DETAILED ACTION
Remarks
Claims 1-18 have been examined and rejected. This Office Action is responsive to the amendment filed on 12/22/2025, which has been entered in the above identified application.
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 .
Claims 1-20 are presented for examination.
Response to Amendment
Applicant’s amendment filed on 12/22/2025 has been entered. Claims 1, 6 and 11 are amended. Claims 19 and 20 are added. Claims 1-20 are pending in the application.
Claim Objections
Claims 1, 6 and 11 are objected to because of the following informalities,
Claim 1 [line 14], claim 6 [line 14] and claim 14 [line 15]: “locally approximate the pre-trained machine learning mode” should be “locally approximate the pre-trained machine learning model”
Appropriate corrections are required.
Claim Rejections - 35 USC § 112
The following is a quotation of 35 U.S.C. 112(b):
(b) CONCLUSION.—The specification shall conclude with one or more claims particularly pointing out and distinctly claiming the subject matter which the inventor or a joint inventor regards as the invention.
The following is a quotation of 35 U.S.C. 112 (pre-AIA ), second paragraph:
The specification shall conclude with one or more claims particularly pointing out and distinctly claiming the subject matter which the applicant regards as his invention.
Claims 1-20 are rejected under 35 U.S.C. 112(b) or 35 U.S.C. 112 (pre-AIA ), second paragraph, as being indefinite for failing to particularly point out and distinctly claim the subject matter which the inventor or a joint inventor (or for applications subject to pre-AIA 35 U.S.C. 112, the applicant), regards as the invention.
With respect to claim 1 [line 13], claim 6 [line 13] and claim 11 [line 14], it is unclear what “the explanatory model being configured to locally approximate the pre-trained machine-learning mode” refers to. The limitation ‘the explanatory model … locally approximate the pre-trained machine-learning mode’ lacks support or clear definition in the Specification. The Specification must provide an adequate written description and enablement of the claimed invention.
With respect to claims 2-5, 7-10 and 12-20, they are rejected based on the virtual dependency of claims 1, 6 and 11.
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 1-20 are rejected under 35 U.S.C. 101 because the claimed invention is directed to an abstract idea without significantly more.
Independent claims
Step 1
Claim 1 is drawn to a non-transitory computer-readable recording medium, claim 6 is drawn to a method of outputting explanatory information, and claim 11 is drawn to an information processing apparatus. Therefore, each of these claim groups falls under one of four categories of statutory subject matter (process/method, machines/product/apparatus, manufactures, and composition of matter).
Step 2A – Prong 1
Claims 1, 6 and 11 are directed to a judicially recognized exception of an abstract idea without significantly more.
Claims 1, 6 and 11 recite further a method of converting at least a portion of the pieces of the attendance data into a tensor format that under its broadest reasonable interpretation enumerates a mathematical concept. A human can perform a mathematical calculation to convert data into a tensor format. Therefore, the step of converting a portion of the pieces of data into tensor format is nothing more than a mathematical concept (MPEP 2106.04(a)(2)(I)).
Claims 1, 6 and 11 recite further a method of generating a core tensor from the attendance data in the tensor format using tensor decomposition that under its broadest reasonable interpretation enumerates a mathematical concept. A human can perform a mathematical calculation to generate a core tensor into a tensor format. Therefore, the step of generating a core tensor from attendance data converted into the tensor format is nothing more than a mathematical concept (MPEP 2106.04(a)(2)(I)).
Claims 1, 6 and 11 recite further a method of predicting leave of absence risk of the employee as a prediction result using the core tensor and a pre-trained machine learning model that under its broadest reasonable interpretation enumerates a mathematical concept. A human can perform a mathematical calculation to predict leave of absence risk using the core tensor. Therefore, the step of predicting leave of absence risk of the employee is nothing more than a mathematical concept (MPEP 2106.04(a)(2)(I)).
Claims 1, 6 and 11 recite further a method of generating an explanatory model by identifying neighborhood data based on a similarity of the core tensor, the explanatory model being configured to locally approximate the pre-trained machine-learning mode that under its broadest reasonable interpretation enumerates a mathematical concept. A human can perform a mathematical calculation to generate explanatory model by identifying neighborhood data. Therefore, the step of generating an explanatory model by identifying neighborhood data is nothing more than a mathematical concept (MPEP 2106.04(a)(2)(I)).
Claims 1, 6 and 11 recite further a method of calculating, using the generated explanatory model, a contribution degree of each of the plurality of pieces of attendance data with respect to the prediction result that under its broadest reasonable interpretation enumerates a mathematical concept. A human can perform a mathematical calculation to calculate contribution degree of each piece of attendance data. Therefore, the step of calculating a contribution degree is nothing more than a mathematical concept (MPEP 2106.04(a)(2)(I)).
Claims 1, 6 and 11 recite further a method of determining specific attendance data among the plurality of pieces of the attendance data based on a value of the priority variable and the calculated contribution degree that under its broadest reasonable interpretation enumerates a mathematical concept. A human can perform a mathematical calculation to determine attendance data based on a variable. Therefore, the step of determining specific attendance data based on a value of the priority variable is nothing more than a mathematical concept (MPEP 2106.04(a)(2)(I)).
Step 2A – Prong 2
Claims 1, 6 and 11 recite further receiving a plurality of pieces of attendance data of an employee, each of the pieces of the attendance data including a plurality of variables that fail to integrate the abstract idea into a practical application. The step of receiving a plurality of pieces of attendance data is a form of insignificant input and output solution activities, where receiving a plurality of attendance data is necessary for all uses of the judicial exception. This additional element does not integrate the abstract idea into a practical application because it does not impose any meaningful limits on practicing the abstract idea (MPEP 2106.05(g)).
Claims 1, 6 and 11 recite further receiving a priority variable that a user considers likely to affect the leave of absence risk from among the plurality of variables that fail to integrate the abstract idea into a practical application. The step of receiving a priority variable is a form of insignificant input and output solution activities, where receiving a priority variable is necessary for all uses of the judicial exception. This additional element does not integrate the abstract idea into a practical application because it does not impose any meaningful limits on practicing the abstract idea (MPEP 2106.05(g)).
Claims 1, 6 and 11 recite further presenting the specific attendance data as explanatory information of the prediction result to the user that fail to integrate the abstract idea into a practical application. The step of presenting attendance data is a form of insignificant input and output solution activities, where presenting the specific attendance data is necessary for all uses of the judicial exception. This additional element does not integrate the abstract idea into a practical application because it does not impose any meaningful limits on practicing the abstract idea (MPEP 2106.05(g)).
Step 2B
The additional elements in step 2A-Prong 2 those are forms of insignificant extra-solution activities, do not amount to significantly more than an abstract idea because the court decision have determined that these additional elements of receiving a plurality of attendance data; receiving priority variable; and presenting the specific attendance data to be well-understood, routine, and conventional when claimed in a merely generic manner (MPEP 2106.05(d)(II)).
As such, claims 1, 6 and 11 are not patent eligible.
Dependent claims
Claims 2-5, 7-10 and 12-20 merely narrow the previously recited abstract idea limitations. For the reasons described above with respect to claims 1, 6 and 11, this judicial exception is not meaningfully integrated into a practical application, or significantly more than the abstract idea. The claims disclose similar limitations described for the independent claims above and do not provide anything more than the mental process that are practically capable of being performed in the human mind with the assistance of pen and paper. Therefore, claims 2-5, 7-10 and 12-20 also recite abstract ideas that do not integrate into a practical application or amount to significantly more than the judicial exception, and are rejected under U.S.C. 101.
Step 1
Claims 2-5, 17, 19 and 20 are drawn to a non-transitory computer-readable recording medium, claims 7-10, 18 are drawn to a method of outputting explanatory information, and claims 12-16 are drawn to an information processing apparatus. Therefore, each of these claim groups falls under one of four categories of statutory subject matter (process/method, machines/product/apparatus, manufactures, and composition of matter).
Step 2A – Prong 1
Dependent claims 2, 7 and 12 recite further the mathematical concept by selecting the priority variable from among the plurality of variables based on a priority variable corresponding to a user that is based on one or more features of the ML project (MPEP 2106.04(a)(2)(I)).
Dependent claims 3, 8 and 13 recite further the mathematical concept by in a case that the priority variable is numerical data, selecting a plurality of pieces of presentation candidate data each having the contribution degree larger than a predetermined threshold from among the plurality of pieces of data; and determining the specific attendance data based on the value of the priority variable and the contribution degree from among the plurality of pieces of presentation candidate data those are based on one or more features of the ML project (MPEP 2106.04(a)(2)(I)).
Dependent claims 5, 10 and 15 recite further the mathematical concept by in a case that the priority variable is categorical data, selecting a plurality of pieces of presentation candidate data in which the value of the priority variable indicates a predetermined category and the contribution degree is larger than a predetermined threshold, from among the plurality of pieces of data; and determining, as the specific attendance data, data with the highest contribution degree from among the plurality of pieces of presentation candidate data those are based on one or more features of the ML project (MPEP 2106.04(a)(2)(I)).
Dependent claim 19 recites further the mathematical concept by identifying neighborhood data includes determining data having a difference from the attendance data that is equal to or smaller than a threshold that is based on one or more features of the ML project (MPEP 2106.04(a)(2)(I)).
Step 2A – Prong 2
Dependent claims 4, 9 and 14 recites further the insignificant extra solution activities by obtaining, for the plurality of pieces of presentation candidate data, a first value by normalizing the value of the priority variable and a second value by normalizing the contribution degree; and determining, as the specific attendance data, data in which a sum of the first value and the second value is the largest among the plurality of pieces of presentation candidate data. These additional elements do not integrate the abstract idea into a practical application because they do not impose any meaningful limits on practicing the abstract idea (MPEP 2106.05(g)).
Dependent claims 16-18 recites further the insignificant extra solution activities by outputting a user-oriented data input screen to be displayed that presents a plurality of the variables as selectable choices to the user. This additional element does not integrate the abstract idea into a practical application because it does not impose any meaningful limits on practicing the abstract idea (MPEP 2106.05(g)).
Dependent claim 20 recites further the insignificant extra solution activities by wherein the similarity of the core tensor is defined by a distance between tensors. This additional element does not integrate the abstract idea into a practical application because it does not impose any meaningful limits on practicing the abstract idea (MPEP 2106.05(g)).
As such, dependent claims 2-5, 7-10 and 12-20 are not patent eligible.
Claim Rejections - 35 USC § 103
The following is a quotation of 35 U.S.C. 103 which forms the basis for all obviousness rejections set forth in this Office action:
A patent for a claimed invention may not be obtained, notwithstanding that the claimed invention is not identically disclosed as set forth in section 102, if the differences between the claimed invention and the prior art are such that the claimed invention as a whole would have been obvious before the effective filing date of the claimed invention to a person having ordinary skill in the art to which the claimed invention pertains. Patentability shall not be negated by the manner in which the invention was made.
Claims 1-20 are rejected under 35 U.S.C. 103 as being unpatentable over Baskaran et al (US 20170168991 A1) hereafter Baskaran, further in view of Harris et al (US 20210176262 A1) hereafter Harris, further in view of Tanizawa et al (US 20210081781 A1) hereafter Tanizawa, and further in view of Motohashi et al (US 20190034945 A1) hereafter Motohashi.
With respect to claim 1, Baskaran teaches a non-transitory computer-readable recording medium storing a program for causing a computer to execute a process (a non-transitory storage medium having stored instructions when executed to decompose tensors so as to facilitate extraction of information from the tensors [par. 0021]), the process comprising:
receiving a plurality of pieces of attendance data of an employee, each of the pieces of the attendance data including a plurality of variables (a tensor representing health-related data at a hospital and dates, and another tensor mode may represent a total number of physicians trained, and a tensor mode may represent a number of physicians in attendance on a particular day. All tensor components may be identified by the decomposition process to reveal useful information about the data represented by the tensor [par. 0006-0011]):
converting at least a portion of the pieces of the attendance data into a tensor format (a tensor component used in providing adequate care relating to the attended physicians can be an array or a vector, a matrix, another tensor, or another suitable data structure. The CP decomposition decomposes a tensor into a sum of component rank one tensors. The tensor is decomposed into R components, wherein each component may identify a pattern in the data or a highly correlated cluster of information in the data [par. 0006-0008, 0029-0031]);
generating a core tensor from the attendance data in the tensor format using tensor decomposition (CANDECOMP/PARAFAC(CP) decomposition and Tucker decomposition are two commonly employed techniques for tensor decomposition. Tucker decomposition technique represents multiple values those are generally specified by the user and the product of these values determines the total number of components resulting from the Tucker decomposition technique [par. 0006-0008, 0029-0032]).
However, Baskaran does not disclose predicting leave of absence risk of the employee as a prediction result using the core tensor and a pre-trained machine learning model; generating an explanatory model by identifying neighborhood data based on a similarity of the core tensor, the explanatory model being configured to locally approximate the pre-trained machine-learning mode; calculating, using the generated explanatory model, a contribution degree of each of the plurality of pieces of attendance data with respect to the prediction result; receiving a priority variable that a user considers likely to affect the leave of absence risk from among the plurality of variables; determining specific attendance data among the plurality of pieces of the attendance data based on a value of the priority variable and the calculated contribution degree; and presenting the specific attendance data as explanatory information of the prediction result to the user.
In the same field of endeavor, Harris teaches predicting leave of absence risk of the employee as a prediction result using the core tensor and a pre-trained machine learning model (tensor factorization may be used to predict the classified data. For example, a plurality of network data may indicate that a hurricane is approaching Florida and car traffic has increased significantly as people able to leave the hurricane area. The predictions may be made as to the decrease in the population in different geographic areas and the potential duration of population decrease in response to the weather network data [par. 0048, 0160]).
It would have been obvious to one of ordinary skill in the art before the effective filing date of the invention to have incorporated the concept of generating matrices based on the plurality of network data and performing tensor factorization on the matrices to obtain latent values in the network data as suggested by Harris into the concept of performing tensor decomposition in a selective expansive and/or recursive manner as suggested by Baskaran because both of these systems addressing the process of training a prediction model including some variables and the uses of tensor decomposition. Doing so would be desirable because the concept of Baskaran would be more efficient by using a tensor decomposition technique to predict outcomes of an event related to a plurality of network data using the tensor decomposition (Harris, [par. 0006-0009]).
However, the combination of Baskaran and Harris does not particularly disclose generating an explanatory model by identifying neighborhood data based on a similarity of the core tensor, the explanatory model being configured to locally approximate the pre-trained machine-learning mode; and calculating, using the generated explanatory model, a contribution degree of each of the plurality of pieces of attendance data with respect to the prediction result; receiving a priority variable that a user considers likely to affect the leave of absence risk from among the plurality of variables; determining specific attendance data among the plurality of pieces of the attendance data based on a value of the priority variable and the calculated contribution degree; and presenting the specific attendance data as explanatory information of the prediction result to the user.
In the same field of endeavor, Tazinawa teaches generating an explanatory model by identifying neighborhood data based on a similarity of the core tensor, the explanatory model being configured to locally approximate the pre-trained machine-learning mode (In XAI and machine learning (ML), an explanatory model can be identified as an interpretable model or an approximation model. The approximate unit generates the width models as the approximate models, and the approximate unit approximates all the weights in the weight matrix W (neighborhood weights) having a size of mxn. A singular value decomposition is used as an approximation method. The vectors used in the approximation method may be selected in descending order of the contribution degree determined based on the singular value [par. 0069-0080, 0131-0144]); and
calculating, using the generated explanatory model, a contribution degree of each of the plurality of pieces of attendance data with respect to the prediction result (the ML model is trained using a tensor decomposition method, a tensor of weighting coefficients in each layer used in a neural network into two or more decomposed tensors. A contribution degree is calculated based on a level of a singular value, wherein the singular value is determined based on the decomposition method in the approximation model. The prediction result is generated based on a result using the core tensor and a pretrained ML model. Here, the approximation model is trained using the decomposition tensor with multiple layers with each layer has multiple weighting coefficients [par. 0069-0080, 0131-0144]).
It would have been obvious to one of ordinary skill in the art before the effective filing date of the invention to have incorporated the concept of using deep learning in which the neural network includes multiple layers that each layer is calculated by summing values obtained by multiplying values of respective nodes in a preceding layer by a weighting coefficient as suggested by Tanizawa into the combination of Baskaran and Harris because all of these systems addressing the process of training a prediction model including some variables and their degree of contribution. Doing so would be desirable because the combination of Baskaran and Harris would be more efficient by training an approximation model by using a tensor decomposition method to generate a contribution degree based on the method using neighborhood weights in the weight matrices (Tanizawa, [par. 0069-0080, 0131-0144]).
However, the combination of Baskaran, Harris and Tanizawa does not particularly disclose receiving a priority variable that a user considers likely to affect the leave of absence riskcalculated contribution degree; and presenting the specific attendance data as explanatory information of the prediction result to the user.
In the same field of endeavor, Motohashi teaches receiving a priority variable that a user considers likely to affect the leave of absence risk from among the plurality of variables (an accepting unit accepts a classification that specifies a prediction target for which a factor related to a user to be analyzed. One or more classifications may be selected by the user, wherein there is a lowest classification and a highest classification in the hierarchical structure [par. 0052-0059]);
determining specific attendance data among the plurality of pieces of the attendance data based on a value of the priority variable and the calculated contribution degree (the total sum of the weights for each explanatory variable included in the prediction model is referred to as a first degree of contribution, wherein the absolute value of the coefficient is used as the weight in order to indicate the degree of contribution of each explanatory variable [par. 0065-0072]); and
presenting the specific attendance data as explanatory information of the prediction result to the user (the aggregating unit 20 may provide a new explanatory variable indicating a difference between the prediction result and the actual measurement result and employ the difference as the degree of contribution of the new explanatory variable. When the user selects classification and aggregation method, the accepting unit and the aggregating unit perform the aggregating process and the output unit outputs an aggregation result to the screen [par. 0105-0107, 0143, 0144]).
It would have been obvious to one of ordinary skill in the art before the effective filing date of the invention to have incorporated the concept of predicting a prediction target specified by a plurality of classifications related to a user as suggested by Motohashi into the combination of Baskaran, Harris and Tanizawa because all of these systems addressing the process of training a prediction model including some variables and their degree of contribution. Doing so would be desirable because the combination of Baskaran, Harris and Tanizawa would be more efficient by including a variable that may affect the prediction target with the degree of contribution determined by the prediction model (Motohashi, [par. 0009-0012]).
With respect to claim 2, the combination of Baskaran, Harris, Tanizawa and Motohashi teaches the process further comprising:
selecting the priority variable from among the plurality of variables based on a priority variable corresponding to a user (Motohashi, an accepting unit accepts a classification that specifies a prediction target for which a factor related to a user to be analyzed. One or more classifications may be selected by the user, wherein there is a lowest classification and a highest classification in the hierarchical structure [par. 0052-0059]).
With respect to claim 3, the combination of Baskaran, Harris, Tanizawa and Motohashi teaches the process further comprising:
in a case that the priority variable is numerical data (Motohashi, the prediction model may predict a numerical value of the prediction target, and the prediction model may output a variable describing a probability distribution of the objective variable [par. 0048]), selecting a plurality of pieces of presentation candidate data each having the contribution degree larger than a predetermined threshold from among the plurality of pieces of data (Motohashi, the explanatory variables with larger weight has a higher degree of contribution. The weight specified for an explanatory variable or an aggregated value of the weights from a predetermined viewpoint is referred to degree of contribution of the explanatory variable [par. 0065-0068]); and
determining the specific attendance data based on the value of the priority variable and the contribution degree from among the plurality of pieces of presentation candidate data (Motohashi, the total sum of the weights for each explanatory variable included in the prediction model is referred to as a first degree of contribution, wherein the absolute value of the coefficient is used as the weight in order to indicate the degree of contribution of each explanatory variable [par. 0065-0072]).
With respect to claim 4, the combination of Baskaran, Harris, Tanizawa and Motohashi teaches the process further comprising:
obtaining, for the plurality of pieces of presentation candidate data, a first value by normalizing the value of the priority variable and a second value by normalizing the contribution degree (Motohashi, the aggregating unit may standardize the coefficients included in the prediction formula that may correct each coefficient in order to sum the total value of the coefficients. The aggregating unit may also standardize the degrees of contribution for each explanatory variable [par. 0065-0072]); and
determining, as the specific attendance data, data in which a sum of the first value and the second value is the largest among the plurality of pieces of presentation candidate data (Motohashi, the aggregating unit standardizes the first degrees of contribution of multiple variable, such as w1, w2, w3 and w4, for each explanatory variable, so that there is a possibility that the standardized degrees of contribution on different scale may be compared [par. 0065-0072 and Fig. 7]).
With respect to claim 5, the combination of Baskaran, Harris, Tanizawa and Motohashi teaches the process further comprising:
in a case that the priority variable is categorical data (Motohashi, when explanatory variables are included in the prediction model, a certain category is set in these explanatory variables [par. 0115-0117]), selecting a plurality of pieces of presentation candidate data in which the value of the priority variable indicates a predetermined category and the contribution degree is larger than a predetermined threshold, from among the plurality of pieces of data (Motohashi, for example, a category of “TV ads”, “calendar”, “weather” or “price” may be set in the explanatory variables. The aggregating unit calculates the degree of contribution by summarizing the explanatory variables into each category. The aggregating unit may then standardize the degree of contribution (known as the third degree of contribution) of the explanatory variables according to the categories. In an example, “calendar” is dominant among other categories [par. 0115-0126 and FIG. 29]); and
determining, as the specific attendance data, data with the highest contribution degree from among the plurality of pieces of presentation candidate data (Motohashi, the aggregating unit may standardize the degrees of contribution aggregated for each category by the respective prediction formulas, and may correct each degree of contribution, for example, the category “calendar” may be dominant among other categories based on the factors of explanatory variables [par. 0120-0126]).
With respect to claim 6, it is a method of outputting explanatory information claim that is corresponding to the non-transitory computer-readable medium of claim 1. Therefore, it is rejected for the same reason as claimed in claim 1 above.
With respect to claim 7, it is a method of outputting explanatory information claim that is corresponding to the non-transitory computer-readable medium of claim 2. Therefore, it is rejected for the same reason as claimed in claim 2 above.
With respect to claim 8, it is a method of outputting explanatory information claim that is corresponding to the non-transitory computer-readable medium of claim 3. Therefore, it is rejected for the same reason as claimed in claim 3 above.
With respect to claim 9, it is a method of outputting explanatory information claim that is corresponding to the non-transitory computer-readable medium of claim 4. Therefore, it is rejected for the same reason as claimed in claim 4 above.
With respect to claim 10, it is a method of outputting explanatory information claim that is corresponding to the non-transitory computer-readable medium of claim 5. Therefore, it is rejected for the same reason as claimed in claim 5 above.
With respect to claim 11, it is an information processing apparatus claim that is corresponding to the non-transitory computer-readable medium of claim 1. Therefore, it is rejected for the same reason as claimed in claim 1 above.
With respect to claim 12, it is an information processing apparatus claim that is corresponding to the non-transitory computer-readable medium of claim 2. Therefore, it is rejected for the same reason as claimed in claim 2 above.
With respect to claim 13, it is an information processing apparatus claim that is corresponding to the non-transitory computer-readable medium of claim 3. Therefore, it is rejected for the same reason as claimed in claim 3 above.
With respect to claim 14, it is an information processing apparatus claim that is corresponding to the non-transitory computer-readable medium of claim 4. Therefore, it is rejected for the same reason as claimed in claim 4 above.
With respect to claim 15, it is an information processing apparatus claim that is corresponding to the non-transitory computer-readable medium of claim 5. Therefore, it is rejected for the same reason as claimed in claim 5 above.
With respect to claim 16, the combination of Baskaran, Harris, Tanizawa and Motohashi teaches wherein the processor is further configured to output a user-oriented data input screen to be displayed that presents a plurality of the variables as selectable choices to the user (Monohashi, an output unit configured to accept input from user and includes a display device to display candidate classifications, and the output unit may accept one or more classifications selected by the user [par. 0058, 0059, 0139-0149]).
With respect to claim 17, it is a non-transitory computer-readable medium claim that is corresponding to the non-transitory computer-readable medium of claim 16. Therefore, it is rejected for the same reason as claimed in claim 16 above.
With respect to claim 18, it is a method of outputting explanatory information claim that is corresponding to the non-transitory computer-readable medium of claim 16. Therefore, it is rejected for the same reason as claimed in claim 16 above.
With respect to claim 19, the combination of Baskaran, Harris, Tanizawa and Motohashi teaches wherein identifying neighborhood data includes determining data having a difference from the attendance data that is equal to or smaller than a threshold (Tanizawa, a tensor may have many weighting coefficients, and the weight matrix is decomposed by using a singular value decomposition technique. A contribution degree is calculated based on the singular value. Each layer may have different weighting coefficients that has lower/higher contribution degree [par. 0069-0079]).
With respect to claim 20, the combination of Baskaran, Harris, Tanizawa and Motohashi teaches wherein the similarity of the core tensor is defined by a distance between tensors (Tanizawa, the extraction unit sets the width of the decomposed tensors of the tensor of the weighting coefficients in accordance with a rank R. The extraction unit selects r basis vectors from the R basis vectors to set the width of the decomposed tensors [par. 0069-0080]).
Response to Arguments
The examiner respectfully acknowledges the applicant’s amendments to claims 1, 6 and 11.
Applicant’s amendments filed on 12/22/2025 regarding claim rejections under 35 USC 112(b) to claims 1-18 have been considered and are consequently withdrawn. However, new matter has been added regarding the rejections to claims 1-20 under 35 USC 112(b) (see rejections above).
Applicant’s arguments filed on 12/22/2025 regarding claim rejections under 35 U.S.C. 101 to claims 1-18 have been fully considered but are not persuasive.
Applicant argued that “The present claims require converting attendance data into a tensor format and generating a core tensor through tensor decomposition. These high-dimensional computational steps are not "thinking" and cannot be performed with pen and paper. Further, "generating an explanatory model by identifying neighborhood data based on a similarity of the core tensor" is a specific improvement to the functioning of the computer in the field of Explainable AI (XAI). It solves the technical problem of "black box" AI not by generic rules, but by utilizing the internal data structure (core tensor) of the model itself. This is not merely "organizing human activity.”
Examiner respectfully disagrees.
Based on what is recited in claim 1, the limitations “converting at least a portion of the pieces of the attendance data into a tensor format”, “generating a core tensor from the attendance data in the tensor format using tensor decomposition”, “predicting leave of absence risk of the employee as a prediction result using the core tensor and a pre-trained machine learning model”, “generating an explanatory model by identifying neighborhood data based on a similarity of the core tensor, the explanatory model being configured to locally approximate the pre-trained machine-learning mode”, “calculating, using the generated explanatory model, a contribution degree of each of the plurality of pieces of attendance data with respect to the prediction result”, and “determining specific attendance data among the plurality of pieces of the attendance data based on a value of the priority variable and the calculated contribution degree” may not recite mental processes. However, these limitations still recite the mathematical concept including converting data in tensor format, tensor decomposition, similarity calculations, local approximation of a ML model, and contribution calculations. These are clearly considered mathematical relationships and computations, which fall within the abstract idea grouping of mathematical concepts under 2019 PEG.
Applicant argued that “The claims integrate any alleged abstract idea into a practical application by providing a specific technological solution to the "black box" problem of AI interpretability. In particular, the generation of a core tensor and an explanatory model are not generic calculations, but a specific technical solution to a tecimical problem inherent in computer-based AI systems: the "black box" nature of complex models. The claims are directed to improving the computer's functionality by creating a specific data structure (the core tensor), and then leveraging a unique property of that structure (the core tensor similarity) to generate a new, more faithful data structure (the explanatory model). This specific process enables the computer to generate explanations that are tailored to both the internal logic of the AI model and the user's perspective, an improvement that is deeply rooted in computer technology.”
Examiner respectfully disagrees.
The claimed improvement is directed to generating explanatory information regarding a prediction result rather than improving the operation of the computer itself. The claimed tensor decomposition, similarity analysis, local approximation, and contribution determination are mathematical techniques used to analyze data and generate explanatory output. The claim does not improve processor efficiency, memory utilization, network operation, model training architecture, or other computer functionality. Rather, the claim improves a user’s understanding of a prediction by generating explanatory information. Accordingly, the additional elements merely use a computer as a tool to perform mathematical analysis and present the results and therefore do not integrate the judicial exception into a practical application.
Therefore, amended claim 1 and its corresponding claims 6 and 11 are not patent-eligible for at least the reasons mentioned above. Accordingly, dependent claims 2-5, 7-10 and 12-20 are not patent-eligible as well based on their virtual dependency of claims 1, 6 and 11, respectively.
Applicant’s arguments filed on 12/22/2025 regarding claim rejections under 35 U.S.C. 103 to claims 1-18 have been fully considered and moot in view of new ground of rejection (see rejection above).
Conclusion
The prior art made of record and not relied upon is considered pertinent to applicant's disclosure.
Hatamizadeh et al (US 20230145535 A1) disclosed apparatuses, systems, and techniques to train a neural network to infer a condition based on an image. In at least one embodiment, a first portion of a neural network is trained to infer a condition from an image using a first dataset, and a second portion of the neural network is trained using a second dataset.
Spurr et al (US 20210233273 A1) disclosed apparatuses, systems, and techniques that determine the pose of a human hand from a 2-D image are described herein. In at least one embodiment, training of a neural network is augmented using weakly labeled or unlabeled pose data which is augmented with losses based on a human hand model.
Baier et al (US 20200171657 A1) disclosed a method for computing joint torques applied by actuators to perform a control of a movement of a robot arm having several degrees of freedom is provided. The method includes the act of providing, by a trajectory generator, trajectory vectors specifying a desired trajectory of the robot arm for each degree of freedom. The trajectory vectors are mapped to corresponding latent representation vectors that capture inherent properties of the robot arm using basis functions with trained parameters. The latent representation vectors are multiplied with trained core tensors to compute the joint torques for each degree of freedom.
Principe et al (US 20160242690 A1) disclosed brain state advisory systems using calibrated metrics and optimal time-series decompositions. In one embodiment, a method includes monitoring brainwave activity of a subject and classifying a brain state of the subject based upon the brainwave activity and a model of the brainwave activity. In another embodiment, brain states of the subject are modeled based upon the brainwave activity of the subject.
Any inquiry concerning this communication or earlier communications from the examiner should be directed to Quoc Phung whose telephone number is (703) 756 1330. The examiner can normally be reached on Monday through Friday from 9am to 5pm PT.
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) athttp://www.uspto.gov/interviewpractice.
If attempts to reach the examiner by telephone are unsuccessful, the examiner’s supervisor, Jennifer Welch can be reached on 571-272-7212. The fax phone number for the organization where this application or proceeding is assigned is 571-273-8300.
Information regarding the status of an application may be obtained from the Patent Application Information Retrieval (PAIR) system. 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 questions on access to the Private PAIR system, contact the Electronic Business Center (EBC) at 866-217-9197 (toll-free). If you would like assistance from a USPTO Customer Service Representative or access to the automated information system, call 800-786-9199 (IN USA OR CANADA) or 571-272-1000.
/Q.L.P./Examiner, Art Unit 2143
/JENNIFER N WELCH/Supervisory Patent Examiner, Art Unit 2143
Prosecution Insights