I-Corps: Probabilistically Detecting Controversy

University Of Massachusetts Amherst, Amherst MA

Investigators

Abstract

The broader impact/commercial potential of this I-Corps project is in leveraging past research on detecting controversy and bringing this technology to a commercial setting. The project provides actionable insights to organizations interested in securities trading, through the usage of sophisticated machine learning and information retrieval techniques for controversial topics. The project will initially explore markets in the financial services sector that derive some portion of the value for their clients from acquired data streams. Besides the financial industry, there are applications of the technology in reputation and crisis management, as well as scope for social impact by opening analysis and explanatory power of controversies to individual users. The project has the potential to help users become better informed and more capable of critically evaluating the often-overwhelming stream of content. This I-Corps project technology relies on sophisticated machine learning and information retrieval techniques in order to automatically detect controversial topics. The enabling NSF-supported research recognized controversy by mapping a webpage to algorithmically-identified controversial topics. The proposed commercialization effort is unique because it addresses controversy and crisis situations. The project's hypothesis is that this technology can be successfully commercialized to the financial, equity-trading industry given their revenue model and cost structure. The project will use the lean startup method to perform customer interviews and explore the viability of a scalable business model.