecasting”, Korea, 2020
This study presents a novel optimization algorithm, GLAdam, aimed at addressing the limitations of conventional gradient-based optimization methods. GLAdam incorporates a heuristic mechanism that leverages past search experiences, resulting in a more efficient exploration-exploitation trade-off during the optimization process. The results of experiments on time series numerical forecasting and image classification datasets show that GLAdam outperforms popular optimization algorithms such as Adagrad, RMSprop, and Adam, with an improvement in accuracy of 5.37% compared to the best performing algorithm.
ACM ICEA, “An Effective Multi-Swarm Algorithm for Optimizing Hyperparameters of DNN”, Korea, 2020
This study proposes an improved Multi-Swarm Particle Swarm Optimization (MSPSO) algorithm for optimizing hyperparameters of Deep Neural Networks (DNNs). The proposed algorithm outperforms traditional methods and was evaluated on Taipei passenger data, demonstrating improved accuracy in predicting the number of passengers for Taipei metro stations compared to other machine learning algorithms, DNN, and PSO with DNN.
Ministry of Science and Technology Program
A High-Efficiency Smart Grid Management System Combining Deep learning and Meta-heuristic Algorithms — 2020
- Using particle swarm optimization algorithm and search economic algorithm to improve the optimizer in deep learning to provide an accurate electric load forecasting model
- Using genetic algorithms to adaptively adjust the convolutional neural network structure and feature extraction of abnormal power consumption in smart grids
Towards Deep Learning for Next-Generation Automation: A Case Study of Intelligent Traffic Control Systems — 2021
- Using AutoML to predict traffic flow on plane roads and predict people flow in mass transit systems
- Using federated learning to control traffic lights at multiple intersections
- Road Travel Recommendation Using Reinforcement Learning