Abstract. This paper presents an empirical study on four techniques of language model adaptation, including a maximum a posteriori (MAP) method and three discriminative training mo...
We present an approximation to the Bayesian hierarchical PitmanYor process language model which maintains the power law distribution over word tokens, while not requiring a comput...
Abstract—Careful modeling of the radio channel characteristics is an important issue for system level simulations. While a number of detailed path loss, shadowing and fast fading...
Siegfried Klein, Serkan Uygungelen, Christian M. M...
We explore unsupervised approaches to relation extraction between two named entities; for instance, the semantic bornIn relation between a person and location entity. Concretely, ...
Limin Yao, Aria Haghighi, Sebastian Riedel, Andrew...
We present a novel fuzzy region-based hidden Markov model (frbHMM) for unsupervised partial-volume classification in brain magnetic resonance images (MRIs). The primary contributio...