Currently the group is active in the following research areas, which also represent our priorities:
- Conformal prediction: How good is your prediction? In risk-sensitive applications, it is crucial to be able to assess the quality of a prediction, however, traditional classification and regression models don't provide their users with any information regarding prediction trustworthiness. In contrast, conformal classification and regression models associate each of their multi-valued predictions with a measure of statistically valid confidence, and let their users specify a maximal threshold of the model's error rate — the price to be paid is that predictions made with a higher confidence cover a larger area of the possible output space. This tutorial, held at IJCNN 2015, aims to provide its attendees with the knowledge necessary to implement conformal prediction in their daily data science work, be it research or practice oriented, as well as highlight current research topics on the subject.
- Vector field semantics: To demonstrate and analyze dynamic processes underlying the evolving semantics of scalable datasets, we focus on reinterpreting vector space semantics in terms of a vector field. This is a broad and multidisciplinary research direction where cooperation partners can contribute from many angles, including anthropology, evolutionary computing, semiotics, general and computational linguistics, physics, computational science, and more. You can read about the idea here.
- Data landscaping: Because data is now a cornerstone of business epistemology and the root of organizational learning, in the narrower sense, this concept refers to a decision making tool that anyone can use on any data problem. Whether you are a student, recent graduate or already in business it is imperative to ‘know how you know’ things, on what basis decisions are made and how you will learn new things about the environment you are in. In a broader and different sense and from a data science perspective, we engineer 3D data landscapes with gradients for machine learning. Think of it with semantics as a force underlying knowledge tectonics in a dynamically changing geological environment, and you get the idea, published in JASIST, downloadable from here. For a metaphor, consider a formal model of plate tectonics.