2k citations · 2.1k across the 4 of their papers we have counts for
3 papers
SlowMo: Improving Communication-Efficient Distributed SGD with Slow Momentum
Jianyu Wang, Vinayak Tantia, Nicolas Ballas +1
Distributed optimization is essential for training large models on large datasets. Multiple approaches have been proposed to reduce the communication overhead in distributed traini…
A dataset and exploration of models for understanding video data through fill-in-the-blank question-answering
Tegan Maharaj, Nicolas Ballas, Anna Rohrbach +2
While deep convolutional neural networks frequently approach or exceed human-level performance at benchmark tasks involving static images, extending this success to moving images i…
FitNets: Hints for Thin Deep Nets
Adriana Romero, Nicolas Ballas, Samira Ebrahimi Kahou +3
While depth tends to improve network performances, it also makes gradient-based training more difficult since deeper networks tend to be more non-linear. The recently proposed know…