new TR available

[Alessandro Sperduti]

Technical Report available:

Comments are welcome !!

******************************************************
* FTP-host: ftp.di.unipi.it
  FTP-filename: pub/Papers/perso/SPERDUTI/tr-16-95.ps.gz
******************************************************


@TECHREPORT{tr-16/95,
AUTHOR = {A. Sperduti and A. Starita},
TITLE = {Supervised Neural Networks for the Classification
         of Structures},
INSTITUTION = {Dipartimento di Informatica, Universit\`{a} 
               di Pisa},
YEAR = {1995},
NUMBER = {TR-16/95} }



Abstract:

Up to now, neural networks have been used for classification 
of unstructured patterns and sequences. Dealing with complex 
structures, however, standard  neural  networks,  as well as 
statistical methods, are usually believed to  be  inadequate 
because of their feature-based approach.   In fact, feature-
based approaches usually fail to give satisfactory solutions 
because of the sensitiveness of the approach to the a priori 
selected  features  and  the  incapacity  to  represent  any 
specific  information  on  the   relationships   among   the 
components of the structures.  On the contrary, we show that 
neural  networks  can  represent  and   classify  structured 
patterns.  The key idea underpinning our approach is the use 
of the so called  "complex  recursive  neuron".   A  complex 
recursive neuron can be understood  as  a  generalization to 
structures of a recurrent neuron. By using complex recursive 
neurons, basically all the supervised networks developed for 
the classification of sequences,  such  as  Back-Propagation 
Through Time  networks, Real-Time Recurrent networks, Simple 
Recurrent Networks, Recurrent Cascade Correlation  networks, 
and Neural  Trees  can  be  generalized  to  structures. The 
results obtained by some of the above networks (with complex 
recursive neurons) on  classification  of  logic  terms  are 
presented.



* No hardcopy available.

* FTP procedure:
        unix> ftp ftp.di.unipi.it
        Name: anonymous
        Password:
        ftp> cd pub/Papers/perso/SPERDUTI
        ftp> binary
        ftp> get tr-16-95.ps.gz
        ftp> bye
        unix> gunzip tr-16-95.ps.gz
        unix> lpr tr-16-95.ps.gz (or however you print postscript)



_________________________________________________________________
Alessandro Sperduti

Dipartimento di Informatica,
Corso Italia 40,                             Phone: +39-50-887264
56125 Pisa,                                  Fax:   +39-50-887226
ITALY                                   E-mail: perso at di.unipi.it
_________________________________________________________________