Daigoro Yokozeki

Title: Daigoro Yokozeki: Inventor in Machine Learning Optimization

Introduction

Daigoro Yokozeki is an inventor based in Musashino, Japan, and is associated with NTT Corporation. His patent work concerns methods for optimizing machine learning models.

Latest Patents

Yokozeki is named on 1 patent, titled “Optimizing Machine Learning Models Using Binarized Parameter Matrices and Iterative Re-Randomization.” The invention describes a learning model optimization device that sets a binarization matrix whose elements are 0 or 1, then combines it element by element with a parameter matrix to create a transformed matrix. A learning unit uses that transformed matrix for machine learning and adjusts the binarization matrix so that the results approach teacher data. The device also includes a re-randomization processing unit that changes a parameter of the original parameter matrix again.

Career Highlights

Yokozeki’s listed patent reflects work on machine learning optimization, particularly the use of binarized matrices and iterative parameter changes. His professional association with NTT Corporation connects this work to a major telecommunications and technology company.

Collaborations

Daiki Chijiwa and Kenji Umakoshi are identified as Yokozeki’s coworkers. The available information does not specify their roles in the patent or whether they collaborated with him on this invention.

Conclusion

Daigoro Yokozeki’s patent focuses on optimizing machine learning through binarized parameter matrices, learning-based adjustment, and re-randomization.