Identification of transliterations is aimed at enriching multilingual lexicons and improving performance in various Natural Language Processing (NLP) applications including Cross ...
Ranking is at the heart of many information retrieval applications. Unlike standard regression or classification in which we predict outputs independently, in ranking we are inter...
—In this paper, we present a novel approach for improving the performance of a large class of CPU and memory intensive passive network monitoring applications, such as intrusion ...
Numerous temporal inference tasks such as fault monitoring and anomaly detection exhibit a persistence property: for example, if something breaks, it stays broken until an interve...
A major difficulty in building Bayesian network models is the size of conditional probability tables, which grow exponentially in the number of parents. One way of dealing with th...