5 problems

For a fixed finite collection of network inputs $x_\alpha\in\mathbb{R}^{n_0}$, $\alpha\in\mathcal{A}$, consider the finite-dimensional quantitative central limit theorem for the ve…

The paper considers average treatment effect estimation using deep neural networks, with regression functions having smoothness parameter $s$. Nearly parametric rate conjecture. Th…

Consider a random deep neural network with a target classifier drawn from the same random distribution as the network's initialization, and suppose that the adversarial robustness…

Consider deep neural networks trained on MNIST and CIFAR10 data, and the adversarial robustness guarantee proved for deep neural networks with random weights and biases: for any…

Let $\mathcal{P}_{\bm{t}}$ and $\bm{V}_s$ be as in Proposition. For a coordinate $k$, define the moment … Let $\Delta$, $\boldsymbol{\lambda}$, and the inverse Laplace operators…