Loading...

Automatic Language Model Adaptation

Sharif Razavian, Ali | 2010

492 Viewed
  1. Type of Document: M.Sc. Thesis
  2. Language: Farsi
  3. Document No: 40577 (19)
  4. University: Sharif University of Technology
  5. Department: Computer Engineering
  6. Advisor(s): Sameti, Hossein
  7. Abstract:
  8. Language Model plays a vital role in speech recognition systems. Restricting the search space, language models combine the result of words recognition phase with the assumptions gained from the rules and structure of words’ syntax (inflection and etc…), resulting into a more appropriate sequence of words as output. Since the language model for a specific language is not static and varies under the influence of different conditions, the problem of matching language model is being posed. In general, the aim of an adaptive language model is to offer a model capable of capturing all possible changes in the structure of words’ syntax, inflection and semantic structures. If the matching process occurs without human user intervention, it is called automatic language model adaptation. In this report, the aim is to provide a proper structure for practical automatic language model adaptation for Persian language. This means that in addition to accuracy and high speed in matching for Persian language, the model is supposed to use less memory and processing time compared to the other conventional models. In this report, while introducing the proposed method for automatically matching the language model, a suitable Data Structure and Clustering Algorithm for achieving this goal are proposed and discussed. To train the model, documents of Peykareh corpus in Persian language has been used and the model resulted in the increasing of the speech recognition system accuracy by 2.5 percent.

  9. Keywords:
  10. Language Modeling ; Statistical Language Modeling ; Adaptive Language Modeling ; Unsupervised Language Modeling

 Digital Object List

 Bookmark