Gurumurthy Swaminathan

Title: Gurumurthy Swaminathan: Innovator in Machine Learning

Introduction

Gurumurthy Swaminathan is a prominent inventor based in Redmond, WA (US). He has made significant contributions to the field of machine learning, holding a total of 14 patents. His work focuses on enhancing the capabilities of machine learning models, particularly in the areas of class-incremental learning and neural network compression.

Latest Patents

One of his latest patents is titled "Class-incremental learning with pretrained machine learning models." This patent describes techniques for extending a pretrained machine learning model to recognize additional object classes without significantly reducing its accuracy for the original set of classes, addressing the issue known as catastrophic forgetting. Another notable patent is "Searching compression profiles for trained neural networks." This invention involves searching for compression profiles of trained neural networks through an iterative process, allowing for the generation of different compression profiles based on a search policy. The performance of these compressed versions is tracked, and the search policy is updated based on the evaluation of their performance.

Career Highlights

Gurumurthy Swaminathan is currently employed at Amazon Technologies, Inc., where he continues to innovate in the field of machine learning. His work has been instrumental in advancing the capabilities of artificial intelligence technologies.

Collaborations

He collaborates with notable colleagues, including Vineet Khare and Xiong Zhou, who contribute to his research and development efforts.

Conclusion

Gurumurthy Swaminathan is a key figure in the realm of machine learning, with a focus on improving model capabilities and efficiency. His contributions through patents and collaborations highlight his commitment to innovation in technology.