Notice of Pre-AIA or AIA Status
The present application, filed on or after March 16, 2013, is being examined under the first inventor to file provisions of the AIA .
Claim Objections
Claim 1 is objected to because of the following informalities:
Line 12: limitation “d)” is crossed out and added to limitation “c)”, but the rest of the limitations do not account for this and change the limitation letter.
Correction: limitation “e)” should be changed to “d)”, limitation “f)” should be changed to “e)”, limitation “g)” should be changed to “f)”
Claim Rejections - 35 USC § 112
The following is a quotation of 35 U.S.C. 112(b):
(b) CONCLUSION.—The specification shall conclude with one or more claims particularly pointing out and distinctly claiming the subject matter which the inventor or a joint inventor regards as the invention.
The following is a quotation of 35 U.S.C. 112 (pre-AIA ), second paragraph:
The specification shall conclude with one or more claims particularly pointing out and distinctly claiming the subject matter which the applicant regards as his invention.
Claims 1-10 rejected under 35 U.S.C. 112(b) or 35 U.S.C. 112 (pre-AIA ), second paragraph, as being indefinite for failing to particularly point out and distinctly claim the subject matter which the inventor or a joint inventor (or for applications subject to pre-AIA 35 U.S.C. 112, the applicant), regards as the invention.
Claim 1 recites limitation “a” on line 4. Limitation “a” states: “reading in a plurality of product version-specific training data sets…”. It is unclear where “reading in” is referring to or from, as there is no origin or destination of the data written in the claim. For the purpose of examination, the examiner will interpret the limitation as “reading in a plurality of product version-specific training data sets from a database, each of which…”
Claim 1 recites limitation “c” on line 9. Limitation “c” states on line 10: “versions into both a first product version-specific machine learning module and also a second product version-specific learning module;”. Further into the limitation, on line 13, it states: “of the first machine learning module reproduces the performance values of the first…of the second machine learning module reproduces…”. There is insufficient antecedent basis for this limitation in the claim. For the purpose of examination, the examiner will interpret “first machine learning module” and “second machine learning module” as “first product version-specific learning module” and “second product version-specific learning module” respectively.
Claim 1 recites limitation “c”. More specifically, on line 13: “machine learning modules.” There is insufficient antecedent basis for this limitation in the claim. The claim recites a machine learning module covering all product versions, a first product version-specific machine learning module, and a second product version-specific machine learning module, but never “machine learning modules”. For the purpose of examination, the examiner will interpret “machine learning modules” as “first and second product version-specific machine learning modules”
Claim 1 recites a limitation on line 9:”c) feeding an output signal…”. Claim 1 further recites a limitation on line 17: “f) feeding the resulting output signal…”. It is unclear whether these output signals are the same. Therefore, there is insufficient antecedent basis for this limitation in the claim. For the purpose of examination, the examiner will interpret “feeding an output signal…” “feeding the resulting output signal…” as “feeding a (the) output of the machine learning module covering all versions”.
Claim 1 recites a limitation on line 17: “f) Feeding the resulting output signal of the trained machine learning module covering all product versions into the trained first machine learning module”. It is unclear what constitutes “the trained first machine learning module”, as it is unclear whether this first machine learning module is the same as or different than the “a first product version-specific machine learning module” in limitation c of the same claim. For the purpose of examination, the examiner will interpret “the trained first machine learning module” as “the first product version specific machine learning module”.
Claim 3 recites a limitation starting on line 6: “and that when the resulting output signal is fed into the trained first machine learning module, the first product specifications are also fed into the trained first machine module.” It is unclear what “the resulting output signal” refers to. For the purpose of examination, the examiner will interpret “the resulting output signal” as “the output of generating a plurality of synthetic design data sets”.
Claims 2, 4-10 are dependent on claim 1 and do not add enough to overcome these rejections. Therefore, they are rejected under the same rationale.
Claim Rejections - 35 USC § 101
Claim 8 is rejected under 35 U.S.C. 101 because the claimed invention is directed to non-statutory subject matter. The claim does not fall within at least one of the four categories of patent eligible subject matter because the broadest reasonable interpretation of the “A system for controlling a production system…configured for implementing a method as claimed in claim 1” encompasses software per se. The claim as it is recited performs the method of claim 1 without disclosing the system comprising any hardware elements. Therefore, the claimed system under broadest reasonable interpretation can be a software system, thus software per se.
Claim 10 is rejected under 35 U.S.C. 101 because the claimed invention is directed to non-statutory subject matter. The claim does not fall within at least one of the four categories of patent eligible subject matter because the broadest reasonable interpretation of the “A machine-readable storage medium having a computer program as claimed in claim 9.” encompasses signals per se. The specification does not further disclose signals within the medium outside of the output signals claimed in claim 1. A claim whose BRI covers both statutory and non-statutory embodiments embraces subject matter that is not eligible for patent protection and therefore is directed to non-statutory subject matter. See MPEP 2106.03(II). It is suggested that claim 10 be amended to recite a “non-transitory” computer readable medium to overcome this rejection.
Claim Rejections - 35 USC § 103
In the event the determination of the status of the application as subject to AIA 35 U.S.C. 102 and 103 (or as subject to pre-AIA 35 U.S.C. 102 and 103) is incorrect, any correction of the statutory basis (i.e., changing from AIA to pre-AIA ) for the rejection will not be considered a new ground of rejection if the prior art relied upon, and the rationale supporting the rejection, would be the same under either status.
The following is a quotation of 35 U.S.C. 103 which forms the basis for all obviousness rejections set forth in this Office action:
A patent for a claimed invention may not be obtained, notwithstanding that the claimed invention is not identically disclosed as set forth in section 102, if the differences between the claimed invention and the prior art are such that the claimed invention as a whole would have been obvious before the effective filing date of the claimed invention to a person having ordinary skill in the art to which the claimed invention pertains. Patentability shall not be negated by the manner in which the invention was made.
Claim(s) 1-10 is/are rejected under 35 U.S.C. 103 as being unpatentable over Venugopalan et al. (WO 2020197529 A1, hereinafter Venugopalan) in view of Chen et al. (US 20200334520 A1, hereinafter Chen).
Regarding Claim 1:
Venugopalan discloses:
A computer-implemented method for controlling a production system for producing a first product version based on design data sets of the first product version and a second product version, the method comprising: ([Para 1-4; Fig 3; Para 71] Discloses a computer implemented method for a production system to produce design data, using design data as input, for products such as turbines. It is further disclosed that one or more designs are used in the production of design. Paragraph 71 further discloses the designs being stored for production later.)
Reading in a plurality of product version-specific training data sets, each of which comprises a design data set, which specifies a design variant of a respective product version, and a performance value quantifying a performance of the design variant. ([Para 18-20] Discloses a database storing the training data, including designs and performance measures (i.e. design datasets and performance values). Paragraph 20 discloses obtaining the training data (i.e. Reading in).)
Feeding the design data sets into a machine learning module covering all product versions ([Para 50; Para 4] Discloses using designs output by a generator as input to an optimization module for selecting the best version(s). Paragraph 4 discloses the optimization module’s seeds being used in simulation.)
Feeding an output signal of the machine learning module ([Para 49-52] Paragraph 49 discloses product designs. Paragraph 50 discloses the optimization module which takes in generated designs (i.e. product) and selects the ones with the best efficiency values (i.e. performance). Designs from these are selected by the optimization module. These selected designs are then simulated in the simulation prediction model, as disclosed in paragraph 52.)
Generating a plurality of synthetic design data sets which are fed into the trained machine learning module covering all product versions ([Para 35; Para 14; Fig 2-4; Para 52] Discloses generating one or more synthetic designs. Paragraph 14 discloses the design being new from machine learning (i.e. synthetic). Figures 2-4 and Paragraph 52 discloses the generated designs being input into the forward model to simulate performance. (i.e. fed). Paragraph 52 further discloses the model doing this for each input design (i.e. all versions))
and depending on the resulting output data of the trained first machine learning module, deriving a performance-optimized design data set from the synthetic design data sets, which is output for producing the first product version ([Para 48] discloses checking the output designs and selecting them based on efficiency values (i.e. performance-optimized). N output designs may be selected (i.e. data set).)
Venugopalan does not explicitly disclose:
version;
Feeding an output signal of the machine learning module covering all product versions into both a first product version-specific machine learning module and also a second product version-specific machine learning module, wherein the machine learning modules are jointly trained so that output data of the first machine learning module reproduces the performance values of the first product version and output data of the second machine learning module reproduces the performance values of the second product version;
Feeding the resulting output signal of the trained machine learning module covering all product versions into the trained first machine learning module;
However, Chen discloses in the same field of endeavor as the invention:
Feeding an output signal of the machine learning module covering all ([Para 16; Para 44-45; Para 21; Para 25; Para 27; Para 35] discloses first and second task-specific layers (i.e. version specific modules). Paragraph 44-45 discloses different training data for different tasks (i.e. version specific/independent). Paragraph 21 teaches the shared layer processing input and sending task specific input to specific task layers (i.e. module covering all versions). Paragraph 25 discloses joint training of task-specific labeled data (i.e. modules being jointly trained). Paragraph 27 discloses the task specific layers producing task specific results. Paragraph 35 discloses that every task specific layer can produce corresponding task-specific output (i.e. reproduce values of input).)
e) Feeding the resulting output signal of the trained machine learning module covering all product versions into the trained first machine learning module, ([Para 21] discloses the shared layer processing input to output to task specific layers. An output can be made structurally to a first task specific layer (i.e. first machine learning module). Additionally, structurally, the model covering all product versions can be the shared layer, which in the scope of the primary reference is the optimization module. The optimization module in the primary sends the selected designs to a prediction model, which structurally can be task specific layers.)
Venugopalan and Chen are both analogous to the art of the present invention because they are from the same fields of endeavor of producing product design generatively, and domain specific development and task handling in layers respectively.
It would have been obvious for one of ordinary skill in the art before the effective filing date of the claimed invention to have combined the design development system disclosed by Venugopalan with the task layering system disclosed by Chen. One would be motivated to combine the design development system with the task layering system to separate the production of designs into versions, while also allowing versions of designs to learn from other versions if necessary.
Regarding Claim 2:
Venugopalan in view of Chen discloses the method as claimed in claim 1.
Venugopalan further discloses:
The method as claimed in claim 1, wherein for a plurality of design data sets, a performance value for the design variant specified by the respective design data set is determined by simulation, and that, together with the respective predicted performance value, a respective design data set is used as a training data set. ([Para 18; Para 21] discloses a plurality of design data sets with performance values, used as training data. Paragraph 21 discloses training data having been previously simulated.)
Regarding Claim 3:
Venugopalan in view of Chen discloses the method as claimed in claim 1.
Venugopalan further discloses:
The method as claimed in claim 1, wherein first product specifications, to be fulfilled by the first product version, and second product specifications, to be fulfilled by the second product version, are read in, ([Para 22; Para 41; Para 29-33; Para 35; Para 66] discloses specifications of the operation of the design is provided with performance information that is used to train the machine learning modules. Additionally, paragraph 41 discloses user provided product specifications during the process of creating optimized data sets, that are fed into the generative model. Paragraphs 29-33 disclose the architecture of the network having layers to communicate between, which includes feedback (i.e. feeding the output back into the first module (i.e. layer)). Paragraph 35 discloses how after generating designs, and being input into the optimization module (i.e. covering all versions structurally). The output of the optimization module is compared by the predictive model (i.e. version specific or first version) using requirements or constraints to optimize the design (i.e. specifications) in simulation. Paragraph 66 discloses specification input to compare the simulated values of the performance of generated designs in the forward simulation model (i.e. specification read into version specific module). Paragraph 32 discloses a GAN with an encoder-decoder system, which predicts the accuracy of the generated designs. The paragraph further discloses the architecture being fully connected for simulation.)
Venugopalan does not explicitly disclose:
However, Chen discloses in the same field of endeavor as the invention:
([Para 17] discloses the task specific layers (i.e. version specific modules) are fed training data specific to the layer, and only that specific layer. The product specifications are included in the training data as disclosed earlier.)
Regarding Claim 4:
Venugopalan in view of Chen discloses the method as claimed in claim 1.
Venugopalan further discloses:
The method as claimed in claim 1, wherein the machine learning module covering all product versions implements a Gaussian process and the product version-specific machine learning modules each implement a linear regression model. ([Para 30-33] Paragraph 30 discloses layers that the architecture can have that can provide abstraction (Convolutional Neural Network), and a dense layer connecting the other layers (i.e. can be a module covering all product versions). Paragraph 33 further discloses the mean-squared error loss that is typically used in linear regression models, which is used on the network including the layers which can convolute the input (i.e. version specific). In addition, paragraph 33 also discloses that all networks used in the system are trained with any optimization, including the linear regression optimization through the mean-squared error loss, and Gaussian methods, as disclosed in paragraph 30.)
Regarding Claim 5:
Venugopalan in view of Chen discloses the method as claimed in claim 1.
Venugopalan further discloses:
The method as claimed in claim 1, wherein the machine learning modules each comprise a neural network. ([Para 30] discloses the architectures of the networks including a convolutional neural network, which uses layers for abstraction. (i.e. modules))
Regarding Claim 6:
Venugopalan in view of Chen discloses the method as claimed in claim 1.
Venugopalan further discloses:
([Para 71] discloses the method storing design data sets that are selected. These data sets are used for production (i.e. product version data))
Venugopalan does not explicitly disclose:
The method as claimed in claim 1, wherein multiple sets of machine learning modules, each with a machine learning module covering all
However, Chen further discloses in the same field of endeavor:
The method as claimed in claim 1, wherein multiple sets of machine learning modules, each with a machine learning module covering all ([Para 6; Para 17; Para 42] discloses multiple teacher instances (i.e. multiple sets of machine learning modules) that each have shared layers with task-specific layers (i.e. module covering all versions and version specific modules). Paragraph 17 further discloses training data being provided for the specific layers. Paragraph 42, alongside the algorithm right after it, discloses determining a number of training datasets “T” and selecting a mini batch of “T” and training based on the task specific training data. The algorithm shows how this method is performed: “t” is selected from the larger “T” datasets into a batch. All of the mini batch of data is merged for all epochs. Then for each “b” task in the data sets, the appropriate training data is used to compute loss. Once loss is computed, the model is updated with the new gradient achieved from the loss calculations. Paragraph 42 additionally states that that the implementation of this training can optimized the different tasks in the data set (i.e. performance optimized data set).)
Regarding Claim 7:
Venugopalan in view of Chen discloses the method as claimed in claim 1.
Chen further discloses:
The method as claimed in claim 6, further comprising checking whether product specifications are available for the first and for the second product version, and that, if the result of the check is positive, a set is selected, into product version-specific machine learning modules of which product specifications are to be fed. ([Para 67] discloses selecting only tasks (i.e. versions) that have available task specific training datasets stored (i.e. specifications). Then for each task, models are trained for said task.(i.e. fed in))
Regarding Claim 8:
Venugopalan in view of Chen discloses the method as claimed in claim 1.
Venugopalan further discloses:
A system for controlling a production system for producing a first product version based on design data sets of the first product version and a second product version, configured for implementing a method as claimed in claim 1. ([Para 3] discloses the production method being implemented on a system)
Regarding Claim 9:
Venugopalan in view of Chen discloses the method as claimed in claim 1.
Venugopalan further discloses:
A computer program product, comprising a computer readable hardware storage device having computer readable program code stored therein, said program code executable by a processor of a computer system to implement a method as claimed in claim 1. ([Para 79] discloses a computer program product with computer-readable storage and program code implementing the production method.)
Regarding Claim 10:
Venugopalan in view of Chen discloses the computer program as claimed in claim 9
Venugopalan further discloses:
A machine-readable storage medium having a computer program as claimed in claim 9. ([Para 79] discloses a computer program product with computer-readable storage and program code implementing the production method.)
Conclusion
Any inquiry concerning this communication or earlier communications from the examiner should be directed to SUMAIR R CHOWDHURY whose telephone number is (571)270-0523. The examiner can normally be reached Monday-Friday 7:30am - 5pm.
Examiner interviews are available via telephone, in-person, and video conferencing using a USPTO supplied web-based collaboration tool. To schedule an interview, applicant is encouraged to use the USPTO Automated Interview Request (AIR) at http://www.uspto.gov/interviewpractice.
If attempts to reach the examiner by telephone are unsuccessful, the examiner’s supervisor, ABDULLAH AL KAWSAR can be reached at (571) 270 3169. The fax phone number for the organization where this application or proceeding is assigned is 571-273-8300.
Information regarding the status of published or unpublished applications may be obtained from Patent Center. Unpublished application information in Patent Center is available to registered users. To file and manage patent submissions in Patent Center, visit: https://patentcenter.uspto.gov. Visit https://www.uspto.gov/patents/apply/patent-center for more information about Patent Center and https://www.uspto.gov/patents/docx for information about filing in DOCX format. For additional questions, contact the Electronic Business Center (EBC) at 866-217-9197 (toll-free). If you would like assistance from a USPTO Customer Service Representative, call 800-786-9199 (IN USA OR CANADA) or 571-272-1000.
/SUMAIR RASHED CHOWDHURY/
Examiner, Art Unit 2127
/ABDULLAH AL KAWSAR/ Supervisory Patent Examiner, Art Unit 2127
Prosecution Insights