For details on current products, see my personal website.
How can we identify those who may potentially do something harmful? This questions is more relevant than ever but has largely been ignored by the research community. Recent developments in verbal deception detection are promising but are not applicable on a large scale. I am working on the development of a non-intrusive, real-time system that can be used as a tool for prospective airport passenger screening. One aim is to develop a chatbot-based information elicitation system that uses machine learning to inform us about the veracity of information given by prospective passengers.
The availability of data (crime data, texts, videos) in the open domain is greater than ever before. Computational methods allow us to grasp concepts that have previously been excluded from academic research. I am particularly interested in using machine learning for the classification of deceptive vs. truthful texts, in the development of methodological tools for more transparent research and data sharing, and in the data-driven analysis of crime (e.g. geographical modeling of burglaries and violent crimes).
How can we detect whether someone possesses critical information or not? Reaction time tasks like the Concealed Information Tes or the Implicit Association Test are promising tools to detect someone's memory trace. I am interested in how these tools can be improved technically (i.e. mobile use, an accurate web-browser timing of reaction times) and how we can use statistical modeling to reverse-engineer the tests towards real memory detection.
If you are interested in my research in the form of a thesis, dissertation or an internship, please write me an email.
--> There are several interesting internship possibilities for 2016/2017 (ranging from experimental lab research to machine learning implementation). Get in touch if you're interested.