education

Education


Round Corner
Department of Computer and Information Science

Oppgaveforslag

Towards more natural explanations of user preferences

In recent work, we introduced a method that can translate a user’s movie preferences to plain English; see the paper. Essentially, a user summary comprises of a set of templated sentences, such as:

  • You like movies that are tagged as ‘action’, especially those that are tagged as ‘violent’, such as Aliens.
  • You like movies that are tagged as ‘twist ending’, such as A Pure Formality.
  • You don’t like movies that are tagged as ‘adventure’, unless they are tagged as ‘thriller’, such as Twister.

The objective of this project is to make these user summaries sound more natural, e.g., say "filled with action" instead of "tagged as action'". Two main lines of possible approaches include (i) mining movie reviews for more colorful expressions and (ii) employing neural (deep learning) language generation techniques.

 

Faglærer

Krisztian Balog Krisztian Balog
Adjunct Professor
248 IT-bygget
 
NTNU logo