The broader impact/commercial potential of this Small Business Innovation Research (SBIR) Phase I project is to use controversy detection to support financial institutions' ability to reduce their risks and increase profits. This is part of a growing trend towards "alternative data" products relying on artificial intelligence and machine learning. Besides the financial industry, there are numerous potential applications of controversy detection technology in a variety of market verticals, such as crisis management, defense applications, and advertising technology. Beyond commercial applications, there is scope for social impact by opening analysis and explanatory power of controversies to individual users. Controversies have a massive impact on civic society and the so-called "filter bubble" exacerbates polarization, both political and otherwise; fake news on both sides of the political spectrum has recently captured public attention and generated political concern. Positive impact on society from commercializing this technology includes helping users become better informed and more capable of critically evaluating the often-overwhelming stream of online content. Proving the feasibility of this innovation in a highly quantifiable space such as finance could answer a customer need in that space and create new jobs for the economy, while enabling social good applications that can improve civic society.
This Small Business Innovation Research (SBIR) Phase I project relies on sophisticated machine learning and information retrieval techniques to automatically detect controversial topics. The initial data were collected at the University of Massachusetts Amherst, using NSF-supported research, which recognized controversy by mapping the text of a webpage to algorithmically-identified controversial topics. Research has demonstrated that controversy cannot be detected using existing methods of sentiment analysis, a widely-adopted natural language processing method. This project bridges the gap between the current capabilities of this nascent technology and the clear user need in the financial domain. It will evaluate the feasibility of controversy detection by applying a real-time controversy detection signal to financial data to reduce risk and increase returns.
This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.
