Platform for a Neural Machine Translation System Demo and User Data Collection
Name
Dmytro Chasovskyi
Abstract
The central problem this work focuses on is that there are no open-source platforms for machine translation systems, which also allow collecting data from end-users.
The existence of such a system will greatly simplify the testing of hypotheses and the conduction experiments both by groups and individual researchers in the field of machine translation as well a receiving feedback from endusers.
The author creates a platform from scratch, validates and continually improves its performance based on feedback from users. The author also thinks of the architecture design of the platform based on the best software engineering practices.
As a result, there is now an open-source platform for machine translation systems, which also allows collecting data from users.
The existence of such a system will greatly simplify the testing of hypotheses and the conduction experiments both by groups and individual researchers in the field of machine translation as well a receiving feedback from endusers.
The author creates a platform from scratch, validates and continually improves its performance based on feedback from users. The author also thinks of the architecture design of the platform based on the best software engineering practices.
As a result, there is now an open-source platform for machine translation systems, which also allows collecting data from users.
Graduation Thesis language
English
Graduation Thesis type
Master - Software Engineering
Supervisor(s)
Mark Fishel
Defence year
2017