Last January 2019, Mariam Farda-Sarbas (FU Berlin), Cristina Sarasua (DDIS, UZH), Claudia Müller-Birn (FU Berlin) and Abraham Bernstein (DDIS, UZH), co-organized a workshop on data quality in Wikidata, hosted at Wikimedia Deutschland. Data growth is more visible than ever before and with this growth trustworthiness of data is becoming a challenge. Wikidata is the central source of structured knowledge for many projects and is developed to be used by anyone anywhere in the world, thus, ensuring data quality is, of utmost importance. Despite development of many methods and tools to improve different dimensions of quality in Wikidata, further efforts are needed to achieve high quality data. One effort in this regard was the workshop “Data Quality Management on Wikidata” which was held on January 18th 2019 at Wikimedia Germany. It brought together scholars interested and working on monitoring and improving data quality on Wikidata as well as members of the Wikidata community.

More details about the workshop can be found in this blogpost.

Coming May, another DDIS paper will be presented at the Web Conference 2019 in San Fransisco. The paper, titled ‘Iteratively Learning Embeddings and Rules for Knowledge Graph Reasoning’, is a collaboration between – among others – DDIS’s own Bibek Paudel and former visiting PhD Student Wen Zhang.

Research

2018 has been a turbulent and exciting year, and has provided many opportunities to do exciting research. More and more data is being produced, which calls for data mining & machine learning to make sense of this data. Technology is becoming more ubiquitous, and research on human computer interaction & computer supported cooperative work helps transition more pleasantly into this technological area. When data has semantic knowledge captured in it, such as in the semantic web, large graph processing allows to use this enriched data. Increasingly, data is not just static anymore; stream processing deals with the continuous stream of new data being generated.

Events

There were many events DDIS organised or helped, including the following:

Conferences

• CI 2018 & HCOMP 2018 were cohosted at the University of Zurich. 
• At Wikicon, Cristina Sarasua gave a keynote talk and participated in the panel discussion, and was interviewed by the Luzerner Zeitung for the occasion. 

Workshops

• Invitation based workshop on stream reasoning , co-organized by Daniel Dell’Aglio
• CrowdBias 2018 @ HCOMP: Cristina Sarasua co-organized this workshop and edited the proceedings
• Web Stream Processing workshop @ WWW: Daniele Dell’Aglio co-organized this workshop and edited the proceedings.  
• Linked Data Experience workshop at DINAcon, co-organized by Cristina Sarasua

Hackathons

• Wikidata Datathon
• Wikidata Hackathon
• Swiss digital day hackathon
• Twist Hackdays co-located with Swiss Statistics Day, where Cristina Sarasua was part of the organization team led by the statistical offices of the canton and the city of Zurich. 

Other events

• SRF theme night: Dataland – the future of Switzerland, featuring Prof. Abraham Bernstein Ph.D.
• Wikidata meetup

Employees and Students

DDIS consists of senior staff, regular PhD students and PhD students in a joined cooperation with the Swiss Federal Institute for Forest, Snow and Landscape Research (WSL). Over the last year, DDIS has decreased its size due to successful PhD students graduating: Shen Gao, Michael Feldman, Tobias Grubenmann and Bibek Paudel. However, the trend will be increasing again, after two new PhD students joined: Narges Ashena and Romana Pernischová. In fact, we are currently looking for a postdoc and PhD student so, do not hesitate to message us if you are interested or forward it to someone who is!

Various courses were supported or provided by DDIS:

• Informatik und Wirtschaft (German, Bachelor course)
• Big data analytics (English, Master course)
• Practical artificial intelligence (English, Master course)
• Social computing (English, Bachelor course)
• Big Data: Technology and Law (English, Master course)
• Semantic web engineering (English, Master course)

Besides courses, the group offers the possibility to pursue master and bachelor theses, as well as master projects (which are part of the master’s program at the faculty). 

On Tuesday the 11th of December 2018, Bibek Paudel successfully defended his thesis ‘Improving recommendation diversity and identifying cultural biases for personalized ranking in large networks‘. As is customary, the team created a graduation hat symbolizing his time at DDIS. We wish him good luck and good research at his next position as postdoctoral researcher at Stanford University!

Yesterday some of our (former) group members had their official graduation ceremony. Congratulations to them on the official graduation!

Our PhD graduates are doing well: Michael Feldman had joined another former graduate from DDIS, Patrick de Boer, to start their company PeakData. Tobias Grubenmann is currently a postdoctoral researcher at the University of Hong Kong. Bibek Paudel will soon start as a postdoctoral researcher at Stanford University.

Romana Pernischová is a fast track PhD student with our group, and has graduated her master top of her class. She will continue her PhD at DDIS for the coming years.

Two papers have been accepted at the conference of Web Search and Data Mining 2019 (WSDM2019). 

Our colleague Cristina Sarasua coauthored a paper together with Lei Han (first author), Kevin Roitero, Ujwal Gadiraju, Alessandro Checco, Eddy Maddalena and Gianluca Demartini called ‘All Those Wasted Hours: On Task Abandonment in Crowdsourcing’ (full paper can be found here). 

Bibek Paudel coauthored a paper together with Wen Zhang (first author), Wei Zhang, Abraham Bernstein and Huajun Chen titled ‘Interaction Embeddings for Prediction and Explanation in Knowledge Graphs’ (full paper can be found here). 

Daniele Dell’Aglio, postdoc at the Department of Informatics in the DDIS group, won the best poster award at the International Semantic Web Conference 2018, held in Monterey, California, from 8.10 to 12.10. The poster is a joint effort with researchers from Politecnico di Milano, Insight – NUI Galway, TU Berlin, and University of Applied Sciences and Arts Western Switzerland,

The poster presents VoCaLS: it is one of the first proposals of vocabularies to describe dynamic information on the web. By using VoCaLS, data publishers can describe properties of the streams, such as information about the stream frequency and the schema of its content, as well as information about the stream content provenance. The abstract of the poster is available at [1]; a more detailed description of VoCaLS is available in the resource paper published at the same conference [2].

Best Poster at ISWC

[1] https://www.zora.uzh.ch/id/eprint/156866/
[2] https://www.zora.uzh.ch/id/eprint/156931/

DDIS members gave two successful presentations at this year’s International Semantic Web Conference ISWC’2018.

First, DDIS PhD student Romana Pernischova gave her talk on the Stream Processing: The Matrix Revolutions at the 12th International Workshop on Scalable Semantic Web Knowledge Base System. The paper presents an approach that integrates matrices, tuples, and triples in stream processing. Such an approach is especially useful in settings, where various data sources need to be combined such as when processing videos, statistics, and structured data in the IPTV environment.

Romana Pernishova giving her Presentation

Second, Matthias Baumgartner presented his paper Aligning Knowledge Base and Document Embedding Models Using Regularized Multi-Task Learning at the main conference. The paper presents an approach to jointly learn two different embeddings (i.e., representations) that can be used to solve tasks both within each of the domains as well as when alining two domains.

Matthias Baumgartner giving his presentation

There is more DDIS @ ISWC to come, when Tobias Grubenmann presents his demo Make restaurants pay your server bills and Daniele Dell’Aglio’s co-authored papers get presented. But more on that tomorrow evening,

Congratulations to DDIS PhD student Bibek Paudel. He was nominated for the best review award at this year’s RecSys (Recommendation Systems) Conference (See https://twitter.com/alansaid/status/1047523315890380800) !

The twelfth ACM Recommender Systems Conference (RecSys) Sys is being held in Vancouver, Canada from October 2-7. According to ACM, RecSys 2018 is “the most important annual conference for the presentation and discussion of recommender systems research.” RecSys will bring together the main international research groups working in recommender systems, along with many of the world’s leading e-commerce and media companies.

This summer, Cristina Sarasua, from DDIS, wrote a piece for the Digital Society Initiative (DSI). In this article, she tells about the potential of the crowd, and how use this to the fullest (article in German).

Imagine using the wisdom of the crowd – such potential! [source: Unsplash]