27/08/2026
Research Spotlight: Continual AI Learning for Urdu SMS Spam Detection
Congratulations to Mr. Qanmber Jawad on his innovative MS Data Science research titled "A Lightweight Agentic Workflow for Continual Learning in SMS Spam Detection for Urdu Script." Mobile spam and fraudulent messages across Pakistan are becoming increasingly sophisticated, often leveraging Roman Urdu variations, orthographic tricks, and rapidly changing vocabulary to evade traditional static filters. Qamber's research solves this challenge by introducing an agentic continual learning workflow. By delegating tasks across lightweight preprocessing, sub-millisecond edge classification, and adaptive memory updates, his framework detects evolving spam patterns in real time without the heavy computational cost of retraining large language models. This work marks a major advancement in localized NLP, mobile security, and edge-AI deployment for Urdu digital communications.
Supervisor: Dr. Hufsa Mohsin
Internal Examiner: Dr. Muhammad Fahad Khan
External Examiner: Dr. Zobia Rehman, Associate Professor (Comsats University Islamabad)