Paper available

[Sebastian Thrun]

This is to announce the following technical report (22 pages):






                        LEARNING ONE MORE THING
    
                by Sebastian Thrun and Tom M. Mitchell
    
    
    Most research on machine learning has focused on scenarios in which a
    learner faces a single, isolated learning task.  The lifelong learning
    framework assumes instead that the learner encounters a multitude of
    related learning tasks over its lifetime, providing the opportunity
    for the transfer of knowledge.
    
    This paper studies lifelong learning in the context of binary
    classification.  It presents the invariance approach, in which
    knowledge is transferred via a learned model of the invariances of the
    domain.  Results on learning to recognize objects from color images
    demonstrate superior generalization capabilities if invariances are
    learned and used to bias subsequent learning.



-----------------------------------------------------------------


The paper can be retrieved via anonymous ftp by following these
instructions:

        unix>      ftp reports.adm.cs.cmu.edu
        ftp:name>  anonymous
        Password:> mickey at mouse.gov
        ftp>       cd 1994
        ftp>       get CMU-CS-94-184.ps       (sorry, it's not compressed)
        ftp>       bye
        unix>      lpr CMU-CS-94-184.ps

The paper uses Backprop/Tangentprop as the main learning engine, but
it also addresses general machine learning issues.  Your feedback is
very much appreciated (thrun at cs.cmu.edu).

Sebastian