Memorization of training data is an active research area, yet our
unders...
Overparameterized neural networks (NNs) are observed to generalize well ...
Reconstructing samples from the training set of trained neural networks ...
Despite a great deal of research, it is still not well-understood why tr...
Understanding to what extent neural networks memorize training data is a...
Despite a great deal of research, it is still unclear why neural network...
We solve an open question from Lu et al. (2017), by showing that any tar...
We study the memorization power of feedforward ReLU neural networks. We ...
We theoretically study the fundamental problem of learning a single neur...
Several recent works have shown separation results between deep neural
n...
Graph neural networks (GNNs) can process graphs of different sizes but t...
We study the effects of mild over-parameterization on the optimization
l...
The lottery ticket hypothesis (Frankle and Carbin, 2018), states that a
...
We consider the fundamental problem of learning a single neuron x
σ(w^ x...
Recently, a spate of papers have provided positive theoretical results f...