Welcome to the Data Science & Machine Learning Lab (DSML Lab) at SINES, NUST. Our lab is equipped with modern computational infrastructure and interdisciplinary expertise. We specialize in designing, training, and deploying machine learning and data science solutions in areas such as NLP, recommender systems, security, behavioral/social media analytics, and fairness & privacy.

PI’s Profile

Dr. Usman Shahid is PhD in Electrical and Computer Engineering from North Dakota State University, Fargo, ND, USA. Dr. Khan is Principal Investigator of the DSML Lab. His research spans network traffic analysis, encrypted traffic fingerprinting, adversarial settings, social media behavior, explainability in AI, and observing how digital platforms mediate content and information. He has supervised numerous MS and PhD students, led several national/international grants, and collaborates on interdisciplinary projects.

Dr. Muhammad Iqbal, Assistant Professor at SINES, NUST, specializes in energy harvesting, renewable energy technologies, self-powered sensors, IoT systems, and artificial intelligence. My research interests include the design and development of smart systems for sustainable energy applications, with a focus on wearable electronics, health monitoring, and structural health monitoring.

Expertise/Capabilities of Lab

• Network Traffic Analysis and Privacy / Security (encrypted traffic fingerprinting, side-channels)
• Natural Language Processing & Text Analytics (hate speech, offensive content, sentiment, spam vs expert detection)
• Behavioral and Observational Data Studies (platform content moderation, advertising observatory style research)
• Recommender Systems & Adaptive Machine Learning (concept drift, robustness, user fairness)
• Explainable / Interpretable AI (especially in health, social & policy domains)
• Out-of-Distribution Detection & Robust ML under changing data distributions

Human Resource

Typical Enrollment of lab researchers: 8-10 MS/PhD researchers

Lab Equipment/Resources

• GPU & deep learning hardware (RTX 4060) for training large models
• Programming & modeling tools: Python (TensorFlow, PyTorch, Scikit-Learn), R; SQL / NoSQL databases
• Data visualization and dashboarding tools
• Full-time human resources: MS/PhD students, research assistants, visiting scholars

Projects

• Encrypted Traffic Fingerprinting for Video Identification
  Inferring video content from encrypted network traffic while maintaining privacy constraints.

• HateClassify: Hate & Offensive Speech Detection
  Classifying social media content with a focus on fairness, bias reduction, and error analysis.

• Observational Studies of Content Moderation & Advertising Platforms
  Measuring how platforms filter or promote content, including ad targeting behavior and policy impact.

• Human Activity Recognition via IoT Sensors
  Using sensor networks to detect human activities with emphasis on privacy and real-world robustness.

• Adaptive Batch Size & Gradient Techniques for Stable ML Training
  Methods to improve convergence, reduce overfitting, and minimize oscillations during training.

• Explainable AI for Health & Behavioral Data
  Building interpretable machine learning models for medical and behavioral datasets to ensure transparency and trust.

Contact Us

Dr. Muhammad Usman Shahid Khan

Contact No.: 5190855747
Email: [email protected]

Dr. Muhammad Iqbal

Contact No.: 5190855746
Email: [email protected]