I am a Data Scientist, Artificial Intelligence Researcher, interested in applied artificial intelligence with focus on data mining and machine learning. I am particularly interested in application of artificial intelligence algorithms in fields such as intelligent systems, ambient intelligence and wearable sensors computing. Specifically, my current research is focused on development and application of machine learning (data mining) methods on sensors data (accelerometers, gyroscopes, location, heart-rate, temperature, etc.) in order to extract some useful information about the user or the environment in general. Applications on which I have focused in the last several years are: human activity recognition, fall detection, energy expenditure estimation, and stress detection.
Currently, I am a Postodctoral Research Fellow at University of Sussex, UK, at the Wearable Technologies Laboratory, which is lead by Dr. Daniel Roggen. My current research project consists in devising new methods to enable the discovery and recognition of new activities, which are not part of the training process, through a combination of new temporal time-series clustering exploiting assumptions of continuity of human behaviour. Within this context, I intend to use Deep Learning (supervised and unsupervised) to learn rich general representations/features which can enable a wider range of activity discovery, compared to engineered features.
Previously, I was a Researcher at the Department of Intelligent Systems at the Jozef Stefan Institute. First as a PhD student and later as a full-time researcher. I have completed my PhD studies in January 2015 at the Jožef Stefan International Postgraduate School in Ljubljana, Slovenia, by defending my thesis entitled: “Context-based Reasoning in Ambient Intelligence” supervised by Prof. Dr. Matjaž Gams and co-supervised by Dr. Mitja Luštrek.
The main scientific contribution for my PhD was the development and application of a novel reasoning method, called CoReAmI: multi-view context based reasoning method. That is, the method includes context in the multi-view reasoning process and this way improves the reasoning performance. More details about the context-based reasoning approach can be found on the “PhD Thesis” page in the “Research” menu item.