In this work, we propose two extensions of standard word lexicons in statistical machine translation: A discriminative word lexicon that uses sentence-level source information to ...
We investigate language model (LM) adaptation in a meeting recognition application, where the LM is adapted based on recognition output from relevant prior meetings and partial ma...
Text classification is one of the most actual among the natural language processing problems. In this paper the application of word-based PPM (Prediction by Partial Matching) mode...
Abstract. In many applications of natural language processing (NLP) it is necessary to determine the likelihood of a given word combination. For example, a speech recognizer may ne...
In Chinese texts, words composed of single or multiple characters are not separated by spaces, unlike most western languages. Therefore Chinese word segmentation is considered an ...