The tree sequence based translation model allows the violation of syntactic boundaries in a rule to capture non-syntactic phrases, where a tree sequence is a contiguous sequence o...
We perform Noun Phrase Bracketing by using a local, maximum entropy-based tagging model, which produces bracketing hypotheses. These hypotheses are subsequently fed into a reranki...
This paper describes a framework for defining domain specific Feature Functions in a user friendly form to be used in a Maximum Entropy Markov Model (MEMM) for the Named Entity Re...
We present a hierarchical phrase-based statistical machine translation in which a target sentence is efficiently generated in left-to-right order. The model is a class of synchron...
In this paper, we propose the use of the Maximum Entropy approach for the task of automatic image annotation. Given labeled training data, Maximum Entropy is a statistical techniqu...