Loading...

A Persian Dialog System with Sequence to Sequence Learning

Ghafourian, Mohammad | 2018

942 Viewed
  1. Type of Document: M.Sc. Thesis
  2. Language: Farsi
  3. Document No: 50940 (19)
  4. University: Sharif University of Technology
  5. Department: Computer Engineering
  6. Advisor(s): Sameti, Hossein
  7. Abstract:
  8. Conversation modeling is one of the most important goals in the field of understanding natural language and machine intelligence. Recently, with the enormous growth of the Internet and social networks, the amount of available data on the Web has increased significantly.This makes it possible to use data-driven approaches to solve the modeling problem of conversation.One of the most recent data-driven methods is the sequence to sequence modeling. In this document, after providing the necessary prerequisites, we examined the various models that have used the sequence to sequence approach for conversation modeling. We further examined the ways of improving the efficiency of this modeling technique by introducing more advanced models such as attention. By performing several tests on several English and Persian databases, we measured the quality of implemented models and compared them.Using the sequence to sequence method and the attention mechanism, despite the desirable outcome in machine translation, there are problems in generating responses in a conversational system such as generating generic responses and the impossibility of using external data. For this reason, we introduced a new structure that combines the memory network approach and attention and with the help of word embeddings, the responses generated by the model are getting more accurate. Also, using the maximum mutual information as an objective function in optimization, we improved the ability of the model to generate more diverse responses. The results of the experiments showed that the model presented by us could have a higher likelihood in the production of test data. Also, responses are less generic
  9. Keywords:
  10. Recurrent Neural Networks ; Machine Learning ; Dialog System ; Sequence to Sequence Learning ; Attention Mechanism ; Memory Networks

 Digital Object List

 Bookmark

...see more