Department of Computer Engineering
CONDITIONED TEXT GENERATION FOR QUERY SUGGESTION IN NLIDBs
Mousa Farshkar Azari
(Supervisor: Prof. Dr. Özgür Ulusoy)
Computer Engineering Department
With the advancement of deep learning in recent years, text generation has undergone significant changes and now provides a wide range of applications for users. Despite the fact that transformer-based models have performed well on a variety of NLP tasks in recent times, text generation remains an extraordinary situation. However, we know that the existing methods now can generate satisfactory results and for that reason, recent studies are trying to generate more humanlike outputs which is also known as conditional text generation which adds some specific purpose or emotion to the generated text. Therefore, Conditional text generation is now a hot research topic and more researchers are focusing on this field. The goal of this work is to come up with a new methodology and make with a novel solution and provide a first of a kind model for query suggestion in NLIDBs: we can theoretically get greater control of the style and generated content by choosing the right data set. On the other hand, NLIDBs have gained great attention in field of database in recent years trying to provide a more intelligent database system and it allows user to ask flexible and more complex queries from the database. This work is trying to provide a model for Conditioned text generation for query suggestion in NLIDBs.
DATE: 29 March 2021, Monday @ 16:00