Learn More
Deep neural nets have caused a revolution in many classification tasks. A related ongoing revolution—also theoretically not understood—concerns their ability to serve as generative models for complicated types of data such as images and texts. These models are trained using ideas like variational autoencoders and Generative Adversarial Networks. We take a(More)
In the last several years, provable guarantees for iterative optimization algorithms like gradient descent and expectation-maximization in non-convex settings have become a topic of intense research in the machine learning community. These works have shed light on the practical success of these algorithms in many unsupervised learning settings such as(More)
An open problem in complexity theory is to find the minimal degree of a polynomial representing the n-bit OR function modulo composite m. This problem is related to understanding the power of circuits with MODm gates where m is composite. The OR function is of particular interest because it is the simplest function not amenable to bounds from communication(More)
  • 1