GIST Develops Technology to Accelerate Semiconductor Device Simulation Using Artificial Neural Networks

The research team led by Professor Hong Seong-min of the Department of Electrical, Electronics and Computer Engineering at Gwangju Institute of Science and Technology (GIST) has developed a technology that can perform semiconductor device simulations much faster through artificial neural networks, in collaboration with Professor Choi Jong-hyun from the AI Graduate School. The photo shows, from left to right, Professor Hong Seong-min, Professor Choi Jong-hyun, and master's student Han Seung-cheol.

[Asia Economy Honam Reporting Headquarters Reporter Park Seon-gang] The research team led by Professor Hong Seong-min of the Department of Electrical Engineering and Computer Science at Gwangju Institute of Science and Technology (GIST) announced on the 24th that, together with Professor Choi Jong-hyun from the AI Graduate School, they have developed a technology that can perform semiconductor device simulations much faster through artificial neural networks.

The research team focused on the fact that most of the semiconductor device simulation time is spent calculating unnecessary intermediate process answers, and succeeded in reducing the simulation time by nearly 10 times by generating excellent approximate solutions using trained artificial neural networks.

Recently, as a global semiconductor shortage has emerged and semiconductor manufacturing technology has attracted significant attention, semiconductor device technology is especially important to complete development in a short time, thus expectations for semiconductor device simulation are high.

However, semiconductor device simulation programs typically require a lot of time to run, which itself becomes a bottleneck in technology development. Although techniques such as parallel computing are used to address this, handling numerous device design candidates requires enormous computing resources, which is a drawback.

The research team shortened the simulation time by directly obtaining answers only for the voltage conditions the user wants to know.

Since semiconductor device simulation involves solving nonlinear equations, it is necessary to have an excellent approximate solution close to the correct answer.

However, it is difficult to know in advance an excellent approximate solution for the voltage condition the user wants to know (around approximately 0.7V), so it is inevitably started from 0V and the voltage is gradually increased.

The research team introduced artificial neural networks to directly obtain answers for the desired voltage conditions.

This artificial neural network performs supervised learning on existing simulation results and generates the potential distribution inside the semiconductor device corresponding to the desired situation. Using this predicted potential distribution as an approximate solution allows finding the correct answer in a short time.

To verify the proposed method, a speed comparison with the existing method was conducted.

Compared to the result of setting the simulation control parameters of the existing method to optimal values, a speed improvement of more than 8.4 times was achieved. Since the optimal values of the simulation control parameters cannot be known before performing the simulation directly, the expected speed improvement in actual application is more than 10 times.

Professor Hong Seong-min said, “The significance of this research achievement lies in confirming for the first time that the execution time of semiconductor device simulations can be greatly reduced by utilizing artificial neural networks,” and added, “It is expected to be actively used in the development of next-generation semiconductor devices through follow-up research.”

This research, led by Professor Hong Seong-min of the Department of Electrical Engineering and Computer Science at GIST, with participation from master’s student Han Seung-cheol and Professor Choi Jong-hyun of the AI Graduate School, was conducted with support from the Individual Basic Research Project (Mid-career), the Materials Innovation Leading Project, and the Institute for Information & Communications Technology Planning & Evaluation (IITP). The research results were published online on the 7th in the world-renowned semiconductor device journal ‘IEEE Transactions on Electron Devices.’