��˾��ֱ��

Richard Zemel has received from the – its highest honour – in recognition of his research contributions to machine learning.

A professor in the department of computer science in the University of Toronto’s Faculty of Arts & Science, Zemel helped build one of the world’s preeminent machine learning groups at ��˾��ֱ��. He was appointed the inaugural research director of in 2017.

Zemel developed a method known as an autoencoder, which trains a neural network to predict its own inputs. Autoencoders have led to arguably the most significant progress in machine learning – the ability to learn representations that can be used for multiple tasks.

Another important dimension of Zemel’s research is algorithmic fairness. His 2012 paper “Fairness through awareness” formalized desirable fairness properties of a classifier. His group followed up by showing how existing learning algorithms could be modified to eliminate discriminatory biases.

Zemel’s group has also published some of the foundational papers on few-shot learning, which considers how new concepts and classes can be learned given only a few labelled examples – and his work has shown how a neural network could be formulated to address this problem. He has also shown how neural networks could be applied to graphs, developing a message-passing approach that has been applied to many important areas, including drug discovery, health and social networks.