The main research areas of the Knowledge Discovery Group are anomaly detection and graph mining. The group also focuses on pre-processing methods for data mining, including deep learning and frequent patterns as new features, as well as on text mining—particularly text summarization and the detection of anomalous documents. Collaboration with industry and the public sector is emphasized. Members of the group teach most courses on machine learning and data mining, including natural language learning and text mining.
Areas of Interest
rare events detection
Anomaly detection and explanation represents the main research domain of the laboratory. Besides standard multidimensional data, the research is also concerned with time series, image, and graph data.
graph mining
We develop methods for mining subgraph patterns in networks that are primarily attributed and time-evolving. The focus is set on frequent as well as anomalous and discriminative patterns.
recognition of vehicle attributes
Applications of machine learning methods in identifying the individual vehicle attributes based on camera images from the real environment. Machine learning algorithms for feature classification (e.g. color, maker, type).
Outliers detection
in traffic images
Anomaly detection in images, mostly using methods of machine learning. Detection of unusual traffic situations, vehicle visual defects or unauthorized manipulation with camera.
Joint Image and Text Learning
Texts and images are humans natural source of information. Therefore, teaching computers to understand information from multimodal sources is necessary step towards artificial intelligence goals.
Multimodal Anomaly detection
Enriching texts with other non-textual features has been proven useful in some areas of machine learning. We are convinced, that anomaly detection is on of them.
