2002, Vol.5, No.2, pp.108-128
A mathematical framework is presented for development of systems of AI
intended to be endowed with ability to understand the meaning of information. The crux
of the approach is representation of information by vectors of some Hilbert space
treated as a semantic space and construction of frames whose vectors are associated
with semantic units corresponding to particular relevant meanings. Two definitions and
main properties of frames in a vector space are outlined. It is shown that frames can
be used to solve such tasks of information processing as least squares approximation
of experimental data, forecasting time series and probability density estimation.
Frames are also applied to tackle some problems of AI pertaining to semantic retrieval
and classification of (hyper) texts, medical diagnostics, control of operation of the
Internet and manufacturing enterprises, etc.
Key words:
artificial intelligence, information processing, frame, vector space,
Hilbert space, functional analysis, data processing, approximation, forecasting,
probability estimation, meaning of information, semantic retrieval, text
classification, Internet search, medical diagnostics, neural field, learning,
mathematical modelling, distributed complex system, Internet traffic, manufacturing
enterprise, intelligent agent.
Full text:
Acrobat PDF (221KB)
PostScript (971KB)
PostScript.gz (227KB)
Copyright © Nonlinear Phenomena in Complex Systems. Last updated: July 16, 2002