School: Computer Science & Engineering Department,
Arizona State University, Tempe, AZ, USA

email: fagelgi-at-asu-dot-edu
web: http://www.public.asu.edu/~fgelgi

Department of Computer Science & Engineering
Arizona State University, Tempe, AZ, USA
Ph.D. degree, 2007
GPA: 4.00/4.00

Department of Computer Engineering
Middle East Tecnical University, Ankara, TURKEY
B.S. degree, 2003
GPA: 3.17/4.00

• F.Gelgi, H.Davulcu, "Baum-Welch Style EM Approach on Simple Bayesian Models for Web Data Annotation", The 2007 IEEE/WIC/ACM International Conference on Web Intelligence (WI 2007), November 2-5, 2007, Silicon Valley, USA
• S.S.Yau, H.Davulcu, S.Mukhopadhyay, D.Huang, H.Gong, P.Singh, F.Gelgi, "Automated Situation-Aware Service Composition in Service-Oriented Computing", International Journal of Web Services Research (IJWSR), v.4 (4), pp. 59-82, 2007.
• F.Gelgi, H.Davulcu and S.Vadrevu, "Term Ranking for Clustering Web Search Results", 10th International Workshop on the Web and Databases (WebDB'07), June, 2007, Beijing, China.
• F.Gelgi, S.Vadrevu and H.Davulcu, "Fixing Weakly Annotated Web Data using Relational Models", 7th International Conference on Web Engineering (ICWE'07), July 16-20, 2007, Como, Italy (to appear).
• F.Gelgi, S.Vadrevu and H.Davulcu, "Relational Model Based Annotation of Web Data", 5th Atlantic Web Intelligence Conference (AWIC'07), June 25-27, 2007, Fontainebleau, France (to appear).
• S.Vadrevu, F.Gelgi and H.Davulcu. "Information Extraction from Web Pages using Presentation Regularities and Domain Knowledge". World Wide Web: Internet and Web Information Systems Journal, v.10 (2), pp. 157-179, 2007.
• F.Gelgi, S.Vadrevu and H.Davulcu, "ScubaDiver: Subspace Clustering of Web Search Result", 3rd International Conference on Web Information Systems and Technologies (WebIST'07), March 3-6, 2007, Barcelona, Spain.
• F.Gelgi and M.Onus, "Heuristics for Minimum Brauer Chain Problem", 21st International Symposium on Computer and Information Sciences (ISCIS'06) , November 1-3 2006, Istanbul, Turkey.
• S.Vadrevu, F.Gelgi, S. Nagarajan and H.Davulcu. "Gathering Metadata and Instances from Object Referral Lists on the Web". Online Information Review , v.30 (3), pp.278-296, 2006.
• S.Vadrevu, F.Gelgi, S.Nagarajan and H.Davulcu. "METEOR: Metadata and Instance Extraction from Object Referral Lists on the Web". 1st Online Metadata and Semantics Research Conference, 2005.
• F.Gelgi, S.Vadrevu and H.Davulcu. "Improving Web Data Annotations with Spreading Activation", The 6th International Conference on Web Information Systems Engineering (WISE 2005), November 20-22 2005, New York City, NY, USA.
• S.Vadrevu, F.Gelgi and H.Davulcu. "Semantic Partitioning of Web pages", The 6th International Conference on Web Information Systems Engineering (WISE 2005), November 20-22 2005, New York City, NY, USA.
• S.Vadrevu, S.Nagarajan, F.Gelgi, and H.Davulcu. "Automated Metadata and Instance Extraction from News Web Sites", The 2005 IEEE/WIC/ACM International Conference on Web Intelligence (WI 2005), September 19-22 2005, Compiegne University of Technology, France.
• H.Davulcu, S.Vadrevu, S.Nagarajan, F.Gelgi. "METEOR: Metadata and Instance Extraction from Object
  Referral Lists on the Web". Accepted for poster presentation at the 13th International World Wide Web Conference (WWW'05) May 10-14 2005, Chiba, Japan.

• Towards the Future Search Engine Technology, July 1, 2005, Middle East Technical University (METU), Ankara, Turkey.
• Image Recognition: Methods and Applications, December 5, 2001, Middle East Technical University (METU), Ankara, Turkey.

• Member of Turkish American Scientists and Scholars Association (TASSA), 2006.
• External reviewer, Very Large Data Bases (VLDB 2005), Aug 30 - Sept 2, 2005, Trondheim, Norway.
• Member of Scientific Committee, National Olympiads in Informatics, 2002, Ankara, Turkey.
• Member of Scientific Committee, National Olympiads in Informatics, 2001, Ankara, Turkey.

Introduction to Theory of Computation, Combinatorial Algorithms and Intractability, Combinatorial Algorithms II, Graph Theory, Randomized Algorithms, Probabilistic Methods, Advanced Optimization Topics, Database Management Systems, Data Mining, Semantic Web Mining, Data on the Web, Planning & Learning in AI, Machine Learning.

Algorithms, Discrete Mathematics, Logic for Computer Science, Logic and Database, Neurocomputing, Lisp and Prolog, Artificial Intelligence, Machine Learning, Formal Languages and Abstract Machines, Natural Language Processing.

• 1999 Spring, University Entrance Exam, Ranked in top 1000 among 1.5 million high school students
• 1996 Spring, High School Examination in TURKEY, 73. place

• 1999 Fall, International Olympiads in Informatics, TURKEY, Bronze medal
• 1999 Summer, Balkan Olympiads in Informatics, GREECE, Honor prize
• 1998 Winter, National Computer Olympiad,TURKEY, Bronze medal
• 1998 Summer, Balkan Olympiads in Informatics, TURKEY, 4.place
• 1997 Spring, National Computer Olympiad, 3.place in Aegean Region

• 1999 Spring, TUBITAK(*) High School Project Competition, TURKEY, Gold medal
• 1999 Spring, TUBITAK(*) High School Project Competition, TURKEY, Bronze medal
• 1999 Summer, International Project Competition, KAZAKHSTAN, Silver medal

• 1999 Spring, City High School Tournaments, Karsiyaka, IZMIR, Gold medal
• 1999 Spring, State College Tournaments, IZMIR, Silver medal
• 1991 Fall, State Primary School Tournaments, ZONGULDAK, 4.place

• 2000 Winter, TZV(**), 5.National Mental Games, participated semifinals
• 1999 Winter, TZV(**), 4.National Mental Games, 4. place
• 1998 Winter, TZV(**), 3.National Mental Games, qualified for semifinals
• 1997 Winter, TZV(**), 2.National Mental Games, qualified for semifinals
• 1996 Winter, TZV(**), 1.National Mental Games, participated semifinals

We present a fast and scalable Spreading Activation Network (SAN) framework for improving weakly annotated data which is typically generated by an automated information extraction system from Web documents. Weakly annotated data suffers from two major problems; (i) might contain incorrect ontological role assignments, and (ii) might have many missing attributes. The SAN model described here is shown to substantially improve the Web data annotations forWeb document collections. Our experimental evaluations with the TAP data set indicate that our model can improve the accuracy of role assignments up to 75% even with 60% and (35%,35%) distortion, and can recover more than half of the 35% missing attributes.
http://www.public.asu.edu/~fgelgi/mining/mining.htm

In general, scheduling problems are characterized by the weight (wj), the processing time (pj), the completion time (Cj) and the due date (dj) for the job j. The objective function for Single Machine Total Weighted Tardiness (SMTWT) is SUMj{wjTj} where Tj = max{Cj - dj,0}. In this project, we present a novel approach that unifies the advantages of both Integer Programming (IP) and Artificial Intelligence (AI) areas. At some nodes that have sufficient IP constraints during the IP search, we will immediately solve the sub-problem using AI planning techniques instead of branching. Here, we used the IP constraints in the sub-problem as search control information in TLPlan.
http://www.public.asu.edu/~fgelgi/ai/ai.htm

In this project, our aim is to prepare a system that is highly adaptable to rapid situation changes in combats of military domain. The project contains a domain dependent replanning system such that given time and cost constraints, i.e. the critical time and cost bound, it will find a plan (mostly depending on the previous plan) as efficient as possible. We made an implementation of military domain with a reasonable replanner that is able to prepare plans using temporal actions. To test the efficiency of the system, we also implemented a Combat Simulator which verifies generated plans and throws random failures during execution.
http://www.public.asu.edu/~fgelgi/ai/ai.htm

This project focuses on identifying missing and similar labels in ontologies built from web pages. We start by converting HTML pages into DOM trees via a semantic partitioner. Next, we extract a set of value-category sets, based on the structure of the trees. Using these sets, we then attempt to identify missing and similar labels by using a clustering algorithm over the entire domain. To do this, we calculate a vector of syntactic statistics for the values in each category set. The vectors are then clustered using the K-means algorithm, with the hopes that both good and bad categories will be combined, thus allowing us to identify missing labels. To improve performance over simple statistical analysis, we also perform some preprocessing to improve our clustering results, including scanning for regular expressions and identifying enumerated types.
http://www.public.asu.edu/~fgelgi/mining/mining.htm

A Brauer chain for a positive integer n is a sequence of integers 1 = a₀, a₁, a₂, ... , a_r = n such that a_i = a_i-1 + a_k for some 0 <= k < i and 1 <= i <= n. Brauer chain simply solves one of the essential problems of arithmetic in computer; "calculating xⁿ using minimum number of multiplications". In this paper, several heuristics for approximating minimum length Brauer chain for a given number n, is discussed. These heuristics are based on some greedy approaches and dynamic programming. Experimental results show that the approximation ratio is smaller than 1.5 for these heuristics and even better we achieved a very good approximation ratio.
http://www.public.asu.edu/~fgelgi/algor/algor.htm

Project is concerned with developing an algorithm to learn opponent stategy in a two player repeated game. The opponents' stategy is to be expressed as a DFA. The learning agent will be maintaining and enhancing the opponents' model during the repeated encounters. Hence the player will select its next move based on its estimation of the opponents' move by looking at the DFA.
http://www.public.asu.edu/~fgelgi/ai/op_model/op_model.htm

We prepared a full functioning language in ELI (a compiler design tool) having C type syntax. It has also some extra features like array operations.
http://cips.eas.asu.edu/fagelgi/studies/others/others.htm

Prepared in C++. The aim is to make learn the computer play basic games. Given just the moves of a game, software starts to play the game without knowing any rules and any thing about the game. By using reinforcement like complex mechanism with some optimizations software learns to play several games.
http://www.public.asu.edu/~fgelgi/ai/ai_contest/ai_contest.htm

Software is developed in Delphi in order to read a written Turkish text with sentence stressing. In a written Turkish passage, software finds necessary parts of the sentences to give stress and by using concatenative synthesis, speech is constructed.
http://cips.eas.asu.edu/fagelgi/studies/tts/tts.htm

The software is developed in Delphi and simulates the calculationa of a basic mobile shape and shows the results of the moving of the tool.

The software, which has been developed in Delphi, can compare images drawn with a pen with each other or with a database of images using graph data type. This could be used in OCRs, in signature, iris and fingerprint recognition or in mapping systems. The project has been presented at nationwide TUBITAK(*) project competition and awarded with a Gold Medal.
http://cips.eas.asu.edu/fagelgi/studies/img_rec/iris/iris.htm

2006 Fall, Arizona State University, Computer Science & Engineering Department

2003-2007, Arizona State University, Computer Science & Engineering Department

• 2002 Summer, TUBITAK(*), Ankara, TURKEY
  Research Intern: Design and Analysis of Algorithms
• 2001 Summer, TUBITAK(*), Ankara, TURKEY
  Research Intern: Design and Analysis of Algorithms

Has attended as an instructor in the following camps, which were organized by TUBITAK (*):

1. 2003 Winter, Computer Olympiad Advanced Camp, Ankara, Turkey
2. 2002 Summer, Computer Olympiad Beginner Camp, Ankara, Turkey
3. 2002 Winter, Computer Olympiad Advanced Camp, Ankara, Turkey
4. 2001 Summer, Computer Olympiad Beginner Camp, Ankara, Turkey
5. 2000 Summer, Computer Olympiad Beginner Camp, Ankara, Turkey
6. 2000 Winter, Computer Olympiad Advanced Camp, Ankara, Turkey

Scholarship from TUBITAK(*) for achievements in international competitions.

Scholarship from Private Yamanlar High School during undergradute study.

Has gained fellowship from Private Yamanlar High School for high degree in high school enterance exam.

C (expert), C++, Java, Pascal (expert), Delphi & Kylix(expert), Prolog, Lisp, Scheme, ML, Basic, Visual Basic

HTML, PHP, CGI, FrontPage, Dreamweaver

Word (Macro Programming), Excel (Macro Programming)

OpenGL, GLUT

Windows Family, MS-DOS (expert), Linux

CPUs: 8086
Sound Hardware: Sound Blaster and FM Chips, PC Speaker
Graphics: VGA, SVGA (Vesa based)
Other: Parallel & Serial Port, System Timer, Mouse, FAT File System

Delphi, Visual Basic (for DOS and windows)

SQL Language, Borland Database Engine, Microsoft Access, Interbase

Turkish (Native), English (Proficient)

Listening to music, reading, cycling, internet and computers, swimming, skating, skiing, playing football, volleyball, chess, tennis, going to the cinema, drawing, calligraphy.

• 2004-2007, Mentor and Volunteer at Sonoran Science Academy.
• 2004-2007, Vice President, Fellowship of Faiths and Cultures Club at ASU.
• 2002-2003, Chair of Scientific Workgroups of Computer Club of METU
• 2001-2002, Chair of Artificial Intelligence Workgroup of Computer Club of METU
• 1999 Winter, TUBITAK(*) Advanced Training Camp, ANTALYA
• 1999 Fall, High School Amateur Photographing Club, IZMIR
• 1999, Captain of the High School Chess Team, IZMIR
• 1998 Winter, TUBITAK(*) Advanced Training Camp, ANTALYA
• 1997-1998, High School Chess Club, IZMIR
• 1997 Summer, TUBITAK(*) Beginner Training Camp, ANKARA