The Cologne Institute for Information Systems (CIIS) of the WiSo Faculty of the University of Cologne, headed by Prof. Dr Detlef Schoder, and the Bonn-based Big Data pioneer Stockpulse GmbH will jointly launch a comprehensive research project on the automatic detection of false news related to the financial market in October 2021. The project aims to develop intelligent software that automatically and practically in real-time detects, categorises and reports so-called "fake news", which are primarily generated and spread by "malicious" social bots or fake accounts, to appropriate bodies such as supervisory and regulatory authorities, internal monitoring units of stock exchange operators as well as to banks and financial service providers.
Within the framework of the German government's National Strategy for Artificial Intelligence (AI) and the High-Tech Strategy 2025, the joint project is funded to a large extent by the German Federal Ministry of Education and Research (BMBF). The collaborative project will run for 30 months and operates under the official title "AFFIN: Automated and Digital Forensics of Publicly Available Data in Finance."
"We are pleased to be able to contribute to a highly relevant societal problem with our expertise in building AI-based, highly innovative and practical information systems," says Prof. Dr Schoder, managing director of CIIS. "To reliably filter manipulative posts in social media, we make use of the recent significant advances in the field of machine learning, especially in Natural Language Processing methods - NLP for short," the WiSo professor further explains.
Professor Schoder's research focuses on digital transformation, NLP-based analysis of data, data-centric business model innovation and applications of AI for decision support in air freight logistics and financial markets. The research of his AI working group "Machine Learning & Natural Language Processing" is based on further developments in the field of Transfer Learning, which in turn is partly based on Deep Learning - a special method of machine learning.
"Stockpulse has specialised in collecting, structuring and analysing news and opinions from social media with a clear financial market connection for more than ten years. This is a unique history for the relatively young format of social media," says Dr Stefan Nann, Managing Director of Stockpulse GmbH. "Especially since 2014, we have noticed a disproportionate increase in dubious posts and obvious fake news aimed at manipulating share prices." Stockpulse has already been working for the trading watchdogs of the world's largest stock exchange operators since 2019, giving it a unique wealth of experience in the field of automated fake news detection; Stockpulse also has an extensive database that can provide extremely valuable services for the collaborative project to detect patterns in fake news. The new research project will massively improve the detection of fake news and, in particular, the real-time forwarding to relevant authorities or monitoring agencies. In addition, the system to be developed - slightly adapted - can also be used for other, socially relevant areas to be able to recognise the fake news spread there as automatically as possible.