Evolutionary Behavior of Textual Semantics (EBOTS)
Participant: Hemant Joshi
Introduction
Text data has been a very powerful source of communication. Most of the
information is expressed in text documents. This directs towards better
organization and understanding meaning of the knowledge the system has.
Evolutionary approach of finding semantics of the data is the state of
the art research to provide insight into this emerging field. Information
should be understood in the context it was established in. A novel approach
has been proposed to implement machine learning algorithms to understand
semantics of the information. Text data shows behavioral properties during
its evolution. These properties should be studied in order to determine
importance and relevance of the information. Semantics of the information
should be represented by a homogeneous system that evolves with growing
knowledge and adapts according to the properties of information.
Description
EBOTS project has a primary focus on evolution of text data in terms of
meaning. Better information organization can be achieved provided computers
can understand the information with its relevance and importance. Information
is understood in the given context. The human mind is the best example
of this process where associative memory is used to express any new information
that is provided. Information can be extracted from huge knowledge bases
by identifying the various properties of the information. Information
expressed in textual data is very common. The majority of documents express
this information and text is the most common media for information exchange.
The EBOTS project is an effort to understand the semantics of the information
in the established context. The context is established using the inherent
properties of information. Any information has temporal properties associated
with it. Temporal properties indicate the chronological history of information
evolution over a certain period of time. Contextual properties of the
information provide insight into the meaning and hidden meaning of the
information. Natural language processing is used to find these inter-domain
relationships. Information domains are defined for any context in which
information is expressed. Semantic aspects of this research field require
collaboration of various fields of study such as philosophy, neuroscience,
computer science, psychology, sociology and anthropology and linguistics.
Such diverse backgrounds contribute to core of the EBOTS.
Layout
-
[1] Hemant Joshi, Coskun Bayrak, “Semantic Information Evolution”, ANNIE 2004
-
[2] Hemant Joshi, Coskun Bayrak, “Learning Contextual Behavior of Text Data”, ICMLA 2005“ - “Presentation “[ppt] “[pdf]
-
[3] James D. Jones, Hemant Joshi, Umit Topaloglu, Eric Nelson: A Child's Story to Illustrate Automated Reasoning Systems Using Opportunity and History. IEEE International conference on Intelligence and Security Informatics ISI 2006 San Diego, CA USA Vol 3975 pp 668-670
-
[4] James D. Jones, Hemant Joshi, Umit Topaloglu, Eric Nelson: Who Stole the Bat? Deception Detection on the Basis of Actions, IEEE International Conference on Systems, Man, and Cybernetics (SMC 2006), Taipei, Taiwan
-
[5] H. Joshi, C. Bayrak, X. Xu, UALR at blog Track, TREC report 15th Text Retrieval Conference proceedings, NIST, 2006
-
[6] H. Joshi, C. Bayrak, “ Automatic Dissemination of Text Information using the EBOTS system,”, International joint conferences on Computer, Information and Systems Sciences, and Engineering (CISSE) 2006 - Presentation [ppt] [pdf]
-
[1] Hemant Joshi, Coskun Bayrak, Xiaowei Xu “Tech Report # UALR05-01”, Matrix Dimensionality Reduction for LSI using Spherical K-means
-
[2] Hemant Joshi, Xiaowei Xu “Tech Report # UALR06-02”, Using Active Learning with Integrated Feature Selection
-
[1] Hemant Joshi “Blogs 101”
Last Update : January 27, 2007