Algorithms of summarization and retrieval of code fragments are frequently used by developers for re-use of code from online repositories. Summarization algorithms can be used to generate proper documentation and retrieval algorithms can be used to find correct code fragment by natural language description.
At the seminar we will see how variational autoencoders work and how they can be used in aforementioned tasks. We will compare this approach with others and see how this architecture can use unlabeled code data for better performance.
Speaker: Yaroslav Sokolov.
Presentation language: Russian.
Date and time: November 28th, 20:00-21:30.
Location: Times, 204.