© The Institution of Electrical Engineers
A simple mathematical technique to evaluate the performance in terms of capacity of two versions of the Hopfield network is demonstrated. The calculated performance is compared with values derived from simulation. An important feature of the analysis is that it can be easily extended to other Hebbian trained associative memories.
References
-
-
1)
-
R. Mceliece ,
E. Posner ,
E. Rodemich ,
S. Venkatesh
.
The capacity of the Hopfield associative memory.
IEEE Trans.
,
4
-
2)
-
P. Houselander ,
J.T. Taylor
.
Measuring the error correcting capability of the Hopfield network.
Electron. Lett.
,
2 ,
85 -
87
-
3)
-
Houselander, P., Taylor, J.T.: `Unifying the Hopfield and Hamming binary associative memories', IEEE ISCAS 90, 1 May 1990, New Orleans.
-
4)
-
S. Oh ,
T. Yoon ,
J.C. Kim
.
Associative-memory model based on neural networks: modification of Hopfield model.
Opt. Lett.
,
1 ,
74 -
76
-
5)
-
C. Weatherburn
.
(1946)
, A first course in mathematical statistics.
-
6)
-
J.J. Hopfield
.
Neural networks and physical systems with emergent collectivecomputational abilities.
Proc. Natl. Acad. Sci.
,
2554 -
2558
http://iet.metastore.ingenta.com/content/journals/10.1049/el_19900815
Related content
content/journals/10.1049/el_19900815
pub_keyword,iet_inspecKeyword,pub_concept
6
6