This inventor holds 1 USPTO granted patent. Top assignee: Mcgill University. Active years: 2025.
Company Filing History:
Years Active: 2025
Title: Arash Ardakani: Innovator in Neural Network Training
Introduction
Arash Ardakani is a notable inventor based in Montreal, Canada. He has made significant contributions to the field of artificial intelligence, particularly in the training of neural networks. His innovative approach has led to the development of a unique method that enhances the efficiency of neural network training.
Latest Patents
Ardakani holds a patent for a "Method and system for training a neural network." This patent describes a computer-implemented method that involves a plurality of interconnected nodes arranged in layers, with associated weights. The method includes obtaining a training dataset from a database, performing forward propagation to produce an output, and executing backward propagation to update the weights. This process utilizes an estimation of polarity and neutrality of gradients, which is crucial for optimizing the neural network's performance.
Career Highlights
Arash Ardakani is affiliated with the Royal Institution for the Advancement of Learning/McGill University, where he contributes to research and development in artificial intelligence. His work focuses on improving neural network methodologies, which are essential for various applications in technology and data analysis.
Collaborations
Ardakani collaborates with Amir Ardakani, enhancing the research output and innovation in their projects. Their combined expertise fosters a productive environment for advancing neural network technologies.
Conclusion
Arash Ardakani's contributions to neural network training represent a significant advancement in artificial intelligence. His innovative methods and dedication to research continue to influence the field positively.