Shih-Jen Chu

Title: Shih-Jen Chu: Innovator in Model Training and Anomaly Detection

Introduction

Shih-Jen Chu is a prominent inventor based in Taipei, Taiwan. She has made significant contributions to the fields of model training and anomaly detection. With a total of 2 patents, her work is recognized for its innovative approaches and practical applications.

Latest Patents

Shih-Jen Chu's latest patents include a model training method and a model training system. The model training method involves a first iteration training where first training data and second training data are input to a target model. The method evaluates the image identification rates of the target model for both sub-tasks and adjusts the sampling rates accordingly. This innovative approach enhances the efficiency of model training.

Another notable patent is the training method for an anomaly detection model and an electronic device using the same. This method utilizes a generative model and a discriminative model to calculate network loss results based on training samples. If the original image does not match the task information, a first loss function is derived, allowing for effective training of the generative model.

Career Highlights

Shih-Jen Chu is currently associated with Pegatron Corporation, where she applies her expertise in model training and anomaly detection. Her work at Pegatron has contributed to advancements in technology and innovation within the company.

Collaborations

[This section has been skipped due to space constraints.]

Conclusion

Shih-Jen Chu is a remarkable inventor whose contributions to model training and anomaly detection are paving the way for future innovations. Her patents reflect her dedication to advancing technology and improving systems in her field.