Plan of Study: MSIA/MSE | Concentration:
Harmony
of MT and IT
|
Graduate: August, Lucky 2001 | Interests: Projects / Papers |
2. Entrepreneur ship originally at Business School (GSIA)
- Sep. 2000 - Nov. 2000
Client: E-Transport
Development Team: E-Bravo!
- Dec. 2000 - Aug. 2001
Client: CarnegieExchange
Development Team: Bravo!Applying SEI approaches: My personal mentor: Len Bass / SEI, the author of Software Architecture in Practice
- Capability Maturity Model (CMM), Team/Personal Software Process (TSP/PSP)
- Architecture Tradeoff Analysis Method (ATAM)
- Survivability Network Analysis (SNA) from CERT Coordination Center (CERT/CC)
- Software Risk Evaluation (SRE)
3. Robot projects at Robotics Institute (RI), the supporter of Super bowl, NASA and more
Project: CarnegieExchange
Entrepreneurs: CMU entrepreneurs
- Since May '00, I have worked with my classmate, Atila Omer (CMU/MBA class of 2001), who originally came up with this idea and whom I met at GSIA on the day of our on-campus admission interview in November '98 after over 20 hour trip from Japan. Ver. 1.0 released on Oct. '00
- Since Dec. '00, ver 2.0 is under development as one of MSE Studio projects because my Studio team unfortunately or fortunately lost the first client in November and I could bring CarnegieExchange as our new Studio project with approvals of other team members and instructors.
- Atila and I graduated from GSIA in May '01 but still working at CMU ^^;
The winning sumobot (desc.in Japanese) of Sumo wrestling robot competition at Intro. to Mobile Robotics class in 2001. My partner and main contributor:
- Single SV203 Servo motor controller board, 3 Servo motors, 2 CdS sensors and 2 LEDs integrated with a neat mechanism, an electric circuit and simple AI.
Chief Mechanic, Masataka Oka (CMU/MSE Dec., 2001)My soccer team for TeamBots / Robot World Cup (RoboCup)
1. Performance: 13.57 vs 1 (against BasicTeam)
2. My strategy (description of my team)
3. Simulation environment (TeamBots 2.0 + Java 1.2)
4. Commands to compile and run at ~/teams directoryHyper-YuzoTeam beating all Teambots 2.0 sample teams
Performance: 36 vs 1 (against BasicTeam with seed 1)
4. Natural Language Processing advised by Language Technology Institute (LTI) inventing |
Return to Top
Information Retrieval (IR) and Extraction (IE) Technology Combined with real software engineering development (practices):
- Java Based Search Engine (Lucene) w/ Java Compiler Compiler (JavaCC)
- Speech Tagging (Brill Tagger) integrated through Socket and Java Native Interface (JNI)
- Lexical Reference system (WordNet)
- Corpus generation integrating Perl scripts through Runtime / Process class of Java
- Search engine parser (WWW::Search) : for Google (WWW::Search::Google)
- Web page downloader (WebFS::FileCopy) w/ a supporting library (LWpng-alpha)
- HTML to text converter (html2txt)
- Syntactic parser of English, based on link grammer (Experiment)
On RedHat Linux powered by Apache Web Server and Jakarta projects :
- Development methodology / approach / Integrated Development Environment (IDE)
- UML components especially for business interface design
- Unified Process (Use Case driven) w/ Rational Rose from Rational Software
Servlet/JSP engine (Tomcat), Framework (Turbine) and Template (Velocity)
applying Model-View-Controller (MVC) model for the parallel development with HCIJamie Callan (CMU/LTI, CALD Prof.) inspired my interest on NLP with some advice
His recommended books in IR world:
1) Foundations of Statistical Natural Language Processing
2) Modern Information Retrieval
and a recommended paper based upon my research focus
3) A Mutually Beneficial Integration of Data Mining and Information Extraction (Texas)Rosie Jones (CMU/LTI Ph.D.) gave me significant insights on research
"Learning Dictionaries for Information Extraction by Multi-level Bootstrapping" by Rosie led
me to her. The co-writer of the Rosie's paper as well as a former colleague of Jamie during
UMass, Ellen Riloff (Utah/CS Prof.), opened the door to a self-learning dictionary for me by
"A Corpus-Based Approach for Building Semantic Lexicons".Several proceedings of the National Conference of Artificial Intelligence (AAAI press) from
American Association for Artificial Intelligence (AAAI) were also valuable. The most of these
papers are available at ResearchIndex with a collection of others (NLP and Machine Learing)
including Feature Engineering for Text Classfication.