Appeal No. 1997-1573
Application No. 08/276,551
348 (CCPA 1958).
Notwithstanding the Examiner’s conclusion of art
recognized equivalence, the evidence of record before us
reveals fundamental distinctions between the back propagation
neural network of Bozich and Appellant’s claimed Hopfield
neural network. The back propagation neural network of Bozich
employs feedforward architecture (Bozich, column 11, line 57
through column 12, line 30) while the Hopfield neural network
described beginning at page 23 of Appellant’s specification
utilizes feedback architecture in which the output of each
neuron is fed back to itself as well as to other neurons in
the network. The only evidence to support any conclusion of
art recognized equivalence of these two types of neural
networks is the Examiner’s own unsubstantiated statements in
the Answer. We are not inclined to dispense with proof by
evidence when the proposition at issue is not supported by a
teaching in a prior art reference, common knowledge or capable
of unquestionable demonstration. Our reviewing court requires
this evidence in order to establish a prima facie case. In re
Knapp-Monarch Co., 296 F.2d 230, 232, 132 USPQ 6, 8 (CCPA
1961); In re Cofer, 354 F.2d 664, 668, 148 USPQ 268, 271-72
9
Page: Previous 1 2 3 4 5 6 7 8 9 10 11 Next
Last modified: November 3, 2007