Identification of transliterations is aimed at enriching multilingual lexicons and improving performance in various Natural Language Processing (NLP) applications including Cross ...
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...
Particle filters (PFs) are powerful samplingbased inference/learning algorithms for dynamic Bayesian networks (DBNs). They allow us to treat, in a principled way, any type of prob...
Arnaud Doucet, Nando de Freitas, Kevin P. Murphy, ...
We investigate in this paper the combination of DBN (Dynamic Bayesian Network) classifiers, either independent or coupled, for the recognition of degraded characters. The independ...
In the context of spoken language interpretation, this paper introduces a stochastic approach to infer and compose semantic structures. Semantic frame structures are directly deri...