The 2019 Antidote Scholarship Awarded to Tristan Deleu

As part of the Druide Fund for Research in Text Analysis, created through a donation from Druide informatique, the MILA Laboratory has chosen the winner of the 2019 Antidote Scholarship. The lucky recipient of the $20,000 scholarship is Mr. Tristan Deleu, a Ph. D. student under the supervision of Professor Yoshua Bengio.

Mr. Deleu holds a Master's in Computer Vision and Machine Learning from l'École Normale Supérieure de Cachan (ENS Cachan) in France and has also worked on natural language processing in privacy-focused voice assistants. Mr. Deleu's doctoral research at MILA focuses on deep learning and reinforcement learning, with particular emphasis on what is known as "meta learning". Meta learning involves "learning to learn" so that algorithms can adapt to new tasks and new situations. Mr. Deleu has applied meta learning techniques to automatically infer cause and effect relationships.

In 2017, to mark the 20-year anniversary of its Antidote software, as well as 50 years of Université de Montréal’s Department of Computer Science and Operations Research (DIRO), Druide informatique announced the creation of a million-dollar fund for research into deep learning-based text analysis. In addition to establishing the Fund, Druide also set up the Antidote Scholarship, which awards $20,000 to one deserving student each year.

The Druide Fund for Research in Text Analysis is awarded to MILA, a laboratory led by Professor Yoshua Bengio, globally renowned for his spectacular breakthroughs in deep learning-based artificial intelligence. With 300 researchers from all over the world, MILA is a leader in its field. Drawing inspiration from Silicon Valley, Professor Bengio and his team strive to make Montreal into a hub for artificial intelligence, a goal that Druide informatique is proud to support through the Druide Fund and the Antidote Scholarship.

Antidote was already a remarkable mainstream application of artificial intelligence when it was launched in 1996. Its corrector, guides and dictionaries have since benefited from more than 20 years of continuous technological and linguistic development. Today, over a million people use Antidote in both English and French. Druide is methodically pursuing the development of Antidote and aims to be among the first to pass on the benefits of deep learning research to its users.