On learning discretized geometric concepts
Bshouty, N.H.
Zhixiang Chen
Homer, S.
Dept. of Comput. Sci., Calgary Univ., Alta.;
This paper appears in: Foundations of Computer Science, 1994 Proceedings., 35th Annual Symposium on
Publication Date: 20-22 Nov 1994
On page(s): 54-63
Meeting Date: 11/20/1994 - 11/22/1994
Location: Santa Fe, NM, USA
ISBN: 0-8186-6580-7
References Cited: 20
INSPEC Accession Number: 4865015
Digital Object Identifier: 10.1109/SFCS.1994.365705
Current Version Published: 2002-08-06
Abstract
We present a polynomial time online learning algorithm that learns
any discretized geometric concept generated from any number of
halfspaces with any number of known (to the learner) slopes in a
constant dimensional space. In particular, our algorithm learns (from
equivalence queries only) unions of discretized axis-parallel rectangles
in a constant dimensional space in polynomial time. The algorithm also
runs in polynomial time in l if the teacher lies on l counterexamples.
We then show a PAC-learning algorithm for the above discretized
geometric concept when the example oracle lies on the labels of the
examples with a fixed probability p⩽½-1/r that runs in
polynomial time also with r. We use these methods, as well as a bounded
version of the finite injury priority method, to construct algorithms
for learning several classes of rectangles. In particular we design
efficient algorithms for learning several classes of unions of
discretized axis-parallel rectangles in either arbitrary dimensional
spaces or constant dimensional spaces
Index
Terms
Available to subscribers and IEEE members.
References
Available to subscribers and IEEE members.
Citing Documents
Available to subscribers and IEEE members.
Cart
