Title: Don’t call me just “Outlier”

Abstract: Outliers are data samples that deviate significantly from the normal objects. Their automatic detection is particularly appealing as enables the discovery of unexpected and potentially harmful patterns. Outliers are often confused with noise even if their generation process is typically not random. This simple observation prompts the use of outlier detection to support advanced knowledge discovery from data of various types and modalities. In this talk, I will discuss recent research lines aimed to leverage outlier detection to cluster structured data, label time series segments, and detect out-of-distribution samples in textual data.

Bio: Luca Cagliero has been associate professor at the Dipartimento di Automatica e Informatica of the Politecnico di Torino since January 2020. He is currently the coordinator of the SmartData@Polito Inter-Departmental center, a research center blending interdisciplinary people and competences from different domains and focused on Big Data technologies, Data Science and Machine Learning approaches. Luca Cagliero teaches B.Sc., M.Sc., Master-level, and Ph.D. courses in the areas of databases, data mining, Deep Learning, and Natural Language Processing. His current research interests are mainly in the fields of Natural Language Processing and Deep Learning, coordinating the activities of several PhD students and PostDoc researchers. He has worked on text summarization, classification and pattern mining and published 150+ papers in international journals, book chapters, and conference proceedings. He is currently Associate Editor of two Elsevier journals (ESWA, MLWA), TPC member of the major data mining conference, among which ACM SIGMOD, IEEE ICDM, ECML PKDD, and reviewer of the most renowned international journals in the fields (e.g., IEEE TKDE, ACM TIST, ACM TOIS, ACM TKDD, IEEE TPAMI).