Volodymyr Kuleshov

Learn More
Tensor factorization arises in many machine learning applications, such knowledge base modeling and parameter estimation in latent variable models. However, numerical methods for tensor factorization have not reached the level of maturity of matrix factorization methods. In this paper, we propose a new method for CP tensor factorization that uses random(More)
The rapid growth of sequencing technologies has greatly contributed to our understanding of human genetics. Yet, despite this growth, mainstream technologies have not been fully able to resolve the diploid nature of the human genome. Here we describe statistically aided, long-read haplotyping (SLRH), a rapid, accurate method that uses a statistical(More)
We introduce new algorithms for sparse principal component analysis (sPCA), a variation of PCA which aims to represent data in a sparse low-dimensional basis. Our algorithms possess a cubic rate of convergence and can compute principal components with k non-zero elements at a cost of O(nk + k 3) flops per iteration. We observe in numerical experiments that(More)
In user-facing applications, displaying calibrated confidence measures— probabilities that correspond to true frequency—can be as important as obtaining high accuracy. We are interested in calibration for structured prediction problems such as speech recognition, optical character recognition, and medical diagnosis. Structured prediction presents new(More)
We analyze the performance of single-parameter mechanisms for markets in which there is competition amongst both consumers and suppliers (namely, two-sided markets). Specifically, we examine the proportional allocation mechanism for two-sided markets. This mechanism is the natural generalization of both Kelly's proportional allocation mechanism for(More)
Simultaneous matrix diagonalization is used as a subroutine in many machine learning problems, including blind source separation and paramater estimation in latent variable models. Here, we extend algorithms for performing joint diagonalization to low-rank and asymmet-ric matrices, and we also provide extensions to the perturbation analysis of these(More)