Introduction
Anomaly detection is the task of automatically determining instances that are dissimilar to others considered as normals. Such instances are called anomalies, as they do not conform to the training data distribution [1]. Automatic detection of anomalies allows users to save time examining lots of normal cases in order to find outliers and finds application in several fields, such as surveillance and medicine. Different kinds of deep learning models have been introduced to solve this task. Recently, a new architecture based on deep learning for anomaly detection in the context of industrial control systems has been proposed [2]. Other approaches rely on attention mechanisms [3] or propose architectures based on Transformers [4] to highlight the anomalies.
The goal of this thesis is to study recently proposed models for detection of anomalous instances and to apply them to an industrial problem.
Planned Activities
- Initial research on deep learning methods for anomaly detection;
- Research on Transformers and their application to anomaly detection;
- Implementation of state-of-the-art anomaly detection methods;
- Application of these models to a real problem.
Who we’re looking for
Students that are about to get their Master Degree in: computer science, computer engineering, mechatronic engineering, mathematical engineering, mathematics, physics, informatics.
Required skills:
- Proficiency in at least one programming language (Python, Lua, Matlab, C++, Java), Python is preferred;
- Basic knowledge of machine learning and Deep Learning algorithms.
Duration of this Projects: 6-8 months
Check these links before moving on
- Anomalous Example Detection in Deep Learning: A Survey (https://ieeexplore.ieee.org/stamp/stamp.jsp?arnumber=9144212)
- DAICS: A Deep Learning Solution for Anomaly Detection in Industrial Control Systems (https://arxiv.org/pdf/2009.06299.pdf)
- HAN-ECG: An Interpretable Atrial Fibrillation Detection Model Using Hierarchical Attention Networks (https://arxiv.org/pdf/2002.05262.pdf)
- Spacecraft Anomaly Detection via Transformer Reconstruction Error (https://link.springer.com/chapter/10.1007%2F978-981-15-1773-0_28)
Contact Us
Directly by email to: [email protected]
By LinkedIn: linkedin.com/in/cannavò-sonia-66a95467
