Addressing Data Scarcity in anomaly detection: One-Class Classification Methods
Talk Abstract:
Machine learning deals with discovering the knowledge that governs the learning process. The science of machine learning helps create techniques that enhance the capabilities of a system using data. Typical machine learning techniques identify or predict different patterns in the data. In classification tasks, a machine learning model is trained using some training data to identify the unknown function that maps the input data to the output labels. The classification task gets challenging if the data from some categories are either unavailable or so diverse that they cannot be modelled statistically. For example, to train a model for anomaly detection, it is usually challenging to collect anomalous data for training, but the normal data is available in abundance. In such cases, it is possible to use One-Class Classification (OCC) techniques where the model is trained by using data only from one class. OCC algorithms are practical in situations where it is vital to identify one of the categories, but the examples from that specific category are scarce. Numerous OCC techniques have been proposed in the literature that model the data in the given feature space; however, such data can be high-dimensional or may not provide discriminative information for classification. To avoid the curse of dimensionality, standard dimensionality reduction techniques are commonly used as a preprocessing step in many machine learning algorithms. In this talk we present a new paradigm that jointly optimizes a subspace and data description for OCC via Support Vector Data Description (SVDD). We will focus on how the data scarcity in many different applications can be addressed by using OCC methods.
Date and Time
Location
Hosts
Registration
- Date: 15 Apr 2025
- Time: 11:00 AM UTC to 12:30 PM UTC
-
Add Event to Calendar
Speakers
Dr. Fahad Sohrab
Biography:
Dr. Fahad Sohrab (Member, IEEE) received his BS degree in telecommunication engineering from the National University of Computer and Emerging Sciences, Peshawar, Pakistan, in 2012. He received his MS degree from Sabanci University, Istanbul, Turkey, in 2016 and his Ph.D. from Tampere University, Finland, in 2022. He has developed several novel machine-learning methods published in reputable journals and conferences. Notably, he was honored with awards from the Nokia Foundation in 2020 and the Finnish Foundation for Technology Promotion in 2022 for his research contributions. He received an honorable mention from the City of Tampere for Technical Creativity of the Year 2022 for developing a framework for early heart attack detection using multi-view echocardiography. He was awarded the Post Docs in Companies (PoDoCo) 2023 grant by the Foundation for Economic Education, Finland, for "Brain neural-like analysis of multimodal data via groundbreaking Machine Learning and AI solutions," together with Haltian. More recently he was part of the team who won the Tampere City Challenge Hackathon 2024 for integrating AI and bioelectronics for capsule endoscopy. Dr. Sohrab is currently a Postdoctoral Researcher at the University of Eastern Finland, while also affiliated with Tampere University. His research interests include machine learning, pattern recognition, subspace learning, and one-class classification.
Address:Finland
Agenda
Tentative agenda:
14:00: Start
14:02: Welcome notes by IEEE CASS Finland Chair
14:10 - 15:00: Technical talk by Dr. Fahad Sohrab
15:00 - 15:15: Q&A and closing