DETAILED ACTION
Claims 21-28 and 33-40 have been examined.
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 Rejections - 35 U.S.C. § 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.
The term “positive surprise” in claims 21-28 and 33-40 is a relative term which renders the claim indefinite. The term “positive surprise” is not defined by the claim, the specification does not provide a standard for ascertaining the requisite degree, and one of ordinary skill in the art would not be reasonably apprised of the scope of the invention.
Claim Rejections - 35 U.S.C. § 101
35 U.S.C. § 101 reads as follows:
Whoever invents or discovers any new and useful process, machine, manufacture, or composition of matter, or any new and useful improvement thereof, may obtain a patent therefor, subject to the conditions and requirements of this title.
The invention, as taught in Claims 21-28 and 33-40, is directed to “mental steps” and “mathematical steps” without significantly more.
The claims recite:
• authorization policies defining who can view data samples and meta information in the secure environment (i.e., mental steps)
• generating an anonymized dataset that does not include personally identifiable information (i.e., mental steps)
• generating an outcome using an adversarial network with sequentially updated model weights comprising a pair of deep neural networks (i.e., mathematical steps)
• pair of deep neural networks including a generator and a critic (i.e., mathematical steps)
• generating using the generator,..., a first result (i.e., mathematical steps)
• providing the first result to the critic (i.e., mathematical steps)
• generating a first surprise factor to stabilize coupled model training (i.e., mathematical steps)
• generate a second result (i.e., mathematical steps)
• providing the second result to the critic (i.e., mathematical steps)
• generating a second surprise factor to stabilize coupled model training based on providing the second result to the critic (i.e., mathematical steps)
• determining, based on the second surprise factor, that the generator has generated a positively surprising result (i.e., mental steps)
• code verification/password verification (i.e., mental steps)
• validating the first result and the second result (i.e., mental steps)
Claim 21
Step 1 inquiry: Does this claim fall within a statutory category?
The preamble of the claim recites “21. A method, comprising…” Therefore, it is a “method” (or “process”), which is a statutory category of invention. Therefore, the answer to the inquiry is: “YES”.
Step 2A (Prong One) inquiry:
Are there limitations in Claim 21 that recite abstract ideas?
YES. The following limitations in Claim 21 recite abstract ideas that fall within at least one of the groupings of abstract ideas enumerated in the 2019 PEG. Specifically, they are “mental steps” and “mathematical steps”:
• authorization policies defining who can view data samples and meta information in the secure environment (i.e., mental steps)
• generating an anonymized dataset that does not include personally identifiable information (i.e., mental steps)
• generating an outcome using an adversarial network with sequentially updated model weights comprising a pair of deep neural networks (i.e., mathematical steps)
• pair of deep neural networks including a generator and a critic (i.e., mathematical steps)
• generating using the generator,..., a first result (i.e., mathematical steps)
• providing the first result to the critic (i.e., mathematical steps)
• generating a first surprise factor to stabilize coupled model training (i.e., mathematical steps)
• generate a second result (i.e., mathematical steps)
• providing the second result to the critic (i.e., mathematical steps)
• generating a second surprise factor to stabilize coupled model training based on providing the second result to the critic (i.e., mathematical steps)
• determining, based on the second surprise factor, that the generator has generated a positively surprising result (i.e., mental steps)
• code verification/password verification (i.e., mental steps)
• validating the first result and the second result (i.e., mental steps)
Step 2A (Prong Two) inquiry:
Are there additional elements or a combination of elements in the claim that apply, rely on, or use the judicial exception in a manner that imposes a meaningful limit on the judicial exception, such that it is more than a drafting effort designed to monopolize the exception?
Applicant’s claims contain the following “additional elements”:
(1) A blockchain
(2) A use of the generator
(3) A presentation of the positively surprising result in a graphical user interface
(4) An access control
(5) A Docker image or Kubernetes cluster
A “blockchain” is a broad term which is described at a high level. Ahmed, page 2, last full paragraph, where it recites:
Blockchain is shorthand for a suite of distributed ledger technologies that can be programmed to record and track anything of value, such as financial transactions, medical records, land titles, and so on. Blockchain technology is based on the centuries-old method of the general financial ledger. In simplified language, it is a digital ledger which holds the records of all sorts of transactions that happen in a peer-to-peer network. This technology is assumed to ‘cut out the middleman’ from any sort of transaction or transfer of digital assets. This is a much more secure and decentralized medium. Financial institutions are exploring the possibilities of using this technology to ensure secure transactions.
This “blockchain” limitation does not integrate the additional element into a practical application and represents “insignificant extra-solution activity”. (See, M.P.E.P. § 2106.05(I)(A)).
A “use of the generator” is a broad term which is described at a high level. Applicant’s Specification recites:
[0097] In some embodiments, a desired outcome may be generated by using adversarial networks, such as generative neural networks. For example, the desired outcome may be determining a global minimum (i.e. optimization) in a given field, such machine learning model pipeline optimization, parameter or combinatorial optimization, or even determining the most efficient financial portfolio (i.e. stocks, options, currencies, etc.) given certain market performances and histories. In some embodiments, an adversarial network may include a pair of deep neural networks (DNN), called a generator and a critic.
This “use of the generator” limitation does not integrate the additional element into a practical application and represents “insignificant extra-solution activity”. (See, M.P.E.P. § 2106.05(I)(A)).
A “presentation of the positively surprising result in a graphical user interface” is a broad term which is described at a high level. M.P.E.P. § 2106.05 (h) recites in part:
Examples of limitations that the courts have described as merely indicating a field of use or technological environment in which to apply a judicial exception include:
***
vi. Limiting the abstract idea of collecting information, analyzing it, and displaying certain results of the collection and analysis to data related to the electric power grid, because limiting application of the abstract idea to power-grid monitoring is simply an attempt to limit the use of the abstract idea to a particular technological environment, Electric Power Group, LLC v. Alstom S.A., 830 F.3d 1350, 1354, 119 USPQ2d 1739, 1742 (Fed. Cir. 2016);
This “presentation of the positively surprising result in a graphical user interface” limitation does not integrate the additional element into a practical application and represents “insignificant extra-solution activity”. (See, M.P.E.P. § 2106.05(I)(A)).
A “access control” is a broad term which is described at a high level. Applicant’s Specification recites:
[0094] In some embodiments, a representative dataset of the dataset may be created, such that the entity creating a model can use the representative dataset to create the model. Since the representative dataset will have similar properties to the original dataset, the model created via the representative dataset should be similar, if not identical, to a model created via the original dataset. This representative dataset may be created via various methods, in accordance with some embodiments. For example, as illustrated in FIG. 7, a Synthetic Data Generator 712 may be used to generate synthetic data. Data Statistics and Visualizations 714 may be used to describe the original dataset, or Subset Selector and Data Anonymizer 716 may be used to generate an anonymized dataset (i.e. free of personally identifiable information). In some embodiments, one, at least one, or all three of the foregoing may be used to describe the original dataset, the resulting dataset being shown as Anonymous Sample Dataset 716. AlSpecialist 726 may in some embodiments utilize Dev Studio 728 to access Cloud VPS (Virtual Private Server) 730, which has Anonymous Sample Dataset 716. AlSpecialist 716, in 732, may Run or Debug Model on Anonymous Sample Dataset 716. AlSpecialist 726, in some embodiments, creates the code that is required to train a model, including model architecture and training scripts, in Cloud VPS 730. AlSpecialist 726 validates the created code on Anonymous Sample Dataset 716, and debugs the created code before executing it on the original data (i.e. step 732). The created code is then run on the original data in a secure environment, where AlSpecialist 726 does not have any access to. In some embodiments, this secure environment can be Compiled Docker Image for Model 734. Compiled Docker Image for Model 734 may be, for example, a Kubernetes cluster that executes the code created by AlSpecialist 716. Compiled Docker Image for Model 734 may include features such as Code Verification and Access Control 736, which may be used to log code and data usage via blockchain in Log Code & Sample Use (on Blockchain) 738. Compiled Docker Image for Model 734 may also include a secure and private cloud computing environment that is separate from Cloud VPS 730, such as Secure and Private Cloud Computing 738.
Note that “access control” is a feature of the conventional “Compiled Docker Image.”
This “access control” limitation does not integrate the additional element into a practical application and represents “insignificant extra-solution activity”. (See, M.P.E.P. § 2106.05(I)(A)).
A “Docker image or Kubernetes cluster” is a broad term which is described at a high level. Applicant’s Specification recites:
[0094] In some embodiments, a representative dataset of the dataset may be created, such that the entity creating a model can use the representative dataset to create the model. Since the representative dataset will have similar properties to the original dataset, the model created via the representative dataset should be similar, if not identical, to a model created via the original dataset. This representative dataset may be created via various methods, in accordance with some embodiments. For example, as illustrated in FIG. 7, a Synthetic Data Generator 712 may be used to generate synthetic data. Data Statistics and Visualizations 714 may be used to describe the original dataset, or Subset Selector and Data Anonymizer 716 may be used to generate an anonymized dataset (i.e. free of personally identifiable information). In some embodiments, one, at least one, or all three of the foregoing may be used to describe the original dataset, the resulting dataset being shown as Anonymous Sample Dataset 716. AlSpecialist 726 may in some embodiments utilize Dev Studio 728 to access Cloud VPS (Virtual Private Server) 730, which has Anonymous Sample Dataset 716. AlSpecialist 716, in 732, may Run or Debug Model on Anonymous Sample Dataset 716. AlSpecialist 726, in some embodiments, creates the code that is required to train a model, including model architecture and training scripts, in Cloud VPS 730. AlSpecialist 726 validates the created code on Anonymous Sample Dataset 716, and debugs the created code before executing it on the original data (i.e. step 732). The created code is then run on the original data in a secure environment, where AlSpecialist 726 does not have any access to. In some embodiments, this secure environment can be Compiled Docker Image for Model 734. Compiled Docker Image for Model 734 may be, for example, a Kubernetes cluster that executes the code created by AlSpecialist 716. Compiled Docker Image for Model 734 may include features such as Code Verification and Access Control 736, which may be used to log code and data usage via blockchain in Log Code & Sample Use (on Blockchain) 738. Compiled Docker Image for Model 734 may also include a secure and private cloud computing environment that is separate from Cloud VPS 730, such as Secure and Private Cloud Computing 738.
This “Docker image or Kubernetes cluster” limitation does not integrate the additional element into a practical application and represents “insignificant extra-solution activity”. (See, M.P.E.P. § 2106.05(I)(A)).
The answer to the inquiry is “NO”, no additional elements integrate the claimed abstract idea into a practical application.
Step 2B inquiry:
Does the claim provide an inventive concept, i.e., does the claim recite additional element(s) or a combination of elements that amount to significantly more than the judicial exception in the claim?
Applicant’s claims contain the following “additional elements”:
(1) A blockchain
(2) A use of the generator
(3) A presentation of the positively surprising result in a graphical user interface
(4) An access control
(5) A Docker image or Kubernetes cluster
A “blockchain” is a broad term which is described at a high level. Ahmed, page 2, last full paragraph, where it recites:
Blockchain is shorthand for a suite of distributed ledger technologies that can be programmed to record and track anything of value, such as financial transactions, medical records, land titles, and so on. Blockchain technology is based on the centuries-old method of the general financial ledger. In simplified language, it is a digital ledger which holds the records of all sorts of transactions that happen in a peer-to-peer network. This technology is assumed to ‘cut out the middleman’ from any sort of transaction or transfer of digital assets. This is a much more secure and decentralized medium. Financial institutions are exploring the possibilities of using this technology to ensure secure transactions.
Therefore, the claim as a whole does not amount to significantly more than the exception itself (i.e., there is no inventive concept in the claim). (See, M.P.E.P. § 2106.05(II)).
A “use of the generator” is a broad term which is described at a high level. Applicant’s Specification recites:
[0097] In some embodiments, a desired outcome may be generated by using adversarial networks, such as generative neural networks. For example, the desired outcome may be determining a global minimum (i.e. optimization) in a given field, such machine learning model pipeline optimization, parameter or combinatorial optimization, or even determining the most efficient financial portfolio (i.e. stocks, options, currencies, etc.) given certain market performances and histories. In some embodiments, an adversarial network may include a pair of deep neural networks (DNN), called a generator and a critic.
Therefore, the claim as a whole does not amount to significantly more than the exception itself (i.e., there is no inventive concept in the claim). (See, M.P.E.P. § 2106.05(II)).
A “presentation of the positively surprising result in a graphical user interface” is a broad term which is described at a high level. M.P.E.P. § 2106.05 (h) recites in part:
Examples of limitations that the courts have described as merely indicating a field of use or technological environment in which to apply a judicial exception include:
***
vi. Limiting the abstract idea of collecting information, analyzing it, and displaying certain results of the collection and analysis to data related to the electric power grid, because limiting application of the abstract idea to power-grid monitoring is simply an attempt to limit the use of the abstract idea to a particular technological environment, Electric Power Group, LLC v. Alstom S.A., 830 F.3d 1350, 1354, 119 USPQ2d 1739, 1742 (Fed. Cir. 2016);
Therefore, the claim as a whole does not amount to significantly more than the exception itself (i.e., there is no inventive concept in the claim). (See, M.P.E.P. § 2106.05(II)).
A “access control” is a broad term which is described at a high level. Applicant’s Specification recites:
[0094] In some embodiments, a representative dataset of the dataset may be created, such that the entity creating a model can use the representative dataset to create the model. Since the representative dataset will have similar properties to the original dataset, the model created via the representative dataset should be similar, if not identical, to a model created via the original dataset. This representative dataset may be created via various methods, in accordance with some embodiments. For example, as illustrated in FIG. 7, a Synthetic Data Generator 712 may be used to generate synthetic data. Data Statistics and Visualizations 714 may be used to describe the original dataset, or Subset Selector and Data Anonymizer 716 may be used to generate an anonymized dataset (i.e. free of personally identifiable information). In some embodiments, one, at least one, or all three of the foregoing may be used to describe the original dataset, the resulting dataset being shown as Anonymous Sample Dataset 716. AlSpecialist 726 may in some embodiments utilize Dev Studio 728 to access Cloud VPS (Virtual Private Server) 730, which has Anonymous Sample Dataset 716. AlSpecialist 716, in 732, may Run or Debug Model on Anonymous Sample Dataset 716. AlSpecialist 726, in some embodiments, creates the code that is required to train a model, including model architecture and training scripts, in Cloud VPS 730. AlSpecialist 726 validates the created code on Anonymous Sample Dataset 716, and debugs the created code before executing it on the original data (i.e. step 732). The created code is then run on the original data in a secure environment, where AlSpecialist 726 does not have any access to. In some embodiments, this secure environment can be Compiled Docker Image for Model 734. Compiled Docker Image for Model 734 may be, for example, a Kubernetes cluster that executes the code created by AlSpecialist 716. Compiled Docker Image for Model 734 may include features such as Code Verification and Access Control 736, which may be used to log code and data usage via blockchain in Log Code & Sample Use (on Blockchain) 738. Compiled Docker Image for Model 734 may also include a secure and private cloud computing environment that is separate from Cloud VPS 730, such as Secure and Private Cloud Computing 738.
Note that “access control” is a feature of the conventional “Compiled Docker Image.”
Therefore, the claim as a whole does not amount to significantly more than the exception itself (i.e., there is no inventive concept in the claim). (See, M.P.E.P. § 2106.05(II)).
A “Docker image or Kubernetes cluster” is a broad term which is described at a high level. Applicant’s Specification recites:
[0094] In some embodiments, a representative dataset of the dataset may be created, such that the entity creating a model can use the representative dataset to create the model. Since the representative dataset will have similar properties to the original dataset, the model created via the representative dataset should be similar, if not identical, to a model created via the original dataset. This representative dataset may be created via various methods, in accordance with some embodiments. For example, as illustrated in FIG. 7, a Synthetic Data Generator 712 may be used to generate synthetic data. Data Statistics and Visualizations 714 may be used to describe the original dataset, or Subset Selector and Data Anonymizer 716 may be used to generate an anonymized dataset (i.e. free of personally identifiable information). In some embodiments, one, at least one, or all three of the foregoing may be used to describe the original dataset, the resulting dataset being shown as Anonymous Sample Dataset 716. AlSpecialist 726 may in some embodiments utilize Dev Studio 728 to access Cloud VPS (Virtual Private Server) 730, which has Anonymous Sample Dataset 716. AlSpecialist 716, in 732, may Run or Debug Model on Anonymous Sample Dataset 716. AlSpecialist 726, in some embodiments, creates the code that is required to train a model, including model architecture and training scripts, in Cloud VPS 730. AlSpecialist 726 validates the created code on Anonymous Sample Dataset 716, and debugs the created code before executing it on the original data (i.e. step 732). The created code is then run on the original data in a secure environment, where AlSpecialist 726 does not have any access to. In some embodiments, this secure environment can be Compiled Docker Image for Model 734. Compiled Docker Image for Model 734 may be, for example, a Kubernetes cluster that executes the code created by AlSpecialist 716. Compiled Docker Image for Model 734 may include features such as Code Verification and Access Control 736, which may be used to log code and data usage via blockchain in Log Code & Sample Use (on Blockchain) 738. Compiled Docker Image for Model 734 may also include a secure and private cloud computing environment that is separate from Cloud VPS 730, such as Secure and Private Cloud Computing 738.
Therefore, the claim as a whole does not amount to significantly more than the exception itself (i.e., there is no inventive concept in the claim). (See, M.P.E.P. § 2106.05(II)).
Therefore, the answer to the inquiry is “NO”, no additional elements provide an inventive concept that is significantly more than the claimed abstract ideas the claimed abstract idea into a practical application.
Claim 21 is, therefore, NOT ELIGIBLE subject matter under 35 U.S.C. § 101.
Claim 22
Claim 22 recites:
22. The method of claim 21, further comprising:
determining that the critic is sufficiently surprised by the generator based on a surprise factor threshold; and
based on the critic indicating to the generator that the critic is sufficiently surprised, the generator continues exploring an area that resulted in the critic being sufficiently surprised.
Applicant’s Claim 22 merely teaches “determining that the critic is sufficiently surprised by the generator based on a surprise factor threshold” (i.e. a mathematical threshold calculation) and “the generator continues exploring an area that resulted in the critic being sufficiently surprised” (i.e., continued operation of the mathematical “generator”). It does not integrate the abstract idea to a practical application, nor is it anything significantly more than the abstract idea. (See, 2106.05(a)(II).)
Claim 22 is, therefore, NOT ELIGIBLE subject matter under 35 U.S.C. § 101.
Claim 23
Claim 23 recites:
23. The method of claim 21, further comprising:
determining that the critic is insufficiently surprised by the generator based on a surprise factor threshold; and
based on the critic indicating to the generator that the critic is insufficiently surprised, the generator discontinues an area that resulted in the critic being insufficiently surprised and explores a different area.
Applicant’s Claim 23 merely teaches “determining that the critic is insufficiently surprised by the generator based on a surprise factor threshold” (i.e. a mathematical threshold calculation) and “based on the critic indicating to the generator that the critic is insufficiently surprised, the generator discontinues an area” (i.e., discontinued operation of the mathematical “generator”). It does not integrate the abstract idea to a practical application, nor is it anything significantly more than the abstract idea. (See, 2106.05(a)(II).)
Claim 23 is, therefore, NOT ELIGIBLE subject matter under 35 U.S.C. § 101.
Claim 24
Claim 24 recites:
24. The method of claim 21, further comprising: based on determining that a surprise factor threshold has not been met a predetermined number of times, lowering the surprise factor threshold.
Applicant’s Claim 24 merely teaches “lowering the surprise factor threshold” (i.e., mathematical steps). It does not integrate the abstract idea to a practical application, nor is it anything significantly more than the abstract idea. (See, 2106.05(a)(II).)
Claim 24 is, therefore, NOT ELIGIBLE subject matter under 35 U.S.C. § 101.
Claim 25
Claim 25 recites:
25. The method of claim 21, further comprising: based on determining that a surprise factor threshold has been met a predetermined number of times, increasing the surprise factor threshold.
Applicant’s Claim 25 merely teaches “increasing the surprise factor threshold” (i.e., mathematical steps). It does not integrate the abstract idea to a practical application, nor is it anything significantly more than the abstract idea. (See, 2106.05(a)(II).)
Claim 25 is, therefore, NOT ELIGIBLE subject matter under 35 U.S.C. § 101.
Claim 26
Claim 26 recites:
26. The method of claim 21, further comprising:
taking, by the generator, a prior comprising aggregated latent features from each individual of a previous generation to generate a population;
evaluating individuals of the current population for an objective function for which the critic also estimates each individual's result;
a prior for a next generation is extracted from an aggregated latent space of the critic for each individual; and
the prior is saved as a best-known prior of an optimization process if it includes a best individual regarding the objective function.
Applicant’s Claim 26 merely teaches a probabilistic “a prior” (i.e., a mathematical step) and “evaluating individuals of the current population for an objective function” (i.e., a mathematical step) and “a prior for a next generation is extracted from an aggregated latent space of the critic for each individual” (i.e., a mathematical step) and “the prior is saved” (i.e., a well-understood, routine, and conventional “saving” operation). It does not integrate the abstract idea to a practical application, nor is it anything significantly more than the abstract idea. (See, 2106.05(a)(II).)
Claim 26 is, therefore, NOT ELIGIBLE subject matter under 35 U.S.C. § 101.
Claim 27
Claim 27 recites:
27. The method of claim 24, further comprising: extracting the prior from the population either by expert defined statistics or by using a deep neural network trained with the generator for feature extraction, the prior including the best individual's latent coordinates in a previous population.
Applicant’s Claim 27 merely teaches “extracting the prior from the population either by expert defined statistics or by using a deep neural network” (i.e., mathematical steps). It does not integrate the abstract idea to a practical application, nor is it anything significantly more than the abstract idea. (See, 2106.05(a)(II).)
Claim 27 is, therefore, NOT ELIGIBLE subject matter under 35 U.S.C. § 101.
Claim 28
Step 1 inquiry: Does this claim fall within a statutory category?
The preamble of the claim recites “28. (New) A system, comprising…” Therefore, it is a “system” (or “apparatus”), which is a statutory category of invention. Therefore, the answer to the inquiry is: “YES”.
Step 2A (Prong One) inquiry:
Are there limitations in Claim 28 that recite abstract ideas?
YES. The following limitations in Claim 28 recite abstract ideas that fall within at least one of the groupings of abstract ideas enumerated in the 2019 PEG. Specifically, they are “mental steps” and “mathematical steps”:
• authorization policies defining who can view data samples and meta information in the secure environment (i.e., mental steps)
• generating an anonymized dataset that does not include personally identifiable information (i.e., mental steps)
• generating an outcome using an adversarial network comprising a pair of deep neural networks (i.e., mathematical steps)
• pair of deep neural networks including a generator and a critic (i.e., mathematical steps)
• generating using the generator,..., a first result (i.e., mathematical steps)
• providing the first result to the critic (i.e., mathematical steps)
• generating a first surprise factor (i.e., mathematical steps)
• generate a second result (i.e., mathematical steps)
• providing the second result to the critic (i.e., mathematical steps)
• generating a second surprise factor (i.e., mathematical steps)
• determining, based on the second surprise factor, that the generator has generated a positively surprising result (i.e., mental steps)
• code verification/password verification (i.e., mental steps)
• validating the first result and the second result (i.e., mental steps)
Step 2A (Prong Two) inquiry:
Are there additional elements or a combination of elements in the claim that apply, rely on, or use the judicial exception in a manner that imposes a meaningful limit on the judicial exception, such that it is more than a drafting effort designed to monopolize the exception?
Applicant’s claims contain the following “additional elements”:
(1) A blockchain
(2) A use of the generator
(3) A presentation of the positively surprising result in a graphical user interface
(4) An access control
(5) A Docker image or Kubernetes cluster
(6) A processor
(7) A memory
A “blockchain” is a broad term which is described at a high level. Ahmed, page 2, last full paragraph, where it recites:
Blockchain is shorthand for a suite of distributed ledger technologies that can be programmed to record and track anything of value, such as financial transactions, medical records, land titles, and so on. Blockchain technology is based on the centuries-old method of the general financial ledger. In simplified language, it is a digital ledger which holds the records of all sorts of transactions that happen in a peer-to-peer network. This technology is assumed to ‘cut out the middleman’ from any sort of transaction or transfer of digital assets. This is a much more secure and decentralized medium. Financial institutions are exploring the possibilities of using this technology to ensure secure transactions.
This “blockchain” limitation does not integrate the additional element into a practical application and represents “insignificant extra-solution activity”. (See, M.P.E.P. § 2106.05(I)(A)).
A “use of the generator” is a broad term which is described at a high level. Applicant’s Specification recites:
[0097] In some embodiments, a desired outcome may be generated by using adversarial networks, such as generative neural networks. For example, the desired outcome may be determining a global minimum (i.e. optimization) in a given field, such machine learning model pipeline optimization, parameter or combinatorial optimization, or even determining the most efficient financial portfolio (i.e. stocks, options, currencies, etc.) given certain market performances and histories. In some embodiments, an adversarial network may include a pair of deep neural networks (DNN), called a generator and a critic.
This “use of the generator” limitation does not integrate the additional element into a practical application and represents “insignificant extra-solution activity”. (See, M.P.E.P. § 2106.05(I)(A)).
A “presentation of the positively surprising result in a graphical user interface” is a broad term which is described at a high level. M.P.E.P. § 2106.05 (h) recites in part:
Examples of limitations that the courts have described as merely indicating a field of use or technological environment in which to apply a judicial exception include:
***
vi. Limiting the abstract idea of collecting information, analyzing it, and displaying certain results of the collection and analysis to data related to the electric power grid, because limiting application of the abstract idea to power-grid monitoring is simply an attempt to limit the use of the abstract idea to a particular technological environment, Electric Power Group, LLC v. Alstom S.A., 830 F.3d 1350, 1354, 119 USPQ2d 1739, 1742 (Fed. Cir. 2016);
This “presentation of the positively surprising result in a graphical user interface” limitation does not integrate the additional element into a practical application and represents “insignificant extra-solution activity”. (See, M.P.E.P. § 2106.05(I)(A)).
A “access control” is a broad term which is described at a high level. Applicant’s Specification recites:
[0094] In some embodiments, a representative dataset of the dataset may be created, such that the entity creating a model can use the representative dataset to create the model. Since the representative dataset will have similar properties to the original dataset, the model created via the representative dataset should be similar, if not identical, to a model created via the original dataset. This representative dataset may be created via various methods, in accordance with some embodiments. For example, as illustrated in FIG. 7, a Synthetic Data Generator 712 may be used to generate synthetic data. Data Statistics and Visualizations 714 may be used to describe the original dataset, or Subset Selector and Data Anonymizer 716 may be used to generate an anonymized dataset (i.e. free of personally identifiable information). In some embodiments, one, at least one, or all three of the foregoing may be used to describe the original dataset, the resulting dataset being shown as Anonymous Sample Dataset 716. AlSpecialist 726 may in some embodiments utilize Dev Studio 728 to access Cloud VPS (Virtual Private Server) 730, which has Anonymous Sample Dataset 716. AlSpecialist 716, in 732, may Run or Debug Model on Anonymous Sample Dataset 716. AlSpecialist 726, in some embodiments, creates the code that is required to train a model, including model architecture and training scripts, in Cloud VPS 730. AlSpecialist 726 validates the created code on Anonymous Sample Dataset 716, and debugs the created code before executing it on the original data (i.e. step 732). The created code is then run on the original data in a secure environment, where AlSpecialist 726 does not have any access to. In some embodiments, this secure environment can be Compiled Docker Image for Model 734. Compiled Docker Image for Model 734 may be, for example, a Kubernetes cluster that executes the code created by AlSpecialist 716. Compiled Docker Image for Model 734 may include features such as Code Verification and Access Control 736, which may be used to log code and data usage via blockchain in Log Code & Sample Use (on Blockchain) 738. Compiled Docker Image for Model 734 may also include a secure and private cloud computing environment that is separate from Cloud VPS 730, such as Secure and Private Cloud Computing 738.
Note that “access control” is a feature of the conventional “Compiled Docker Image.”
This “access control” limitation does not integrate the additional element into a practical application and represents “insignificant extra-solution activity”. (See, M.P.E.P. § 2106.05(I)(A)).
A “Docker image or Kubernetes cluster” is a broad term which is described at a high level. Applicant’s Specification recites:
[0094] In some embodiments, a representative dataset of the dataset may be created, such that the entity creating a model can use the representative dataset to create the model. Since the representative dataset will have similar properties to the original dataset, the model created via the representative dataset should be similar, if not identical, to a model created via the original dataset. This representative dataset may be created via various methods, in accordance with some embodiments. For example, as illustrated in FIG. 7, a Synthetic Data Generator 712 may be used to generate synthetic data. Data Statistics and Visualizations 714 may be used to describe the original dataset, or Subset Selector and Data Anonymizer 716 may be used to generate an anonymized dataset (i.e. free of personally identifiable information). In some embodiments, one, at least one, or all three of the foregoing may be used to describe the original dataset, the resulting dataset being shown as Anonymous Sample Dataset 716. AlSpecialist 726 may in some embodiments utilize Dev Studio 728 to access Cloud VPS (Virtual Private Server) 730, which has Anonymous Sample Dataset 716. AlSpecialist 716, in 732, may Run or Debug Model on Anonymous Sample Dataset 716. AlSpecialist 726, in some embodiments, creates the code that is required to train a model, including model architecture and training scripts, in Cloud VPS 730. AlSpecialist 726 validates the created code on Anonymous Sample Dataset 716, and debugs the created code before executing it on the original data (i.e. step 732). The created code is then run on the original data in a secure environment, where AlSpecialist 726 does not have any access to. In some embodiments, this secure environment can be Compiled Docker Image for Model 734. Compiled Docker Image for Model 734 may be, for example, a Kubernetes cluster that executes the code created by AlSpecialist 716. Compiled Docker Image for Model 734 may include features such as Code Verification and Access Control 736, which may be used to log code and data usage via blockchain in Log Code & Sample Use (on Blockchain) 738. Compiled Docker Image for Model 734 may also include a secure and private cloud computing environment that is separate from Cloud VPS 730, such as Secure and Private Cloud Computing 738.
This “Docker image or Kubernetes cluster” limitation does not integrate the additional element into a practical application and represents “insignificant extra-solution activity”. (See, M.P.E.P. § 2106.05(I)(A)).
A “processor” is a broad term which is described at a high level and includes general purpose computers. M.P.E.P. § 2016.05(f) recites:
2106.05(f) Mere Instructions To Apply An Exception [R-10.2019]
Another consideration when determining whether a claim integrates a judicial exception into a practical application in Step 2A Prong Two or recites significantly more than a judicial exception in Step 2B is whether the additional elements amount to more than a recitation of the words “apply it” (or an equivalent) or are more than mere instructions to implement an abstract idea or other exception on a computer. As explained by the Supreme Court, in order to make a claim directed to a judicial exception patent-eligible, the additional element or combination of elements must do “‘more than simply stat[e] the [judicial exception] while adding the words ‘apply it’”. Alice Corp. v. CLS Bank, 573 U.S. 208, 221, 110 USPQ2d 1976, 1982-83 (2014) (quoting Mayo Collaborative Servs. V. Prometheus Labs., Inc., 566 U.S. 66, 72, 101 USPQ2d 1961, 1965). Thus, for example, claims that amount to nothing more than an instruction to apply the abstract idea using a generic computer do not render an abstract idea eligible. Alice Corp., 573 U.S. at 223, 110 USPQ2d at 1983. See also 573 U.S. at 224, 110 USPQ2d at 1984 (warning against a § 101 analysis that turns on “the draftsman’s art”).
Further, M.P.E.P. § 2106.05(f)(2) recites:
(2) Whether the claim invokes computers or other machinery merely as a tool to perform an existing process. Use of a computer or other machinery in its ordinary capacity for economic or other tasks (e.g., to receive, store, or transmit data) or simply adding a general purpose computer or computer components after the fact to an abstract idea (e.g., a fundamental economic practice or mathematical equation) does not integrate a judicial exception into a practical application or provide significantly more. See Affinity Labs v. DirecTV, 838 F.3d 1253, 1262, 120 USPQ2d 1201, 1207 (Fed. Cir. 2016) (cellular telephone); TLI Communications LLC v. AV Auto, LLC, 823 F.3d 607, 613, 118 USPQ2d 1744, 1748 (Fed. Cir. 2016) (computer server and telephone unit). Similarly, “claiming the improved speed or efficiency inherent with applying the abstract idea on a computer” does not integrate a judicial exception into a practical application or provide an inventive concept. Intellectual Ventures I LLC v. Capital One Bank (USA), 792 F.3d 1363, 1367, 115 USPQ2d 1636, 1639 (Fed. Cir. 2015). In contrast, a claim that purports to improve computer capabilities or to improve an existing technology may integrate a judicial exception into a practical application or provide significantly more. McRO, Inc. v. Bandai Namco Games Am. Inc., 837 F.3d 1299, 1314-15, 120 USPQ2d 1091, 1101-02 (Fed. Cir. 2016); Enfish, LLC v. Microsoft Corp., 822 F.3d 1327, 1335-36, 118 USPQ2d 1684, 1688-89 (Fed. Cir. 2016). See MPEP §§ 2106.04(d)(1) and 2106.05(a) for a discussion of improvements to the functioning of a computer or to another technology or technical field.
This “processor” limitation does not integrate the additional element into a practical application and represents “insignificant extra-solution activity”. (See, M.P.E.P. § 2106.05(I)(A)).
A “memory” is a broad term which is described at a high level.
M.P.E.P. § 2106.05(d)(II) recites:
The courts have recognized the following computer functions as well‐understood, routine, and conventional functions when they are claimed in a merely generic manner (e.g., at a high level of generality) or as insignificant extra-solution activity.
***
iv. Storing and retrieving information in memory, Versata Dev. Group, Inc. v. SAP Am., Inc., 793 F.3d 1306, 1334, 115 USPQ2d 1681, 1701 (Fed. Cir. 2015); OIP Techs., 788 F.3d at 1363, 115 USPQ2d at 1092-93;
This “memory” limitation does not integrate the additional element into a practical application and represents “insignificant extra-solution activity”. (See, M.P.E.P. § 2106.05(I)(A)).
The answer to the inquiry is “NO”, no additional elements integrate the claimed abstract idea into a practical application.
Step 2B inquiry:
Does the claim provide an inventive concept, i.e., does the claim recite additional element(s) or a combination of elements that amount to significantly more than the judicial exception in the claim?
Applicant’s claims contain the following “additional elements”:
(1) A blockchain
(2) A use of the generator
(3) A presentation of the positively surprising result in a graphical user interface
(4) An access control
(5) A Docker image or Kubernetes cluster
A “blockchain” is a broad term which is described at a high level. Ahmed, page 2, last full paragraph, where it recites:
Blockchain is shorthand for a suite of distributed ledger technologies that can be programmed to record and track anything of value, such as financial transactions, medical records, land titles, and so on. Blockchain technology is based on the centuries-old method of the general financial ledger. In simplified language, it is a digital ledger which holds the records of all sorts of transactions that happen in a peer-to-peer network. This technology is assumed to ‘cut out the middleman’ from any sort of transaction or transfer of digital assets. This is a much more secure and decentralized medium. Financial institutions are exploring the possibilities of using this technology to ensure secure transactions.
This “blockchain” limitation does not integrate the additional element into a practical application and represents “insignificant extra-solution activity”. (See, M.P.E.P. § 2106.05(I)(A)).
A “use of the generator” is a broad term which is described at a high level. Applicant’s Specification recites:
[0097] In some embodiments, a desired outcome may be generated by using adversarial networks, such as generative neural networks. For example, the desired outcome may be determining a global minimum (i.e. optimization) in a given field, such machine learning model pipeline optimization, parameter or combinatorial optimization, or even determining the most efficient financial portfolio (i.e. stocks, options, currencies, etc.) given certain market performances and histories. In some embodiments, an adversarial network may include a pair of deep neural networks (DNN), called a generator and a critic.
This “use of the generator” limitation does not integrate the additional element into a practical application and represents “insignificant extra-solution activity”. (See, M.P.E.P. § 2106.05(I)(A)).
A “presentation of the positively surprising result in a graphical user interface” is a broad term which is described at a high level. M.P.E.P. § 2106.05 (h) recites in part:
Examples of limitations that the courts have described as merely indicating a field of use or technological environment in which to apply a judicial exception include:
***
vi. Limiting the abstract idea of collecting information, analyzing it, and displaying certain results of the collection and analysis to data related to the electric power grid, because limiting application of the abstract idea to power-grid monitoring is simply an attempt to limit the use of the abstract idea to a particular technological environment, Electric Power Group, LLC v. Alstom S.A., 830 F.3d 1350, 1354, 119 USPQ2d 1739, 1742 (Fed. Cir. 2016);
This “presentation of the positively surprising result in a graphical user interface” limitation does not integrate the additional element into a practical application and represents “insignificant extra-solution activity”. (See, M.P.E.P. § 2106.05(I)(A)).
A “access control” is a broad term which is described at a high level. Applicant’s Specification recites:
[0094] In some embodiments, a representative dataset of the dataset may be created, such that the entity creating a model can use the representative dataset to create the model. Since the representative dataset will have similar properties to the original dataset, the model created via the representative dataset should be similar, if not identical, to a model created via the original dataset. This representative dataset may be created via various methods, in accordance with some embodiments. For example, as illustrated in FIG. 7, a Synthetic Data Generator 712 may be used to generate synthetic data. Data Statistics and Visualizations 714 may be used to describe the original dataset, or Subset Selector and Data Anonymizer 716 may be used to generate an anonymized dataset (i.e. free of personally identifiable information). In some embodiments, one, at least one, or all three of the foregoing may be used to describe the original dataset, the resulting dataset being shown as Anonymous Sample Dataset 716. AlSpecialist 726 may in some embodiments utilize Dev Studio 728 to access Cloud VPS (Virtual Private Server) 730, which has Anonymous Sample Dataset 716. AlSpecialist 716, in 732, may Run or Debug Model on Anonymous Sample Dataset 716. AlSpecialist 726, in some embodiments, creates the code that is required to train a model, including model architecture and training scripts, in Cloud VPS 730. AlSpecialist 726 validates the created code on Anonymous Sample Dataset 716, and debugs the created code before executing it on the original data (i.e. step 732). The created code is then run on the original data in a secure environment, where AlSpecialist 726 does not have any access to. In some embodiments, this secure environment can be Compiled Docker Image for Model 734. Compiled Docker Image for Model 734 may be, for example, a Kubernetes cluster that executes the code created by AlSpecialist 716. Compiled Docker Image for Model 734 may include features such as Code Verification and Access Control 736, which may be used to log code and data usage via blockchain in Log Code & Sample Use (on Blockchain) 738. Compiled Docker Image for Model 734 may also include a secure and private cloud computing environment that is separate from Cloud VPS 730, such as Secure and Private Cloud Computing 738.
Note that “access control” is a feature of the conventional “Compiled Docker Image.”
Therefore, the claim as a whole does not amount to significantly more than the exception itself (i.e., there is no inventive concept in the claim). (See, M.P.E.P. § 2106.05(II)).
A “Docker image or Kubernetes cluster” is a broad term which is described at a high level. Applicant’s Specification recites:
[0094] In some embodiments, a representative dataset of the dataset may be created, such that the entity creating a model can use the representative dataset to create the model. Since the representative dataset will have similar properties to the original dataset, the model created via the representative dataset should be similar, if not identical, to a model created via the original dataset. This representative dataset may be created via various methods, in accordance with some embodiments. For example, as illustrated in FIG. 7, a Synthetic Data Generator 712 may be used to generate synthetic data. Data Statistics and Visualizations 714 may be used to describe the original dataset, or Subset Selector and Data Anonymizer 716 may be used to generate an anonymized dataset (i.e. free of personally identifiable information). In some embodiments, one, at least one, or all three of the foregoing may be used to describe the original dataset, the resulting dataset being shown as Anonymous Sample Dataset 716. AlSpecialist 726 may in some embodiments utilize Dev Studio 728 to access Cloud VPS (Virtual Private Server) 730, which has Anonymous Sample Dataset 716. AlSpecialist 716, in 732, may Run or Debug Model on Anonymous Sample Dataset 716. AlSpecialist 726, in some embodiments, creates the code that is required to train a model, including model architecture and training scripts, in Cloud VPS 730. AlSpecialist 726 validates the created code on Anonymous Sample Dataset 716, and debugs the created code before executing it on the original data (i.e. step 732). The created code is then run on the original data in a secure environment, where AlSpecialist 726 does not have any access to. In some embodiments, this secure environment can be Compiled Docker Image for Model 734. Compiled Docker Image for Model 734 may be, for example, a Kubernetes cluster that executes the code created by AlSpecialist 716. Compiled Docker Image for Model 734 may include features such as Code Verification and Access Control 736, which may be used to log code and data usage via blockchain in Log Code & Sample Use (on Blockchain) 738. Compiled Docker Image for Model 734 may also include a secure and private cloud computing environment that is separate from Cloud VPS 730, such as Secure and Private Cloud Computing 738.
Therefore, the claim as a whole does not amount to significantly more than the exception itself (i.e., there is no inventive concept in the claim). (See, M.P.E.P. § 2106.05(II)).
A “processor” is a broad term which is described at a high level and includes general purpose computers. M.P.E.P. § 2016.05(f) recites:
2106.05(f) Mere Instructions To Apply An Exception [R-10.2019]
Another consideration when determining whether a claim integrates a judicial exception into a practical application in Step 2A Prong Two or recites significantly more than a judicial exception in Step 2B is whether the additional elements amount to more than a recitation of the words “apply it” (or an equivalent) or are more than mere instructions to implement an abstract idea or other exception on a computer. As explained by the Supreme Court, in order to make a claim directed to a judicial exception patent-eligible, the additional element or combination of elements must do “‘more than simply stat[e] the [judicial exception] while adding the words ‘apply it’”. Alice Corp. v. CLS Bank, 573 U.S. 208, 221, 110 USPQ2d 1976, 1982-83 (2014) (quoting Mayo Collaborative Servs. V. Prometheus Labs., Inc., 566 U.S. 66, 72, 101 USPQ2d 1961, 1965). Thus, for example, claims that amount to nothing more than an instruction to apply the abstract idea using a generic computer do not render an abstract idea eligible. Alice Corp., 573 U.S. at 223, 110 USPQ2d at 1983. See also 573 U.S. at 224, 110 USPQ2d at 1984 (warning against a § 101 analysis that turns on “the draftsman’s art”).
Further, M.P.E.P. § 2106.05(f)(2) recites:
(2) Whether the claim invokes computers or other machinery merely as a tool to perform an existing process. Use of a computer or other machinery in its ordinary capacity for economic or other tasks (e.g., to receive, store, or transmit data) or simply adding a general purpose computer or computer components after the fact to an abstract idea (e.g., a fundamental economic practice or mathematical equation) does not integrate a judicial exception into a practical application or provide significantly more. See Affinity Labs v. DirecTV, 838 F.3d 1253, 1262, 120 USPQ2d 1201, 1207 (Fed. Cir. 2016) (cellular telephone); TLI Communications LLC v. AV Auto, LLC, 823 F.3d 607, 613, 118 USPQ2d 1744, 1748 (Fed. Cir. 2016) (computer server and telephone unit). Similarly, “claiming the improved speed or efficiency inherent with applying the abstract idea on a computer” does not integrate a judicial exception into a practical application or provide an inventive concept. Intellectual Ventures I LLC v. Capital One Bank (USA), 792 F.3d 1363, 1367, 115 USPQ2d 1636, 1639 (Fed. Cir. 2015). In contrast, a claim that purports to improve computer capabilities or to improve an existing technology may integrate a judicial exception into a practical application or provide significantly more. McRO, Inc. v. Bandai Namco Games Am. Inc., 837 F.3d 1299, 1314-15, 120 USPQ2d 1091, 1101-02 (Fed. Cir. 2016); Enfish, LLC v. Microsoft Corp., 822 F.3d 1327, 1335-36, 118 USPQ2d 1684, 1688-89 (Fed. Cir. 2016). See MPEP §§ 2106.04(d)(1) and 2106.05(a) for a discussion of improvements to the functioning of a computer or to another technology or technical field.
Therefore, the claim as a whole does not amount to significantly more than the exception itself (i.e., there is no inventive concept in the claim). (See, M.P.E.P. § 2106.05(II)).
A “memory” is a broad term which is described at a high level.
M.P.E.P. § 2106.05(d)(II) recites:
The courts have recognized the following computer functions as well‐understood, routine, and conventional functions when they are claimed in a merely generic manner (e.g., at a high level of generality) or as insignificant extra-solution activity.
***
iv. Storing and retrieving information in memory, Versata Dev. Group, Inc. v. SAP Am., Inc., 793 F.3d 1306, 1334, 115 USPQ2d 1681, 1701 (Fed. Cir. 2015); OIP Techs., 788 F.3d at 1363, 115 USPQ2d at 1092-93;
Therefore, the claim as a whole does not amount to significantly more than the exception itself (i.e., there is no inventive concept in the claim). (See, M.P.E.P. § 2106.05(II)).
Therefore, the answer to the inquiry is “NO”, no additional elements provide an inventive concept that is significantly more than the claimed abstract ideas the claimed abstract idea into a practical application.
Claim 28 is, therefore, NOT ELIGIBLE subject matter under 35 U.S.C. § 101.
Claim 33
Claim 33 recites:
33. The system of claim 28, further comprising:
deploying an application programming interface endpoint; and
controlling who has access to the application programming interface endpoint using the one or more authorization policies.
Applicant’s Claim 33 merely teaches standard computing functions of “deploying an application programming interface endpoint” and “controlling who has access”. It does not integrate the abstract idea to a practical application, nor is it anything significantly more than the abstract idea. (See, 2106.05(a)(II).)
Claim 33 is, therefore, NOT ELIGIBLE subject matter under 35 U.S.C. § 101.
Claim 34
Claim 34 recites:
34. The system of claim 28, wherein: the generator and the critic collaborate through auto-regressive feedback.
Applicant’s Claim 34 merely teaches “generator and the critic collaborate through auto-regressive feedback” (i.e., mathematical steps). It does not integrate the abstract idea to a practical application, nor is it anything significantly more than the abstract idea. (See, 2106.05(a)(II).)
Claim 34 is, therefore, NOT ELIGIBLE subject matter under 35 U.S.C. § 101.
Claim 35
Step 1 inquiry: Does this claim fall within a statutory category?
The preamble of the claim recites “35. A non-transitory computer-readable storage medium comprising instructions, which when executed by a processing device, cause the processing device to…” Therefore, it is a “non-transitory computer-readable storage medium” (or “product of manufacture”). Therefore, the answer to the inquiry is: “YES”.
Step 2A (Prong One) inquiry:
Are there limitations in Claim 35 that recite abstract ideas?
YES. The following limitations in Claim 35 recite abstract ideas that fall within at least one of the groupings of abstract ideas enumerated in the 2019 PEG. Specifically, they are “mental steps” and “mathematical steps”:
• authorization policies defining who can view data samples and meta information in the secure environment (i.e., mental steps)
• generating an anonymized dataset that does not include personally identifiable information (i.e., mental steps)
• generating an outcome using an adversarial network comprising a pair of deep neural networks (i.e., mathematical steps)
• pair of deep neural networks including a generator and a critic (i.e., mathematical steps)
• generating using the generator,..., a first result (i.e., mathematical steps)
• providing the first result to the critic (i.e., mathematical steps)
• generating a first surprise factor (i.e., mathematical steps)
• generate a second result (i.e., mathematical steps)
• providing the second result to the critic (i.e., mathematical steps)
• generating a second surprise factor (i.e., mathematical steps)
• determining, based on the second surprise factor, that the generator has generated a positively surprising result (i.e., mental steps)
• code verification/password verification (i.e., mental steps)
• validating the first result and the second result (i.e., mental steps)
Step 2A (Prong Two) inquiry:
Are there additional elements or a combination of elements in the claim that apply, rely on, or use the judicial exception in a manner that imposes a meaningful limit on the judicial exception, such that it is more than a drafting effort designed to monopolize the exception?
Applicant’s claims contain the following “additional elements”:
(1) A blockchain
(2) A use of the generator
(3) A presentation of the positively surprising result in a graphical user interface
(4) An access control
(5) A Docker image or Kubernetes cluster
A “blockchain” is a broad term which is described at a high level. Ahmed, page 2, last full paragraph, where it recites:
Blockchain is shorthand for a suite of distributed ledger technologies that can be programmed to record and track anything of value, such as financial transactions, medical records, land titles, and so on. Blockchain technology is based on the centuries-old method of the general financial ledger. In simplified language, it is a digital ledger which holds the records of all sorts of transactions that happen in a peer-to-peer network. This technology is assumed to ‘cut out the middleman’ from any sort of transaction or transfer of digital assets. This is a much more secure and decentralized medium. Financial institutions are exploring the possibilities of using this technology to ensure secure transactions.
This “blockchain” limitation does not integrate the additional element into a practical application and represents “insignificant extra-solution activity”. (See, M.P.E.P. § 2106.05(I)(A)).
A “use of the generator” is a broad term which is described at a high level. Applicant’s Specification recites:
[0097] In some embodiments, a desired outcome may be generated by using adversarial networks, such as generative neural networks. For example, the desired outcome may be determining a global minimum (i.e. optimization) in a given field, such machine learning model pipeline optimization, parameter or combinatorial optimization, or even determining the most efficient financial portfolio (i.e. stocks, options, currencies, etc.) given certain market performances and histories. In some embodiments, an adversarial network may include a pair of deep neural networks (DNN), called a generator and a critic.
This “use of the generator” limitation does not integrate the additional element into a practical application and represents “insignificant extra-solution activity”. (See, M.P.E.P. § 2106.05(I)(A)).
A “presentation of the positively surprising result in a graphical user interface” is a broad term which is described at a high level. M.P.E.P. § 2106.05 (h) recites in part:
Examples of limitations that the courts have described as merely indicating a field of use or technological environment in which to apply a judicial exception include:
***
vi. Limiting the abstract idea of collecting information, analyzing it, and displaying certain results of the collection and analysis to data related to the electric power grid, because limiting application of the abstract idea to power-grid monitoring is simply an attempt to limit the use of the abstract idea to a particular technological environment, Electric Power Group, LLC v. Alstom S.A., 830 F.3d 1350, 1354, 119 USPQ2d 1739, 1742 (Fed. Cir. 2016);
This “presentation of the positively surprising result in a graphical user interface” limitation does not integrate the additional element into a practical application and represents “insignificant extra-solution activity”. (See, M.P.E.P. § 2106.05(I)(A)).
A “access control” is a broad term which is described at a high level. Applicant’s Specification recites:
[0094] In some embodiments, a representative dataset of the dataset may be created, such that the entity creating a model can use the representative dataset to create the model. Since the representative dataset will have similar properties to the original dataset, the model created via the representative dataset should be similar, if not identical, to a model created via the original dataset. This representative dataset may be created via various methods, in accordance with some embodiments. For example, as illustrated in FIG. 7, a Synthetic Data Generator 712 may be used to generate synthetic data. Data Statistics and Visualizations 714 may be used to describe the original dataset, or Subset Selector and Data Anonymizer 716 may be used to generate an anonymized dataset (i.e. free of personally identifiable information). In some embodiments, one, at least one, or all three of the foregoing may be used to describe the original dataset, the resulting dataset being shown as Anonymous Sample Dataset 716. AlSpecialist 726 may in some embodiments utilize Dev Studio 728 to access Cloud VPS (Virtual Private Server) 730, which has Anonymous Sample Dataset 716. AlSpecialist 716, in 732, may Run or Debug Model on Anonymous Sample Dataset 716. AlSpecialist 726, in some embodiments, creates the code that is required to train a model, including model architecture and training scripts, in Cloud VPS 730. AlSpecialist 726 validates the created code on Anonymous Sample Dataset 716, and debugs the created code before executing it on the original data (i.e. step 732). The created code is then run on the original data in a secure environment, where AlSpecialist 726 does not have any access to. In some embodiments, this secure environment can be Compiled Docker Image for Model 734. Compiled Docker Image for Model 734 may be, for example, a Kubernetes cluster that executes the code created by AlSpecialist 716. Compiled Docker Image for Model 734 may include features such as Code Verification and Access Control 736, which may be used to log code and data usage via blockchain in Log Code & Sample Use (on Blockchain) 738. Compiled Docker Image for Model 734 may also include a secure and private cloud computing environment that is separate from Cloud VPS 730, such as Secure and Private Cloud Computing 738.
Note that “access control” is a feature of the conventional “Compiled Docker Image.”
This “access control” limitation does not integrate the additional element into a practical application and represents “insignificant extra-solution activity”. (See, M.P.E.P. § 2106.05(I)(A)).
A “Docker image or Kubernetes cluster” is a broad term which is described at a high level. Applicant’s Specification recites:
[0094] In some embodiments, a representative dataset of the dataset may be created, such that the entity creating a model can use the representative dataset to create the model. Since the representative dataset will have similar properties to the original dataset, the model created via the representative dataset should be similar, if not identical, to a model created via the original dataset. This representative dataset may be created via various methods, in accordance with some embodiments. For example, as illustrated in FIG. 7, a Synthetic Data Generator 712 may be used to generate synthetic data. Data Statistics and Visualizations 714 may be used to describe the original dataset, or Subset Selector and Data Anonymizer 716 may be used to generate an anonymized dataset (i.e. free of personally identifiable information). In some embodiments, one, at least one, or all three of the foregoing may be used to describe the original dataset, the resulting dataset being shown as Anonymous Sample Dataset 716. AlSpecialist 726 may in some embodiments utilize Dev Studio 728 to access Cloud VPS (Virtual Private Server) 730, which has Anonymous Sample Dataset 716. AlSpecialist 716, in 732, may Run or Debug Model on Anonymous Sample Dataset 716. AlSpecialist 726, in some embodiments, creates the code that is required to train a model, including model architecture and training scripts, in Cloud VPS 730. AlSpecialist 726 validates the created code on Anonymous Sample Dataset 716, and debugs the created code before executing it on the original data (i.e. step 732). The created code is then run on the original data in a secure environment, where AlSpecialist 726 does not have any access to. In some embodiments, this secure environment can be Compiled Docker Image for Model 734. Compiled Docker Image for Model 734 may be, for example, a Kubernetes cluster that executes the code created by AlSpecialist 716. Compiled Docker Image for Model 734 may include features such as Code Verification and Access Control 736, which may be used to log code and data usage via blockchain in Log Code & Sample Use (on Blockchain) 738. Compiled Docker Image for Model 734 may also include a secure and private cloud computing environment that is separate from Cloud VPS 730, such as Secure and Private Cloud Computing 738.
This “Docker image or Kubernetes cluster” limitation does not integrate the additional element into a practical application and represents “insignificant extra-solution activity”. (See, M.P.E.P. § 2106.05(I)(A)).
The answer to the inquiry is “NO”, no additional elements integrate the claimed abstract idea into a practical application.
Step 2B inquiry:
Does the claim provide an inventive concept, i.e., does the claim recite additional element(s) or a combination of elements that amount to significantly more than the judicial exception in the claim?
Applicant’s claims contain the following “additional elements”:
(1) A blockchain
(2) A use of the generator
(3) A presentation of the positively surprising result in a graphical user interface
(4) An access control
(5) A Docker image or Kubernetes cluster
A “blockchain” is a broad term which is described at a high level. Ahmed, page 2, last full paragraph, where it recites:
Blockchain is shorthand for a suite of distributed ledger technologies that can be programmed to record and track anything of value, such as financial transactions, medical records, land titles, and so on. Blockchain technology is based on the centuries-old method of the general financial ledger. In simplified language, it is a digital ledger which holds the records of all sorts of transactions that happen in a peer-to-peer network. This technology is assumed to ‘cut out the middleman’ from any sort of transaction or transfer of digital assets. This is a much more secure and decentralized medium. Financial institutions are exploring the possibilities of using this technology to ensure secure transactions.
Therefore, the claim as a whole does not amount to significantly more than the exception itself (i.e., there is no inventive concept in the claim). (See, M.P.E.P. § 2106.05(II)).
A “use of the generator” is a broad term which is described at a high level. Applicant’s Specification recites:
[0097] In some embodiments, a desired outcome may be generated by using adversarial networks, such as generative neural networks. For example, the desired outcome may be determining a global minimum (i.e. optimization) in a given field, such machine learning model pipeline optimization, parameter or combinatorial optimization, or even determining the most efficient financial portfolio (i.e. stocks, options, currencies, etc.) given certain market performances and histories. In some embodiments, an adversarial network may include a pair of deep neural networks (DNN), called a generator and a critic.
Therefore, the claim as a whole does not amount to significantly more than the exception itself (i.e., there is no inventive concept in the claim). (See, M.P.E.P. § 2106.05(II)).
A “presentation of the positively surprising result in a graphical user interface” is a broad term which is described at a high level. M.P.E.P. § 2106.05 (h) recites in part:
Examples of limitations that the courts have described as merely indicating a field of use or technological environment in which to apply a judicial exception include:
***
vi. Limiting the abstract idea of collecting information, analyzing it, and displaying certain results of the collection and analysis to data related to the electric power grid, because limiting application of the abstract idea to power-grid monitoring is simply an attempt to limit the use of the abstract idea to a particular technological environment, Electric Power Group, LLC v. Alstom S.A., 830 F.3d 1350, 1354, 119 USPQ2d 1739, 1742 (Fed. Cir. 2016);
Therefore, the claim as a whole does not amount to significantly more than the exception itself (i.e., there is no inventive concept in the claim). (See, M.P.E.P. § 2106.05(II)).
A “access control” is a broad term which is described at a high level. Applicant’s Specification recites:
[0094] In some embodiments, a representative dataset of the dataset may be created, such that the entity creating a model can use the representative dataset to create the model. Since the representative dataset will have similar properties to the original dataset, the model created via the representative dataset should be similar, if not identical, to a model created via the original dataset. This representative dataset may be created via various methods, in accordance with some embodiments. For example, as illustrated in FIG. 7, a Synthetic Data Generator 712 may be used to generate synthetic data. Data Statistics and Visualizations 714 may be used to describe the original dataset, or Subset Selector and Data Anonymizer 716 may be used to generate an anonymized dataset (i.e. free of personally identifiable information). In some embodiments, one, at least one, or all three of the foregoing may be used to describe the original dataset, the resulting dataset being shown as Anonymous Sample Dataset 716. AlSpecialist 726 may in some embodiments utilize Dev Studio 728 to access Cloud VPS (Virtual Private Server) 730, which has Anonymous Sample Dataset 716. AlSpecialist 716, in 732, may Run or Debug Model on Anonymous Sample Dataset 716. AlSpecialist 726, in some embodiments, creates the code that is required to train a model, including model architecture and training scripts, in Cloud VPS 730. AlSpecialist 726 validates the created code on Anonymous Sample Dataset 716, and debugs the created code before executing it on the original data (i.e. step 732). The created code is then run on the original data in a secure environment, where AlSpecialist 726 does not have any access to. In some embodiments, this secure environment can be Compiled Docker Image for Model 734. Compiled Docker Image for Model 734 may be, for example, a Kubernetes cluster that executes the code created by AlSpecialist 716. Compiled Docker Image for Model 734 may include features such as Code Verification and Access Control 736, which may be used to log code and data usage via blockchain in Log Code & Sample Use (on Blockchain) 738. Compiled Docker Image for Model 734 may also include a secure and private cloud computing environment that is separate from Cloud VPS 730, such as Secure and Private Cloud Computing 738.
Note that “access control” is a feature of the conventional “Compiled Docker Image.”
Therefore, the claim as a whole does not amount to significantly more than the exception itself (i.e., there is no inventive concept in the claim). (See, M.P.E.P. § 2106.05(II)).
A “Docker image or Kubernetes cluster” is a broad term which is described at a high level. Applicant’s Specification recites:
[0094] In some embodiments, a representative dataset of the dataset may be created, such that the entity creating a model can use the representative dataset to create the model. Since the representative dataset will have similar properties to the original dataset, the model created via the representative dataset should be similar, if not identical, to a model created via the original dataset. This representative dataset may be created via various methods, in accordance with some embodiments. For example, as illustrated in FIG. 7, a Synthetic Data Generator 712 may be used to generate synthetic data. Data Statistics and Visualizations 714 may be used to describe the original dataset, or Subset Selector and Data Anonymizer 716 may be used to generate an anonymized dataset (i.e. free of personally identifiable information). In some embodiments, one, at least one, or all three of the foregoing may be used to describe the original dataset, the resulting dataset being shown as Anonymous Sample Dataset 716. AlSpecialist 726 may in some embodiments utilize Dev Studio 728 to access Cloud VPS (Virtual Private Server) 730, which has Anonymous Sample Dataset 716. AlSpecialist 716, in 732, may Run or Debug Model on Anonymous Sample Dataset 716. AlSpecialist 726, in some embodiments, creates the code that is required to train a model, including model architecture and training scripts, in Cloud VPS 730. AlSpecialist 726 validates the created code on Anonymous Sample Dataset 716, and debugs the created code before executing it on the original data (i.e. step 732). The created code is then run on the original data in a secure environment, where AlSpecialist 726 does not have any access to. In some embodiments, this secure environment can be Compiled Docker Image for Model 734. Compiled Docker Image for Model 734 may be, for example, a Kubernetes cluster that executes the code created by AlSpecialist 716. Compiled Docker Image for Model 734 may include features such as Code Verification and Access Control 736, which may be used to log code and data usage via blockchain in Log Code & Sample Use (on Blockchain) 738. Compiled Docker Image for Model 734 may also include a secure and private cloud computing environment that is separate from Cloud VPS 730, such as Secure and Private Cloud Computing 738.
Therefore, the claim as a whole does not amount to significantly more than the exception itself (i.e., there is no inventive concept in the claim). (See, M.P.E.P. § 2106.05(II)).
Therefore, the answer to the inquiry is “NO”, no additional elements provide an inventive concept that is significantly more than the claimed abstract ideas the claimed abstract idea into a practical application.
Claim 35 is, therefore, NOT ELIGIBLE subject matter under 35 U.S.C. § 101.
Claim 36
Claim 36 recites:
36. The non-transitory computer-readable storage medium of claim 35, wherein the generator has a goal to generate a new population of individuals that are seen as underdogs by the critic and will surprise the critic.
Applicant’s Claim 36 merely teaches the mental step of “a goal”. It does not integrate the abstract idea to a practical application, nor is it anything significantly more than the abstract idea. (See, 2106.05(a)(II).)
Claim 36 is, therefore, NOT ELIGIBLE subject matter under 35 U.S.C. § 101.
Claim 37
Claim 37 recites:
37. The non-transitory computer-readable storage medium of claim 36, wherein the critic undergoes a training iteration to learn rewards that emerge from individual members of the population allowing continuously balancing between exploration and exploitation.
Applicant’s Claim 37 merely teaches “a training iteration” (i.e., mathematical steps). It does not integrate the abstract idea to a practical application, nor is it anything significantly more than the abstract idea. (See, 2106.05(a)(II).)
Claim 37 is, therefore, NOT ELIGIBLE subject matter under 35 U.S.C. § 101.
Claim 38
Claim 38 recites:
38. The non-transitory computer-readable storage medium of claim 35, wherein the critic gives the generator feedback by sharing a perspective on a generated population.
Applicant’s Claim 38 merely teaches “the critic gives the generator feedback by sharing a perspective” (i.e., mathematical step). It does not integrate the abstract idea to a practical application, nor is it anything significantly more than the abstract idea. (See, 2106.05(a)(II).)
Claim 38 is, therefore, NOT ELIGIBLE subject matter under 35 U.S.C. § 101.
Claim 39
Claim 39 recites:
39. The non-transitory computer-readable storage medium of claim 38, wherein when the generator accomplishes surprising the critic with surprising individuals that exceed the critic's expectations, the generator allocates more resources on exploiting areas from where the surprises originated.
Applicant’s Claim 39 merely teaches “the generator allocates more resources on exploiting areas from where the surprises originated” (i.e., mathematical step). It does not integrate the abstract idea to a practical application, nor is it anything significantly more than the abstract idea. (See, 2106.05(a)(II).)
Claim 39 is, therefore, NOT ELIGIBLE subject matter under 35 U.S.C. § 101.
Claim 40
Claim 40 recites:
40. The non-transitory computer-readable storage medium of claim 35, wherein when specific candidates perform worse than the critic's expectation, the critic does not risk spending more resources and allocates the resources to explore areas with surprising individuals.
Note that the term “resource” is defined in Applicant's Specification, paragraph [0098]:
[0098] In some embodiments, an evolutionary swarm optimization method ensures the quality and diversity of a population of individual trials during an optimization process by continuously seeking surprising individuals through a relationship constructed between the generator and the critic. In this way, the generator seeks to positively surprise the critic, rather than deceiving a discriminator, and they are more friendly as the two collaborate through auto-regressive feedback, with the critic helping the generator in its goal of improving the quality-diversity of the population. In each iteration, the generator tries generating a new population of individuals who would be seen as underdogs by the critic and will outperform its expectations (i.e. positively surprise the critic). The critic then undergoes training iteration separately for learning the rewards that emerge from individual members of the population. This allows continuously balancing between exploration and exploitation during the optimization process. If the generator keeps generating the same individuals, the critic would quickly develop a reasonable estimate of their behavior, and hence they could not be considered underdogs anymore (i.e. the critic is not surprised anymore). In each iteration, the critic also gives the generator feedback by sharing its perspective on the generated population. If the generator accomplishes surprising the critic with novel or surprising individuals with their exceeding the critic's expectations, the generator would allocate more resources (population members) on exploiting those areas where those surprises originate. If specific candidates or a candidate class performs worse than the critic's expectation, the critic would not risk spending more resources on such candidates or class, and allocate such resources to explore or exploit areas with surprising individuals.
Note that the claimed resources are mathematical population members in a mathematical machine learning space (i.e., evolutionary swarm optimization).
Applicant’s Claim 40 merely teaches the allocation of more mathematical “resources” (i.e., mathematical particles in a “swarm optimization)”. It does not integrate the abstract idea to a practical application, nor is it anything significantly more than the abstract idea. (See, 2106.05(a)(II).)
Claim 40 is, therefore, NOT ELIGIBLE subject matter under 35 U.S.C. § 101.
Response to Arguments
Applicant's arguments filed 12 MAR 2026 have been fully considered but they are not persuasive. Specifically, Applicant argues:
Argument 1
…claim 27 has been amended to correct an antecedent basis issue (dependency changed from claim 24 to claim 26).
Amendment noted.
Argument 2
Rejections of Claims 21-28 and 33-40 under 35 U.S.C. § 112(b)
The Office Action rejected claims 21-28 and 33-40 under 35 U.S.C. § 112(b) as indefinite, specifically asserting that the term "surprise" is not defined by the claim, the specification does not provide a standard for ascertaining the requisite degree, and one of ordinary skill in the art would not be reasonably apprised of the scope of the invention. See Office Action, pp. 2-3.
Applicant respectfully traverses this rejection.
I. The Claims Have Been Amended
Applicant has amended the independent claims to recite "positively surprised," which clarifies the directional and functional nature of the surprise relationship between the generator and critic. Support for this amendment can be found in the filed specification at paragraph [0098]. Applicant submits that these amendments will overcome the § 112(b) rejection. Nevertheless, Applicant further addresses the examiner's position on the merits below.
II. The Legal Standard
To comply with 35 U.S.C. § 112(b), claim terms must inform those skilled in the art about the scope of the invention with reasonable certainty, read in light of the specification and prosecution history. Nautilus, Inc. V. Biosig Instruments, Inc., 572 U.S. 898, 910 (2014). The standard does not require mathematical precision or explicit numerical values - it requires only that a person of ordinary skill in the art (POSITA) be able to ascertain the scope of the claimed subject matter with reasonable certainty. Id.
Further, in making a prima facie case of indefiniteness, the examiner must point out the specific term or phrase that is indefinite, explain in detail why the term renders the metes and bounds of the claim scope unclear, and, whenever practicable, indicate how the indefiniteness issues may be resolved. MPEP § 2173.02. Critically, claim terms must be analyzed in view of the specification from the perspective of those skilled in the relevant art, as a term used in one application may carry a precise technical meaning not immediately apparent from its ordinary dictionary definition. Medrad, Inc. v. MRI Devices Corp., 401 F.3d 1313, 1318, 74 USPQ2d 1184, 1188 (Fed. Cir. 2005).
III. "Surprise" Is a Term of Art with Defined Mathematical Meaning in the Context of Adversarial Networks
The examiner's rejection treats "surprise" as a subjective, human-facing term. Applicant respectfully submits this mischaracterizes both the specification and the technical field.
First, the specification defines "surprise" operationally and algorithmically through the generator/critic relationship. Paragraph [0098] discloses that the generator seeks to produce individuals that "outperform [the critic's] expectations" - i.e., exceed the critic's predicted reward score. This is a mathematical comparison between a predicted value and an actual computed output, not a subjective human judgment.
Applicant argues that:
…the generator seeks to produce individuals that "outperform [the critic's] expectations" - i.e., exceed the critic's predicted reward score.
Note that Applicant does not define “performance” or “outperform”. There is no clear standard of measurement.
Further, Paragraph [0098] discusses “reward” separately from “surprise” and the term “reward” is not used to define the term “surprise”.
Further, Applicant argues that the determination of a “surprise” is a “mathematical comparison”:
This is a mathematical comparison between a predicted value and an actual computed output, not a subjective human judgment.
Though it is clear that Applicant intends the “surprise” to be a calculated mathematical quantity (and therefore, abstract), the actual definition of how to clearly determine the term is missing. The closest thing in the Specification to what Applicant argues is “prediction error”, but that is not used in the Specification to define “surprise,” either.
Further, Applicant's argued “surprise” is not the only possible formulation for the term. “Surprise” could also be represented on an information theoretic basis by Entropy = - p log(p), where in one extreme, p=1, causing the equation to be -log(1) = 0 (i.e., no surprise); and in the other extreme, p = 0, causing the equation to go to infinity (i.e., infinite surprise.) Though this is a possible meaning of the term, this definition is not in the Specification either.
There are other possible formulations, as well…even fuzzy number representations. There is nothing in the claims or Specification defining what measure of surprise is calculated, how it is calculated, etc.
Though Applicant argues a particular method, even that method is not actually shown in the Specification. The method of measurement could even be subjective. Therefore, Applicant does not provide a clear standard of measurement for “surprise”. The term is indefinite.
Applicant's argument is unpersuasive.
The rejections stand.
Argument 3
Second, the specification discloses a concrete, objective "surprise factor threshold" mechanism that provides the precise standard the examiner claims is missing. Paragraphs [0101]-[0102] disclose that: (a) a surprise factor is computed and compared against a threshold; (b) if the threshold is met, the generator continues exploiting that area; (c) if the threshold is not met, the generator discontinues that area and explores new ones; (d) if the threshold is not met multiple times, the threshold is lowered; and (e) if the threshold is met multiple times consecutively, the threshold is raised. This is an explicit, rule-based, computable mechanism - precisely the kind of objective standard that satisfies § 112(b).
Third, in the field of adversarial and evolutionary machine learning - the relevant art here - "surprise" carries a well-understood technical meaning. It corresponds to the degree by which an individual candidate's performance exceeds the model's predicted estimate, analogous to the concept of "surprisal" in information theory (-log P(x)). A POSITA in this field would understand "surprise factor" as a computable value derived from the adversarial interaction between a generator and a critic, not as an indeterminate human emotion. The examiner's own § 101 analysis acknowledges as much by characterizing these same limitations as "mathematical steps" - a characterization that is fundamentally inconsistent with a finding of indefiniteness on the grounds of subjective imprecision.
Fourth, the Federal Circuit has consistently held that relative terms are not indefinite per se where the specification provides context sufficient for a POSITA to determine scope with reasonable certainty. Sonix Tech. Co. V. Publ'ns Int'l, Ltd., 844 F.3d 1370, 1377-78 (Fed. Cir. 2017) (holding "visually negligible" definite where the specification provided objective guidance). Here, Applicant's disclosure goes well beyond the guidance found sufficient in Sonix - the specification provides objective function calculations, loss functions, threshold adjustment rules, and a fully described adversarial network architecture from which a POSITA would understand precisely how "surprise" is calculated and determined.
For at least these reasons, the term "surprise," read in light of the specification, informs a POSITA of the scope of the claimed invention with reasonable certainty and does not render the claims indefinite under § 112(b).
Accordingly, withdrawal and/or reconsideration of the rejection of claims 21-28 and 33-40 under 35 U.S.C. § 112(b) is respectfully requested.
Though Applicant argued a particular method in the various arguments above, even that method is not clearly shown in the Specification. The method of measurement could even be subjective. Therefore, Applicant does not provide a clear standard of measurement for “surprise”. The term is indefinite.
Applicant's argument is unpersuasive.
The rejections stand.
Argument 4
II. The Specification Discloses Improvements to ML Training
The specification discloses two interlocking technical improvements that distinguish this invention from the prior art and establish its eligibility under §101.
First-Sequentially Updated Model Weights in an Adversarial-Coupled Architecture: Paragraph [00100] of the specification discloses the specific training mechanism at the core of the invention. After computing respective losses, both models' weights are sequentially updated: the generator minimizes a prediction-error-based novelty loss (Eq. 3), and the critic separately minimizes a mean-squared-error loss on its objective function estimation (Eq. 4), with gradient-norm rescaling applied to stabilize both models' training. The specification further explains at paragraph [0098] that the critic undergoes its training iteration separately from the generator. This sequential, alternating weight-update architecture is a operational parameter of how the adversarial network functions as a machine-it is not an abstract concept.
The only support for this in the Specification, paragraph [00100] where it recites:
…both models' weights are sequentially updated using Quasi-Hyperbolic Adam…
Firstly, the Quasi-Hyperbolic Adam is fundamentally a mathematical process, specifically an optimization algorithm used in machine learning for training neural networks. It operates by performing a weighted average of two different optimization approaches—plain Stochastic Gradient Descent (SGD) and momentum SGD—to update model parameters based on gradients.
Secondly, even if sequentially updated weights were to be regarded as an “additional element”, it is well-known, routine, and conventional, as shown in page 5, last full paragraph of the following 2019 evidence: Ma, et al., "Quasi-Hyperbolic Momentum and Adam for Deep Learning", arXiv:1810.06801v4, 02 MAY 2019, pp. 1-38.
Applicant's argument is unpersuasive.
The rejections stand.
Argument 5
Second-Surprise Factor as a Stabilization Mechanism: The specification at paragraph [00100] further discloses that the surprise factor-computed from the critic's prediction error on the generator's output-functions as the operational signal that drives and stabilizes the sequential coupled training of the two-network system. The generator's loss depends on the critic's estimation error for evaluating novelty, and the specification expressly distinguishes this mechanism from the prior art: "using the prediction-error is superior as each individual's novelty is not only measured against the current population or previous ones that are memorized; but also against those that could be predicted by learning them." Spec. [00100]. This explicit distinction over conventional k-nearest-neighbor diversity heuristics establishes that the surprise factor mechanism is not routine or conventional.
Nothing in paragraph [00100] clearly, expressly discusses surprise factor.
Applicant's argument is unpersuasive.
The rejections stand.
Argument 6
Third-Blockchain-Enforced Secure Containerized Execution: The specification at paragraphs [0022] and [0094] discloses that the invention enables an AI specialist to develop and refine a model against an anonymized dataset in a Cloud VPS environment, after which validated training code is compiled into a containerized runtime (Docker image or Kubernetes cluster) and executed against the original sensitive data in a blockchain-enforced secure environment that the specialist cannot directly access. Paragraph [0023] explains that the containerized environment enables efficient packaging of the AI pipeline's training code and dependencies as modular, portable components. The blockchain logs code and data usage immutably, ensuring integrity of the training provenance. This two-stage workflow is a specific solution to the problem of enabling ML model optimization over sensitive data without exposing that data to the model developer.
Applicant's argument regarding a “containerized environment” is generic and may refer to a “docker image” or a “Kubernetes cluster”, among other options. Similarly, a claim to a “docker image” could be a “Kubernetes cluster”, among other options (see paragraph [0094] in the Specification).
[0094] In some embodiments, a representative dataset of the dataset may be created, such that the entity creating a model can use the representative dataset to create the model.
***
In some embodiments, this secure environment can be Compiled Docker Image for Model 734. Compiled Docker Image for Model 734 may be, for example, a Kubernetes cluster that executes the code created by AlSpecialist 716. Compiled Docker Image for Model 734 may include features such as Code Verification and Access Control 736, which may be used to log code and data usage via blockchain in Log Code & Sample Use (on Blockchain) 738. Compiled Docker Image for Model 734 may also include a secure and private cloud computing environment that is separate from Cloud VPS 730, such as Secure and Private Cloud Computing 738.
This means that claim to a “docker image” is generic, well understood, routine, and conventional.
Applicant's argument is unpersuasive.
The rejections stand.
Argument 7
III. The Amended Claims Integrate the Improvements into a Practical Application
Independent claims 21, 28, and 35 have been amended to add two coordinated limitations that directly tie the claims to the concrete technical improvements disclosed in the specification:(1) the adversarial network limitation now recites an adversarial network "with sequentially updated model weights"; and (2) each surprise factor generation step now recites the purpose of stabilizing coupled model training-each surprise factor is generated "to stabilize coupled model training" based on the respective result. These two amendments work in concert to reflect the technical improvement of the invention as disclosed in the specification. Additionally, Claim 35 was amended to recite "non-transitory"; to avoid threshold issues Examiner indicated.
The amended claims, read as an ordered combination, recite the following specific technical sequence: a blockchain-enforced secure environment is created based on authorization policies (enabling data-isolated ML execution); an anonymized dataset is generated; an adversarial network specifically configured with sequentially updated model weights generates results within a containerized runtime environment; those results are provided to a critic and evaluated; the resulting surprise factor-generated to stabilize coupled model training-is used to guide the generator in producing further results through an iterative adversarial loop; and the process terminates when the most surprising result is identified, validated on the blockchain, and presented in a graphical user interface. This ordered combination implements a specific, concrete ML training architecture. It is not the abstract idea of optimization, or of using a neural network, or of computing a mathematical value. It is a particular adversarial training system in which the specific operational mechanism (sequential weight updates stabilized by the surprise factor) solves the concrete technical problem of balancing exploration and exploitation in global optimization without divergence of the coupled training loop.
This is precisely what Desjardins requires: the claim must "include the components or steps of the invention that provide the improvement described in the specification." Desjardins at *9. The improvement described in the specification is the specific adversarially-coupled training architecture in which two deep neural networks are jointly trained through sequential weight updates, driven by the surprise factor as a stabilization signal. The amended claims now recite exactly that architecture. The claims reflect the improvement. Per MPEP §2106.05(a)(I), Example xiv, claims directed to improvements to computer performance based on adjustments to parameters of an ML model associated with specific tasks are eligible, and the sequential weight update architecture with surprise-factor-driven stabilization is directly analogous.
Regarding the argued “sequentially updated model weights" and the argued “surprise factor generation step”, those issues were discussed in prior arguments regarding claim 21. Similar limitations in similar claims 28, and 35 are similarly unpersuasive.
The sequentially updated weights are purely mathematical.
Even if sequentially updated weights were to be regarded as “additional elements”, they are well-known, routine, and conventional, as shown in page 5, last full paragraph of the following 2019 evidence: Ma, et al., "Quasi-Hyperbolic Momentum and Adam for Deep Learning", arXiv:1810.06801v4, 02 MAY 2019, pp. 1-38.
Regarding the other independent claims 28, and 35, similar arguments for similar claims are similarly unpersuasive.
Applicant's argument is unpersuasive.
The rejections stand.
Argument 8
Furthermore, the Office Action analyzed the claim elements in isolation, dismissing each as individually abstract or as generic computer components, without performing the ordered-combination analysis required by Desjardins and MPEP §2106. The Examiner characterized the blockchain as "insignificant extra-solution activity," the containerized runtime as a generic computer component, and the generator/critic pair as abstract mathematical operations-all evaluated in isolation. Desjardins expressly prohibits this approach: "examiners and panels should not evaluate claims at such a high level of generality that potentially meaningful technical limitations are dismissed without adequate explanation." Desjardins at *8. The blockchain and containerized runtime are not incidental-they are structural elements of the secure execution environment in which the improved training architecture operates. They are part of the particular solution, not mere post-solution activity.
The claims merely teach operating on “a blockchain”. They do not improve the blockchain and do not specify which type of blockchain, since there are many. Applicant's reference to the blockchain is generic, well-understood, routine, and conventional.
Applicant's argument regarding a “containerized environment” is generic and may refer to a “docker image” or a “Kubernetes cluster”, among other options. Similarly, a claim to a “docker image” could be a “Kubernetes cluster”, among other options (see paragraph [0094] in the Specification).
[0094] In some embodiments, a representative dataset of the dataset may be created, such that the entity creating a model can use the representative dataset to create the model.
***
In some embodiments, this secure environment can be Compiled Docker Image for Model 734. Compiled Docker Image for Model 734 may be, for example, a Kubernetes cluster that executes the code created by AlSpecialist 716. Compiled Docker Image for Model 734 may include features such as Code Verification and Access Control 736, which may be used to log code and data usage via blockchain in Log Code & Sample Use (on Blockchain) 738. Compiled Docker Image for Model 734 may also include a secure and private cloud computing environment that is separate from Cloud VPS 730, such as Secure and Private Cloud Computing 738.
This means that a claim to a “docker image” is generic, well understood, routine, and conventional.
The claimed the generator/critic pair (i.e., GAN) is pure mathematical operations and, if regarded as an additional element, would be well-understood, routine and conventional.
Applicant's argument is unpersuasive.
The rejections stand.
Argument 9
IV. The Claimed Elements Are Not Well-Understood, Routine, or Conventional
Even assuming arguendo that the claims were not found to integrate the exception into a practical application at Step 2A Prong Two-which Applicant disputes-the rejection would still fail at Step 2B. The Examiner's assertion that the combination of elements is routine and conventional is unsupported by evidence. Under Berkheimer, 890 F.3d at 1374, whether additional elements are well-understood, routine, and conventional is a question of fact. The Examiner must identify evidentiary support for such a finding. No such evidence has been provided.
The specification expressly identifies the claimed combination as an improvement to the conventional approach. Specifically, with respect to the sequential weight update architecture driven by the surprise factor, the specification states at paragraph [00100] that the prediction- error-based novelty mechanism "is superior as each individual's novelty is not only measured against the current population or previous ones that are memorized; but also against those that could be predicted by learning them." Spec. [00100]. This affirmative disclosure that the claimed mechanism outperforms the conventional k-nearest-neighbor diversity heuristic constitutes intrinsic evidence that the claimed combination was not well-understood, routine, or conventional. The Examiner has not identified any prior art or other evidence to the contrary. Accordingly, the rejection cannot be sustained at Step 2B.
For all of the above reasons, the rejection of claims 21-28 and 33-40 under 35 U.S.C. §101 is respectfully traversed, and withdrawal of the rejection is requested.
Firstly, Applicant does not clearly define a standard of measurement for a “surprise factor.” The method of measurement could even be subjective. Therefore, since Applicant does not provide a clear standard of measurement for “surprise”, The term is indefinite.
The standard is not whether some “combination of elements is routine and conventional”, as Applicant argued. It is whether “additional elements” to the abstract ideas are well-understood, routine and conventional. Examiner has shown this in response to arguments, above, as well as in the rejections.
Applicant's argument is unpersuasive.
The rejections stand.
Conclusion
THIS ACTION IS MADE FINAL. Applicant is reminded of the extension of time policy as set forth in 37 CFR 1.136(a).
A shortened statutory period for reply to this final action is set to expire THREE MONTHS from the mailing date of this action. In the event a first reply is filed within TWO MONTHS of the mailing date of this final action and the advisory action is not mailed until after the end of the THREE-MONTH shortened statutory period, then the shortened statutory period will expire on the date the advisory action is mailed, and any nonprovisional extension fee (37 CFR 1.17(a)) pursuant to 37 CFR 1.136(a) will be calculated from the mailing date of the advisory action. In no event, however, will the statutory period for reply expire later than SIX MONTHS from the mailing date of this final action.
Any inquiries concerning this communication or earlier communications from the examiner should be directed to Wilbert L. Starks, Jr., who may be reached Monday through Friday, between 8:00 a.m. and 5:00 p.m. EST. or via telephone at (571) 272-3691 or email: Wilbert.Starks@uspto.gov.
If you need to send an Official facsimile transmission, please send it to (571) 273-8300.
If attempts to reach the examiner are unsuccessful the Examiner’s Supervisor (SPE), Kakali Chaki, may be reached at (571) 272-3719.
Hand-delivered responses should be delivered to the Receptionist @ (Customer Service Window Randolph Building 401 Dulany Street, Alexandria, VA 22313), located on the first floor of the south side of the Randolph Building.
Finally, information regarding the status of an application may be obtained from the Patent Application Information Retrieval (PAIR) system. Moreover, status information for published applications may be obtained from either Private PAIR or Public PAIR. Status information for unpublished applications is available through Private PAIR only. For more information about the PAIR system, see http://pair-direct.uspto.gov. Should you have any questions on access to the Private PAIR system, contact the Electronic Business Center (EBC) toll-free @ 1-866-217-9197.
/WILBERT L STARKS/
Primary Examiner, Art Unit 2122
WLS
27 MAR 2026
Prosecution Insights