Primary supervisor

Mahsa Salehi

This project aims to develop foundation models for detecting anomalies in time series data. Anomalies, such as unusual patterns or unexpected events, can signal critical issues in systems like healthcare, finance, or cybersecurity. Current methods are often limited by the fact that they reuire long training before one can test the model on a new time series due to complexity and variability of real-world time series data. By leveraging advanced machine learning techniques, this project seeks to create robust and adaptable models that can generalize across diverse time series scenarios. The expected outcomes include improved accuracy in detecting anomalies. The benefits span various sectors, including cybersecurity and healthcare.

Student cohort

Double Semester

Aim/outline

Building upon a successful foundation model [3] developed in our research group, we will extend the model for the problem of time series anomaly detection [1,2]. 

URLs/references

[1] Darban, Z. Z., Webb, G. I., Pan, S., Aggarwal, C. C., & Salehi, M. (2025). CARLA: Self-supervised contrastive representation learning for time series anomaly detection. Pattern Recognition, 157, 110874.

[2] Zamanzadeh Darban, Z., Webb, G. I., Pan, S., Aggarwal, C., & Salehi, M. (2022). Deep learning for time series anomaly detection: A survey. ACM Computing Surveys.

[3] Mohammadi Foumani, N., Mackellar, G., Ghane, S., Irtza, S., Nguyen, N., & Salehi, M. (2024, August). Eeg2rep: enhancing self-supervised EEG representation through informative masked inputs. In Proceedings of the 30th ACM SIGKDD Conference on Knowledge Discovery and Data Mining (pp. 5544-5555).

 

Required knowledge

Machine learning

Python programming

PyTorch