Company Filing History:
Years Active: 2023-2025
Title: Thibaut Durand: Innovator in Machine Learning
Introduction
Thibaut Durand is a prominent inventor based in Vancouver, Canada. He has made significant contributions to the field of machine learning, particularly in developing systems that can handle partially-observed data. With a total of 4 patents, Durand is recognized for his innovative approaches to complex data modeling.
Latest Patents
One of Durand's latest patents is titled "System and method for machine learning architecture for partially-observed multimodal data." This invention addresses the challenges of learning Variational Autoencoders (VAEs) from partially-observed data. The proposed framework allows for effective learning conditioned on a fully-observed mask, enabling the model to learn a proper proposal distribution based on missing data. The framework has been evaluated for both high-dimensional multimodal data and low-dimensional tabular data.
Another notable patent is "System and method for a convolutional neural network for multi-label classification with partial annotations." This invention tackles the practical problem of training machine learning systems with incomplete labels. Durand's approach modifies loss functions on a proportionality basis and introduces a graph neural network to identify correlations between categories. Additionally, the prediction approach described in this patent aims to increase the proportion of labeled data relative to all labels.
Career Highlights
Thibaut Durand currently works at the Royal Bank of Canada, where he applies his expertise in machine learning to develop innovative solutions. His work has garnered attention for its practical applications in various industries.
Collaborations
Durand collaborates with talented individuals such as Gregory Mori and Nazanin Mehrasa, contributing to a dynamic and innovative work environment.
Conclusion
Thibaut Durand is a key figure in the field of machine learning, with a focus on addressing the challenges of partially-observed data. His patents reflect a commitment to advancing technology and improving machine learning systems.