Prosecution Insights
Last updated: November 20, 2026
Application No. 18/640,364

CONVERSION OF QUANTITATIVE ENGINES TO FUZZY QUALITATIVE ENGINES

Non-Final OA §101§103
Filed
Apr 19, 2024
Examiner
KARTHOLY, REJI P
Art Unit
2143
Tech Center
2100 — Computer Architecture & Software
Assignee
International Business Machines Corporation
OA Round
1 (Non-Final)
63%
Grant Probability
Moderate
1-2
OA Rounds
7m
Est. Remaining
99%
With Interview

Examiner Intelligence

Grants 63% of resolved cases
63%
Career Allowance Rate
105 granted / 166 resolved
+8.3% vs TC avg
Strong +70% interview lift
Without
With
+69.6%
Interview Lift
resolved cases with interview
Typical timeline
3y 2m
Avg Prosecution
16 currently pending
Career history
186
Total Applications
across all art units

Statute-Specific Performance

§101
14.3%
-25.7% vs TC avg
§103
61.3%
+21.3% vs TC avg
§102
10.2%
-29.8% vs TC avg
§112
12.8%
-27.2% vs TC avg
Black line = Tech Center average estimate • Based on career data from 166 resolved cases

Office Action

§101 §103
DETAILED ACTION This Office Action is in response to Applicant's Communication received on 04/19/2024 for application number 18/640,364. Claims 1-20 are presented for examination. Claims 1, 8, and 15 are independent claims. 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 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. Step 1 Claims 1-7 are directed to a method, Claims 8-14 are directed to a system and Claims 15-20 are directed to a computer program product and fall within one of the statutory categories (process, machine, articles of manufacture) and are eligible under Step 1. Step 2A Prong 1 Independent Claims Claims 1, 8, 15 recite: processing one or more quantitative values in the received one or more payloads; executing a table mapping between the processed one or more quantitative values and one or more qualitative values - these limitations recite mental processes, i.e., observation, evaluation, judgement, and opinion that a person can practically perform in mind or with pen and paper: a person can look at a set of numbers/ numeric values and evaluate them, make a judgement to pair each numeric value with a word/ label, and write the resulting pars in a two-column table with pen and paper. Accordingly, the claims recites an abstract idea that falls under “mental process” grouping. Step 2A Prong 2 Independent Claims Additional elements Claims 1, 8, 15: receiving one or more payloads including one or more data values - these limitations amount to insignificant extra-solution activity of mere data gathering (see MPEP § 2106.05(g)). training a rule engine with the table mapping as a first input to the rule engine - these limitations merely recite an idea of a solution or outcome with no restriction on how the result is accomplished (see MPEP § 2106.05(f)). These limitations can also be viewed as generally linking the use of a judicial exception to the field of automated rule based decision system (see MPEP § 2106.05(h)). Claim 1: a computer-based method of converting quantitative engines to fuzzy qualitative engines - these limitations are recited at a high-level of generality such that it amount to no more than using generic computer components to apply the judicial exception (see MPEP § 2106.05(f)). Claim 8: a computer system, the computer system comprising: one or more processors, one or more computer-readable memories, one or more computer-readable tangible storage medium, and program instructions stored on at least one of the one or more computer-readable tangible storage medium for execution by at least one of the one or more processors via at least one of the one or more computer-readable memories, wherein the computer system is capable of performing a method - these limitations are recited at a high-level of generality such that it amount to no more than using generic computer components to apply the judicial exception (see MPEP § 2106.05(f)). Claim 15: a computer program product, the computer program product comprising: one or more computer-readable tangible storage medium and program instructions stored on at least one of the one or more computer-readable tangible storage medium, the program instructions executable by a processor capable of performing a method - these limitations are recited at a high-level of generality such that it amount to no more than using generic computer components to apply the judicial exception (see MPEP § 2106.05(f)). Accordingly, these additional elements do not integrate the judicial exception into a practical application because they do not impose any meaningful limits on practicing the abstract idea. The claim is directed to the abstract idea. Step 2B Independent Claims Additional elements Claims 1, 8, 15: receiving one or more payloads including one or more data values - these limitations amount to insignificant extra-solution activity of mere data gathering , which is well-understood, routine, and conventional activity (see MPEP § 2106.05(d), “receiving/ transmitting data”). training a rule engine with the table mapping as a first input to the rule engine - these limitations merely recite an idea of a solution or outcome with no restriction on how the result is accomplished (see MPEP § 2106.05(f)). These limitations can also be viewed as generally linking the use of a judicial exception to the field of automated rule based decision system (see MPEP § 2106.05(h)). Claim 1: a computer-based method of converting quantitative engines to fuzzy qualitative engines - these limitations are recited at a high-level of generality such that it amount to no more than using generic computer components to apply the judicial exception (see MPEP § 2106.05(f)). Claim 8: a computer system, the computer system comprising: one or more processors, one or more computer-readable memories, one or more computer-readable tangible storage medium, and program instructions stored on at least one of the one or more computer-readable tangible storage medium for execution by at least one of the one or more processors via at least one of the one or more computer-readable memories, wherein the computer system is capable of performing a method - these limitations are recited at a high-level of generality such that it amount to no more than using generic computer components to apply the judicial exception (see MPEP § 2106.05(f)). Claim 15: a computer program product, the computer program product comprising: one or more computer-readable tangible storage medium and program instructions stored on at least one of the one or more computer-readable tangible storage medium, the program instructions executable by a processor capable of performing a method - these limitations are recited at a high-level of generality such that it amount to no more than using generic computer components to apply the judicial exception (see MPEP § 2106.05(f)). Accordingly, these additional elements do not amount to significantly more than the judicial exception. As such, the claim is patent ineligible. Step 2A Prong 1 Dependent Claims Claims 2, 9, 16: determining whether the query contains at least one qualitative value; based on determining the query contains the at least one qualitative value, translating the at least one qualitative value into a corresponding quantitative value based on the table mapping -these limitations recite mental processes, i.e., observation, evaluation, judgement, and opinion that a person can practically perform in mind or with pen and paper. Claims 3, 10, 17: extracting one or more structured values from the received query; converting the extracted one or more structured values into one or more unstructured values; executing an updated table mapping between the extracted one or more structured values and the one or more unstructured values - these limitations recite mental processes, i.e., observation, evaluation, judgement, and opinion that a person can practically perform in mind or with pen and paper. Claims 4, 11, 18: based on determining that the one or more data values include simulation data, executing the table mapping between the processed one or more quantitative values and the one or more qualitative values in the simulation data - these limitations recite mental processes, i.e., observation, evaluation, judgement, and opinion that a person can practically perform in mind or with pen and paper. Claims 5, 12, 19: based on determining that the one or more data values include a live quantitative value and a live qualitative value, executing the table mapping between the live quantitative value and the live qualitative value - these limitations recite mental processes, i.e., observation, evaluation, judgement, and opinion that a person can practically perform in mind or with pen and paper. Claims 6, 13, 20: based on determining that the one or more data values include a live quantitative value, ingesting the live quantitative value; generating a live qualitative value corresponding to the live quantitative value; and executing the table mapping between the live qualitative value and the live quantitative value - these limitations recite mental processes, i.e., observation, evaluation, judgement, and opinion that a person can practically perform in mind or with pen and paper. Claims 7 and 14: creating a fuzzy logic threshold of acceptance for the rule engine; executing a string to pseudo-sentiment transversal on a first qualitative value and a second qualitative value, wherein the first qualitative value includes a fuzzy logic expected value; generating a fuzzy logic score for the second qualitative value based on the string to pseudo-sentiment transversal; and based on determining the fuzzy logic score is within the fuzzy logic threshold of acceptance, assigning the fuzzy logic expected value to the second qualitative value - these limitations recite mental processes, i.e., observation, evaluation, judgement, and opinion that a person can practically perform in mind or with pen and paper. Thus, the claims recite the abstract idea. Step 2A Prong 2 Dependent Claims Additional elements Claims 2, 9, 16: receiving a query from a use; outputting, by the rule engine, a response to the query by processing the corresponding quantitative value - these limitations amount to insignificant extra-solution activity of mere data gathering and outputting (see MPEP § 2106.05(g)). Claims 3, 10, 17: by a generative pre-trained transformer - these limitations are recited at a high-level of generality such that it amount to no more than using generic class of machine learning models to apply the judicial exception (see MPEP § 2106.05(f)). retraining the rule engine with the updated table mapping as a second input to the rule engine - these limitations merely recite an idea of a solution or outcome with no restriction on how the result is accomplished (see MPEP § 2106.05(f)). These limitations can also be viewed as generally linking the use of a judicial exception to the field of automated rule based decision system (see MPEP § 2106.05(h)). Claims 4, 11, 18: by the rule engine - these limitations are recited at a high-level of generality such that it amount to no more than using generic automated rule based decision system to apply the judicial exception (see MPEP § 2106.05(f)). Claims 5, 12, 19: by the rule engine - these limitations are recited at a high-level of generality such that it amount to no more than using generic automated rule based decision system to apply the judicial exception (see MPEP § 2106.05(f)). Claims 6, 13, 20: by the rule engine, by a large language model - these limitations are recited at a high-level of generality such that it amount to no more than using generic class of automated rule based decision system/ machine learning models to apply the judicial exception (see MPEP § 2106.05(f)). Accordingly, these additional elements do not integrate the judicial exception into a practical application because they do not impose any meaningful limits on practicing the abstract idea. The claims are directed to the abstract idea. Step 2B Dependent Claims Additional elements Claims 2, 9, 16: receiving a query from a use; outputting, by the rule engine, a response to the query by processing the corresponding quantitative value - these limitations amount to insignificant extra-solution activity of mere data gathering and outputting, which is well-understood, routine, and conventional activity (see MPEP § 2106.05(d), “receiving/ transmitting data”, “presenting offers”). Claims 3, 10, 17: by a generative pre-trained transformer - these limitations are recited at a high-level of generality such that it amount to no more than using generic class of machine learning models to apply the judicial exception (see MPEP § 2106.05(f)). retraining the rule engine with the updated table mapping as a second input to the rule engine - these limitations merely recite an idea of a solution or outcome with no restriction on how the result is accomplished (see MPEP § 2106.05(f)). These limitations can also be viewed as generally linking the use of a judicial exception to the field of automated rule based decision system (see MPEP § 2106.05(h)). Claims 4, 11, 18: by the rule engine - these limitations are recited at a high-level of generality such that it amount to no more than using generic automated rule based decision system to apply the judicial exception (see MPEP § 2106.05(f)). Claims 5, 12, 19: by the rule engine - these limitations are recited at a high-level of generality such that it amount to no more than using generic automated rule based decision system to apply the judicial exception (see MPEP § 2106.05(f)). Claims 6, 13, 20: by the rule engine, by a large language model - these limitations are recited at a high-level of generality such that it amount to no more than using generic class of automated rule based decision system/ machine learning models to apply the judicial exception (see MPEP § 2106.05(f)). Accordingly, these additional elements do not amount to significantly more than the judicial exception. As such, the claims are patent ineligible. 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, 5, 8, 12, 15, and 19 are rejected under 35 U.S.C. 103 as being unpatentable over Maeda et al. (US 5,802,509 A hereinafter Maeda) in view of Kofman et al. (US 2009/0276391 A1 hereinafter Kofman). Regarding Claim 1, Maeda teaches a computer-based method of converting quantitative engines to fuzzy qualitative engines (Abstract - a rule generation apparatus… numeric data of the training data is converted into categorical data expressed by symbols to generate an instance table; column 1, lines 10 to 12 - automatically generating fuzzy rules from training data including accumulated numerical data; column 5, lines 26 to 40 -automatic rule generation apparatus of FIG. 1 includes a label presenter 103, a rule induction (RI) device 106, and a rule converter 107; they may be constructed of a central processing unit (CPU) and a memory), the method comprising: receiving one or more payloads including one or more data values (column 1, lines 28 to 30 - the label presenter 103 receives as inputs thereto training data 101 and a division method option 102; column 1, lines 48 to 50 - the training data 101 includes a plurality of data items (records) and each data item includes attribute values related to several attributes (i.e., payloads including one or more data values)); processing one or more quantitative values in the received one or more payloads (column 1, lines 48 to 60 - the training data 101 includes a plurality of data items (records) and each data item includes attribute values related to several attributes; the attribute value represented by numeric data; the label presenter 103 converts numeric data selected from the attribute values of the training data 101 into categorical data; data assigned to a label is referred to as categorical data; numeric data belonging to the attribute are divided at several predetermined points (values) into groups (intervals) so as to assign labels to the respective intervals - thus, processing the quantitative values in the payloads/ records containing attribute values); executing a table mapping between the processed one or more quantitative values and one or more qualitative values (column 5, lines 46 to 48 - the label presenter 103 assigns labels "small", "middle", "large ", etc. to values of numeric data of the training data 101; column 7, lines 66 to 67 and column 8, lines 1 to 12 - the instance table 104 is obtained by converting the training data 101 by the label presenter 103; the numeric data of the training data 101 is replaced with categorical data corresponding thereto; the division information 105 indicates a definition domain and dividing points for each of the input and output variables; FIG. 7 shows the structure of the division information 105; a column 701 represents an input/output variable name; in columns 702, ai and bi indicate respectively the minimum and maximum values in the definition domain of the i-th input variable - thus, the correspondence between numeric ranges and labels is held in the division information, construed as the table mapping, which tabulates for each variable the minimum and maximum values and the diving points of the labeled intervals; column 14, lines 44 to 50 - in the label code correspondence table 2501, a reference code 2502 indicates that labels "small", "middle", and "large" defined for the attribute "balance" respectively correspond to an attribute value less than 0, an attribute value equal to or more than 0 and less than 1000, and an attribute value equal to or more than 1000 - thus, the instance table is the result of executing the mapping between the processed numerical attribute values in the training data and one or more qualitative values/ labels). Maeda does not explicitly disclose wherein training a rule engine with the table mapping as a first input to the rule engine. However, Maeda discloses generating fuzzy rules from the table mapping; the RI device 106 generates RI rules 109 from the instance table 104, which records the result of executing the table mapping/ division information 105, and the rule converter 107 generates fuzzy rules 110 from the division information 105, which is the table mapping itself (column 5, lines 28 to 35 - the label presenter 103 receives as inputs thereto training data 101 and a division method option 102 and produces an RI instance table 104; the RI device 106 receives as inputs thereto the RI instance table 104 and the RI option 108 and generates therefrom RI rules 109; the rule converter 107 receive as inputs thereto the division information 105 and the RI rules 109 and creates fuzzy rules 110) and column 15, lines 19 to 21 - RI rules produced by the rule induction are transformed into fuzzy rules according to division information and/or the label code correspondence table) and the membership function of fuzzy rules are further trained by learning (column 9, lines 63 to 67 - membership function of the fuzzy rules 110 thus prepared, the position and contour can be automatically improved through learning in accordance with the training data 101); Maeda applies the resulting fuzzy rules to infer outputs from numeric inputs (column 10, lines 7 to 9 - output values are inferred from numeric data according to the fuzzy rules 110), which implies training with the table mapping. In the same field of endeavor, Kofman teaches wherein training a rule engine with the table mapping as a first input to the rule engine ([0097] Expert system—program utilizing Knowledge Base and Inference Engine to respond to a user's request for advice; [0114] Inference Engine—software providing reasoning mechanism in an Expert System; [0121] Knowledge Base—part of an Expert System containing collection of facts and “if . . . then . . . ” rules (The expert system consisting of knowledge base of rules and inference engine is construed as rule engine); [0075] Continuous Data Type—numeric data; [0172] inference Engine 603 utilizing Rule Base 606 deduces output Term Sets; [0187] in step 716… continuous data is clustered; an Input/Output Variable Term Set is assigned to each cluster (e.g. High Sales); the Term Sets and their Membership Functions are used in the step 717; [0188] in step 717, an initial Neural Network structure is generated based on the Clustering performed - thus, qualitative term sets are assigned to clusters of the continuous data/ numeric data from which the expert system/ rule engine is built (i.e., first input to the rule engine); [0189] in step 718, Neuro-Fuzzy Expert System 304 is Automatically trained on the OLAP Cube's Dataset imported in step 710; the Expert System's structure and parameters are determined by Training its underlying Neural Network on the imported Dataset; [0023] term Sets' parameters on input and output universes of discourse are trained by a Backpropagation and Genetic algorithms; [0011] Fuzzy Logic Expert System implemented as a Neural Network derives its Rule Base from data). It would have been obvious to a person having ordinary skill in the art before the effective filing date of the claimed invention to have incorporated wherein training a rule engine with the table mapping as a first input to the rule engine, as taught by Kofman into Maeda. Doing so would be desirable because it would allow for system learning, noise tolerance, and infusing of high level thinking into Neural Networks (Kofman [0011]). As to dependent Claim 5, Maeda and Kofman teach all the limitations of Claim 1. Maeda further teaches wherein based on determining that the one or more data values include a live quantitative value and a live qualitative value, executing the table mapping between the live quantitative value and the live qualitative value (column 17, claim 1 - the training data including both numeric data and symbol data; column 16, lines 64 to 67 and column 17, lines 1 to 4 - the training data is converted into an instance table according to the label record correspondence table; any attribute symbolically expressed in the table of FIG. 25 is replaced with a label code assigned with an attribute value; whereas, any attribute numerically expressed in the table is classified according to a numeric range specified in the table so as to be substituted for a label code; column 13, lines 36 to 40 - personal credit assignment is decided according to personal information in a bank; column 14 - FIG. 21B shows correspondences between the deposit balance (small/middle/large/unknown) and label codes (0/1/2/3); FIG. 21D shows relationships between decision for credit (non-grantable/grantable) and label codes (0/1) - the figures 21A-D and 25 show credit decision example in which the numeric deposit balance and the qualitative decision for credit (grantable/ non-grandable); column 10, lines 60 to 67 - the amount of order has been determined by experts for the stock control problem according to experiences accumulated in a long period of time and then training data is already prepared (i.e., the training data are actual accumulated records rather than artificially generated data)). Kofman further teaches the rule engine ([0097] Expert system (i.e., rule engine)—program utilizing Knowledge Base and Inference Engine to respond to a user's request for advice; [0023] describe data type (crisp continuous, crisp categorical, fuzzy); Expert System generation continues with Clustering Automatically performed on each of the input and output Attributes' data with all Clustering parameters determined from the data; [0187] continuous data is clustered; an Input/Output Variable Term Set is assigned to each cluster (e.g. High Sales); in case of categorical data a Term Set is assigned for each category). Claims 8 and 12 are system claims corresponding to the method claims 1 and 5 above and therefore, rejected for the same reasons. Maeda further teaches wherein a computer system, the computer system comprising: one or more processors, one or more computer-readable memories, one or more computer-readable tangible storage medium, and program instructions stored on at least one of the one or more computer-readable tangible storage medium for execution by at least one of the one or more processors via at least one of the one or more computer-readable memories, wherein the computer system is capable of performing a method (column 5, lines 20 to 40 -FIG. 1 shows the configuration of the fuzzy rule automatic generation system; automatic rule generation apparatus of FIG. 1 includes a label presenter 103, a rule induction (RI) device 106, and a rule converter 107; they may be constructed of a central processing unit (CPU) and a memory; the training data 104, division information 105, instance table 104, RI rules 109, and fuzzy rules 110 are stored on a disk or in a memory to be respectively accessed from the apparatuses 103, 106 and 107 as necessary; column 10, lines 50 to 54 - the above processing procedure can be automatically accomplished by a processor). Claims 15 and 19 are product claims corresponding to the method claims 1 and 5 above and therefore, rejected for the same reasons. Maeda further teaches wherein a computer program product, the computer program product comprising: one or more computer-readable tangible storage medium and program instructions stored on at least one of the one or more computer-readable tangible storage medium, the program instructions executable by a processor capable of performing a method (column 5, lines 20 to 40 -FIG. 1 shows the configuration of the fuzzy rule automatic generation system; automatic rule generation apparatus of FIG. 1 includes a label presenter 103, a rule induction (RI) device 106, and a rule converter 107; they may be constructed of a central processing unit (CPU) and a memory; the training data 104, division information 105, instance table 104, RI rules 109, and fuzzy rules 110 are stored on a disk or in a memory to be respectively accessed from the apparatuses 103, 106 and 107 as necessary; column 10, lines 50 to 54 - the above processing procedure can be automatically accomplished by a processor). Claims 2, 9, and 16 are rejected under 35 U.S.C. 103 as being unpatentable over Maeda in view of Kofman, further in view of Setlur et al. (US 2020/0110779 A1 hereinafter Setlur). As to dependent Claim 2, Maeda and Kofman teach all the limitations of Claim 1. Maeda further teaches the table mapping (column 14, lines 44 to 50 - in the label code correspondence table 2501, a reference code 2502 indicates that labels "small", "middle", and "large" defined for the attribute "balance" respectively correspond to an attribute value less than 0, an attribute value equal to or more than 0 and less than 1000, and an attribute value equal to or more than 1000). However, Maeda and Kofman fail to expressly teach wherein receiving a query from a user; determining whether the query contains at least one qualitative value; based on determining the query contains the at least one qualitative value, translating the at least one qualitative value into a corresponding quantitative value; and outputting, by the rule engine, a response to the query by processing the corresponding quantitative value. In the same field of endeavor, Setlur teaches wherein receiving a query from a user ([0007] receiving user selection of a data source; receiving a first user input to specify a natural language command directed to the data source; [0147] user input of a natural language command “wineries with high points”); determining whether the query contains at least one qualitative value ([0059] inferencing module includes algorithms for inferring reasonable defaults for vague (e.g., ambiguous) concepts such as “low,” “expensive,” and “popular’; [0147] the command includes a vague concept “high points”); based on determining the query contains the at least one qualitative value, translating the at least one qualitative value into a corresponding quantitative value ([0145] the application infers, using the inferencing module, the vague concept “expensive” to range from [avg+1SD, max], where avg, SD and max are the average, standard deviation and maximum values for the numerical field “Price”; [0147] a range of values (91.52 to 100) corresponding to default values for “high Points”); and outputting, by the rule engine, a response to the query by processing the corresponding quantitative value (Abstract - executes the database queries to retrieve data sets from the data source, then generates and displays a data visualization of the retrieved data sets; [0147] a range of values (91.52 to 100) corresponding to default values for “high Points”; FIG. 8D shows a user modification 812 of the default range of values, from “91.52 to 100” in FIG. 8C to “89.62 to 100” in FIG. 8D; [0148] an updated visualization 816 in accordance with the modified values). It would have been obvious to a person having ordinary skill in the art before the effective filing date of the claimed invention to have incorporated wherein receiving a query from a user; determining whether the query contains at least one qualitative value; based on determining the query contains the at least one qualitative value, translating the at least one qualitative value into a corresponding quantitative value; and outputting, by the rule engine, a response to the query by processing the corresponding quantitative value, as taught by Setlur into Maeda and Kofman. Doing so would be desirable because it would improve the effectiveness of natural language interfaces on data visualization platforms by resolving ambiguous (e.g., vague) natural language utterances directed to a data source (Setlur [0044]). Claim 9 is a system claim corresponding to the method claim 2 above and therefore, rejected for the same reasons. Claim 16 is a product claim corresponding to the method claim 2 above and therefore, rejected for the same reasons. Claims 3, 10, and 17 are rejected under 35 U.S.C. 103 as being unpatentable over Maeda in view of Kofman and Setlur, further in view of Harkous et al. (US 11,847,424 B1 hereinafter Harkous). As to dependent Claim 3, Maeda, Kofman, and Setlur teach all the limitations of Claim 2. Setlur further teaches wherein extracting one or more structured values from the received query ([0096] the user enters a command “price over 200 by variety” in the command box; the natural language processing module combines adjacent words “price over 200” into a phrase and determines a function for the phrase; the natural language processing module interprets the phrase “price over 200” as having a filter concept). Maeda further teaches wherein retraining the rule engine with the updated table mapping (column 14, lines 15 to 20 - the label code correspondence table is created when the user inputs a label minimum values and a label maximum value for each of the variables or according to a portion of training data and a division method option; column 15, lines 19 to 21 - RIrules produced by the rule induction are transformed into fuzzy rules according to division information and the label code correspondence table; column 9, lines 63 to 66 - for each membership function of the fuzzy rules prepared, the position and contour can be automatically improved through learning in accordance with the training data). Kofman further teaches wherein updated table mapping as a second input to the rule engine ([0172] inference Engine 603 utilizing Rule Base 606 deduces output Term Sets; [0187] in step 716… continuous data is clustered; an Input/Output Variable Term Set is assigned to each cluster (e.g. High Sales); the Term Sets and their Membership Functions are used in the step 717; [0188] in step 717, an initial Neural Network structure is generated based on the Clustering performed; [0024] an interactive loop wherein user decides whether to create another view of the Cube's data and send a different set of data to the Application for Neuro-Fuzzy Expert System generation for analysis). However, Maeda, Kofman, and Setlur fail to expressly teach wherein converting, by a generative pre-trained transformer, the extracted one or more structured values into one or more unstructured values; executing an updated table mapping between the extracted one or more structured values and the one or more unstructured values. In the same field of endeavor, Harkous teaches wherein converting, by a generative pre-trained transformer, the extracted one or more structured values into one or more unstructured values (column 5, lines 6 to 13 - the data-to-text language model may be a multi-layer autoregressive language model, such as the OpenAI GPT-2 model; the data-to-text language model may be any neural language model, such as a transformer decoder language model, such as a causal transformer decoder configured to receive a structured data input and output text data that semantically represents the structured data); executing an updated table mapping between the extracted one or more structured values and the one or more unstructured values (column 5, lines 9 to 13 - the data-to-text language model may be any neural language model, such as a transformer decoder language model, such as a causal transformer decoder configured to receive a structured data input and output text data that semantically represents the structured data; column 5, lines 43 to 46 - parameters of the data-to-text language model may be updated using a dataset data, text tuples; the data. text tuples stored in a training data repository - thus, the structured data and corresponding text/ unstructured values are stored as paired tuples used to update the model). It would have been obvious to a person having ordinary skill in the art before the effective filing date of the claimed invention to have incorporated wherein converting, by a generative pre-trained transformer, the extracted one or more structured values into one or more unstructured values; executing an updated table mapping between the extracted one or more structured values and the one or more unstructured values, as taught by Harkous into Maeda, Kofman, and Setlur. Doing so would be desirable because it would improve accuracy of the model (Harkous, column 3, lines 1 to 10). Claim 10 is a system claim corresponding to the method claim 3 above and therefore, rejected for the same reasons. Claim 17 is a product claim corresponding to the method claim 3 above and therefore, rejected for the same reasons. Claims 4, 11, and 18 are rejected under 35 U.S.C. 103 as being unpatentable over Maeda in view of Kofman, further in view of Mitsubuchi et al. (US 5,608,846 A hereinafter Mitsubuchi). As to dependent Claim 4, Maeda and Kofman teach all the limitations of Claim 1. Maeda further teaches wherein executing the table mapping between the processed one or more quantitative values and the one or more qualitative values in the data (column 7, lines 66 to 67 and column 8, lines 1 to 12 - the instance table 104 is obtained by converting the training data 101 by the label presenter 103; the numeric data of the training data 101 is replaced with categorical data corresponding thereto; the division information 105 indicates a definition domain and dividing points for each of the input and output variables; FIG. 7 shows the structure of the division information 105; a column 701 represents an input/output variable name; in columns 702, ai and bi indicate respectively the minimum and maximum values in the definition domain of the i-th input variable; column 14, lines 44 to 50 - in the label code correspondence table 2501, a reference code 2502 indicates that labels "small", "middle", and "large" defined for the attribute "balance" respectively correspond to an attribute value less than 0, an attribute value equal to or more than 0 and less than 1000, and an attribute value equal to or more than 1000 - thus, executing the mapping between the processed numerical attribute values in the data and one or more qualitative values/ labels). However, Maeda and Kofman fail to expressly teach wherein based on determining, by the rule engine, that the one or more data values include simulation data, executing the values in the simulation data. In the same field of endeavor, Mitsubuchi teaches wherein based on determining, by the rule engine, that the one or more data values include simulation data, executing the values in the simulation data (column 10, lines 43 to 50 - a data generator 801 provides sample input data; the input data are fed both into a neural network and into the fuzzy rule generator; the neural network then generates corresponding output data which synthesizes a fuzzy system simulation of the neural network's input/ output characteristics based upon the sampled input data from data generator and the output characteristics of the neural network; membership functions and rules are then generated which simulate the neural network's characteristics to be employed in a fuzzy system - thus, the fuzzy rule generator received artificially generated sample data and generates membership functions and rules from it; Abstract - fuzzy rules and membership functions are synthesized by observing a sample output/input data array). It would have been obvious to a person having ordinary skill in the art before the effective filing date of the claimed invention to have incorporated wherein based on determining, by the rule engine, that the one or more data values include simulation data, executing the values in the simulation data, as taught by Mitsubuchi into Maeda and Kofman. Doing so would be desirable because it would provide systematic ways to design and tune fuzzy systems (Mitsubuchi, column 1, lines 29 to 30). Claim 11 is a system claim corresponding to the method claim 4 above and therefore, rejected for the same reasons. Claim 18 is a product claim corresponding to the method claim 4 above and therefore, rejected for the same reasons. Claims 6, 13, and 20 are rejected under 35 U.S.C. 103 as being unpatentable over Maeda in view of Kofman, further in view of Harkous et al. (US 11,847,424 B1 hereinafter Harkous). As to dependent Claim 6, Maeda and Kofman teach all the limitations of Claim 1. Maeda further teaches wherein based on determining, by the rule engine, that the one or more data values include a live quantitative value, and executing the table mapping between the live qualitative value and the live quantitative value (column 17, claim 1 - the training data including both numeric data and symbol data; column 16, lines 64 to 67 and column 17, lines 1 to 4 - the training data is converted into an instance table according to the label record correspondence table; any attribute symbolically expressed in the table of FIG. 25 is replaced with a label code assigned with an attribute value; whereas, any attribute numerically expressed in the table is classified according to a numeric range specified in the table so as to be substituted for a label code; column 13, lines 36 to 40 - personal credit assignment is decided according to personal information in a bank; column 14 - FIG. 21B shows correspondences between the deposit balance (small/middle/large/unknown) and label codes (0/1/2/3); FIG. 21D shows relationships between decision for credit (non-grantable/grantable) and label codes (0/1) - the figures 21A-D and 25 show credit decision example in which the numeric deposit balance and the qualitative decision for credit (grantable/ non-grandable); column 10, lines 60 to 67 - the amount of order has been determined by experts for the stock control problem according to experiences accumulated in a long period of time and then training data is already prepared (i.e., the training data are actual accumulated records rather than artificially generated data)). Kofman further teaches the rule engine ([0097] Expert system (i.e., rule engine)—program utilizing Knowledge Base and Inference Engine to respond to a user's request for advice; [0023] describe data type (crisp continuous, crisp categorical, fuzzy); Expert System generation continues with Clustering Automatically performed on each of the input and output Attributes' data with all Clustering parameters determined from the data; [0187] continuous data is clustered; an Input/Output Variable Term Set is assigned to each cluster (e.g. High Sales); in case of categorical data a Term Set is assigned for each category). However, Maeda and Kofman fail to expressly teach wherein ingesting, by a large language model, the live quantitative value; generating a live qualitative value corresponding to the live quantitative value. In the same field of endeavor, Harkous teaches wherein ingesting, by a large language model, the live quantitative value (column 5, lines 6 to 13 - the data-to-text language model may be a multi-layer autoregressive language model, such as the OpenAI GPT-2 model; the data-to-text language model may be any neural language model, such as a transformer decoder language model, such as a causal transformer decoder configured to receive a structured data input and output text data that semantically represents the structured data; column 4, lines 25 to 28 - for the query. "how Old is Barack Obama?" a knowledge graph may be used to generate structured result data comprising the triple [barack_obama] [his_age] [age:integer["58"]] - thus, the structured data includes numeric values); generating a live qualitative value corresponding to the live quantitative value (column 5, lines 9 to 13 - the data-to-text language model may be any neural language model, such as a transformer decoder language model, such as a causal transformer decoder configured to receive a structured data input and output text data that semantically represents the structured data (i.e., corresponding qualitative value)). It would have been obvious to a person having ordinary skill in the art before the effective filing date of the claimed invention to have incorporated wherein ingesting, by a large language model, the live quantitative value; generating a live qualitative value corresponding to the live quantitative value, as taught by Harkous into Maeda and Kofman. Doing so would be desirable because it would improve accuracy of the model (Harkous, column 3, lines 1 to 10). Claim 13 is a system claim corresponding to the method claim 6 above and therefore, rejected for the same reasons. Claim 20 is a product claim corresponding to the method claim 6 above and therefore, rejected for the same reasons. Claims 7 and 14 are rejected under 35 U.S.C. 103 as being unpatentable over Maeda in view of Kofman, further in view of Yu (US 2005/0262210 A1). As to dependent Claim 7, Maeda and Kofman teach all the limitations of Claim 1. Kofman further teaches wherein the first qualitative value includes a fuzzy logic expected value ([0096] average value of a Term Set with Gaussian Membership Function (i.e., each qualitative term has an expected value); [0168] the Term Set High of the Output Variable Sales is defined as Gaussian shaped Membership Function with Expectation of $56,260.64; [0195] input type..fuzzy (e.g. Temperature=Warm); [0187] in case of categorical data a Term Set is assigned for each category; [0197] input source data is Fuzzified: corresponding input Term Sets and Memberships are determined for each Input Variable's value - thus, qualitative values are entered as input to the rule engine and matched to their term sets). However, Maeda and Kofman fail to expressly teach wherein creating a fuzzy logic threshold of acceptance for the rule engine; executing a string to pseudo-sentiment transversal on a first qualitative value and a second qualitative value; generating a fuzzy logic score for the second qualitative value based on the string to pseudo-sentiment transversal; and based on determining the fuzzy logic score is within the fuzzy logic threshold of acceptance, assigning the fuzzy logic expected value to the second qualitative value. In the same field of endeavor, Yu teaches wherein creating a fuzzy logic threshold of acceptance for the rule engine ([0107] if the distance is within a threshold then detect as matched; [0038]-[0039] if it is below a threshold then assume that it is close enough (i.e., match threshold set for accepting fuzzy match)); executing a string to pseudo-sentiment transversal on a first qualitative value and a second qualitative value ([0034]-[0036] compute the number and type of transformations (e.g. letter substitutions) required to convert the word/phrase in the spam message to the word/phrase in the spam dictionary (i.e., each dictionary term/ first qualitative value is compared with words of the text/ second qualitative value by computing the transformations needed to convert one string into another); [0106] Compute distance from each of the spam words - thus, the comparison of two text strings produces a measure of similarity/ string to pseudo-sentiment transversal); generating a fuzzy logic score for the second qualitative value based on the string to pseudo-sentiment transversal ([0034]-[0036] compute the number and type of transformations (e.g. letter substitutions) required to convert the word/phrase in the spam message to the word/phrase in the spam dictionary; [0037] each type of transformation has a weight; sum the weight of all transformations required (i.e., weighted transformation score/ fuzzy logic score is computed)); and based on determining the fuzzy logic score is within the fuzzy logic threshold of acceptance, assigning the fuzzy logic expected value to the second qualitative value ([0107] if the distance is within a threshold then detect as matched; [0062-[0063] determine whether text has been purposely obfuscated versus legit use of the word or an accidental misspelling of the word; map the obfuscated text to the intended text (i.e., an accidentally misspelled word is mapped to the intended dictionary term and handled as that term)). It would have been obvious to a person having ordinary skill in the art before the effective filing date of the claimed invention to have incorporated wherein creating a fuzzy logic threshold of acceptance for the rule engine; executing a string to pseudo-sentiment transversal on a first qualitative value and a second qualitative value; generating a fuzzy logic score for the second qualitative value based on the string to pseudo-sentiment transversal; and based on determining the fuzzy logic score is within the fuzzy logic threshold of acceptance, assigning the fuzzy logic expected value to the second qualitative value, as taught by Yu into Maeda and Kofman. Doing so would be desirable because it would provide features for text and message analysis that work effectively even on obfuscated text and messages (Yu [0007]). Claim 14 is a system claim corresponding to the method claim 7 above and therefore, rejected for the same reasons. Conclusion The prior art made of record and not relied upon is considered pertinent to applicant's disclosure. Applicant is required under 37 CFR § 1.111(c) to consider these references fully when responding to this action. Ganti et al. (US 2011/0314010 A1) teaches: The query 18 may be received as a phrase, such as “small HiTech notebook,” and may be partitioned into a series of keywords 32, such as “small,” “HiTech,” and “notebook”; for each keyword 32, the keyword map 48 may be consulted to retrieve an associated query predicate 44 and confidence score 46; the keyword 20 “portable” may be associated with the fragment “weight <7.0”; a query predicate 44 that targets a numeric attribute 14 may specify this query restriction in various ways, such as a numeric range (e.g., the keyword 20 “light” might be translated as the query predicate 44 “weight <7.0” (see [0031], [0077]). Any inquiry concerning this communication or earlier communications from the examiner should be directed to REJI KARTHOLY whose telephone number is (571)272-3432. The examiner can normally be reached on Monday - Thursday from 7:30 am to 3:30 pm. If attempts to reach the examiner by telephone are unsuccessful, the examiner’s supervisor, Jennifer Welch, can be reached at telephone number 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 Patent Center. Status information for published applications may be obtained from Patent Center. Status information for unpublished applications is available through Patent Center for authorized users only. Should you have questions about access to Patent Center, contact the Electronic Business Center (EBC) at 866-217-9197 (toll-free). 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) Form at https://www.uspto.gov/patents/uspto-automated- interview-request-air-form. /REJI KARTHOLY/Primary Examiner, Art Unit 2143
Read full office action

Prosecution Timeline

Apr 19, 2024
Application Filed
Oct 28, 2026
Non-Final Rejection mailed — §101, §103 (current)

Precedent Cases

Applications granted by this same examiner with similar technology

Patent 12775820
OPTIMIZING METHOD AND COMPUTING APPARATUS FOR DEEP LEARNING NETWORK AND COMPUTER-READABLE STORAGE MEDIUM
4y 1m to grant Granted Oct 27, 2026
Patent 12744835
CROSS-PLATFORM DATA REFRESH FOR COMMUNICATION PROCESS FLOWS
3y 9m to grant Granted Sep 22, 2026
Patent 12725035
METHOD OF AND APPARATUS FOR MACHINE LEARNING IN A RADIO NETWORK
3y 10m to grant Granted Sep 01, 2026
Patent 12694287
PRIVACY-PRESERVING FEDERATED MACHINE LEARNING
4y 3m to grant Granted Jul 28, 2026
Patent 12694308
NEIGHBORHOOD-BASED LINK PREDICTION FOR RECOMMENDATION SYSTEMS
3y 7m to grant Granted Jul 28, 2026
Study what changed to get past this examiner. Based on 5 most recent grants.

Strategy Recommendation AI-generated — please review before filing

Get a prosecution strategy drawn from examiner precedents, rejection analysis, and claim mapping.
Typically takes 5-10 seconds — AI-generated, attorney review required before filing

Prosecution Projections

1-2
Expected OA Rounds
63%
Grant Probability
99%
With Interview (+69.6%)
3y 2m (~7m remaining)
Median Time to Grant
Low
PTA Risk
Based on 166 resolved cases by this examiner. Grant probability derived from career allowance rate.

Sign in with your work email

Enter your email to receive a magic link. No password needed.

Personal email addresses (Gmail, Yahoo, etc.) are not accepted.

Free tier: 3 strategy analyses per month