TY - GEN
T1 - Noninvertibility in neural networks
AU - Rico-Martinez, Ramiro
AU - Kevrekidis, Ioannis G.
AU - Adomaitis, Raymond A.
N1 - Funding Information:
The authors would like to acknowledge the support of the National Science Foundation and UTRC. One of the authors (R.R.M.) was partially supported by CONACyT; the hospitality of the Center for Nonlinear Studies at the Los Alamos National Laboratory is gratefully acknowledged.
PY - 1993
Y1 - 1993
N2 - We present a method for assessing certain validity aspects of predictions made by neural networks used to approximate continuous (in time) dynamical systems. This method searches for noninvertibility (non-uniqueness of the reverse time dynamics) of the fitted model, an indication of breakdown of 'proper' dynamical behavior. It is useful for computing bounds on the valid range of network predictions.
AB - We present a method for assessing certain validity aspects of predictions made by neural networks used to approximate continuous (in time) dynamical systems. This method searches for noninvertibility (non-uniqueness of the reverse time dynamics) of the fitted model, an indication of breakdown of 'proper' dynamical behavior. It is useful for computing bounds on the valid range of network predictions.
UR - http://www.scopus.com/inward/record.url?scp=84943259587&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=84943259587&partnerID=8YFLogxK
M3 - Conference contribution
AN - SCOPUS:84943259587
SN - 0780312007
T3 - 1993 IEEE International Conference on Neural Networks
SP - 382
EP - 386
BT - 1993 IEEE International Conference on Neural Networks
A2 - Anon, null
PB - Publ by IEEE
T2 - 1993 IEEE International Conference on Neural Networks
Y2 - 28 March 1993 through 1 April 1993
ER -