Richard Alvarez

Title: Richard Alvarez and Automated Machine-Learning Model Development.

Introduction

Richard Alvarez is an inventor based in Montpelier, Virginia, whose patent concerns tools for searching, designing, and developing machine-learning models. The disclosed approach aims to make model development more accessible to users who may not have specialized expertise in particular model types.

Latest Patents

Alvarez’s listed patent, “Heuristic Operational Network for Deep Learning Search and Machine Learning Model Search, Design and Development,” describes a process that first applies a gradient-boosting grid search to source data and a target. This step produces a reduced search space and prioritizes features according to their importance.

The system then uses a genetic algorithm to explore supervised classification problems across different model choices and feature-engineering methods. It can test models, hyperparameters, and network architectures, while generating features through combinations, scaling, normalization, and convolutions. Model development may use crossover, mutation, and selection, and promote features that meet a defined fitness threshold. Iterative testing can produce pretrained models and code for deploying production models.

Career Highlights

The provided information identifies Citicorp Credit Services, Inc. (USA) as Alvarez’s company affiliation. His listed patent focuses on automating parts of machine-learning model search and development.

Collaborations

The available information does not identify specific collaborators or research partnerships associated with Alvarez’s patent.

Conclusion

Richard Alvarez’s listed invention applies heuristic search, genetic algorithms, and feature engineering to streamline machine-learning model development. The record provided lists 1 patent associated with him.