Sergio Dorado

Title: Sergio Dorado and Recurrent Neural Network Innovation.

Introduction

Sergio Dorado is an inventor associated with Yorktown Heights, New York. His work includes a patented approach to recurrent neural network architectures inspired by discrete-time dynamical systems. He has been associated with International Business Machines Corporation and Rensselaer Polytechnic Institute.

Latest Patents

Dorado is credited with 1 patent. The patent, “Designing Ladder and Laguerre Orthogonal Recurrent Neural Network Architectures Inspired by Discrete-Time Dynamical Systems,” describes a method for making predictions with a recurrent neural network (RNN) modeling a discrete-time linear time-invariant system.

The RNN receives memory, external input, state, and hidden-output flows. It updates the state flow through a linear combination of the external input and memory flow, constrained by matrices corresponding to a Laguerre orthogonal basis. A coordinate matrix is then used to compute an updated memory flow. The system applies an activation function to update the hidden output and produces an output flow from a weighted sum of the hidden output, updated state, and updated memory flows.

Career Highlights

Dorado’s listed professional affiliations include International Business Machines Corporation and Rensselaer Polytechnic Institute. His patent reflects work at the intersection of neural network design and discrete-time dynamical systems.

Collaborations

Dorado’s listed coworker is Luigi Vanfretti, with whom he is associated in the context of his professional and technical work.

Conclusion

Sergio Dorado’s patented RNN approach uses Laguerre orthogonal structures and state and memory flows to support prediction in discrete-time systems.