Feasibility of using YouTube Conversations for Pair Programming Intent ClassificationPoster
Pair programming conversational agents demand vast amounts of data for training. Recently a benchmark dataset of developer-developer and developer-agent pair programming conversations, from lab studies (8,324 utterances) was released for training natural language unit of a pair programming conversational agent. Unfortunately, this dataset is limited to a single domain and language. To investigate if it was feasible to utilize already available pair programming conversations from online video hosting platforms (i.e. YouTube), we collected five Youtube videos (roughly 350 minutes) with 4,822 utterances. We used transformer-based language model BERT to compare the lab studies with online videos. We found that a transfer learning approach, first training BERT on online videos and then fine-tuning with specific developer-agent data, resulted in the best performance.
Thu 15 SepDisplayed time zone: Amsterdam, Berlin, Bern, Rome, Stockholm, Vienna change
10:30 - 15:30 | Poster PresentationPosters and Showpieces / Graduate Consortium at Corridor of San Francesco Room Authors will attend only during breaks. | ||
10:30 5hPoster | Making the Invisible Visible in Computational NotebooksPoster Posters and Showpieces Mauricio Verano Merino Vrije Universiteit Amsterdam, L. Thomas van Binsbergen University of Amsterdam, Mazyar Seraj Eindhoven University of Technology DOI | ||
10:30 5hPoster | A technique to improve text editing on smartphonesPoster Posters and Showpieces Maria Giovanna Albanese Dipartimento di Informatica, Università di Salerno, Gennaro Costagliola Università di Salerno, Mattia De Rosa University of Salerno, Vittorio Fuccella University of Salerno DOI | ||
10:30 5hPoster | Chaldene: Towards Visual Programming Image Processing in Jupyter NotebooksPoster Posters and Showpieces Fei Chen , Philipp Slusallek German Research Center for Artificial Intelligence, Saarland University, Martin Muller Saarland University, Tim Dahmen German Research Center for Artificial Intelligence DOI | ||
10:30 5hPoster | Feasibility of using YouTube Conversations for Pair Programming Intent ClassificationPoster Posters and Showpieces Jacob Hart University of Tulsa, Jake AuBuchon University of Tulsa, Sandeep Kuttal The University of Tulsa DOI | ||
10:30 5hPoster | Evaluating Gender Bias in Pair Programming Conversations with an AgentPoster Posters and Showpieces Alex McAuliffe The University of Tulsa, Jacob Hart University of Tulsa, Sandeep Kuttal The University of Tulsa DOI | ||
10:30 5hPoster | Estimating Foraging Values and Costs in Stack OverflowPoster Posters and Showpieces Abim Sedhain The University of Tulsa, Sruti Srinivasa Ragavan Microsoft Research; School of EECS, Oregon State University, Brett McKinney The University of Tulsa, Sandeep Kuttal The University of Tulsa DOI | ||
10:30 5hPoster | Information Seeking Behavior for Bugs on GitHub: An Information Foraging PerspectivePoster Posters and Showpieces DOI | ||
10:30 5hPoster | Developers’ Foraging Behavior on Stack OverflowPoster Posters and Showpieces Vaishvi Diwanji The University of Tulsa, Abim Sedhain The University of Tulsa, Grey Bodi The University of Tulsa, Sandeep Kuttal The University of Tulsa DOI | ||
10:30 5hPoster | Which Technologies are Most Frequently Used by Data Scientists?Poster Posters and Showpieces Paula Pereira University of Minho, João Paulo Fernandes LIACC, Universidade do Porto, Porto, Portugal, Jácome Cunha University of Porto DOI | ||
10:30 5h | Tools for Creating UI Automation MacrosGC Poster Graduate Consortium Rebecca Krosnick University of Michigan DOI | ||