Recently Added

Much real-world data is sampled at irregular intervals, but most time series models require regularly-sampled data. Continuous-time models address this problem, but until now only deterministic (ODE) models or linear-Gaussian models were efficiently trainable with millions of parameters. We construct a scalable algorithm for computing gradients of samples from stochastic differential equations (SDEs), and for gradient-based stochastic variational inference in function space, all with the use of adaptive black-box SDE solvers.

We investigate the approximability of the following optimization problem, whose input is an
n-by-n matrix A and an origin symmetric convex set C that is given by a membership oracle:
"Maximize the quadratic form as x ranges over C."

This is a rich and expressive family of optimization problems; for different choices of forms A
and convex bodies C it includes a diverse range of interesting combinatorial and continuous
optimization problems. To name some examples, max-cut, Grothendieck's inequality, the

A weighted graph H is a sparsifier of a graph G if H has much fewer edges than G and, in an appropriate technical sense, H "approximates" G. Sparsifiers are useful as compressed representations of graphs and to speed up certain graph algorithms. In a "cut sparsifier," the notion of approximation is that every cut is crossed by approximately the same number of edges in G as in H. In a "spectral sparsifier" a stronger, linear-algebraic, notion of approximation holds. Similar definitions can be given for hypergraphs.

Modern deep generative models like GANs, VAEs and invertible flows are showing amazing results on modeling high-dimensional distributions, especially for images. We will show how they can be used to solve inverse problems by generalizing compressed sensing beyond sparsity. We will present the general framework, new results and open problems in this space.