Category Research
Privacy and 1d Histograms
Suppose there is an underlying 1 dimensional histogram stored in the cloud. As a concrete example, consider the distribution of bank deposits (x-axis is the amount of dollars, and the y-axis is the count of accounts). For simplicity let’s assume that all the amounts of money deposited are integers within the range . The histogram is queried by […]
Dense subgraph discovery applications
Recently, I compiled a collection of applications that rely on dense subgraph discovery for my KDD’15 tutorial with Aris Gionis. In general, dense subgraph discovery is a key graph mining primitive. While by “dense” we generally mean subgraphs which are large enough and contain many edges, the exact notion of dense is application dependent. The […]
Provably Fast Inference of Latent Features from Networks
Latent feature learning: where overlapping communities, correlation clustering, binary matrix factorization and extremal graph theory meet Motivation: Suppose we are given the following. Agents: Five agents, represented by . Binary vertex features. An agent can either be interested or not in each out of three news categories business, entertainment, sports. Each interest is represented by a […]
tsourakakis.com 