Ex Parte Creger et al - Page 4

                Appeal 2007-1088                                                                             
                Application 10/006,959                                                                       


                      machine and the delivered neural network model on the test machine)                    
                      are the same before further updating the delivered neural network                      
                      model on the test machine.  Furthermore, “the model development                        
                      machine 104 may function as a test machine 606” implies that they                      
                      may be the same machine.  In other words, during the baseline                          
                      training stage, a machine having a trained baseline neural network                     
                      would be a mode development machine.  Later on, when the same                          
                      machine having the same trained baseline neural network is ready to                    
                      have a fine-tuning at the specific work site it functions as a test                    
                      machine.  Accordingly, in view of the specification, Jelley’s steps for                
                      training a neural network anticipate the argued limitation.                            
                      We are not persuaded by Appellants’ remarks beginning at page 2 of                     
                the Reply Brief that the Examiner has mischaracterized Appellants’ claimed                   
                invention.  Appellants argue Appellants’ disclosure rather than the claimed                  
                invention when they characterize claims 1 and 7 as requiring separate                        
                machines and separate models.  The claim language does not preclude the                      
                Examiner’s view that the same machine and model can function at different                    
                times as the claimed “model development machine having a first at least one                  
                model” and as the claimed “at least one test machine having a second at least                
                one model.”  Appellants (Reply Br. 3) take issue with the Examiner’s above-                  
                quoted interpretation of “[i]n some circumstances, the model development                     
                machine 104 may function as a test machine 606” (Specification para. 29) to                  
                mean “these two models (i.e., the neural network model on the model                          
                development machine and the delivered neural network model on the test                       
                machine) are the same” (Answer 10).  However, Appellants have not                            
                provided an alternative explanation of the meaning of that statement in the                  
                Specification.  We also agree with the Examiner’s above-noted conclusion                     
                that the language “delivering a neural network model from the model                          

                                                     4                                                       

Page:  Previous  1  2  3  4  5  6  7  8  Next

Last modified: September 9, 2013