DETAILED ACTION
Notice of Pre-AIA or AIA Status
The present application, filed on or after March 16, 2013, is being examined under the first inventor to file provisions of the AIA .
Response to Amendment
The Amendment filed July 24, 2026, has been entered. Claims 1 – 20 are pending in the application. Applicant’s amendments to the Drawings and Specification have overcome each and every objection previously set forth in the Non-Final Office Action mailed April 24, 2026.
Response to Arguments
Applicant’s arguments, filed July 24, 2026, with respect to claims 1 – 20 have been considered but are moot because the new ground of rejection does not rely on any reference applied in the prior rejection of record for any teaching or matter specifically challenged in the argument.
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, 15 – 18 and 20 are rejected under 35 U.S.C. 103 as being unpatentable over Saha et al. ("MURMUR: Modular Multi-Step Reasoning for Semi-Structured Data-to-Text Generation"), hereinafter Saha, in view of Dale ("Navigating the text generation revolution: Traditional data-to-text NLG companies and the rise of ChatGPT").
Regarding claim 1, Saha discloses a method for generating an electronic document containing natural language text with a natural language generation (NLG) system using a plurality of text segment configurations including a first text segment configuration and a second text segment configuration, the NLG system communicatively coupled to at least one data store, the NLG system including a symbolic NLG engine and a machine learning (ML) NLG engine (Abstract, lines 7-16, "We propose MURMUR, a neuro-symbolic modular approach to text generation from semi-structured data with multi-step reasoning. MURMUR is a best-first search method that generates reasoning paths using: (1) neural and symbolic modules with specific linguistic and logical skills, (2) a grammar whose production rules define valid compositions of modules, and (3) value functions that assess the quality of each reasoning step.”), the method comprising:
using at least one computer hardware processor (Abstract, lines 17-21, "We conduct experiments on two diverse data-to-text generation tasks like WebNLG and LogicNLG. The tasks differ in their data representations (graphs and tables) and span multiple linguistic and logical skills."; Conducting experiments on data-to-text generation tasks demonstrates the use of a processor.) to perform:
generating a first portion of the electronic document by using the first text segment configuration and the symbolic NLG engine to generate a first natural language text segment to include in the first portion of the document (Section 3.1, lines 1-6, "MURMUR defines a set of modules
{
M
i
}
i
=
1
m
that perform specialized reasoning skills for the corresponding task. Formally, each module
M
i
is defined as a multi-variate function
M
i
: X → y that maps an n-tuple input X = (
x
1
,· · · ,
x
n
) to an output y."; Section 3.1, lines 14-19, "The modules are implemented as few-shot neural models or symbolic functions. We choose few-shot neural modules for linguistic skills that LLMs typically excel at and symbolic modules for logical operations that LLMs mostly struggle with"; Section 3.2, lines 1-3, "The role of the grammar is to determine a set of plausible modules in a reasoning step and how they should be composed."; A symbolic module reads on a symbolic NLG engine, the module output reads on a natural language text segment, and the grammar reads on the text segment configuration.), wherein generating the first portion comprises:
obtaining a first text segment configuration, the first text segment configuration specifying a first set of one or more data variables (Section 3.2, lines 24-31, "Generating logical summaries from a table is a more challenging task. Based on the types of modules introduced previously, we define a grammar, as shown in Table 2. As an instance, the first rule encodes the knowledge that given an input of type Table, one can output a Table, a Row of the table, a Number, or a Boolean."; A grammar reads on a text segment configuration, and an input of type Table reads on a first set of one or more data variables.);
obtaining values of at least some of the first set of data variables using first data obtained from the at least one data store (Section 1, lines 85-91, "Our findings are: MURMUR can perform multistep generative reasoning on simple to complex semi-structured data-to-text generation tasks including WebNLG (Gardent et al., 2017), a graph-to-text task (§5) and LogicNLG (Chen et al., 2020a), a table-to-text task (§6)."; Section 2, lines 1-6, "A Reasoning Step is a triple (M, X, y) where a module M performs a certain skill by conditioning on an input X to generate an output y. For example, in Fig. 2, the module argmin takes a table and a column (points) as input and outputs the row with the minimum points."; Taking data from a table as input reads on obtaining values of a set of data variables, and a table reads on a data store.);
generating the first natural language text segment using the first text segment configuration, the values of the at least some of the first set of one or more data variables, and the symbolic NLG engine (Section 2, lines 1-6, "A Reasoning Step is a triple (M,X, y) where a module M performs a certain skill by conditioning on an input X to generate an output y. For example, in Fig. 2, the module argmin takes a table and a column (points) as input and outputs the row with the minimum points."; Section 3.1, lines 44-48, "For LogicNLG, drawing motivation from prior work (Chen et al., 2020c), we define different categories of symbolic modules that perform logical operations over tables (see Table 1 and refer to Table 8 for the detailed list)."; A module performing a certain skill by conditioning on an input to generate an output reads on generating a natural language text segment, a symbolic module reads on a symbolic NLG engine, and data from a table reads on values of a set of data variables.);
generating a second portion of the electronic document by using the second text segment configuration and the ML NLG engine to generate a second natural language text segment to include in the second portion of the document (Section 1, lines 74-79, "Neural modules perform linguistic skills that LLMs are good at (e.g., the Surface Realization module in Fig. 1 converts a reasoning path to a natural language summary) and symbolic modules perform logical skills that they mostly struggle with"; Section 3.1, lines 29-33, "In any modular data-to-text generation approach, one of the modules is responsible for the transition from structured data to unstructured text. We call it Surface Realization."; A neural module reads on a ML NLG engine, a module configured to perform surface realization reads on a second text segment configuration, and transitioning from structured data to unstructured text reads on generating a second portion of the electronic document.), wherein generating the second portion comprises:
obtaining an initial text segment (Section 3.1, lines 29-33, "In any modular data-to-text generation approach, one of the modules is responsible for the transition from structured data to unstructured text. We call it Surface Realization."; Structured data reads on an initial text segment.);
obtaining a second text segment configuration, the second text segment configuration specifying information to use for generating the second natural language text segment using the ML NLG engine from the initial text segment (Section 3.1, lines 29-39, "In any modular data-to-text generation approach, one of the modules is responsible for the transition from structured data to unstructured text. We call it Surface Realization. In particular, for WebNLG, we define it as
M
s
r
: t → s that converts a triple t (with data type Triple) into a short sentence s (with data type String). For LogicNLG, we define it as
M
s
r
: (t, p) → s that takes a table t (with data type Table) and a reasoning path p as input and converts it into a summary s (with data type String)."; A module configured to perform surface realization reads on a second text segment configuration.);
generating the second natural language text segment by using the initial text segment, the second text segment configuration, and the ML NLG engine (Section 1, lines 74-79, "Neural modules perform linguistic skills that LLMs are good at (e.g., the Surface Realization module in Fig. 1 converts a reasoning path to a natural language summary) and symbolic modules perform logical skills that they mostly struggle with"; Section 3.1, lines 29-33, "In any modular data-to-text generation approach, one of the modules is responsible for the transition from structured data to unstructured text. We call it Surface Realization."; A neural module performing surface realization to transition from structured data to unstructured text reads on generating a second natural language text segment by using the initial text segment, the second text segment configuration, and the ML NLG engine.);
and outputting the second natural language text segment from the ML NLG engine as the second portion of the electronic document (Section 1, lines 74-79, "Neural modules perform linguistic skills that LLMs are good at (e.g., the Surface Realization module in Fig. 1 converts a reasoning path to a natural language summary) and symbolic modules perform logical skills that they mostly struggle with"; Section 2, lines 14-17, "MURMUR generates textual summaries by constructing such reasoning paths that are then converted to the final outputs through a Surface Realization module, as shown in Fig. 2."; A neural module performing surface realization to generate the final outputs of textual summaries reads on outputting the second natural language text segment from the ML NLG engine as the second portion of the electronic document.);
and outputting the generated electronic document (Section 2, lines 14-17, "MURMUR generates textual summaries by constructing such reasoning paths that are then converted to the final outputs through a Surface Realization module, as shown in Fig. 2."; A final output reads on a generated electronic document.).
Saha does not specifically disclose: outputting the first natural language text segment from the symbolic NLG engine as the first portion of the electronic document.
Dale teaches:
outputting the first natural language text segment from the symbolic NLG engine as the first portion of the electronic document (Page 1196, lines 1-9, "Finally, there’s Textual.ai, whose text authoring platform provides the most full-blown integration of GPT as a central component. A key feature is the Copy Assistant, which assists in the development and management of prompts for both category descriptions and product descriptions. A prompt design tool enables reuse of previously constructed prompt elements and allows the user to fine-tune instructions in a multi-shot fashion, adding the actual product data to be added at the last step; once tested, the fine-tuned prompt can then be used for other products in the category. Prompt selection is triggered by specified combinations of metadata values. The text editing tool allows interleaving of template-generated and GPT-generated text blocks; machine translation is used to generate other language versions of templates."; Performing text authoring by interleaving template-generated and GPT-generated text blocks reads on outputting the first natural language text segment from the symbolic NLG engine as the first portion of the electronic document, where generating template-generated text blocks reads on a symbolic natural language generation engine and generating GPT-generated text blocks reads on a machine learning natural language generation engine.).
Dale is considered to be analogous to the claimed invention because it is in the same field of natural language generation. Therefore, it would have been obvious to someone of ordinary skill in the art before the effective filing date of the claimed invention to have modified Saha to incorporate the teachings of Dale to perform text authoring by interleaving template-generated and GPT-generated text blocks. Doing so would allow for text authoring of category descriptions and product descriptions (Dale; Page 1196, lines 1-9).
Regarding claim 2, Saha in view of Dale discloses the method as claimed in claim 1.
Saha further discloses:
wherein the ML NLG engine is configured to generate natural language text using a large language model (LLM) (Section 8, lines 1-10, "We presented MURMUR, a neuro-symbolic modular reasoning approach for data-to-text generation. MURMUR shows the benefits of building interpretable modular text generation systems by breaking a task down into sub-problems and then solving them through separate modules, without requiring module-specific supervision. It utilizes the power of LLMs in solving linguistic sub-tasks through in-context learning while delegating the logical subtasks to symbolic modules.").
Regarding claim 3, Saha in view of Dale discloses the method as claimed in claim 2.
Saha further discloses:
wherein the generating the second natural language text segment using the ML NLG engine further comprises: generating, using the second text segment configuration, a prompt from the initial text segment; providing the prompt as input to the LLM; and processing the prompt with the LLM to obtain the second natural language text segment (Section 3.1, lines 29-39, "In any modular data-to-text generation approach, one of the modules is responsible for the transition from structured data to unstructured text. We call it Surface Realization. In particular, for WebNLG, we define it as
M
s
r
: t → s that converts a triple t (with data type Triple) into a short sentence s (with data type String). For LogicNLG, we define it as
M
s
r
: (t, p) → s that takes a table t (with data type Table) and a reasoning path p as input and converts it into a summary s (with data type String)."; Section 4.1, lines 3-7, "We implement both modules, Surface Realization and Text Fusion as few-shot neural models by prompting OPT-175B (Zhang et al., 2022) with skill-specific prompts (see Appendix D) and greedy decoding."; Section 8, lines 1-10, "We presented MURMUR, a neuro-symbolic modular reasoning approach for data-to-text generation. MURMUR shows the benefits of building interpretable modular text generation systems by breaking a task down into sub-problems and then solving them through separate modules, without requiring module-specific supervision. It utilizes the power of LLMs in solving linguistic sub-tasks through in-context learning while delegating the logical subtasks to symbolic modules.; Implementing neural modules for surface realization and text fusion with skill-specific prompts, where the neural module is a large language model, reads on generating a prompt, providing the prompt as input to an LLM, and processing the prompt with the LLM to obtain a second natural language text segment.).
Regarding claim 4, Saha in view of Dale discloses the method as claimed in claim 1.
Saha further discloses:
wherein the information to use for generating the second natural language text segment specifies at least one transformation to be made to a grammatical aspect, content, tone, and/or style of the initial text segment (Section 3.1, lines 29-33, "In any modular data-to-text generation approach, one of the modules is responsible for the transition from structured data to unstructured text. We call it Surface Realization."; Transition from structured data to unstructured text reads on a transformation to be made to a grammatical aspect of the initial text segment.).
Regarding claim 5, Saha in view of Dale discloses the method as claimed in claim 1.
Saha further discloses:
wherein the initial text segment is the first natural language text segment output from the symbolic NLG engine (Section 2, lines 8-17, "Fig. 2 shows an example of a reasoning path, represented as a nested structure. It consists of three reasoning steps for three modules (argmin, hop, and eq). The argmin module outputs the row in the table with minimum points, which is the input to the next module hop that selects a column from that row. MURMUR generates textual summaries by constructing such reasoning paths that are then converted to the final outputs through a Surface Realization module, as shown in Fig. 2."; Section 3.1, lines 14-19, "The modules are implemented as few-shot neural models or symbolic functions. We choose few-shot neural modules for linguistic skills that LLMs typically excel at and symbolic modules for logical operations that LLMs mostly struggle with"; The argmin module outputting the row in the table with minimum points to be the input to the next module reads on the initial text segment being the first natural language text segment output from the symbolic NLG engine.).
Regarding claim 15, Saha in view of Dale discloses the method as claimed in claim 1.
Saha further discloses:
wherein the plurality of text segment configurations includes a third text segment configuration (Section 3.2, lines 24-31, "Generating logical summaries from a table is a more challenging task. Based on the types of modules introduced previously, we define a grammar, as shown in Table 2. As an instance, the first rule encodes the knowledge that given an input of type Table, one can output a Table, a Row of the table, a Number, or a Boolean."; A grammar reads on a text segment configuration.), the method comprising:
determining, using data in the third text segment configuration, whether to generate a third portion of the electronic document by using the ML NLG engine or the symbolic NLG engine (Section 3.1, lines 14-19, "The modules are implemented as few-shot neural models or symbolic functions. We choose few-shot neural modules for linguistic skills that LLMs typically excel at and symbolic modules for logical operations that LLMs mostly struggle with"; A neural module reads on a ML NLG engine, and a symbolic module reads on a symbolic NLG engine.);
when it is determined to use the ML NLG engine to generate the third portion of electronic document, generating the third portion of the electronic document using the ML NLG engine (Section 1, lines 74-79, "Neural modules perform linguistic skills that LLMs are good at (e.g., the Surface Realization module in Fig. 1 converts a reasoning path to a natural language summary) and symbolic modules perform logical skills that they mostly struggle with"; Section 3.1, lines 29-33, "In any modular data-to-text generation approach, one of the modules is responsible for the transition from structured data to unstructured text. We call it Surface Realization."; A neural module performing surface realization to transition from structured data to unstructured text reads on generating a portion of the electronic document, and a neural module reads on a ML NLG engine.);
and when it is determined to use the symbolic NLG engine to generate the third portion of electronic document, generating the third portion of the electronic document using the symbolic NLG engine (Section 2, lines 1-6, "A Reasoning Step is a triple (M,X, y) where a module M performs a certain skill by conditioning on an input X to generate an output y. For example, in Fig. 2, the module argmin takes a table and a column (points) as input and outputs the row with the minimum points."; Section 3.1, lines 44-48, "For LogicNLG, drawing motivation from prior work (Chen et al., 2020c), we define different categories of symbolic modules that perform logical operations over tables (see Table 1 and refer to Table 8 for the detailed list)."; A module performing a certain skill by conditioning on an input to generate an output reads on generating a portion of electronic document, and a symbolic module reads on a symbolic NLG engine.).
Regarding claim 16, arguments analogous to claim 1 are applicable. In addition, Saha discloses: a system, comprising: at least one computer hardware processor; and at least one non-transitory computer readable storage medium storing processor- executable instructions (Abstract, lines 17-21, "We conduct experiments on two diverse data-to-text generation tasks like WebNLG and LogicNLG. The tasks differ in their data representations (graphs and tables) and span multiple linguistic and logical skills."; Conducting experiments on data-to-text generation tasks demonstrates the use of a processor executing instructions from memory.) that, when executed by the at least one computer hardware processor, cause the at least one computer hardware processor to perform the steps of claim 1.
Regarding claim 17, arguments analogous to claim 3 are applicable.
Regarding claim 18, arguments analogous to claim 1 are applicable. In addition, Saha discloses: at least one non-transitory computer readable storage medium storing processor- executable instructions that, when executed by at least one computer hardware processor (Abstract, lines 17-21, "We conduct experiments on two diverse data-to-text generation tasks like WebNLG and LogicNLG. The tasks differ in their data representations (graphs and tables) and span multiple linguistic and logical skills."; Conducting experiments on data-to-text generation tasks demonstrates the use of a processor executing instructions from memory.), cause the at least one computer hardware processor to perform the steps of claim 1.
Regarding claim 20, arguments analogous to claim 15 are applicable.
Claims 6 – 14 and 19 are rejected under 35 U.S.C. 103 as being unpatentable over Saha in view of Dale, and further in view of Salmon et al. (US Patent No. 11,449,687), hereinafter Salmon.
Regarding claim 6, Saha in view of Dale discloses the method as claimed in claim 1, but does not specifically disclose: wherein the first text segment configuration further specifies first attributes, a first vocabulary, and a first document structure configuration, and generating the first natural language text segment comprises: generating a first intermediate representation of the first text segment configuration using the first document structure configuration, the values of the first set of data variables, and the first attributes; generating a second intermediate representation of the first text segment configuration from the first intermediate representation of the first text segment configuration using the first vocabulary; and generating the first natural language text segment from the second intermediate representation of the first text segment configuration.
Salmon teaches:
wherein the first text segment configuration further specifies first attributes, a first vocabulary, and a first document structure configuration (Column 7, lines 21-27, "The techniques include a method comprising: (1) obtaining a first specification of the first semantic object, the first specification specifying a first set of one or more data variables of the first semantic object, first attributes of the first semantic object, a first vocabulary of the first semantic object, and a first document structure configuration of the first semantic object;"), and generating the first natural language text segment comprises:
generating a first intermediate representation of the first text segment configuration using the first document structure configuration, the values of the first set of data variables, and the first attributes (Column 7, lines 32-39, "generating the natural language text including first natural language text, using the first specification of the first semantic object, the values of at least some of the first set of data variables, and the NLG system, at least in part by: (4a) generating a first intermediate representation of the first semantic object using the first document structure configuration, the values of the first set of data variables, and the first attributes of the first semantic object;");
generating a second intermediate representation of the first text segment configuration from the first intermediate representation of the first text segment configuration using the first vocabulary (Column 7, lines 39-42, "generating a second intermediate representation of the first semantic object from the first intermediate representation using the first vocabulary of the first semantic object;");
and generating the first natural language text segment from the second intermediate representation of the first text segment configuration (Column 7, lines 42-44, "generating the first natural language text from the second intermediate representation of the first semantic object”).
Salmon is considered to be analogous to the claimed invention because it is in the same field of natural language generation. Therefore, it would have been obvious to someone of ordinary skill in the art before the effective filing date of the claimed invention to have modified Saha in view of Dale to incorporate the teachings of Salmon to obtain a specification of a semantic object that specifies a set of data variables, attributes, a vocabulary, and a document structure configuration of the semantic object, generate natural language text using the specification of the semantic object by generating a first intermediate representation of the semantic object using the document structure configuration, the values of the set of data variables, and the attributes of the semantic object, generating a second intermediate representation of the semantic object from the first intermediate representation using the vocabulary of the semantic object, and generating natural language text from the second intermediate representation of the semantic object. Doing so would allow for automating the configuration of various stages of the natural language generation process including lexicalization, aggregation, referential expression generation, and surface realization (Salmon; Column 6, lines 1-21).
Regarding claim 7, Saha in view of Dale, and further in view of Salmon, discloses the method as claimed in claim 6.
Salmon further teaches:
wherein the first document structure configuration specifies multiple document structures, wherein the method further comprises selecting from among the multiple document structures to obtain a selected document structure, and wherein generating the first intermediate representation is performed using the selected document structure (Column 8, lines 28-36, "In some embodiments, the first document structure configuration for the first semantic objects specifies multiple document structures (e.g., as shown in FIG. 3D where two document structures called “first variant” and “second variant” are illustrated), wherein the method further comprises selecting from among the multiple document structures to obtain a selected document structure, and wherein generating the first intermediate representation is performed using the selected document structure.").
Salmon is considered to be analogous to the claimed invention because it is in the same field of natural language generation. Therefore, it would have been obvious to someone of ordinary skill in the art before the effective filing date of the claimed invention to have modified Saha in view of Dale, and further in view of Salmon, to further incorporate the teachings of Salmon to implement a document structure configuration for the first semantic object specifying multiple document structures, select from among the multiple document structures to obtain a selected document structure, and generate the first intermediate representation using the selected document structure. Doing so would allow for automating the configuration of various stages of the natural language generation process including lexicalization, aggregation, referential expression generation, and surface realization (Salmon; Column 6, lines 1-21).
Regarding claim 8, Saha in view of Dale, and further in view of Salmon, discloses the method as claimed in claim 6.
Salmon further teaches:
wherein the first text segment configuration further comprises a content selection configuration indicating a subset of the first attributes to use for generating the natural language text, and wherein generating the first intermediate representation of the first text segment configuration is performed using the content selection configuration (Column 8, lines 44-50, 'In some embodiments, the first specification of the first semantic object further comprises a content selection configuration indicating a subset of the first attributes to use for generating the natural language text, and wherein generating the first intermediate representation of the firs semantic object is performed using the content selection configuration.").
Salmon is considered to be analogous to the claimed invention because it is in the same field of natural language generation. Therefore, it would have been obvious to someone of ordinary skill in the art before the effective filing date of the claimed invention to have modified Saha in view of Dale, and further in view of Salmon, to further incorporate the teachings of Salmon to implement a content selection configuration indicating a subset of attributes to use for generating natural language text, and generate the first intermediate representation of a semantic object using the content selection configuration. Doing so would allow for automating the configuration of various stages of the natural language generation process including lexicalization, aggregation, referential expression generation, and surface realization (Salmon; Column 6, lines 1-21).
Regarding claim 9, Saha in view of Dale, and further in view of Salmon, discloses the method as claimed in claim 6.
Salmon further teaches:
wherein the first text segment configuration further comprises a micro-planning configuration, and wherein the method further comprises: applying automatic aggregation to the first intermediate representation of the first text segment configuration using the micro-planning configuration (Column 8, lines 51-56, "In some embodiments, the first specification of the first semantic object further comprises a micro-planning configuration (e.g., as illustrated in FIG. 3F), and wherein the method further comprises: applying automatic aggregation to the first intermediate representation of the first semantic object using the micro-planning configuration.").
Salmon is considered to be analogous to the claimed invention because it is in the same field of natural language generation. Therefore, it would have been obvious to someone of ordinary skill in the art before the effective filing date of the claimed invention to have modified Saha in view of Dale, and further in view of Salmon, to further incorporate the teachings of Salmon to implement a micro-planning configuration and apply automatic aggregation to a first intermediate representation of a semantic object using the micro-planning configuration. Doing so would allow for automating the configuration of various stages of the natural language generation process including lexicalization, aggregation, referential expression generation, and surface realization (Salmon; Column 6, lines 1-21).
Regarding claim 10, Saha in view of Dale, and further in view of Salmon, discloses the method as claimed in claim 6.
Salmon further teaches:
wherein the first text segment configuration further comprises a micro-planning configuration, and wherein the method further comprises: applying referent generation to the second intermediate representation of the first text segment configuration using the micro-planning configuration (Column 8, lines 57-61, 'In some embodiments, the first specification of the first semantic object further comprises a micro-planning configuration, and the method further comprises: applying referent generation to the second intermediate representation of the first semantic object using the micro-planning configuration.").
Salmon is considered to be analogous to the claimed invention because it is in the same field of natural language generation. Therefore, it would have been obvious to someone of ordinary skill in the art before the effective filing date of the claimed invention to have modified Saha in view of Dale, and further in view of Salmon, to further incorporate the teachings of Salmon to implement a micro-planning configuration and apply referent generation to the second intermediate representation of a semantic object using the micro-planning configuration. Doing so would allow for automating the configuration of various stages of the natural language generation process including lexicalization, aggregation, referential expression generation, and surface realization (Salmon; Column 6, lines 1-21).
Regarding claim 11, Saha in view of Dale discloses the method as claimed in claim 1, but does not specifically disclose: wherein the first text segment configuration further specifies a first analysis configuration, and wherein determining the values of at least one of the first set of data variables comprises processing the data obtained from the at least one data store using the first analysis configuration.
Salmon teaches:
wherein the first text segment configuration further specifies a first analysis configuration, and wherein determining the values of at least one of the first set of data variables comprises processing the data obtained from the at least one data store using the first analysis configuration (Column 8, lines 37-43, "In some embodiments, the first specification of the first semantic object further specifies a first analysis configuration (e.g., as illustrated in FIG. 3B), and wherein determining the values of at least one of the first set of data variables comprises processing the data obtained from the at least one data store using the first analysis configuration (e.g., as illustrated in FIGS. 4B and 4C).").
Salmon is considered to be analogous to the claimed invention because it is in the same field of natural language generation. Therefore, it would have been obvious to someone of ordinary skill in the art before the effective filing date of the claimed invention to have modified Saha in view of Dale to incorporate the teachings of Salmon to implement a first analysis configuration and determine the values of a set of data variables by processing the data obtained from a data store using the first analysis configuration. Doing so would allow for automating the configuration of various stages of the natural language generation process including lexicalization, aggregation, referential expression generation, and surface realization (Salmon; Column 6, lines 1-21).
Regarding claim 12, Saha in view of Dale discloses the method as claimed in claim 1, but does not specifically disclose: wherein the at least one data store is external to the NLG system and the NLG system is communicatively coupled to the at least one data store using a communication network.
Salmon teaches:
wherein the at least one data store is external to the NLG system and the NLG system is communicatively coupled to the at least one data store using a communication network (Column 7, lines 16-21, "Some embodiments are directed to techniques for generating natural language text with a natural language generation (NLG) system using a plurality of semantic objects including a first semantic object, the NLG system communicatively coupled to at least one data store (e.g., a database external to the system). "; Column 11, lines 12-21, "In some embodiments, database interface module 124 may be configured to access business data 113 from business data store(s) 112. This may be done in any suitable way. In some embodiments, the database interface module 124 may be configured to obtain data from (either pull data from or be provided data by) the business data store(s) 112. The data may be provided via a communication network (not shown), such as the Internet or any other suitable network, as aspects of the technology described herein are not limited in this respect.").
Salmon is considered to be analogous to the claimed invention because it is in the same field of natural language generation. Therefore, it would have been obvious to someone of ordinary skill in the art before the effective filing date of the claimed invention to have modified Saha in view of Dale to incorporate the teachings of Salmon to implement a natural language generation system communicatively coupled to a data store external to the natural language generation system via a communication network. Doing so would allow for automating the configuration of various stages of the natural language generation process including lexicalization, aggregation, referential expression generation, and surface realization (Salmon; Column 6, lines 1-21).
Regarding claim 13, Saha in view of Dale discloses the method as claimed in claim 1, but does not specifically disclose: wherein outputting the electronic document comprises providing the natural language text to a publishing system external to the NLG system.
Salmon teaches:
wherein outputting the electronic document comprises providing the natural language text to a publishing system external to the NLG system (Column 9, lines 1-8, "In some embodiments, outputting the natural language text generated by an NLG system comprises providing the natural language text to a publishing system external to the NLG system. The publishing system may be configured to generate an electronic document (e.g., a webpage, a PDF file, a text document, etc.) including the natural language text; and transmit the electronic document over at least one communication network to a user.").
Salmon is considered to be analogous to the claimed invention because it is in the same field of natural language generation. Therefore, it would have been obvious to someone of ordinary skill in the art before the effective filing date of the claimed invention to have modified Saha in view of Dale to incorporate the teachings of Salmon to provide natural language text to a publishing system external to the natural language generation system. Doing so would allow for automating the configuration of various stages of the natural language generation process including lexicalization, aggregation, referential expression generation, and surface realization (Salmon; Column 6, lines 1-21).
Regarding claim 14, Saha in view of Dale discloses the method as claimed in claim 1, but does not specifically disclose: further comprising transmitting the electronic document over at least one communication network to a user.
Salmon teaches:
further comprising transmitting the electronic document over at least one communication network to a user (Column 9, lines 1-8, "In some embodiments, outputting the natural language text generated by an NLG system comprises providing the natural language text to a publishing system external to the NLG system. The publishing system may be configured to generate an electronic document (e.g., a webpage, a PDF file, a text document, etc.) including the natural language text; and transmit the electronic document over at least one communication network to a user.").
Salmon is considered to be analogous to the claimed invention because it is in the same field of natural language generation. Therefore, it would have been obvious to someone of ordinary skill in the art before the effective filing date of the claimed invention to have modified Saha in view of Dale to incorporate the teachings of Salmon to transmit the electronic document over at least one communication network to a user. Doing so would allow for automating the configuration of various stages of the natural language generation process including lexicalization, aggregation, referential expression generation, and surface realization (Salmon; Column 6, lines 1-21).
Regarding claim 19, arguments analogous to claim 6 are applicable.
Conclusion
The prior art made of record and not relied upon is considered pertinent to applicant's disclosure:
Barnes et al. (Barnes, Emily, and James Hutson, "Natural Language Processing and Neurosymbolic AI: The Role of Neural Networks with Knowledge-Guided Symbolic Approaches", January 2024, Journal of Artificial Intelligence and Robotics, Vol. 2, No. 1, pp. 1-13.) teaches performing natural language processing by merging the pattern recognition of neural networks with the structured reasoning of symbolic AI.
Garcez et al. (Garcez, Artur d’Avila, and Luis C. Lamb, "Neurosymbolic AI: the 3rd wave”, March 2023, Artificial Intelligence Review, Vol. 56, No. 11, pp. 12387-12406.) teaches neural network-based learning with symbolic knowledge representation and logical reasoning.
Nye et al. (Nye, Maxwell, Michael Tessler, Josh Tenenbaum, and Brenden M. Lake, "Improving Coherence and Consistency in Neural Sequence Models with Dual-System, Neuro-Symbolic Reasoning", December 2021, 35th Conference on Neural Information Processing Systems (NeurIPS 2021), pp. 1-13.) teaches examining candidate generations from a neural sequence model by a symbolic reasoning module.
Applicant's amendment necessitated the new ground(s) of rejection presented in this Office action. Accordingly, THIS ACTION IS MADE FINAL. See MPEP § 706.07(a). Applicant is reminded of the extension of time policy as set forth in 37 CFR 1.136(a).
A shortened statutory period for reply to this final action is set to expire THREE MONTHS from the mailing date of this action. In the event a first reply is filed within TWO MONTHS of the mailing date of this final action and the advisory action is not mailed until after the end of the THREE-MONTH shortened statutory period, then the shortened statutory period will expire on the date the advisory action is mailed, and any nonprovisional extension fee (37 CFR 1.17(a)) pursuant to 37 CFR 1.136(a) will be calculated from the mailing date of the advisory action. In no event, however, will the statutory period for reply expire later than SIX MONTHS from the mailing date of this final action.
Any inquiry concerning this communication or earlier communications from the examiner should be directed to James Boggs whose telephone number is (571)272-2968. The examiner can normally be reached M-F 8:00 AM - 5:00 PM.
Examiner interviews are available via telephone, in-person, and video conferencing using a USPTO supplied web-based collaboration tool. To schedule an interview, applicant is encouraged to use the USPTO Automated Interview Request (AIR) at http://www.uspto.gov/interviewpractice.
If attempts to reach the examiner by telephone are unsuccessful, the examiner’s supervisor, Daniel Washburn can be reached at (571)272-5551. The fax phone number for the organization where this application or proceeding is assigned is 571-273-8300.
Information regarding the status of published or unpublished applications may be obtained from Patent Center. Unpublished application information in Patent Center is available to registered users. To file and manage patent submissions in Patent Center, visit: https://patentcenter.uspto.gov. Visit https://www.uspto.gov/patents/apply/patent-center for more information about Patent Center and https://www.uspto.gov/patents/docx for information about filing in DOCX format. For additional questions, contact the Electronic Business Center (EBC) at 866-217-9197 (toll-free). If you would like assistance from a USPTO Customer Service Representative, call 800-786-9199 (IN USA OR CANADA) or 571-272-1000.
/JAMES BOGGS/Examiner, Art Unit 2657
Prosecution Insights