Fahad Pervaiz
Professional Researcher in LUMS
Teacher Assistant in LUMS and FAST-NU
PHD(CSE) student in UW from Fall 2012
MS(CS) student in LUMS
BS(CS) from NUCES(FAST-NU)

About Me

I am starting my PhD, in Computer Science and Engineering, this Fall (2012) from University of Washington (UW), Seattle. During my PhD, I will be co-advised by Gaetano Borriello, Richard Anderson and Lakshminarayanan Subramanian. Currently, I am a Professional Researcher and Grad. student in LUMS. I am also Teacher Assistant of "Advance Operating Systems" in LUMS and "Software for Mobile Devices" in FAST-NU. I completed my Bachelors in Computer Science from FAST-NU Lahore Campus. Currently, I am doing my research with Dr. Umar Saif, Dr. Zeeshan Syed and Dr. Lakshminarayanan Subramanian.

My Research Interest is building systems which have large impact. So far, I have developed systems using text mining and health informatics. My research during past couple of years comprises of developing system which mines and extracts structured information from unstructured text like SMS and developing disease outbreak detection and prediction.


Research Projects

I am working on disease surveillance system which works on the internet search query data available from Google. One of the challenges we have is the non-availability of actual number of cases from hospitals and health centers. But that can be derived from the internet search on the disease as many people who get sick search on the disease or its symptoms [Detecting Influenza Epidemics]. We are using Google flu trend as base of our system. Google flu trend gives you the trend of flu without any delay. We are detecting outbreaks over that with very good accuracy. This system is called flubreaks (http://newt.lums.edu.pk/flubreaks)

Second system I am working on is to improve the DEWS (Disease Early Warning System) in Pakistan which is managed by WHO and Pakistan's Health Ministry. We are working to make the system more adaptive in detection of outbreaks. The system should quickly adapt to the disasters like floods and give outbreaks accordingly. For now, we are building this system for Muzaffargarh district of Punjab province. We plan to scale this system for all districts after we have tested our system on one district.

I am also working to determine disease spread through CDR data. After analysis on CDR data, we can infer social network of people. Based on this network, we can build agent simulation model.

I am also working on topic identification and classification of SMS in which Urdu is written in roman letters. This work was done in collaboration with Dr Lakshminarayanan Subramanian from New York University. We worked on an unsupervised algorithm which could detect unigram, bigram and trigram topics along with the challenges of transliteration of urdu, abbreviations in SMS, no grammar and linguistic construction model. We are working on semi-supervised algorithm now, for better results. This has many applications in medical systems and other information based systems.


Contact

Neighbourhood for Emerging World Technologies (NEWT)

Computer Science Department
Lahore University of Management Sciences
Lahore, Pakistan

fahadp(at)lums(dot)edu(dot)pk
+92 (42) 111 115 867, Ext: 4581