The paper, entitled "Nearly tight sample complexity bounds for learning mixtures of Gaussians via sample compression schemes", was awarded a NeurIPS 2018 Best Paper Award, 1 of 4 awards among 4854 submissions. It will be presented at the NeurIPS 2018 conference in Montreal, a flagship machine learning conference.
The paper is authored by Hassan Ashtiani (McMaster), Shai Ben-David (Waterloo), Nicholas Harvey (UBC CS), Christopher Liaw (UBC CS), Abbas Mehrabian (McGill), and Yaniv Plan (UBC Math). The paper gives an optimal sampling bound for learning a mixture of Gaussians in a distribution learning setting.
Congratulations to Professor Nicholas Harvey and PhD student Christopher Liaw!