Christopher Thomas

(706) 254-6861

Resume

Objective

Academic or Industry Position, Spring or Fall 2009

Education

PhD in Computer Science, University of Georgia and Wright State University since 2003

Focus:

· Ontology creation for the Biochemistry domain

· Extending present ontology language formalisms to account for probabilistic and fuzzy reasoning

· Semiautomatic Taxonomy/Ontology generation

· Statistical clustering techniques and NLP

· Automatic Entity recognition/classification/extraction

· Human and Social Computing

Master of Science in Artificial Intelligence, University of Georgia 2001 - 2003

Focus:

· Ontologies and Semantics

· Different representations of First Order Logic

· Rule based systems and decision trees

· Inferencing systems

B.S. in Computer Science from the University of Koblenz, Germany 1998 - 2001

Focus:

· Human-Computer Interaction,

· Image Processing/Recognition,

· 2D/3D Computer Graphics

Studies in Philosophy, German literature and linguistics, Mathematics 1995 - 1998

University of Cologne (Germany ).

Professional Experience

· Research Assistant in the Kno.e.sis Center at Wright State University. Since 2007

· Research Assistant in the LSDIS Lab at the University of Georgia. 2003 - 2006

· Consultant for email and networking 2002 – 2003

· Research Assistant in the Image Recognition Lab at the. 1999 – 2001
University of Koblenz. Work on automatic classification of recycling material for waste management companies.

· Consultant in Human-Computer-Interaction questions 2001
for Soft-Use GmbH, Germany

· WWW-programmer (PHP/JavaScript/HTML) 1995 - 2001

· Instructor for MS Windows-applications 1998 - 2001

Spoken Languages

English, German, French

Technical Skills

Research and Projects

My PhD research has been focused on approaching formal semantics from different angles. For a machine to easily draw deductive conclusions, it needs to have access to a formal representation of the information it is fed. The issues I am concerned with are:

• How to model formal ontologies and how to meaningfully represent information

  • e.g. Modeling of GlycO, an Ontology for the Biochemistry domain. Representations of molecular structure and metabolic pathways.

• How to (semi-)automatically populate ontologies from structured and semistructured sources, such as databases and database driven websites.

  • e.g. Extraction of complex molecule representations from multiple sources

• How to transform and disambiguate the extracted data to gain more meaningful representations

  • e.g. Transforming textual representations of complex molecules into a canonical formal representation used in GlycO; after the transformation, molecules from different sources could be compared

• How to automatically extract more meaningful information from unstructured data

  • e.g. TaxaMiner's semiautomatic taxonomy extraction from biomedical abstracts using clustering techniques and NLP.

• How to find more powerful knowledge representations formalisms that allow for uncertainty and contradictions or are “aware” of the degree of truth/certainty/subjectivity of the knowledge

  • see e.g papers on “Semantics for the Semantic Web: the Implicit, the Formal and the Powerful” or “On the Expressiveness of the Languages for the Semantic Web - Making a Case for 'A Little More'”

• How to harvest individual and symbiotic human intelligence as computing resources, especially:

• How to harvest individual and symbiotic human intelligence to gain formal knowledge representations.

  • In the process of recent web developments, e.g. what is known as Web 2.0, individuals and communities have started to accumulate quasi-formal knowledge in the form of tags and annotations.

  • This information can be analyzed, processed and formalized

    • e.g. augmenting a folksonomy to a taxonomy

  • The contributors can be coaxed into more carefully reviewing the information provided

    • e.g. community review in Wikipedia

  • My focus here is on how to provide a framework that does not only allow users to provide more formal representations, but that a) makes the formalism transparent, so the user does not have to deal with it and b) make it desirable to participate and to provide highest quality. See e.g. Luis von Ahn's work on Human Computaion

In my Masters’ work , the focus of my research was Semantics in general and the question of how to make a computer understand (written) language in particular. I concentrated on automatic ontology learning using hierarchical clustering, natural language processing and analogical reasoning. The system takes unstructured text as input and outputs an ontological representation of the set of documents using OWL ontology modeling language.

Other projects I have been working on during my AI and CS studies were:

- Document clustering and Taxonomy learning using Hierarchical k-means or Naïve Bayes Classifier and NLP techniques

- Ontology creation for the Biochemistry domain, focusing on the representation of Glycan structures and Glycan interactions

- Automatic texture recognition and ontology-based image annotation. This tool trained on several textures and was given a spatial ontology of how objects consist of different textures

- Genetic algorithm to build Hidden Markov Models for predicting protein secondary structure

- Genetic algorithm that compose music.

- Neural Networks that can track a ball for the RoboCup robot soccer.

- Automatic maze generation for the game Mummy Maze, involving algorithms to find solutions for the mazes which used heuristics and forward/backward chaining procedures.

- Artificial Agents whose behavior was dependent on

o Rule Systems

o Decision Trees

o Neural Networks

to demonstrate the strengths and shortcomings of different forms of representations in AI.

Other interests

- Logic:

o First Order Logic and its subsets, such as Description Logics, Frame logic.

o Nonmonotonic logic, such as defeasible logics or modal logics.

o Analogical Reasoning

- Language:

Processing of language in the human brain, especially Metaphors. This was the first step towards my interest and active research in analogical reasoning. There is much evidence that humans see their worlds metaphorically, i.e. we try to see new discoveries in the light of known ones, we tend to describe concepts in technical domains with words that are very close to our everyday experience (e.g. parent-child relations in tree data structures, etc.). Based on this I believe that the ability to do analogical reasoning will assist a computer in learning concepts.

Publications

Conference Publications

· Christopher J. Thomas, Amit P. Sheth: Semantic Convergence of Wikipedia Articles, In Proceedings of the 2007 IEEE/WIC International Conference on Web Intelligence, 2007Semantic Convergence of Wikipedia Articles, In Proceedings of the 2007 IEEE/WIC International Conference on Web Intelligence, 2007

· Satya S. Sahoo, Christopher J. Thomas, Amit P. Sheth, William S. York, Samir Tartir, “Knowledge Modeling and its Application in Life Sciences: A Tale of two Ontologies”, To appear in the Proceedings of WWW 2006, Edinburgh, Scotland, 23rd-26th May 2006 (acceptance rate 11%).

· Prashant Doshi and Christopher J. Thomas, “Inexact Matching of Ontology Graphs Using Expectation-Maximization”, to appear in AAAI, Special track on AI and the Web, 2006

· Christopher J. Thomas, Amit P. Sheth, William S. York, “"Modular Ontology Design Using Canonical Building Blocks in the Biochemistry Domain". (To appear in the proceedings of the International Conference on Formal Ontology in Information Systems 2006)

Journal Publications

· Satya S. Sahoo, Christopher J. Thomas, Amit Sheth, Cory Henson, William S. York, “GLYDE - An expressive XML standard for the representation of glycan structure”, Carbohydrate Research, Volume 340, Issue 18, 30 December 2005, Pages 2802-2807

· Vipul Kashyap, Cartic Ramakrishnan, Christopher J. Thomas and Amit Sheth, “TaxaMiner: An Experimental Framework for Automated Taxonomy Bootstrapping”, International Journal of Web and Grid Services, Special Issue on Semantic Web and Mining Reasoning, September 2005

· Amit Sheth, Cartic Ramakrishnan, and Christopher J. Thomas, “Semantics for the Semantic Web: the Implicit, the Formal and the Powerful”, International Journal on Semantic Web & Information Systems, 1 (1), Jan-Mar 2005, pp. 1-18.

Book Chapters

· Christopher J. Thomas and Amit Sheth, “On the Expressiveness of the Languages for the Semantic Web - Making a Case for 'A Little More'”, in E. Sanchez (Editor), "Fuzzy Logic and the Semantic Web", Elsevier, March 2006

· Nicole Oldham, Christopher J. Thomas, Amit Sheth, Kunal Verma, “METEOR-S Web Service Annotation Framework with Machine Learning Classification”, Lecture Notes in Computer Science, Volume 3387, Jan 2005, Pages 137 - 146

Workshop Publications

· Nicole Oldham, Christopher J. Thomas, Amit Sheth, Kunal Verma, “METEOR-S Web service Annotation Framework with Machine Learning Classification”, First International Workshop on Semantic Web Services and Web Process Composition (SWSWPC 2004), In conjunction with the 2004 IEEE International Conference on Web Services (ICWS'2004).

· William S. York, Amit Sheth, Krystof Kochut, John A. Miller, Christopher J. Thomas, Karthik Gomadam, Meenakshi Nagarajan, Xiaochuan Yi, "Semantic Integration of Glycomics Data and Information", Human Disease Glycomics/Proteome Initiative 1st Workshop 2004: Functional Glycomics in Disease, Osaka, Japan

· Amit Sheth, William York, Christopher J. Thomas, Meenakshi Nagarajan, John A. Miller, Krys Kochut, Satya Sahoo, Xiaochuan Yi, "Semantic Web technology in support of Bioinformatics for Glycan Expression". W3C Workshop on Semantic Web for Life Sciences, 27-28 October 2004, Cambridge, Massachusetts USA.

Technical Reports

· V. Kashyap, C. Ramakrishnan, Christopher J. Thomas, D. Bassu, T. C. Rindflesch and A. Sheth, “TaxaMiner: An Experimentation Framework for Automated Taxonomy Bootstrapping”, Technical Report number UGA-CS-TR-04-005, Computer Science Dept., University of Georgia.

· C. Ramakrishnan, Christopher J. Thomas, V. Kashyap and A. Sheth, “TaxaMiner: Improving Taxonomy Label Quality Using Latent Semantic Indexing”, Technical Report number UGA-CS-TR-04-006, Computer Science Dept., University of Georgia.

· Christopher Thomas, “Predicting Domain Specific Entities with Limited Background Knowledge.” LSDIS Technical Report, August 2006.

Relevant Course Work

Databases, Software Engineering, Distributed and Parallel Computing, Machine Learning, Semantic Enterprise, Natural Language Processing, Logic Programming, Automata and Formal Languages, Graph Theory, Computer Vision and Pattern Recognition, Computational Intelligence

(for a full list of my coursework, including descriptions, please look at http://knoesis.wright.edu/topher/courses.html )