Publications Repository - Gdańsk University of Technology

The paper presents the approach to using tags from Stack Overflow questions as a data source in the process of building domain-specific unsupervised term embeddings. Using a huge dataset of Stack Overflow posts, our solution employs the LSA algorithm to learn latent representations of information technology terms. The paper also presents the Teamy.ai system, currently developed by Scalac company, which serves as a platform that helps match IT project inquiries with potential candidates. The heart of the system is the information retrieval module that searches for the best-matching candidates according to the project requirements. In the paper, we used our pre-trained embeddings to enhance the search queries using the query expansion algorithm from the neural information retrieval domain. The proposed solution improves the precision of the retrieval compared to the basic variant without query expansion.

Authors

• mgr inż. Szymon Olewniczak link open in new tab ,
• dr hab. inż. Julian Szymański link open in new tab ,
• Piotr Malak,
• Robert Komar link open in new tab ,
• Agnieszka Letowska