Chip complexity is growing beyond the capabilities of human intelligence alone.
This is why semiconductor evolution requires electronic design automation (EDA) with AI. Mila, the Montréal Institute for Learning Algorithms, is working closely with industry to develop these IC EDA as it believes breakthroughs will require both AI and domain expertise.
Reinforcement learning can help explore solution spaces when tool parameters and processes can be optimized by analyzing past component failures via Deep Learning. Other use cases include modeling physics environments to reduce compute overhead through Foundation Models, AI-guided Inverse Design to accelerate the discovery of new precursor materials, chip energy efficiency improved with Distributed Learning, and Computer Vision for rapid wafer quality inspection.
Mila, the world's largest academic research center in deep learning, is at the forefront of developing these methods. Founded by Professor Yoshua Bengio of the Université de Montreal in Québec, the research institute in artificial intelligence brings together over 1,200 specialized researchers in machine learning. It is globally recognized for its significant contributions, especially in the fields of language modeling, automatic translation, object recognition, and generative models.
Québec's great AI hub is made up of +18,000 university students in programs related to AI and data analytics. The province is home to the headquarters of Scale AI, the AI supercluster dedicated to Canada's supply chain. This prominent AI ecosystem feeds from surrounding strategic industries and vice versa. This includes life sciences, quantum sciences, optic-photonics, and of course microelectronics among others.
Québec is an immense collaborative space with intense synergy. All levels of government with industry and academia collaborate very tightly for opportunities to foster. Technum is Québec's innovation zone for microelectronics in Bromont. It is home to some of the largest companies – ABB, IBM, Teledyne – as well as innovation collaborative centers such as C2Mi and CMC, providing access to state-of-the-art technology for making micro-nanotechnologies, and for training industry-ready graduates.
On the AI side, Mila plays a central role in the Pan-Canadian Artificial Intelligence Strategy–the first national AI strategy in the world– for which a 2-billion-dollar investment was announced in the April 2024 federal budget. Thanks to applied research and its large talent pool, Mila is a real catalyst for development for industry partners at home and abroad. Just in 2023 and in collaboration with Intel, they published 12 material discovery papers on automating AI-driven discovery of novel materials, applying causal machine learning for climate science, and accelerating the study of molecular drivers of diseases and drug discovery.
AI's already crucial integration with semiconductor manufacturing will grow in every step of the supply chain as shrinking components introduce novel challenges. The length of time from design to manufacture will shrink as AI produces faster designs, fewer manufacturing errors, and quicker recalibrations.
Radical innovations will emerge as a result of growing compute demand, a faster pace of chip design, and new AI-driven material discoveries. This innovation can already be seen in Québec, in which semiconductor and AI ecosystems expand and synchronize.
The Mila is cultivating the next generation of Canadian compute by focusing on three areas relevant to the semiconductor revolution: AI for material discovery through experimental new architecture, tighter co-development between fundamental research and industry for AI-specific chips and firmware optimization and finally, integration of AI in IC EDA. The technological advancements meet the specifics of the latest generation chips and semiconductors.
Invest Québec, the financial arm and economic development agency for the Québec government collaborates closely with foreign companies to kickstart their expansion in the American Northeast corridor where a fully integrated value chain for semiconductors is building itself from the ground up.
To learn more about what Québec's ecosystem can bring to your business contact Rachel Yin at https://www.linkedin.com/in/rachel-yin/, Info@Invest-Quebec.com
