Farhan Khawar

RESEARCH

• Probabilistic Graphical Model based behavior Learning in Minority Games       Advisor:Qin Zengcheng

       Intelligent Computing and Machine Learning Lab (ICMLL)                          Mar 2012-Present

• Introduced a novel framework of using probabilistic graphical model for behavioral decomposition of minority game.
• The novel generative model is based on the design that the collective behavior of the minority game can be decomposed into individual behavior. The individual behaviors are then inferred from the probabilistic graphical model to predict future collective behavior.
• Used this framework to analyze real world stock markets and exchange rate. The prediction accuracy was between 51%-60%.
• Paper submitted to IJCAI2013.

• Using GA for topology evolution and training of Turing Unorganized Machine   Advisor:Qin Zengchang

       Intelligent Computing and Machine Learning Lab (ICMLL)                          Mar 2013-Present

• Using WB-type nodes for forming networks capable of intelligent tasks like Pattern Recognition.
• Using GA (genetic Algorithm) to learn the structure, mix of A-, B-Type nodes and modifier settings of the network.

UNDERGRADUATE PROJECT

Undergraduate Final Project

• Implementation and Performance analysis of MIMO detection using the Sphere decoder and its variants. Implementation of MIMO detection using Sphere Decoder and its variants using USRP and simulating real world impairments like: Spatial and temporal correlation between Tx and Rx antennas, Channel estimation errors, and quantization errors for performance analysis.

• Paper: AA Khan, S Irteza, F Khawar, M Shahab, SI Shah "Performance analysis of detection schemes in MIMO spatially correlated channels", IEEE Symposium on Industrial Electronics & Applications, Vol.  2 page 968 - 972, Oct. 2009, Kuala Lumpur, Malaysia.

Undergradaute Semester Projects

• Simulation of MIMO Detection (Zero Forcing Maximum likelihood hybrid (ZFML)) with Turbo Codes.
• MATLAB Implementation of a cross-modal technique for Video- Assisted Audio Source Separation, in presence of multiple sources and noise, using a single microphone.
• Implementation of an Analog Data-logger circuit board, using microcontroller, along with its PC interface.
• Hardware implementation of data transmission with Encryption, Decryption (Vernam Cipher technique) and Parity check.
• An intelligent 3-D TIC TAC TOE game on the Widows platform, using Windows API and C++, capable of predicting opponent’s moves.
• Hardware Implementation of Line Following Robot using microcontroller.

WORK EXPERIENCE

• VAS Level-II Support engineer at PMCL Mobilink (Orascom group of companies)                                 Oct 2009-Aug 2011
Responsible for providing Level-2 support for Acision (SMSC,MMSC,OMG,PSA),Huawei(EVC and CRBT),Alcatel(GPRS),and other VAS platforms like AVS,ADM,OTA.
• 6-week Internship at PTCL NGN (Huawei softX-3000).                                                                        Aug 2008 - Sept 2008
• Internship at Naval Precision and Electronics Complex (NPEC) .                                                             July 2007 - Sept 2007
Involved Printed Circuit Board (PCB) diagnosis and design using state of the art equipment.