MathWorks, a leading developer of technical computing software, has sponsored the '2023 Sejong AI Challenge' for the second consecutive year, encouraging university students to develop practical skills using MATLAB.

MathWorks announced on the 4th that it recently sponsored the '2023 Sejong AI Challenge', an artificial intelligence (AI) problem-solving ability competition hosted by Sejong University.

A total of 196 students participated in this competition and implemented ideas using MATLAB AI tools. 20 students, including Jaehoon Shim, a student from the Department of Intelligent Mechatronics Engineering who won the grand prize, stood on the podium.

MathWorks Korea has been an official sponsor of Sejong University's 'Sejong AI Challenge' since 2022, providing students with an environment where they can experience the company's software.

In addition, we are participating in Sejong University's ICT Innovation Talent 4.0 project and contributing to the development of the curriculum for fostering talents in unmanned vehicles.

This year's competition was held as a hackathon divided into a MATLAB track and other tracks. MATLAB is used by more than 100,000 companies of all sizes, from corporations, universities, government agencies, to startups, and is cited in more than 4 million studies. This competition provided students with an opportunity to build practical skills using MATLAB.

In this year's '2023 Sejong AI Challenge', a problem was presented of predicting the activity status of a patient wearing an IMU (Inertial Measurement Unit) on the body using GPUs on the MATLAB online server.

Students access MATLAB Online in a cloud environment to rapidly execute and scale computations in machine learning, image processing, and more, and MATLAB’s GPU support allows them to perform computations on GPUs without deep programming knowledge.

Jaehoon Shim, a student from the Department of Intelligent Mechanical Engineering who won the grand prize, used MATLAB's Deep Neural Network Designer app to select sensor value data, and then organized the layers into a convolutional neural network (CNN) structure to solve the problem.

Jaehoon Shim, a student in the Department of Functional Mechanical Engineering at Sejong University, said, “In this competition, I was able to deeply think about the entire process of analyzing problems on my own, preprocessing data, and tuning it to develop an optimized model.”

This competition was run by Professors Kim Seong-han and Choi Yu-kyung of the Department of Intelligent Mechanical Engineering, College of Software Convergence, Sejong University, and judged by Professors Lee Hyeon-seok, Kim Hyeong-seok, Kim Se-won (Department of Intelligent Mechanical Engineering), and Jeon Chang-jae (Department of Artificial Intelligence).

Kim Seong-han, a professor of unmanned vehicle engineering at the Department of Intelligent Mechanical Engineering at Sejong University, said, “In this competition, students were able to apply the MATLAB language to practical projects that can be found in various industries such as automobiles and robots.” He added, “MATLAB’s high convenience allows students to perform more advanced data analysis.”“It was a great help in producing good results through testing and verification,” he said.

Professor Kim Seong-han also said, “For this 2023 Sejong AI Challenge, we presented a problem of predicting a patient’s condition using six data sets,” and “Generally, three data sets are selected to utilize a convolutional neural network (CNN) through transfer learning, but Jaehoon Shim, who won the grand prize, created a new neural network and applied the existing six data sets to the prediction model, demonstrating excellent results with high accuracy.”

“MathWorks provides various learning materials such as examples and educational videos so that college students can easily utilize our software,” said Kim Kyeong-rok, educational sales manager at MathWorks Korea. “We hope that students will be able to build the skills necessary for practical work through MathWorks software not only in lectures but also in various opportunities such as competitions.”

Meanwhile, the scalability of MATLAB Online Server is especially useful for engineers and researchers who perform large-scale calculations but do not have the resources to maintain their own hardware. It is also easy to share scripts and results, making it easy to collaborate, and since it is hosted online, MATLAB work can be accessed from anywhere on any Internet-connected device. When used together with GPUs, it can improve overall productivity by offloading heavy calculations and allocating the CPU to other tasks, and it can use Tesla-based GPGPU (General-Purpose computing on Graphics Processing Units) even on PCs without external GPU cards.