Joan Puigcerver I Perez

Title: Innovations of Joan Puigcerver I Perez

Introduction

Joan Puigcerver I Perez is a prominent inventor based in Zurich, Switzerland. She has made significant contributions to the field of neural networks and computer vision. With a total of 2 patents, her work focuses on advancing the capabilities of artificial intelligence through innovative methodologies.

Latest Patents

Joan's latest patents include groundbreaking methods for training neural networks using transfer learning. One of her patents details systems and apparatus for training neural networks to perform a first prediction task. This involves obtaining trained model parameters for various candidate neural networks, which have been pre-trained for different tasks. The process includes selecting a proper subset of candidate networks and generating fine-tuned neural networks to enhance performance. Another patent focuses on transfer learning between different computer vision tasks, where an initial neural network is pre-trained to share layers with a subsequent network for improved task performance.

Career Highlights

Joan is currently employed at Google Inc., where she continues to push the boundaries of technology. Her work has been instrumental in developing advanced AI systems that leverage transfer learning techniques.

Collaborations

Joan has collaborated with notable colleagues, including Neil Matthew Tinmouth Houlsby and Basil Mustafa. These partnerships have fostered an environment of innovation and creativity in her projects.

Conclusion

Joan Puigcerver I Perez is a trailblazer in the field of neural networks and computer vision. Her innovative patents and contributions to artificial intelligence continue to shape the future of technology.