Yuji Higuchi

Title: Yuji Higuchi: Inventor in Information Processing and Evaluation Methods.

Introduction

Yuji Higuchi is an inventor based in Kawasaki, Japan, and works at Fujitsu Corporation. His patent work includes methods for evaluating estimation-attack resistance and processing image data with machine-learning models. His listed portfolio includes 2 patents.

Latest Patents

Higuchi’s latest patents address evaluation of machine-learning systems and image-based information processing.

“Computer-readable recording medium storing evaluation program, evaluation method, and information processing device” describes a program for assessing a model’s resistance to estimation attacks. It compares attack results across models trained with different amounts of synthetic data, identifies specific data for further testing, and evaluates resistance using the resulting attack outcomes.

“Control method and information processing apparatus” concerns a classification model that calculates confidence scores from image data and a transformation model that converts lower-dimensional feature values into image data. The method updates a feature value so that the resulting image is associated with a higher confidence score, using the probability distribution of the feature values.

Career Highlights

At Fujitsu Corporation, Higuchi’s listed inventions focus on information-processing techniques involving synthetic data, model evaluation, image transformation, and confidence scoring. These subjects reflect practical challenges in developing and assessing machine-learning systems.

Collaborations

Higuchi’s listed coworkers include Toshiya Shimizu and Ikuya Morikawa.

Conclusion

Yuji Higuchi’s patent work at Fujitsu Corporation spans machine-learning evaluation and image-processing methods, contributing technical approaches to the assessment and operation of information-processing systems.