I think, as statistic researchers, we need to reach out to different disciplines. Statistics can solve problems.

Adding ‘statistical flavour’ to machine learning research

For Linglong Kong, machine learning and statistics are forever entwined — to build the best artificial intelligence models means understanding the data they are learning from. Kong’s statistics training is the foundation of his work in areas such as deep learning and distributed reinforcement learning, as well as using artificial intelligence to better analyze medical data and protect patient privacy. His work with quantile regression and robust statistics has him studying how to improve datasets to help optimize machine learning, such as reducing racial and gender bias in data used for social work.

Kong started his statistics training at Beijing Normal University, before getting his M.Sc. in Probability and Statistics from Peking University. He then moved to the University of Alberta to finish his PhD in Statistics. Much of his early work focused on biostatistic research with the U of A, and later, The University of North Carolina at Chapel Hill.

Kong returned to the University of Alberta to teach in the math and statistics department in 2012. In addition to his position as Fellow & Canada CIFAR AI Chair at Amii, he is a Canada Research Chair in Statistical Learning and is active as an associate editor for several journals, including the Canadian Journal of Statistics, the Journal of the American Statistical Association, Applications & Case Studies, and Frontiers in Neuroscience.

Statistics can provide insight into many things machine learning is trying to do; things like dealing with uncertainty or interpretability. Statistics has always tried to do this.