Joao Gomes De Sa E Sousa

Title: Joao Gomes De Sa E Sousa: Innovator in Privacy-Preserving Neural Network Training

Introduction

Joao Gomes De Sa E Sousa is a notable inventor based in Renens, Switzerland. He has made significant contributions to the field of computer science, particularly in the area of privacy-preserving distributed training of neural network models. His innovative approach addresses critical challenges in data privacy and security.

Latest Patents

Joao holds a patent for a system and method for privacy-preserving distributed training of neural network models on distributed datasets. This computer-implemented method involves a distributed computer system that allows for the secure training of a global neural network model while maintaining the privacy of individual datasets. The system includes data providers that are communicatively coupled, each possessing a local training dataset and a vector of output labels. It utilizes a cryptographic distributed secret key and a corresponding collective cryptographic public key from a multiparty fully homomorphic encryption scheme. The weights of the global model are encrypted with this collective public key, ensuring that sensitive data remains protected throughout the training process.

Career Highlights

Joao is affiliated with the École Polytechnique Fédérale de Lausanne (EPFL), where he continues to advance research in neural networks and data privacy. His work is instrumental in developing methodologies that enhance the security of machine learning applications.

Collaborations

Joao collaborates with esteemed colleagues such as Sinem Sav and Juan Ramon Troncoso-Pastoriza, contributing to a dynamic research environment that fosters innovation and discovery.

Conclusion

Joao Gomes De Sa E Sousa exemplifies the spirit of innovation in the realm of computer science, particularly through his pioneering work in privacy-preserving technologies. His contributions are vital for the future of secure machine learning applications.