Mila – Quebec AI Institute, originally known as the Montreal Institute for Learning Algorithms, is a research institute in Montreal, Quebec, dedicated primarily to Machine learning research. As of 2022, it comprised approximately 1000 students and researchers and 100 faculty members. Along with Alberta's Amii and Toronto's Vector Institute, Mila is a key component of the Pan-Canadian Artificial Intelligence Strategy.
The institute traces its origins to 1993, when Yoshua Bengio founded the Laboratoire d'informatique des systèmes adaptatifs (LISA) at the Université de Montréal. Bengio, a Turing Award winner and a pioneer of Deep learning, remains the scientific director. Mila also incorporates the Reasoning and Learning Lab (RL-Lab) at McGill University, established in 2001 and co-directed by Prakash Panangaden, Doina Precup, Joëlle Pineau, and Jackie Chi Kit Cheung. In 2017, Mila was formally established as a partnership between the Université de Montréal and McGill University, with the École Polytechnique de Montréal and HEC Montréal also joining.
History and Evolution
From its beginnings as LISA, Mila grew into a major hub for Artificial intelligence research. In December 2020, Mila partnered with IBM to accelerate AI and machine learning research using open-source technology. This collaboration integrated Mila's open-source software, Oríon, with IBM's Watson Machine Learning Accelerator, a tool for AI model training and inference. The partnership aimed to enhance research capabilities and foster innovation in the field.
Notable Researchers and Alumni
Beyond Bengio, Mila's faculty includes prominent researchers such as Joëlle Pineau and Doina Precup. The institute has trained many influential figures in AI. Notable alumni include Marc Bellemare (MSc 2007, RL-Lab), Hugo Larochelle (PhD 2009, LISA), and Ian Goodfellow (PhD 2014, LISA), each of whom has made significant contributions to Deep learning and related areas.
Research Activities
Research at Mila is largely focused on Deep learning and reinforcement learning, but also spans a wide range of topics. Key areas include generative models, Natural language processing, meta-learning, computer vision, and reinforcement learning. Mila also explores applications of AI to pressing global challenges, such as climate change and COVID-19 research. Additionally, the institute investigates advanced topics like Artificial general intelligence, superintelligence, AI alignment, AGI governance, and frontier models.
Governance and Board of Directors
Mila's strategic directions are overseen by a Board of Directors comprising leaders from academia, industry, and government. Notable members include Alexandre Le Bouthillier, General Partner at Linearis; Anne-Marie Hubert, Managing Partner at Ernst & Young; Benoit Boulet, Associate Dean at McGill University; and Marie-Josée Hébert, Vice Rector at Université de Montréal. Yoshua Bengio serves as an observer and scientific director, while Valerie Pisano is President and CEO. The board ensures that Mila maintains its research excellence and strategic alignment.
Educational and Open-Source Contributions
Mila serves as both a research and educational institution, offering postdoctoral and internship positions, and providing a research environment for graduate students pursuing PhD and Master's degrees in machine learning. The institute has also contributed significantly to open-source software. Theano, one of the early deep learning programming frameworks, originated at Mila. In 2020, active projects included Myia, a deep learning framework for Python, and Baby-AI, a platform for simulating language learning with a human in the loop. Mila also engages in numerous partnerships with private companies, further extending its impact on the AI ecosystem.