Neural Network Implementation and its application

Prof. Kim started researches in the field of neural network implementation from his sabbatical year. For the lightweight neural networks, complex data format and operations are converted into simple data format and operations with negligible performance degradation.

He has interests in applying the approximate computing and stochastic computing to the neural network implementation. Firstly, he developed several iterative structures based on logarithmic multiplications. Secondly, he is studying the application of stochastic processing in the inference system. At this time, it is thought that there are many things to be researched in the training of neural networks with approximate computing and stochastic processing.

On the other hand, the application of neural networks is very interesting for him. He has published a domestic paper related to stock price estimation using neural networks. A patent to be registered is related to data augmentation for motion recognition using convolutional neural networks. Additionally, he extends his research the application of the reinforcement learning to EDA tool and circuit design optimization.

As his future research topic, he has interests in combining in-memory computing with the secrete of computer arithmetic.