Boobesh Rajendran

Title: Boobesh Rajendran: Inventor in Deep Learning

Introduction

Boobesh Rajendran is an inventor based in Navalur, India. His work includes a patent application concerning deep-learning models and the preservation of relationships between response variables and selected explanatory variables. He is associated with Accenture Global Solutions Limited.

Latest Patents

Rajendran is credited with 1 patent. The patent describes implementations for training a denoising stacked autoencoder (DAE) using a noisy training dataset containing both noisy and non-noisy subsets. It also describes providing an artificial neural network (ANN) with multiple hidden layers, including at least one layer that incorporates part of the DAE encoder and parameters determined during DAE training. The ANN is then trained on a training dataset, and a version is provided for inference.

Career Highlights

Rajendran’s listed invention focuses on applying a trained DAE encoder within an ANN. The approach is intended to support deep-learning models that preserve relationships between response variables and selected explanatory variables.

Collaborations

His listed coworkers include Georgios Passalis and Neha Chopra. The available information identifies them as coworkers but does not specify their roles in the patent.

Conclusion

Boobesh Rajendran’s patent work centers on combining denoising autoencoder training with neural-network inference to support deep-learning models that preserve important variable relationships.