Can a student learn optimally from two different teachers


We explore the effects of over-specificity in learning algorithms by investigating the behavior of a student, suited to learn optimally from a teacher B, learning from a teacher B' ? B. We only considered the supervised, on-line learning scenario with teachers selected from a particular family. We found that, in the general case, the application of the optimal algorithm to the wrong teacher produces a residual generalization error, even if the right teacher is harder. By imposing mild conditions to the learning algorithm form, we obtained an approximation for the residual generalization error. Simulations carried out in finite networks validate the estimate found.

Publication DOI:
Divisions: College of Engineering & Physical Sciences > Systems analytics research institute (SARI)
Additional Information: © 2010 IOP Publishing Ltd.
Uncontrolled Keywords: over-specificity in learning algorithms,the supervised,on-line learning scenario,optimal algorithm,approximation,residual generalization error simulations,finite networks,Mathematical Physics,Modelling and Simulation,Statistics and Probability,Physics and Astronomy(all),Statistical and Nonlinear Physics
Publication ISSN: 1751-8121
Last Modified: 24 Jan 2024 08:04
Date Deposited: 14 Dec 2011 12:22
Full Text Link: http://iopscien ... 21/43/1/015101/
Related URLs: http://www.scop ... tnerID=8YFLogxK (Scopus URL)
PURE Output Type: Article
Published Date: 2010
Authors: Neirotti, Juan P. (ORCID Profile 0000-0002-2409-8917)


Export / Share Citation


Additional statistics for this record