The collaborative powerhouse of AI: A glimpse into Mila, Montreal, with Laurent Tillement
Established in the 1990s by a group of professors including the renowned Yoshua Bengio, and Laurent Tillement, OSP was delighted to be invited to visit during a recent trip to Montreal. Tillement explained how Mila has grown rapidly over the last decade to become one of the world's largest AI research institutions. In a comprehensive discussion, he shared insights into the institute’s structure, objectives, and significant contributions to the AI landscape.
Mila’s primary mission revolves around cultivating new AI talent. Tillement said: “Our first main goal is to educate these new professionals,” highlighting the institute’s commitment to nurturing the next generation of AI experts. Most of Mila’s experts are affiliated with universities across Quebec, including the University of Montreal, and they bring with them approximately 1,300 students. This collaboration fosters a rich learning environment where students and professionals can interact and develop cutting-edge research in machine learning and deep learning.
Industry partnerships and real-world impact
Mila’s success is also deeply rooted in its extensive network of industry partnerships. Currently, Mila collaborates with over 130 companies, offering a dynamic exchange of ideas and resources. "We need them to tell us what is their problematic for the students and professors to work on real-world projects, real-world problems," Tillement added. These partnerships not only provide practical challenges for Mila’s researchers but also drive innovation in various industries.
A key aspect of these collaborations is the Entrepreneurship Lab, which supports students interested in launching their own startups. Over the past three years, this initiative has led to the creation of more than 30 companies, underscoring Mila’s role in fostering entrepreneurial spirit within the AI community. “We developed an entrepreneurship lab, where we take students with an idea... and we really explain what it is to develop the company.”
Research and development
Mila’s research is divided into two main streams: academic and applied research. The academic research is characterized by its open science approach, where findings are published and shared broadly to advance the field of AI. “The company needs to think when they work with these people because you permission algorithm when you publish the data set,” Tillement explained, emphasizing the transparency and collaborative nature of this work.
On the other hand, applied research involves more direct collaboration with companies, focusing on specific projects that require confidentiality and proprietary developments. "For the applied research team, it's totally different because they work with the company and only with the company," Tillement said. This flexibility allows companies to retain intellectual property rights while benefiting from Mila’s expertise.
AI for humanity
Mila is not just focused on technological advancement but also on leveraging AI for societal good. One of the most compelling aspects of Mila’s work is its AI for Humanity initiative, which tackles global challenges such as agriculture, health, and modern slavery. For instance, a project in Rwanda aims to predict agricultural outcomes to improve crop yields, while another project seeks to identify signs of modern slavery in corporate documents. “Some projects are really looking for interactions,” Tillement highlighted, indicating a strong emphasis on collaboration for social impact.
Training and ethical AI
Recognizing the importance of responsible AI, Mila has developed programs to ensure that AI technologies are developed ethically. The "Train" program, for example, provides certification in responsible AI, teaching developers how to create algorithms with ethical considerations in mind. "We really have this responsibility by an agency mind,” Tillement said, emphasizing Mila’s commitment to ethical AI practices.
Community and culture
The physical layout of Mila’s facilities is designed to foster interaction and collaboration. With common areas and a single coffee machine for everyone, the institute encourages spontaneous discussions and idea exchanges. “The keyword is interaction... you have a whiteboard with pens, so we can go and stop to think about the project,” Tillement described, illustrating the informal yet innovative culture at Mila.
Mila, through its strategic partnerships, educational initiatives, and groundbreaking research, continues to shape the future of AI. By fostering a collaborative environment and prioritizing ethical considerations, Mila not only advances technological frontiers but also ensures that AI serves the greater good of humanity. As the institute grows and evolves, it remains a beacon of innovation and a hub for the global AI community.
