ZEHRA CATALTEPE

Istanbul Technical University, Computer Engineering Department, Maslak, Istanbul, Turkey

EDUCATION:

• 1998 : Ph.D in Computer Science at Caltech
  Thesis Title: Incorporating Input Information into Learning and Augmented Objective Functions"
  Thesis Advisor: Prof. Yaser S. Abu-Mostafa
  
• 1994 : M.S. in Computer Science at Caltech
  Thesis Title: "The Scheduling Problem in Learning From Hints"
  Thesis Advisor: Prof. Yaser S. Abu-Mostafa
  
• 1991 : B.S. in Computer Engineering and Information Sciences at Bilkent University, Turkey (Graduated first in the class)
  Thesis Title: "Field: A Data Type for Automated Perception in Go"
  Thesis Advisor: Prof. Varol Akman
  

RESEARCH INTERESTS:

• Machine learning applications (time series analysis and prediction, image classification, machine and plant condition monitoring, music genre and composer classification, web mining, text-to-speech).
• Machine learning algorithms (RBF Networks, SVMs, linear models, examplar based models).
• Machine learning theory (improving generalization by regularization or additional information about the problem).
  

PUBLICATIONS:

• (2005) Z. Cataltepe, E. Cetin and T. Pearson, "Identification of Insect Damaged Wheat Kernels Using Transmittance Images" , Electronics Letters (ISSN: 0013-5194), vol 41, no 5, pp 23--24.
  
• (1999) Z. Cataltepe, Y. S. Abu-Mostafa and M. Magdon-Ismail, "No Free Lunch for Early Stopping", Neural Computation, (ISSN:0899-7667) vol 11, no 4, pp 995--1009.
  
• (1997) Z. Cataltepe and M. Magdon-Ismail, "Incorporating Test Inputs into Learning" , Proceedings of the Advances in Neural Information Processing Systems 10 .
  
• (2005) Z. Cataltepe, C. Yuan, C. Neubauer, Meltem Demirkus and H.Gerd-Brummel, "A Robust Method to Identify Faults in Correlated Sensors in Machine Condition Monitoring", accepted to EUSIPCO (European Signal Processing Conference) September 2005, Antalya, Turkey.
  
• (2005) C. Yuan, C. Neubauer, Z. Cataltepe and H.Gerd-Brummel, "Support Vector Methods and Use of Computed Sensors for Power Plant Monitoring" , ICASSP (International Conf. on Acoustics, Speech and Signal Processing), Philadelphia, USA, March 2005.
  
• (2004) Z. Cataltepe, T. Pearson and E. Cetin, "Fast Insect Damage Detection in Wheat Kernels Using Transmittance Images" , IEEE IJCNN (International Joint Conference on Neural Networks) 2004, Budapest, Hungary.
• (2004) Z. Cataltepe, E. Cetin and T. Pearson, "Identification of Insect Damaged Wheat Kernels Using Transmittance Images", IEEE ICIP (International Conference on Image Processing) 2004, Singapore.
• (2003) Z. Cataltepe, M. Chang, T. Stiefmeier, B. Chen and C. Neubauer, "Using ROC Curve In the Absence of Positive Examples ", CINC 6 (The 6th International Conference on Computational Intelligence and Natural Computing) Cary, North Carolina.
• (2003) Z. Cataltepe, P. Moghe, "Characterizing Nature and Location of Congestion on the Public Internet", ISSC' (IEEE Symposium on Computers and Communication), Kemer, Antalya, Turkey.
• (1998) Z. Cataltepe, "Incorporating Input Information into Learning and Augmented Objective Functions", Ph.D. Thesis, California Institute of Techology.
  
• (1997) E. Bax, Z. Cataltepe and J. Sill, "The Central Classifier Bound -- A New Error Bound for the Classifier Chosen by Early Stopping" , IEEE PACRIM, Victoria, B.C. 1997 conference proceedings.
  
• (1997) R. A. Sahan, D. C. Albin, N. K. Sahan and Z. Cataltepe, "Modelling Of Transitional Air Flow System In A Channel Using Artificial Neural Networks", Sixth Turkish Symposium on Artificial Intelligence and Neural Networks ( TAINN 97 ).
• (1994) Z. Cataltepe and Y. S. Abu-Mostafa, "Estimating Learning Performance Using Hints" , Proceedings of the 1993 Connectionist Models Summer School, M. Mozer et. al. (Eds.), Lawrence Erlbaum Associates, Publishers, Hillsdale, NJ. pp.380-386.
  
• (1991) Z. Kok and V. Akman, "Field: A Data Type for Automated Perception in Go", Proceedings of International Symposium on Computer and Information Sciences (ISCIS-VI), Turkey, Elsevier Science Publishers, pp 641-650.
  

PATENTS :

• 2001: Z. Cataltepe, K. Mhatre, P. Moghe, G. Zhou (Streamcenter Inc.), "A method to characterize Internet congestion behavior for streaming audio/video". (pending)
• 2001: Z. Cataltepe, P. Moghe, R. Narayan, G. Zhou (Streamcenter Inc.), "A method for property-based clustering of Internet routes". (pending).

STUDENTS:

• (2005) Abdullah Sonmez, M.Sc. in Computer Engineering, Istanbul Technical University,
  Thesis title: "Music Genre and Composer Identification by Using Kolmogorov Distance" (co-advisor)
  

HONORS AND AWARDS:

• 1992: Earle C. Anthony Graduate Fellowship at Caltech
• 1991: Bilkent University, graduated first in class of Computer Engineering and Information Sciences Department.
• 1986:Ranked 25. at OSS and 35 at OYS, nationwide university entrance exams and got a full scholarship at Bilkent University.
• 1986: Physics and Chemistry awards of Turkish Science Foundation (TUBITAK)

EXPERIENCE:

• 2005-Now: Assistant Professor at Istanbul Technical University, Computer Engineering Department.
  
  
• 2002-2004: Siemens Corporate Research, Intelligent Vision and Reasoning Department, Princeton, NJ
  • Designed and implemented machine learning algorithms and software for monitoring of power plants and AC motors and visualization of sensor data.
  • Published three papers and applied for patents.
  
  
• 2000-2001: Technical Team Member, Stream Center Inc. , Murray Hill, NJ
  
  • Designed experiments to collect internet congestion data, to benchmark the company streaming software and also to understand the location and characteristics of internet congestion as seen by end users when streaming audio/video.
  • Designed and implemented algorithms to analyze the internet congestion data.
  • Participated in developing congestion signatures to identify the amount of congestion as seen by a streaming session and property based clustering to identify the location of congestion.
  • Filed two US patents and published two papers.
• 1998-1999: Postdoctoral member of technical staff, Bell Labs Statistics and Data Mining Research, Murray Hill, NJ
  
  • As a postdoctoral member of tecnical staff, investigated modifications to discriminant analysis taking input information into account.
• 1992-1997: Research and teaching, California Institute of Technology, Computer Science Department, Pasadena, CA
  
  • Showed that extra information, such as input information or expert knowledge (hints), about a machine-learning problem can be incorporated into learning from examples through augmented objective functions.
  • Proved that early stopping can not improve the model, when no additional information is available.
  • Investigated different aspects of learning from examples, such as objective functions, regularization methods, optimization methods.
• 1990-1991: Programmer, Halici Bilgi Islem Ltd., Ankara, Turkey
  
  • Wrote, debugged and maintaned business applications in Pascal.

SKILLS:

• Java, Matlab, C, C++, Pascal, SQL, Mathematica, Mac, Linux/Unix, Windows.

REFERENCES:
Available upon request.