This tutorial–review on applications of artificial neural networks in photonics targets a broad
audience, ranging from optical research and engineering communities to computer science …

The number of 'small'molecules that may be of interest to chemical biologists—chemical
space—is enormous, but the fraction that have ever been made is tiny. Most strategies are …

Modern neural architectures for classification tasks are trained using the cross-entropy loss,
which is widely believed to be empirically superior to the square loss. In this work we …

Deep generative models provide a systematic way to learn nonlinear data distributions,
through a set of latent variables and a nonlinear" generator" function that maps latent points …

We compare classification and regression tasks in an overparameterized linear model with
Gaussian features. On the one hand, we show that with sufficient overparameterization all …

Characterizing the function spaces corresponding to neural networks can provide a way to
understand their properties. In this paper we discuss how the theory of reproducing kernel …

A common assumption in generative models is that the generator immerses the latent space
into a Euclidean ambient space. Instead, we consider the ambient space to be a …

Recent studies indicate that kernel machines can often perform similarly or better than deep
neural networks (DNNs) on small datasets. The interest in kernel machines has been …

Domain adaptation addresses the learning problem where the training data are sampled
from a source joint distribution (source domain), while the test data are sampled from a …

In this work, we address the problem of domain generalization for classification, where the
goal is to learn a classification model on a set of source domains and generalize it to a target …