We will show that the friendship paradox exists on Twitter and applies to other user attributes. The paradox also introduces information overload.
The talk focuses on the load balancing problem in stream processing engines that exists due to the the presence of data skewness.
We discuss efficient algorithms for finding approximate frequent subgraph patterns from large graphs.
We describe the Causality research at Courant, NYU, its historical roots in logic and probability, and applications to biology and social sciences.
The talk will describe SentiStrength for fast unsupervised sentiment strength detection in short informal text with human-level accuracy in many cases.