Yujun Li

Title: Yujun Li and Neural Network Optimization

Introduction

Yujun Li is an inventor based in Shenzhen, China, whose patent work concerns artificial intelligence and neural network optimization. He is associated with Huawei Technologies Co., Limited. The supplied information lists 1 patent for him.

Latest Patents

Li’s listed patent is “Neural Network Optimization Method and Apparatus.” It describes sampling a preset hyperparameter search space to generate multiple combinations and iteratively evaluating their performance. When a combination meets a specified condition, the method reevaluates it to obtain an additional performance result.

The method then identifies an optimal hyperparameter combination. If that combination does not meet a second preset condition, the system updates a preset model using the evaluation results and samples again. If it meets the condition, the combination is used as a neural network’s hyperparameters. This approach is intended to guide the search toward effective settings through repeated evaluation and model updates.

Career Highlights

The patent connects Li’s work with the development of methods and apparatuses for optimizing neural networks. His listed company affiliation is Huawei Technologies Co., Limited, a technology company with research and development activities.

Collaborations

The supplied information names Yimin Huang and Zhenguo Li as Li’s coworkers. Their inclusion indicates professional connections, but no specific joint patent or project is identified in the provided details.

Conclusion

Yujun Li’s listed invention addresses neural network hyperparameter optimization through iterative evaluation, conditional reevaluation, and guided sampling. It reflects a practical focus on improving how neural network settings are selected.