Regina Deak-Meszlenyi

Title: Regina Deak-Meszlenyi: Innovator in Neural Network Architectures

Introduction

Regina Deak-Meszlenyi is a prominent inventor based in Budapest, Hungary. She has made significant contributions to the field of neural networks, particularly in image processing tasks. With a total of 2 patents, her work showcases her innovative approach to technology and machine learning.

Latest Patents

One of her latest patents is titled "Method for determining the encoder architecture of a neural network." This method involves calculating characteristic scale distributions based on training data for various image processing tasks. It generates encoder architecture candidates that include a shared encoder layer and branches for different computational operations. The assessment measures are calculated and compared to select the most suitable encoder architecture.

Another notable patent is "Systems and methods for deep multi-task learning for embedded machine vision applications." This computer-implemented method involves receiving data from vehicle sensors and performing multiple prediction tasks simultaneously using a multi-task neural network. The network is designed with shared parameters to enhance efficiency across different tasks.

Career Highlights

Regina has worked with esteemed organizations such as Nanyang Technological University and Continental Automotive Technologies GmbH. Her experience in these institutions has allowed her to refine her skills and contribute to groundbreaking research in neural networks.

Collaborations

Some of her notable coworkers include Csaba Nemes and Sinno Jialin Pan. Their collaboration has likely enriched her research and development efforts in the field.

Conclusion

Regina Deak-Meszlenyi is a trailblazer in the realm of neural networks, with her innovative patents paving the way for advancements in image processing and multi-task learning. Her contributions continue to influence the technology landscape significantly.