DETAILED ACTION
1. This communication is in response to the Amendments and Arguments filed on 1/5/2026. Claims 1-22 are pending and have been examined.
Response to Amendments and Arguments
2. Regarding applicant’s amended claims, note that these claims necessitate the new ground(s) of rejection presented in this Office action, because these claims introduce new issue and/or change the scope of the claims.
With the amendments, 35 USC 101 rejections (mental process) are maintained, with the rejections further including mathematical operations, where no particular model or training step (such as particular type of neural network and training) are specified for the “machine learning model” to overcome rejections on mental process and mathematical operations. In addition, no particular device or transmission methods are recited for the steps of “receiving” and “sending.” Likewise, “interactive interface” is subject to BRI without reciting any particular device. The applicant is advised to consider adding these limitations, also to overcome the 35 USC 101 rejections on using generic computing devices (not a particular solution or a particular way) for "integrating the judicial exception into a practical application," "improvement of any technical field," and "solves a problem rooted in computer technology and software.”.
With the amendments, 35 USC 103 rejections have been fully considered, but they are not persuasive. In particular, the applicant argues that the cited references do not teach the limitation “dividing the first input into first chunks; converting each of the first chunks into a respective numerical representation; dividing the second input into second chunks; converting each of the second chunks into a respective numerical representation; determining, based on the numerical representations, one or more relevant chunks of the first chunks that contain information associated with the second input ..” In response, the examiner respectfully disagrees.
Note that PARK teaches: [0008] “automatic segmentation of contact center calls by applying natural language processing and machine learning technologies” and [0056] “Once a transcript with segments has been produced, the information can be useful for future calls .. By matching only the “Question” sections of the in-coming and previous calls, one can more easily find calls having identical or similar problems.” SEO teaches: [0014] “performing embedding by representing the identified word by a vector and vectorizing the word through part-of-speech tagging <read on to generate numeric representations>” and [0012] “structuring a relationship <read on ‘determining relevant chunks’> of the named entity based on a word embedding method ..”
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.
3. Claims 1-9, 12-20 are rejected under 35 U.S.C. 101 because the claimed invention is directed to an abstract idea without significantly more. The independent claims 1, 12 recite a system and a method thus relating to a statutory category.
Claims 1, 12 further recite “A system for entity extraction .. each entity is data to be extracted from the first input; creating a tailored input based on the second input .. providing a processed interactive output to the user .. dividing the first input into first chunks; converting each of the first chunks into a respective numerical representation .. determining, based on the numerical representations, one or more relevant chunks of the first chunks that contain information associated with the second input .. wherein the interactive interface comprises the processed output.” The limitations as drafted cover mental process and mathematical operations. More specifically, a human receiving a body of text such as a natural language article can mentally generate a tailored text as a combination of the article and a query for extraction and output of an entity in the article where ‘interactive interface’ is subject to BRI. Note that no particular model or training step (such as particular type of neural network and training) are specified for the “machine learning model” to overcome rejections on mental process and mathematical operations. In addition, no particular device or transmission methods are recited for the steps of “receiving” and “sending.”
This judicial exception is not integrated into a practical application. In particular, independent claim 1 recites additional element of “processor” which amounts to general purpose computing devices, as specified in the Specification [26] “processing device 103 may include any type of single or multi-core processor, mobile device microcontroller, central processing unit, or any circuitry that performs logic operations.” Accordingly, these additional elements do not integrate the abstract idea into a practical application because it does not impose any meaningful limits on practicing the abstract idea. The claim is directed to an abstract idea.
Claims 1, 12 do not include additional elements that are sufficient to amount to significantly more than the judicial exception. Therefore, the claimed limitations are directed towards insignificant solution activity. The claims are not patent eligible.
With respect to dependent claims 2, 13, the claims further recite “the second input comprises free form text” where a human is able to handle and manipulate free-form text by mental process. The claims do not include additional elements that are sufficient to amount to significantly more than the judicial exception.
With respect to dependent claims 3, 14, the claims further recite “each of the at least one element comprises .. a name for the entity” where a human is able to handle and manipulate an entity name by mental process. The claims do not include additional elements that are sufficient to amount to significantly more than the judicial exception.
With respect to dependent claims 4, 15, the claims further recite “the at least one search term comprises at least one of: a first location in the body of text, the first location associated with the entity; or a data format ..” where a human is able to handle and manipulate an entity location by mental process. The claims do not include additional elements that are sufficient to amount to significantly more than the judicial exception.
With respect to dependent claims 5, 16, the claims further recite “the tailored input comprises: a predefined header comprising a description of the first input and instructions that direct the machine learning model to identify, locate, and output information associated with each of the at least one entity in the first input; the first input; and the second input” where a human is able to handle and manipulate a natural language tailored input by mental process. The claims do not include additional elements that are sufficient to amount to significantly more than the judicial exception.
With respect to dependent claims 6, 17, the claims further recite “the machine learning model is a large language model” where a human with mental process can function like a large language model. The claims do not include additional elements that are sufficient to amount to significantly more than the judicial exception.
With respect to dependent claims 7, 18, the claims further recite “the machine learning model is not trained, using training data similar to the first input.” The claims recite “the machine learning model is not trained” and therefore do not include additional elements that are sufficient to amount to significantly more than the judicial exception.
With respect to dependent claims 8, 19, the claims further recite “the output comprises .. extracted information for each of the at least one entity ..” where a human is able to handle and manipulate an extracted entity and its related information by mental process. The claims do not include additional elements that are sufficient to amount to significantly more than the judicial exception.
With respect to dependent claims 9, 20, the claims further recite “converting the output into a predefined format; and validating each extracted information” where a human is able to convert extracted information in a predefined format and verify the information by mental process. The claims do not include additional elements that are sufficient to amount to significantly more than the judicial exception.
Claim Rejections - 35 USC § 103
4. Claims 1-22 are rejected under 35 U.S.C. 103 as being unpatentable over Calapodescu, et al. (US 20170300565; hereinafter Calapodescu) in view of Park (US 20100104086A1; hereinafter PARK), further in view of Seo, et al. (US 20200035228; hereinafter SEO), and further in view of Fan, et al. (US 20090292687; hereinafter FAN).
As per claim 1, Calapodescu (Title: System and method for entity extraction from semi-structured text documents) discloses “A system for entity extraction (Calapodescu, Title) comprising: at least one processor; and at least one non-transitory computer-readable medium containing instructions that, when executed by the system, cause the system to perform operations comprising:
receiving, from a user, a first input comprising a body of text to be processed for information (Calapodescu, [0040], The system receives as input a text document, such as a resume, for processing <read on ‘a first input’ and ‘processor, memory/medium’>);
receiving, from the user, a second input comprising a set of at least one element, wherein each of the at least one element comprises information associated with an entity, and wherein each entity is data to be extracted from the first input (Calapodescu, [0002], In order to classify, search, or sort the resumes, recruiters use dedicated tools called Applicant Tracking Systems (ATS). These tools populate a database with information about candidates, such as personal information, skills, and companies they have worked for, etc. <read on ‘a second input .. comprises information associated with an entity’>);
[ dividing the first input into first chunks; converting each of the first chunks into a respective numerical representation; dividing the second input into second chunks; converting each of the second chunks into a respective numerical representation; determining, based on the numerical representations, one or more relevant chunks of the first chunks that contain information associated with the second input ];
[ creating a tailored input for a machine learning model based on the second input, wherein the tailored input comprises the determined one or more relevant chunk ]; sending the tailored input to the machine learning model; receiving an output from the machine learning model; processing the output; and providing an interactive interface to the user, wherein the interactive interface comprises the processed output (Calapodescu, [0040], The system receives as input a text document, such as a resume, for processing <read on ‘a first input’>; [0057], machine learning component may use the extracted entities for learning a model for extracting other information from resumes; [0019], Information based on the extracted new entities in the resume is output <where ‘The system receives as input .. Information based on the extracted new entities in the resume is output’ reads on ‘interactive interface’ with the user>).”
Calapodescu does not expressly disclose “dividing the first input into first chunks .. dividing the second input into second chunks ..” However, the limitation is taught by PARK (Title: System and method for automatic call segmentation at call center).
In the same field of endeavor, PARK teaches: [0008] “automatic segmentation of contact center calls by applying natural language processing and machine learning technologies” and [0056] “Once a transcript with segments has been produced, the information can be useful for future calls .. By matching only the “Question” sections of the in-coming and previous calls, one can more easily find calls having identical or similar problems.”
Therefore, it would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to incorporate the teachings of PARK in the system (as taught by Calapodescu) to segment the textual script and the corresponding voice recording into sections for matching and further processing.
Calapodescu in view of PARK does not expressly disclose “converting each of the first chunks into a respective numerical representation .. converting each of the second chunks into a respective numerical representation; determining, based on the numerical representations, one or more relevant chunks of the first chunks that contain information associated with the second input ..” However, the limitation is taught by SEO (Title: Method and apparatus for speech recognition).
In the same field of endeavor, SEO teaches: [0014] “performing embedding by representing the identified word by a vector and vectorizing the word through part-of-speech tagging <read on to generate numeric representations>” and [0012] “structuring a relationship <read on ‘determining relevant chunks’> of the named entity based on a word embedding method ..”
Therefore, it would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to incorporate the teachings of SEO in the system (as taught by Calapodescu in view of PARK) to represent textual script and the corresponding voice recording in numerical vector form based on word embedding for efficient calculation of similarity measure for relevance between any two vectors.
Calapodescu in view of PARK and SEO does not explicitly disclose “creating a tailored input .. based on the second input, wherein the tailored input comprises the determined one or more relevant chunk ..” However, this limitation is taught by FAN (Title: System and method for providing question and answers with deferred type evaluation).
In the same field of endeavor, FAN teaches: [Abstract], processing a query <read on ‘second input’> including waiting until a “Type” (i.e. a descriptor) <read on ‘determined one or more relevant chunk’> is determined AND a candidate answer is provided <where the final overall query reads on ‘a tailored input’>).”
Therefore, it would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to incorporate the teachings of FAN in the system (as taught by Calapodescu, PARK and SEO) to combine multiple input data as an integrated input query <read on ‘a tailored input’> for target entity extraction in any text document.
As per claim 2 (dependent on claim 1), Calapodescu in view of PARK, SEO and FAN further discloses “wherein the second input comprises free form text (Calapodescu, [0002], In order to classify, search, or sort the resumes, recruiters use dedicated tools called Applicant Tracking Systems (ATS). These tools populate a database with information about candidates, such as personal information, skills, and companies they have worked for, etc. <read on ‘free form text’>).”
As per claim 3 (dependent on claim 1), Calapodescu in view of PARK, SEO and FAN further discloses “wherein each of the at least one element comprises at least one of: a name for the entity; at least one synonym of the name; at least one keyword associated with the entity; a description of the entity; or at least one search term (Calapodescu, [0002], In order to classify, search, or sort the resumes .. populate a database with information about candidates, such as personal information <read on ‘a name for the entity’>).”
As per claim 4 (dependent on claim 3), Calapodescu in view of PARK, SEO and FAN further discloses “wherein the at least one search term comprises at least one of: a first location in the body of text, the first location associated with the entity; or a data format, the data format associated with the output from the machine learning model (Calapodescu, [0016], At least a subset of the extracted entities in the first set is clustered into clusters, based on locations of the entities in the document).”
As per claim 5 (dependent on claim 1), Calapodescu in view of PARK, SEO and FAN further discloses “wherein the tailored input comprises: a predefined header comprising a description of the first input and instructions that direct the machine learning model to identify, locate, and output information associated with each entity in the first input; the first input; and the second input (Calapodescu, [0040], The system receives as input a text document, such as a resume, for processing <read on ‘a first input’>; FAN, [Abstract], processing a query <read on ‘second input’> including waiting until a “Type” (i.e. a descriptor) <read on ‘a predefined header’> is determined AND a candidate answer is provided <where the final overall query reads on ‘a tailored input’>).”
As per claim 6 (dependent on claim 1), Calapodescu in view of PARK, SEO and FAN further discloses “wherein the machine learning model is a large language model (Calapodescu, Title <read on text and language>; [0057], machine learning component may use the extracted entities for learning a model for extracting other information from resumes <to clarify - claim 6 and claim 7 appear to be contradictory>).”
As per claim 7 (dependent on claim 1), Calapodescu in view of PARK, SEO and FAN further discloses “wherein the machine learning model is not trained, using training data similar to the first input, to extract each entity (to clarify - the machine learning model is to process ‘tailored input’ different from the first input).”
As per claim 8 (dependent on claim 1), Calapodescu in view of PARK, SEO and FAN further discloses “wherein the output comprises at least one of: extracted information for each entity; a second location in the first input where the extracted information is located; or an explanation why each of the extracted information is associated with its respective entity (Calapodescu, [0077], information may be output by the system, based on the entities identified).”
As per claim 9 (dependent on claim 1), Calapodescu in view of PARK, SEO and FAN further discloses “wherein processing the output comprises: converting the output into a predefined format; and validating each extracted information (Calapodescu, [0058], The output component outputs information extracted by the first and second entity extraction components, such as the sequences of entities and their classes, e.g., in table format or other structured data format; [0161], The human annotations of the corpora were cross-validated by different people in the team <read on a ready mechanism to validate any information>; FAN, [Abstract], processing a query .. including waiting until a “Type” (i.e. a descriptor) is determined; SPEC [40], validating the extracted data may involve validating the data type for each entity).”
As per claim 10 (dependent on claim 8), Calapodescu in view of PARK, SEO and FAN further discloses “wherein providing the interactive further interface (?) to the user comprises: displaying the first input; displaying each of the at least one entity of the second input in a list; displaying each extracted information; and creating at least one user-interactive element for each of the at least one entity (Calapodescu, [0077], information may be output by the system, based on the entities identified .. Relevant documents, as identified by the machine learning model, may be displayed <read on displaying any information>; [0040], The system receives as input a text document, such as a resume, for processing; [0019], Information based on the extracted new entities in the resume is output <where ‘The system receives as input .. Information based on the extracted new entities in the resume is output’ reads on ‘interactive’ I/O with the user>).”
As per claim 11 (dependent on claim 10), Calapodescu in view of PARK, SEO and FAN further discloses “wherein the at least one user-interactive element comprises at least one of: a first user-interactive element that is configured to display the explanation; and a second user-interactive element that is configured to navigate the user to the second location in the displayed first input (Calapodescu, [0077], information may be output by the system, based on the entities identified .. Relevant documents, as identified by the machine learning model, may be displayed <read on a ready mechanism for displaying any information>; [0040], The system receives as input a text document, such as a resume, for processing; [0019], Information based on the extracted new entities in the resume is output <where ‘The system receives as input .. Information based on the extracted new entities in the resume is output’ reads on ‘interactive’ I/O with the user>).”
Claims 12-22 (similar in scope to claims 1-11, respectively) are rejected under the same rationale as detailed above for claims 1-11, respectively.
Conclusion
5. Applicant's amendment necessitates 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 extension fee 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 date of this final action.
Any inquiry concerning this communication or earlier communications from the examiner should be directed to FENG-TZER TZENG whose telephone number is 571-272-4609. The examiner can normally be reached on M-F (8:30-5:00). The fax phone number where this application or proceeding is assigned is 571-273-4609.
If attempts to reach the examiner by telephone are unsuccessful, the examiner’s supervisor, Paras Shah (SPE) can be reached on 571-270-1650.
Information regarding the status of an application may be obtained from the Patent Application Information Retrieval (PAIR) system. Status information for published applications may be obtained from either Private PAIR or Public PAIR. Status information for unpublished applications is available through Private PAIR only. For more information about the PAIR system, see http://pair-direct.uspto.gov. Should you have questions on access to the Private PAIR system, contact the Electronic Business Center (EBC) at 866-217-9197 (toll-free). If you would like assistance from a USPTO Customer Service Representative or access to the automated information system, call 800-786-9199 (IN USA OR CANADA) or 571-272-1000.
/FENG-TZER TZENG/ 3/23/2026Primary Examiner, Art Unit 2653