Kenton Michael Lyons

Title: Kenton Michael Lyons: Inventor in Data Science and Preference Prediction.

Introduction

Kenton Michael Lyons is an inventor based in Los Altos, California. His patent work includes methods for predicting preferences and organizing data through machine-learning systems. He holds 3 patents.

Latest Patents

Lyons’s recent patent work includes “Systems and Methods Providing a ConjointNet Architecture for Enhanced Conjoint Analysis for Preference Prediction with Representation Learning.” The disclosed system uses an autoencoder to transform choices and features into a latent representation, then applies a choice classification network to predict preferences from that representation.

Another patent, “System for Labeling a Data Set,” describes a method for improving data labels. It generates related labels, measures how often labels occur, calculates breadth scores, and updates labels when specified conditions are met. The resulting labels are used to annotate the data set.

Career Highlights

Lyons’s professional affiliations include Toyota Research Institute, Inc. and Toyota Jidosha Kabushiki Kaisha. His inventions reflect work at the intersection of data analysis, machine learning, and computational methods.

Collaborations

Among Lyons’s coworkers are Yin-Ying Chen and Yanxia Zhang. Their shared professional context connects his work with broader research and development efforts.

Conclusion

Kenton Michael Lyons’s patents address practical approaches to preference prediction and data labeling, highlighting his contributions to machine-learning methods and applied data science.