A Search Engine for Arguments

Bielefeld University is leading new priority program in cooperation with universities from Bremen, Duisburg-Essen, Leipzig, and Weimar.

Discussions on the Internet are hard to analyze, whether it is controversy over the trade agreement TTIP (Transatlantic Treaty and Investment Partnership between Europe and the United States), or the debate on refugees. These conversations sometimes take place over the course of years, and millions of participants may have their say. A new priority program funded by the German Research Foundation is working to shed light on this problem: starting in 2017, researchers from across Germany will work on a search engine for arguments. These technical systems will be designed to analyze, for instance, how groups such as PEGIDA, the far right, anti-Islamic movement, or anti-globalization activists, argue in online discussions. Bielefeld University is leading the program “Ratio” in cooperation with the Jacobs University Bremen, the University of Duisburg-Essen, Leipzig University, and the Bauhaus University Weimar.

The full name of the German Research Foundation’s priority program is “Robust Argumentation Machines” (adaptive, scalable, and error-tolerant argumentation systems). Under the leadership of Professor Dr. Philipp Cimiano from the Cluster of Excellence Cognitive Interaction Technology (CITEC) at Bielefeld University, the coordinators of the program are currently working on a nationwide call for proposals in which researchers are invited to apply with their project ideas. From this search, up to 15 project proposals will be approved.

“We want to lay the foundation for a new technology that will evaluate digital texts in an automated way, recognize and summarize argumentative structures, and sum up the various pro and con arguments from a discussion at a glance,” explains Philipp Cimiano, who heads the Semantic Computing Group, which is part of the Cluster of Excellence CITEC. His team specializes in the research and development of methods that can extract and summarize the “meaning” from unstructured data, particularly digital texts.

Political researchers, for instance, could use this new generation of systems to analyze what people on the Internet think and say about a topic. In the current debate on refugees, the so-called “closing of the Balkan route” is put forth by many as the method of choice for dealing with the situation. The new systems would not just be able to track how widespread support of this “solution” is, but also identify the opposing arguments made against it. “In principle, these systems could also show which arguments are being repeated, and which new arguments are being introduced into the discussion,” explains Philipp Cimiano.

 For years, industry has been working on systems that automatically comb through and evaluate large volumes of data (“Big Data”). “The problem here is that it is still unclear where these systems are getting their information, or what their suggestions are based on,” says Philipp Cimiano. “Our systems should go beyond what previous projects have done before.” These systems will be designed to analyze unstructured documents, from which they will extract argumentative contexts and make them understandable. And they will not just give recom-mendations for action, but will provide reasoning for their suggestions, pointing out sources and further information. “For us, the goal is to bring transparency to Big Data analysis,” says the CITEC researcher. “We are working on intelligent advisory systems that ultimately allow users make their own decisions.”

Analyzing online discussions is only one possible area in which these argumentation machines could be used. The new systems are meant to help support experts from many different professional fields in making decisions, for instance in finance, medicine, technical documentation, politics, or sociology.

For example, a system could be developed that could be used to give advice to doctors: the system scans millions of medical articles on clinical presentation of disease and therapies, and the doctor can ask the digital advisor a question – thanks to a dialogue assistant – and learn about, for instance, a new therapy that the doctor might otherwise not have known about. This kind of system could also be useful in an industrial factory setting: if a machine is not working, the advisory system could “read” instructions, documentation, or even Internet forums, from which it could derive suggestions to solve the problem. Journalists could also benefit from such systems that take large amounts of information and evaluate it – whether freely available data from public administrations, or leaked documents like the “Panama Papers.”

The priority program brings together researchers from a range of sub-disciplines in computer science, including artificial intelligence, computer linguistics, knowledge representation, search engines, the semantic web, and human-computer interaction. “Ratio” will run from 2017 to 2023, and two million Euro will be made available each year for research funding, for a total of 12 million Euro for the entire duration of the project. The program committee is made up of the following members: Professor Dr. Philipp Cimiano (Bielefeld University), Professor Dr. Gerhard Heyer (Leipzig University), Professor Dr. Benno Stein (Bauhaus University Weimar), Professor Dr. Michael Kohlhase (Jacobs University Bremen), Professor Dr. Jürgen Ziegler (University of Duisburg-Essen).

The German Research Foundation’s priority programs promote basic scientific research, particularly in current or emerging areas of research. “Ratio” is one of the 17 new priority programs that will begin in 2017 and were chosen from among 76 initiatives submitted to the German Research Foundation.

Further information is available online at:
“German Research Foundation Establishes 17 New Priority Programs” (Press release from 21 March 2016): http://www.dfg.de/en/service/press/press_releases/2016/press_release_no_10/index.html

Prof. Dr. Philipp Cimiano, Bielefeld University
Faculty of Technology / Cluster of Excellence Cognitive Interaction Technology (CITEC)
Telephone: +49 521 106-12249
Email: cimiano@techfak.uni-bielefeld.de