Our article "User Models for Culture-Aware Music Recommendation: Fusing Acoustic and Cultural Cues" together with Markus Schedl from JKU Linz, Austria has just been published by the Transactions of the International Society for Music Information Retrieval  journal. In this article, we propose a novel approach to jointly model users by their musical preferences and cultural backgrounds. We describe the musical preferences of users by the acoustic features of the songs the users have listened to and characterize the cultural background of users by culture-related socio-economic features that we infer from the user’s country. 

You can find the article here.

As in previous years, DBIS members are co-chairs of the PAN Style Change Detection task, co-located with CLEF 2020 (Conference and Labs of the Evaluation Forum).

The goal of the style change detection task is to detect changes in writing style to identify text positions within a given multi-author document at which the author switches. Particularly, given a document, we ask participants to answer the following two questions:

• Was the given document written by multiple authors? (task 1)
• For each pair of consecutive paragraphs in the given document: is there a style change between these paragraphs? (task 2)

More information about the task can be found on the task website. 

This week, the thirty-third Conference on Neural Information Processing Systems (NeuIPS) takes place. Together with Suzana Ilic and Asir Saeed, Eva has made a contribution to the Machine Learning for Creativity and Design Workshop, in which we present a generative adversarial network approach for the generation of creative contents such as poems, lyrics or metaphors: Creative GANs for generating poems, lyrics, and metaphors.

This week, Eva and Michael are at the 20th International Society for Music Information Retrieval Conference (ISMIR) in Delft, The Netherlands. At the conference, Eva will be a project guide at the WiMIR 2nd Annual Workshop, a satellite event of ISMIR 2019. She also served as co-chair for tutorials for the conference and together with Michael, she will present their paper "Hit Song Prediction: Leveraging Low- and High-level Features".

This week, Max is at MediaEval 2019 in Nice, France. MediaEval is a workshop consisting of various challenges in the area of multimedia information retrieval. Together with our partners from Academia Sinica and Taiwan AI Labs, DBIS co-authored two papers on the Emotion and Theme Recognition in Music Using Jamendo task. Max will present one of those papers (Recognizing Song Mood and Theme Using Convolutional Recurrent Neural Networks) on Tuesday.

Günther Specht is featured in the Standort magazine, where he discusses Building Information Management (BIM) and particularly, our freeBIM project.

This week, Benjamin and Eva are in Lugano for the CLEF conference (Conference and Labs of the Evaluation Forum, a central conference on evaluating information retrieval tasks) to present their works. Benjamin will present his full paper entitled "Generating Cross-Domain Text Corpora from Social Media Comments" at the main conference on Wednesday. Eva was task chair for the Style Change Detection task of the PAN@CLEF Challenge 2019 and will present the task and the participant's results as part of the PAN lab (Overview of the Style Change Detection Task at PAN 2019, Overview of PAN 2019: Bots and Gender Profiling, Celebrity Profiling, Cross-Domain Authorship Attribution and Style Change Detection). 

We are happy to announce that new topics for Bachelor and Master theses are available now. They cover the whole spectrum of our research. You can choose from music, natural language and database related topics.

Take a look at our theses page for further details.

Hooray - DBIS got three papers accepted at major conferences last week:

• Generating Cross-Domain Text Corpora from Social Media Comments (Benjamin Murauer, Günther Specht) got accepted at the Conference and Labs of the Evaluation Forum (CLEF 2019)
• Hit Song Prediction: Leveraging Low- and High-Level Audio Features (Eva Zangerle, Ramona Huber, Michael Vötter, Yi-Hsuan Yang) got accepted at the 20th Conference of the International Society for Music Information Retrieval (ISMIR 2019)
• Article Quality Classification on Wikipedia: Introducing Document Embeddings and Content Features (Eva Zangerle, Manuel Schmidt) got accepted at the 15th International Symposium on Open Collaboration (OpenSym 2019)

As a result of years of cooperation between the Austrian Standards International Working Group 011.09 "Technical Drawing and Documentation in Building Construction" and the research project freeBIM, which has been supported for years by the Economic Development Department of the state of Tyrol and the FFG, the ASI-PropertyServer is now Open Source and hence freely available to all interested parties, developers and companies. The already free and unlimited use of the ASI-PropertyServer for the Austrian BIM standard A 6241-2 has thus been expanded to include the option to install it locally, to supplement own special libraries and to cooperate in further development. It can be found at https://github.com/asi-propertyserver and is coordinated by the University of Innsbruck, Department of Databases and Information Systems (DBIS) under Univ. Prof. Dr. Günther Specht. 

The Economic Development Department of the state of Tyrol has recently confirmed that it will continue to support the further development of the project "freeBIM.connect" at the University of Innsbruck together with Tyrolean IT and industrial companies for the next two years.
 

We are proud to announce that Benjamin's proposal for a DOC scholarship by the Österreichische Akademie der Wissenschaften has been accepted! This ensures that he can finish his PhD while remaining to do research in and for the department. Starting from July, the scholarship will provide financial support for two years.

This week, Michael is presenting our StyleExplorer tool at the 41st European Conference on Information Retrieval (ECIR 2019) in Cologne, Germany. StyleExplorer was developed by Thibault Gerrier as part of his Bachelor thesis and provides an online toolkit to analyze, visualize and compare arbitrary text documents. During the course of the ECIR conference, many participants expressed their interests and are looking forward to using it for their own research problems.

In cooperation with the GoStudent platform and funded by the Austrian Research Promotion Agency (FFG), we just kicked off a new project. GoStudent provides an interactive learning platform that raises e-learning to a new level. Thousands of students are supported by a large community of validated tutors available over chat or via video sessions to answer student questions. Since the launch of GoStudent, answers were given to more than 500.000 questions. 

The goal of this project is to reuse this knowledge in self-learning Question Answering Systems (QAS) to instantly satisfy students' curiosity. The main problems we aim to tackle are the automatic recognition of similar questions and the extraction of possible answers given in a conversation. Via methodologies such as Deep- and Machine-Learning, the boundaries of a standard bot will be excelled to raise user satisfaction. 

The DBIS group has been busy at the BTW conference (Database systems for Business, Technology and Web) in Rostock in March 2019. Johannes Kessler and Günther Specht presented and demoed our RelaX tool. Furthermore, Günther Specht participated in the Workshop for "Digitale Lehre im Fach Datenbanken", where he also showcased the RelaX tool. RelaX provides an open-source calculator for relational algebra, particularly aiming at support students and teachers for databases courses. You can find out more about RelaX in the demo paper and the GitHub repo of RelaX.