In a new paper, David Rolnick and colleagues argue that problem-oriented AI research will help increase the long-term effectiveness of AI.
The AI Policy Compass program is designed to provide decision-makers, policymakers and professional working in policy with a foundational understanding of AI technology.
Studying at Mila means joining a dynamic community at one of the world's leading AI research institutes.
Mila is one of the world’s largest academic research institutes dedicated to deep learning. For years, its professors and students have been pushing the boundaries of artificial intelligence (AI) research across Canada and around the world.
As a leading AI research institute, Mila collaborates with Canadian universities, industry partners, international organizations and governments, putting AI at the service of society while also contributing to Canada’s vibrant AI ecosystem.
Drawing upon the skills and knowledge of more than 1,200 researchers, Mila is responsibly bridging the gap between state-of-the-art Al and human intelligence, particularly in relation to reasoning, robustness and other cognitive abilities still lacking in Al.
At Mila, students and world-renowned professors are pursuing groundbreaking work on both the foundational aspects of machine learning and many innovative algorithms. Each year, hundreds of Mila’s peer-reviewed scientific papers are presented at major AI conferences, including ICLR, ICML and NeurIPS.
Deep Learning
Reinforcement Learning
Causality
Machine Learning Theory
Natural Language Processing
Optimization
Out-of-Distribution Generalization
Representation Learning
Multimodal Learning
Mila published many of the early papers on deep learning, including the introduction of word embeddings in neural language models; denoising auto-encoders; deep nets with ReLUs instead of sigmoids or tanh; self-attention and the resulting machine translation and NLP revolution; GANs, etc. Mila is also responsible for a popular textbook on deep learning (MIT Press, 2016), and co-created the International Conference on Learning Representations (ICLR).
The institute has been a hothouse for the development of deep reinforcement learning, and the theoretical foundations of deep learning, including why it works, optimization methods and analysis, generalization principles, in- and out-of-distribution data processing, causality, and generative and probabilistic methods.
Mila has developed expertise at the intersection of theoretical neuroscience and deep learning (NeuroAI), as well as in many areas of applied machine learning that focus on AI for social good. Computer vision, NLP and robotics, and applications in health care such as medical imaging and drug discovery have all been front and centre. So, too, has the environment, through Mila's co-founding of the Climate Change AI organization. In addition, Mila has chosen to focus on AI4Science to support modelling and discovery in physics, chemistry and biology, and works actively within the social sciences — fighting sexual exploitation and modern slavery, while also contributing to the legal and philosophical aspects of responsible and safe AI.