14 citations · 48 across the 7 of their papers we have counts for
Showing 2022Show all
2 papers · 1 filter
cs.LG2022★ 1 cited
A picture of the space of typical learnable tasks
Rahul Ramesh, Jialin Mao, Itay Griniasty +5
We develop information geometric techniques to understand the representations learned by deep networks when they are trained on different tasks using supervised, meta-, semi-superv…
cs.LG2022★ 8 cited
Scalable and Efficient Training of Large Convolutional Neural Networks with Differential Privacy
Zhiqi Bu, Jialin Mao, Shiyun Xu
Large convolutional neural networks (CNN) can be difficult to train in the differentially private (DP) regime, since the optimization algorithms require a computationally expensive…