now browsing by month
- Six MAST members attended RE19
- 1 full paper, 1 workshop paper, 2 Poster and Tool demos
- AffectRE workshop organization
- Best Tool and Demo award
Six members of the MAST team attended the 27th IEEE International Requirements Engineering Conference RE19 (Sept. 23rd – 27th) on Jeju Island, South Korea.
We presented 1 conference paper, 1 workshop paper, as well as 2 Poster and Tool demos.
- Extracting and Analyzing Context Information in User-Support Conversations on Twitter, Daniel Martens and Walid Maalej. [preprint]
- Classifying Multilingual User Feedback using Traditional Machine Learning and Deep Learning, Christoph Stanik, Marlo Haering, and Walid Maalej. [preprint]
- Requirements Intelligence with OpenReq Analytics, Christoph Stanik and Walid Maalej. [preprint] [poster] [video]
- OpenReq Issue Link Map: A Tool to Visualize Issue Links in Jira, Clara Marie Lüders, Mikko Raatikainen, Joaquim Motger, and Walid Maalej. [preprint] [poster] [video]
Davide Fucci and Walid Maalej from MAST, as well as Simone Kühn from the Max Planck Institute for Human Development organized the AffectRE workshop.
The topic of the workshop can be summarized as: As software systems are designed and used by human beings which can be characterized by emotions, the aim of Affective Computing is to study the development of software systems and devices that can recognize, interpret, process, and exploit human affects, feelings, emotion, attitudes and personalities.
The program included Jan Ole Johannsen from the Technische Universität München as the keynote speaker, as well as the presentations of the five accepted papers.
Best Tool and Demo Award
We are proud for receiving the best RE’19 Poster and Tool Demo award for our contribution: Requirements Intelligence with OpenReq Analytics.
The tool is being developed in the Horizon 2020 Project OpenReq. The tool is meant to support requirements engineering related tasks by analyzing user feedback from social media and app stores. Currently, it provides three distinct views – a dashboard that reports the general health of a software product or service, focus views for filtering and reading user feedback, and a competitor comparison. The tool is briefly introduced in the linked video. The linked poster further summarizes the contribution.
In collaboration with HITeC, OpenReq and Iteratec, we organized the 2nd Hamburg Requirements Engineering Symposium on September 4th, 2019. This event gathered 65 participants from industry and academia to discuss the state-of-the-art in RE, particularly applying Machine Learning and Artificial Intelligence to Requirements Engineering and vice versa. The event consisted of two keynote speakers, 14 talks in two parallel tracks, and Demos/Posters as well as networking opportunities through the coffee and lunch break. With such a high turnout (twice as many people as last year) and positive responses from the participants, this event was a large success.
The first keynote was by Dr. Prof. Vogelsang from TU Berlin, titled: “Is AI changing RE or is RE changing AI?” started the event off strong. Vogelsang gave the audience a big picture perspective on AI and RE, and how the two fields are interacting with and changing each other. Following Vogelsang, Dr. Stettinger from TU Graz gave a presentation of OpenReq Live, the main tool output of the OpenReq project. To close the event, Prof. Dr. Mobasher from DePaul University in the US gave an in depth keynote on the current trend of Context-Aware Recommender Systems.
The talks at this year’s symposium included talks from OpenReq members, OpenReq OpenCall partners, and members from the RE industry. Besides the keynotes, there were four sessions at this year’s symposium. The topics of the sessions were: User Feedback, Requirements & Developers, RE in Practice: Specs and Beyond, and Machine Learning for Issue Trackers.
User Feedback provided insights into how RE can leverage the magnitude of end-user feedback that currently exists regarding software (and hardware) products. Requirements & Developers discussed how to connect developers with requirements, whether that be through supporting and enhancing their workflow, or building better requirements documents to make their job easier. RE in Practice: Specs and Beyond gave a refreshing perspective on how RE is conducted in practice, current challenges, and how state-of-the-art research is being used to enhance the workflow of requirements engineers. And finally, Machine Learning for Issue Trackers took a look at industry standard issue trackers to manage requirements, and how innovative work within OpenReq is improving how developers and requirements engineers interact with them.
Following the main event, there was a 90 minutes boat trip through the Hamburg harbor. The weather held out long enough for the participants to have a thorough history lesson while enjoying the beautiful views offered by the Elbe river and Hamburg harbor.
Finally, the 2nd Hamburg Requirements Engineering Symposium came to a close with the participants connecting over good food and a relaxed atmosphere.
One of the most vibrant topics in requirements engineering research is the analysis of app reviews. The field is concerned with helping developers and managers to extract information from masses of unstructured and freely-written user reviews. The goal is to better understand user needs such as problems users face and features they wish would be integrated in the app. The Data Challenge was taking place on Monday and Tuesday of the OpenReq Week. We distributed six thousand app reviews labeled as problem report, feature request, and other. The participants were also given ten thousand English tweets containing the same labels, as well as issue tracker data to come up with ideas on how to perform an accurate classification. The winner of the Data Challenge, Tim Pietz, a computer-science student at the University of Hamburg, applied Google’s state-of-the-art deep learning model BERT to perform the classification task. He successfully outperformed approaches from research by up to 7% (f1 score).
The Eclipse Hackathon, in parallel with the Data Challenge, took place on Tuesday of the OpenReq Week. The objective of this Hackathon was to improve the Eclipse IDE with the help of OpenReq. Participants were asked to use the OpenReq Eclipse IDE Bug prioritizer plugin to find bugs to resolve. The plugin shows bug reports and feature requests from bugs.eclipse.org based on a personalized recommendations for the current user. It further shows the most discussed bugs in the last 30 days from bugs.eclipse.org to find bugs with a potentially high impact. The hackathon winner will be announced in a few weeks after we had been able to evaluation all contributions.
This year’s symposium was a huge success and we are very happy to have supported yet another year. If you have interest in the Hamburg Requirements Engineering community (or RE community in general), please reach out to our group to get in touch regarding RE research, industry contacts, the OpenReq project, or anything else you might have in mind.