DETAILED ACTION
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 .
Information Disclosure Statement
The information disclosure statements submitted on February 7, 2023 and November 14, 2023 have been considered by the Examiner.
Claim Rejections - 35 USC § 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.
Claims 1-4 are rejected under 35 U.S.C. 101 because the claimed invention is directed to an abstract idea (e.g. a mental process, mathematical concept) without significantly more.
As described in MPEP § 2106, the analysis as to whether a claim qualifies as eligible subject matter under 35 U.S.C. § 101 includes the following determinations:
(1) Whether the claim is to a statutory category, i.e. to a process, machine, manufacture or composition of matter (“Step 1”) – see MPEP §§ 2106, subsection III, and 2106.03
(2) If the claim is to a statutory category, whether the claim recites any judicial exceptions, including certain groupings of abstract ideas (i.e., mathematical concepts, certain methods of organizing human activity, or mental processes) (“Step 2A, Prong One”) – see MPEP §§ 2106, subsection III, and 2106.04
(3) If the claim recites a judicial exception, whether the claim recites additional elements that integrate the judicial exception into a practical application (“Step 2A, Prong Two”) – see MPEP §§ 2106, subsection III, and 2106.04
(4) If the claim does not recite additional elements that integrate the judicial exception into a practical application, whether the claim recites additional elements that amount to significantly more than the judicial exception (“Step 2B”) – see MPEP §§ 2106, subsection III, and 2106.05
Claims 1, 3 and 4
Regarding “Step 1,” independent claims 1, 3 and 4 are to a statutory category. Claim 1 is directed to a non-transitory computer-readable recording medium, which can be considered a manufacture or composition of matter. Claim 3 is directed to an apparatus, which can be considered a machine or manufacture. Claim 4 is directed to a method, i.e. an apparatus.
Accordingly, the analysis proceeds to “Step 2A, Prong One” to determine if the claims recite a judicial exception. In this case, claims 1, 3 and 4 recite mathematical concepts and mental processes. As noted in MPEP § 2106.04IIB, a claim may recite multiple judicial exceptions, and the same eligibility analysis is to be applied regardless of the number of exceptions recited therein.
In each of claims 1, 3 and 4 the following is particularly considered a recitation of a mathematical concept:
calculating, for each attribute of the plurality of attributes, a processing amount based on a difference between each piece of data of the first plurality of pieces of data and a corresponding piece of data of the second plurality of pieces of data;
…
identifying a magnitude of contribution of the at least one second attribute, the magnitude of contribution indicating a degree how the at least one second attribute affects an inference result obtained by a machine learning model in a case where the machine learning model performs inference in response to inputting data into the machine learning model; and
determining, based on the magnitude of the contribution, an influence degree that indicates a degree how the second plurality of pieces of data affect the machine learning model in a case where the machine learning model is trained using the second plurality of pieces of data.
The first limitation noted above requires a difference (i.e. subtraction) operation, while the latter two limitations require other mathematical operations (see e.g. paragraphs 0043-0045 and 0050 of the published Application, U.S. Patent Application Publication No. 2023/0385690).
Further in each of claims 1, 3 and 4, the recitation of “identifying, from among the plurality of attributes, at least one second attribute for which the processing amount calculated is larger than or equal to a predetermined threshold” is considered a recitation of a mental process. “’[T]he mental processes’ abstract idea grouping in particular is defined as concepts performed in the human mind, and examples of mental processes include observations, evaluations, judgements, and opinions.” MPEP § 2106.04(a)(2), subsection III. “If a claim recites a limitation that can practically be performed in the human mind, with or without the use of a physical aid such as pen and paper, the limitation falls within the mental processes grouping, and the claims recites an abstract idea. MPEP § 2106.04(a)(2), subsection III,B (citations omitted). Identifying if a number (e.g. a processing amount) is larger than another number (e.g. a threshold) can practically be performed in the human mind.
Because the claims recite a judicial exception (i.e. mathematical concepts, mental processes), the analysis proceeds to “Step2A, Prong Two” to determine whether the claims recite additional elements that integrate the judicial exception into a practical application. In this case, other than the above-noted mental processes, each claim recites “obtaining, based on a first plurality of pieces of data each of which includes a plurality of attributes, a second plurality of pieces of data generated by processing the first plurality of pieces of data in accordance with nonuniformity of the first plurality of pieces of data with reference to a first attribute of the plurality of attributes, each of the second plurality of pieces of data including data generated from a corresponding piece of data among the first plurality of pieces of data.” However, this is considered insignificant extra-solution activity, i.e. mere data gathering, and is insufficient to integrate the abstract idea into a practical application. See MPEP § 2106.5(g). This limitation for obtaining a second plurality of pieces of data generated by processing a first plurality of pieces of data can alternatively be considered a recitation of a mathematical concept.
Claim 1 further recites that the above-described tasks are implemented via a “non-transitory computer-readable recording medium storing a determination program.” Claim 3 recites that the tasks are implemented via a “determination apparatus comprising: a memory; and a processor coupled to the memory….” Claim 4 recites that the method is “implemented by a computer.” However, these additional elements are mere instructions to apply the abstract idea on a generic computer and thus fail to integrate the abstract idea into a practical application. See MPEP § 2106.05(f).
Accordingly, as claim 1, 3 and 4 do not recite additional elements that integrate the judicial exception into a practical application, the analysis proceeds to “Step 2B” to determine whether the claims recite additional elements that amount to significantly more than the judicial exception. However, in this case, the claims do not. As noted above, claims 1, 3 and 4 additionally recite “obtaining, based on a first plurality of pieces of data each of which includes a plurality of attributes, a second plurality of pieces of data….” However, as further noted above, these elements amount to mere data gathering, which is considered insignificant extra-solution activity. Such data gathering is also well-understood, routine and conventional. See, e.g., Versata Dev. Group, Inc. v. SAP Am., Inc., 793 F.3d 1306, 1334, 115 USPQ2d 1681, 1701 (Fed. Cir. 2015); buySAFE, Inc. v. Google, Inc., 765 F.3d 1350, 1355, 112 USPQ2d 1093, 1096 (Fed. Cir. 2014).
As further noted above, claim 1 recites that the above-described tasks are implemented via a “non-transitory computer-readable recording medium storing a determination program,” claim 3 recites that the tasks are implemented via a “determination apparatus” comprising a memory and a processor, and claim 4 recites that the method is “implemented by a computer.” However, as further noted above, these additional elements amount to mere instructions to apply the abstract idea on a generic computer. As such, they do not amount to significantly more than the judicial exception. See MPEP § 2106.05(f).
Consequently, claims 1, 3 and 4 recite an abstract idea but do not include additional elements that integrate the abstract idea into a practical application or that amount to significantly more than the abstract idea. As a result, and for the reasons described above, claims 1, 3 and 4 are rejected as being patent ineligible under 35 U.S.C. § 101.
Claim 2
Claim 2 further recites that “the at least one second attributes includes a plurality of second attributes, and the identifying of the at least one second attribute includes identifying, based on the difference between the second plurality of pieces of data and the first plurality of pieces of data, one or more of the plurality of second attributes processed with processing amounts larger than or equal to the predetermined threshold in descending order out of the plurality of attributes.” However, this is considered a recitation of a mental process. Identifying if a number (e.g. a processing amount) is larger than another number (e.g. a threshold) can practically be performed in the human mind, as can sorting a plurality of such numbers. Claim 2 fails to recite any additional elements (i.e. additional to the mental process) that integrate the abstract idea into a practical application or that amount to significantly more than the abstract idea, and as a result, claim 2 is also patent ineligible under 35 U.S.C. § 101.
Claim Rejections - 35 USC § 102
The following is a quotation of the appropriate paragraphs of 35 U.S.C. 102 that form the basis for the rejections under this section made in this Office action:
A person shall be entitled to a patent unless –
(a)(1) the claimed invention was patented, described in a printed publication, or in public use, on sale, or otherwise available to the public before the effective filing date of the claimed invention.
Claims 1-4 are rejected under 35 U.S.C. 102(a)(1) as being anticipated by U.S. Patent Application Publication No. 2019/0043070 to Merrill et al. (“Merrill”).
Regarding claims 1, 3 and 4, Merrill describes a model evaluation system for determining whether a machine learning model is likely to generate results that disparately impact a protected class (see e.g. paragraph 0026). Like claimed, Merrill particularly teaches:
obtaining, based on a first plurality of pieces of data (i.e. a first plurality of data sets) each of which includes a plurality of attributes (i.e. variables), a second plurality of pieces of data generated by processing the first plurality of pieces of data in accordance with nonuniformity of the first plurality of pieces of data with reference to a first attribute of the plurality of attributes, each of the second plurality of pieces of data including data generated from a corresponding piece of data among the first plurality of pieces of data (see e.g. paragraphs 0031-0032 and 0074-0081: Merrill teaches that the model evaluation system determines impactful variables and their associated “score impact values” by, in part, accessing a plurality of data sets, e.g. training data sets used to train the model to be evaluated. Each data set includes a plurality of variables – see e.g. paragraphs 0074-0075. The system then generates a plurality of modified data sets by changing an individual variable value within each data set to one of a plurality of unique nonuniform values for the variable – see e.g. paragraphs 0075-0081. Accordingly, Merrill teaches obtaining, based on a first plurality of data sets, each of which includes a plurality of attributes, i.e. variables, a second plurality of data sets, i.e. the modified data sets, which are generated in part by processing the first plurality of data sets in accordance with nonuniformity of the first plurality of pieces of data with reference to a first variable, wherein each of the plurality of second data sets includes data generated from a corresponding data set among the first plurality of data sets.);
calculating, for each attribute of the plurality of attributes, a processing amount based on a difference between each piece of data of the first plurality of pieces of data and a corresponding piece of data of the second plurality of pieces of data (see e.g. paragraphs 0075 and 0078-0083: Merrill teaches identifying a score for each data set of the original plurality of data sets by transmitting the data set to the modeling system, and identifying a score for each modified data set of the plurality of modified data sets by transmitting the modified data set to the modeling system. Merrill further teaches determining a difference between the score of the original data set and the score of the corresponding modified data set – see e.g. paragraphs 0075 and 0078-0084. The system then determines the impact value for each variable from these score differences for each value of each variable in the plurality of modified data sets – see e.g. paragraphs 0075 and 0084-0086. This impact value for each variable is considered indicative of a processing amount based on a difference between each data set of the first plurality of data sets and a corresponding data set of the second plurality of data sets.);
identifying, from among the plurality of attributes, at least one second attribute for which the processing amount calculated is larger than or equal to a predetermined threshold (see e.g. paragraphs 0031-0032, 0036, 0101 and 0143-0146: Merrill discloses that the impact value for each variable can be compared to a predetermined threshold, whereby variables with impact values below the threshold value can be removed to modify the model. Conversely, variables with impact values above the threshold are relatively important, and would be kept – see e.g. paragraphs 0031-0032, 0101 and 0143-0146. Merrill thus teaches identifying, from among the plurality of variables, at least one second variable for which the processing amount, i.e. impact value, calculated is larger than or equal to a predetermined threshold.);
identifying a magnitude of contribution (i.e. score impact value) of the at least one second attribute, the magnitude of contribution indicating a degree how the at least one second attribute affects an inference result obtained by a machine learning model in a case where the machine learning model performs inference in response to inputting data into the machine learning model (see e.g. see e.g. paragraphs 0031-0032 and 0074-0081: as described above, Merrill teaches that the model evaluation system determines an impact value for each variable. The impact value for a variable indicates a degree how the variable affects an inference result, e.g. score, obtained by the model in a case where the model performs inference in response to inputting data into the machine learning model – see e.g. paragraphs 0031-0033 and 0074-0085. The impact value is thus considered a magnitude of contribution like claimed, and is identified for each variable, including the second attribute described above.); and
determining, based on the magnitude of contribution, an influence degree that indicates a degree how the second plurality of pieces of data affect the machine learning model in a case where the machine learning model is trained by using the second plurality of pieces of data (see e.g. paragraphs 0031-0032, 0036, 0101 and 0143-0146: as noted above, Merrill discloses that the impact value for each variable can be compared to a predetermined threshold, whereby variables with impact values below the threshold value can be removed to modify the model. That is, the model is trained with the data sets having the low impact variables removed. The impact value for a variable thus indicates a degree as to how the variable affects the machine learning model in a case where the machine learning model is trained by datasets including the variable. Accordingly, the impact values can be considered an influence degree like claimed, which indicates how the data sets, e.g. the second plurality of pieces of data, having the variables affect the machine learning model in a case where the machine learning model is trained by using the data sets.).
Accordingly, Merrill teaches a determination method like recited in claim 4. Merrill discloses that such teachings can be implemented via a program stored in the memory of a computer that comprises a processor to execute the program and thereby perform the above-noted tasks (see e.g. paragraphs 0330-0339). The memory of a computer storing a program to implement the above-described teachings of Merrill is considered a non-transitory computer-readable recording medium like that of claim 1, and the computer is considered a determination apparatus like that of claim 3.
Regarding claim 2, Merrill further teaches that the at least one second attribute includes a plurality of second attributes (i.e. a plurality of variables), and the identifying of the at least one second attribute includes identifying, based on the difference between the second plurality of pieces of data and the first plurality of pieces of data (i.e. based on the impact values for each variable), one or the plurality of second attributes processed with processing amounts larger than or equal to the predetermined threshold in descending order out of the plurality of attributes (see e.g. paragraphs 0078-0085 and 0148-0153). Accordingly, Merrill further teaches a non-transitory computer-readable recording medium like that of claim 2.
Conclusion
The prior art made of record on form PTO-892 and not relied upon is considered pertinent to applicant’s disclosure. The applicant is required under 37 C.F.R. §1.111(C) to consider these references fully when responding to this action. In particular, the U.S. Patent Application Publication to Lee et al. cited therein describes a method and system for detecting biases in predictive models whereby a processing device identifies feature groups from training data and generates performance metrics and baseline metrics for a feature group. The U.S. Patent Application Publication to Rink et al. cited therein describes operations for determining one or more adjustment parameters for balancing bias reduction and prediction accuracy in machine learning model output. The article to Adebayo et al. cited therein describes an iterative procedure for enabling interpretability of predictive models, whereby one can quantify the relative dependence of a model on its input attributes.
Any inquiry concerning this communication or earlier communications from the examiner should be directed to BLAINE T BASOM whose telephone number is (571)272-4044. The examiner can normally be reached Monday-Friday, 9:00 am - 5:30 pm, EST.
Examiner interviews are available via telephone, in-person, and video conferencing using a USPTO supplied web-based collaboration tool. To schedule an interview, applicant is encouraged to use the USPTO Automated Interview Request (AIR) at http://www.uspto.gov/interviewpractice.
If attempts to reach the examiner by telephone are unsuccessful, the examiner’s supervisor, Matt Ell can be reached at (571)270-3264. The fax phone number for the organization where this application or proceeding is assigned is 571-273-8300.
Information regarding the status of published or unpublished applications may be obtained from Patent Center. Unpublished application information in Patent Center is available to registered users. To file and manage patent submissions in Patent Center, visit: https://patentcenter.uspto.gov. Visit https://www.uspto.gov/patents/apply/patent-center for more information about Patent Center and https://www.uspto.gov/patents/docx for information about filing in DOCX format. For additional questions, contact the Electronic Business Center (EBC) at 866-217-9197 (toll-free). If you would like assistance from a USPTO Customer Service Representative, call 800-786-9199 (IN USA OR CANADA) or 571-272-1000.
/BTB/
1/9/2027
/MATTHEW ELL/Supervisory Patent Examiner, Art Unit 2141