Appeal 2007-1088
Application 10/006,959
“[n]eural network weights are well known in neural network theory and
applications, and will not be described further.”
Rather repeat the positions of the Appellants and the Examiner,
reference is made to the Brief and Reply Brief for Appellants’ positions, and
to the Answer for the Examiner’s positions.
OPINION
For the reasons set forth by the Examiner in the Answer, as expanded
upon here, we sustain the respective rejections of all claims on appeal.
As reflected beginning at page 13 of the principal Brief on appeal,
Appellants’ principal argument for all the rejections of the claims on appeal
is that Jelley uses only a single model to predict a machine parameter. The
subject matter of independent claims 1 and 7 on appeal is said to require two
machines and two models that predict machine parameters. In response, the
Examiner correctly held in the paragraph bridging pages 9 and 10 of the
Answer that the neural network model of the claimed model development
machine is broad enough to read on providing a trained baseline neural
network as taught in Jelley.
Of major significance in this appeal are the following remarks by the
Examiner at page 10 of the Answer:
Second, as described at page 6, paragraph [29], “In some
circumstances, the model development machine 104 may function as a
test machine 606” and at page 7, paragraph [32], “In a first control
block 402, a neural network model 802 is delivered from the model
development machine 104 to each test machine 106.” In view of the
specification, the neural network model of the test machine is
delivered from the model development machine. Therefore, these two
models (i.e., the neural network model on the model development
3
Page: Previous 1 2 3 4 5 6 7 8 Next
Last modified: September 9, 2013