www.digitimes.com, Sept. 05, 2024 –
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 Montréal 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.