Herna Viktor

Herna Viktor is a Full Professor at the School of Electrical Engineering and Computer Science (EECS) at the University of Ottawa (uOttawa) and the Director of Applied Artificial Intelligence at uOttawa. Her research focuses on data-driven discovery, with an emphasis on advanced machine learning algorithms to extract deep semantic meaning from evolving streams. She is the author of more than 150 journal articles, conference papers and book chapters. The end results of her research have been applied across numerous and diverse domains, including the study of anaemia paediatric patients, in collaboration with the Hospital for Sick Children in Toronto; exploring the evolving media discourse regarding the Alberta oil sands debate; sentiment analysis for opinion mining in elections; and, most recently, a study of content-based detection of online influence campaigns. Her work has received a number of recognitions, including the innovative application award at the 10th European Conference on Practice and Principles of Knowledge Discovery (PKDD 2006), a best paper award at the 8th International Conference Machine Learning and Data Mining in Pattern Recognition (MLDM 2012) and a best student paper runner-up award at the European Conference on Machine Learning and Knowledge Discovery in Databases (ECML/PKDD 2018).

Foundations and applications of adaptive learning methodologies

Data streams are ubiquitous in society, and extracting accurate and timely knowledge from these repositories is one of the most significant challenges that we face today. Indeed, in an increasingly fast-paced world, many of today’s pressing problems present themselves in the form of ubiquitous streams. Learning in such evolving domains presents many challenging and interrelated problems, due to the change in data distributions and feature importance over time, as well as the added complexities when learning in the presence of emerging concepts and highly skewed data. Applications are numerous and include detecting intrusion and phishing attacks in cyberspace, monitoring environmental patterns for emergency awareness and responsiveness, and preventive social media analysis. This talk highlights our current research on adaptive algorithms designed to construct accurate, trustworthy models that seamlessly adapt as data streams ebb and flow.