Ji Eun Lim

Title: Ji Eun Lim: Inventor Advancing Adaptive Neural Network Quantization.

Introduction

Ji Eun Lim is an inventor based in Seoul, South Korea. His patent work addresses how deep learning neural network models can adapt their quantization settings when external conditions change the characteristics of input data.

Latest Patents

Lim is associated with 1 patent: “Method and Device for Quantizing Deep Learning Neural Network Model by Considering Change in External Environment.” The disclosed method detects changes in input-image features caused by shifts in the external environment. It then performs quantization calibration to determine a new parameter suited to those changes and updates at least one preset quantization parameter. This approach is intended to help a quantized neural network respond to changing input conditions.

Career Highlights

Lim’s listed patent work focuses on quantization for deep learning models, a technique that can reduce the computational demands of neural networks. The patent describes a process for adjusting quantization parameters in response to changes in the environment rather than relying only on fixed settings.

Collaborations

Lim is associated with Rebellions Inc. His listed coworker is Won Jae Lee.

Conclusion

Ji Eun Lim’s patent presents an adaptive approach to neural network quantization by recalibrating model parameters when external changes affect input data.