Biography
Mohamed Ragab is a Researcher at the Technology Innovation Institute (TII), UAE, and an Adjunct Researcher at the Agency for Science, Technology, and Research (A*STAR), Singapore. His expertise focuses on developing AI-driven solutions for real-world time series applications, particularly within healthcare, manufacturing, and other scarce labeled data, distribution shifts, and sustainable, privacy-preserving AI technologies. His extensive research portfolio comprises over 20 publications, with selected contributions in top-tier conferences, such as ICML, KDD, and IJCAI, as well as leading journals including IEEE TPAMI and IEEE TNNLS. He recently presented his work at ICML 2024. Dr. Ragab’s research excellence is further demonstrated by securing a competitive $150K research grant aimed at advancing AI methodologies capable of learning effectively from minimal data.
Download my CV .
- Deep Learning
- Transfer Learning
- Domain Adaptation
- Time-series data
- Predictive Maintenance
-
PhD in Computer Science and Engineering, 2022
Nanyang Technological University
-
MSc in Medical Image Processing, 2017
Aswan University
-
BSc in Electrical Engineering, 2014
Aswan University
Recent News
| Date | News |
|---|---|
| May-2025 | Our paper entitled Boosting Time-series Domain Adaptation via A Time-Frequency Consensus Framework has been accepted. This work introduces a novel consensus-based strategy that integrates time- and frequency-domain representations to enhance robustness in domain adaptation for time-series tasks. |
| Feb-2025 | Our paper entitled Evidential Domain Adaptation for Remaining Useful Life Prediction with Incomplete Degradation has been accepted in IEEE Transactions on Instrumentation and Measurement. This work introduces an evidential learning framework to handle incomplete degradation data during domain adaptation for RUL prediction. |
| Jan-2025 | Our paper entitled EverAdapt: Continuous adaptation for dynamic machine fault diagnosis environments has been accepted in Mechanical Systems and Signal Processing. EverAdapt offers a novel framework for lifelong learning in fault diagnosis, continuously adapting to changes in operating conditions. |
| Jan-2025 | Our paper entitled From Inconsistency to Unity: Benchmarking Deep Learning-Based Unsupervised Domain Adaptation for RUL has been accepted in IEEE Transactions on Automation Science and Engineering. It provides a unified benchmark and protocol for evaluating UDA methods in RUL prediction. |
| Jan-2025 | Our paper entitled Augmented Contrastive Clustering with Uncertainty-Aware Prototyping for Time Series Test Time Adaptation has been accepted at KDD 2025. This method improves test-time adaptation by leveraging contrastive learning and prototype uncertainty for time series. |
| Oct-2024 | Our paper entitled Overcoming Negative Transfer by Online Selection: Distant Domain Adaptation for Fault Diagnosis has been accepted in IEEE Transactions on Instrumentation and Measurement. It presents a novel online selection strategy to reduce negative transfer in distant domain scenarios. |
| Jul-2024 | Our paper entitled A Virtual-Label-Based Hierarchical Domain Adaptation Method for Time-Series Classification has been accepted in IEEE Transactions on Neural Networks and Learning Systems (TNNLS). The work introduces a hierarchical adaptation method using virtual labels to enhance classification performance. |
| Jun-2024 | Our survey paper Label-Efficient Time Series Representation Learning: A Review has been accepted in IEEE Transactions on Artificial Intelligence. It comprehensively reviews recent advances in SSL and semi-supervised methods for time series analysis. |
| May-2024 | Our paper entitled TSLANet: rethinking transformers for time series representation learning has been accepted at International Conference on Machine Learning (ICML). This work introduces a novel transformer architecture specifically designed for time series representation learning. |
| Apr-2024 | Our paper entitled Universal Semi-Supervised Domain Adaptation by Mitigating Common-Class Bias has been accepted in Conference on Computer Vision and Pattern Recognition (CVPR). |
| Jan-2024 | Received the Competitive Career Development Fund (CDF) from A*STAR for the project titled Label-Efficient and Resilient Federated Learning Approach for Time Series Applications |
| Aug-2023 | Our paper entitled Self-supervised Contrastive Representation Learning for Semi-supervised Time-Series Classification has been accepted in IEEE Transactions on Pattern Analysis and Machine Intelligence (TPAMI). |
Awards
[Jan-2024] Competitive Career Development Fund (CDF) from A*STAR for the project titled Label-Efficient and Resilient Federated Learning Approach for Time Series Applications
[Jul-2020] Finalist Paper Award at (ICPHM2020).
[Jun-2018] Singapore International Graduate Award.
[Dec-2017] Best Master’s Thesis Award, Aswan University.
[Jul-2014] Bachelor’s Degree with The First-class Honours, Aswan University.
[Jan-2014] Exemplary Student Award, Aswan University
Experience
- Privacy-preserving domain adaptation algorithms for time series data
- Continual and Sustainible AI for time series applications
- Robustness and Uncertainty Quantification for time seris data.
- Anomaly detection using LSTM, CNN and Autoencoder techniques. I have provided an improved arsenal to tackle future component Predictive Maintenance projects.
- In advance detection of failure of various air-crafts engines using automatic feature extraction.
- Implement end-to-end data science pipeline from data collection to machine learning model deployment for predictive maintenance tasks such as Anomaly detection, Fault Diagnosis, and Fault Prognosis
- Design Advanced deep learning algorithms for time series data.
- Develop Transfer Learning and Domain Adaptation techniques to address the challenges of real-world predictive maintenance.
- Assist head faculty member with classroom instruction material, exams, and record keeping
- Guide the development and training of the new graduate assistants
- Lead, supervise, and plan undergraduate laboratory experience
- Assists with labs or discussion sections.
- Attend weekly course staff meetings.
- Perform occasional other tasks such as mentoring student in the E-learning.
