DETAILED ACTION
Status of Claims
This Office action is responsive to communications filed on 2025-11-30. Claim(s) 1-6 is/are pending and are examined herein.
Claim(s) 1-6 is/are objected to.
Claim(s) 4 is/are rejected under 35 USC 112(b).
Claim(s) 1-6 is/are rejected under 35 USC 101.
Claim(s) 1-6 is/are rejected under 35 USC 103.
Notice of Pre-AIA or AIA Status
The present application, filed on or after 2013-03-16, is being examined under the first inventor to file provisions of the AIA .
Response to Arguments
Regarding rejections under 35 USC 112(b), the applicant’s amendments resolve some, but not all, of the issues raised in the previous Office action. The amendments also introduce new issues. Unresolved and newly introduced issues are indicated below. The examiner notes that the amendments help clarify claim scope but do not substantively change the scope of the claims.
Regarding the rejections under 35 USC 101, the applicant’s remarks have been fully considered but they are not persuasive:
The applicant states that “the method of the claims should performed by the computer” [remarks, page 2; sic]. While the claims do not at present explicitly recite a computer (indicating only that the method is “automatic”), even if the claimed method is interpreted as being implemented on a computer, the applicant appears to be asserting, as best understood by the examiner, that the mere fact that the claimed method is performed on a computer prevents it from being directed to an abstract idea. This assertion is inconsistent with the MPEP, which indicates in numerous places that steps that are performable by a human mind are not integrated into a practical application, and do not amount to significantly more than an abstract idea, merely by virtue of being performed on a computer. For example, MPEP 2106.04(a)(2)(III) indicates that “[c]ourts have examined claims that required the use of a computer and still found that the underlying, patent-ineligible invention could be performed via pen and paper or in a person’s mind”. Similarly, MPEP 2106.04(a)(2)(III)(C) indicates that “[c]laims can recite a mental process even if they are claimed as being performed on a computer”. Similarly, MPEP 2106.05(f) indicates that the “[u]se of a computer or other machinery in its ordinary capacity… or simply adding a general purpose computer or computer components after the fact to an abstract idea… does not integrate a judicial exception into a practical application or provide significantly more”.
The applicant asserts that the “claims recite an improvement to the field of natural language processing” and points specifically to “introducing prior knowledge” and “introducing knowledge connection” as the elements which provide the purported improvement [remarks, page 3]. However, both prior knowledge and knowledge connection are themselves abstract ideas: a human mind has prior knowledge, and a human mind can make connections between pieces of knowledge. MPEP 2106.05(a) indicates that “judicial exception alone cannot provide the improvement”. In other words, any purported improvement provided by the “prior knowledge” and the “knowledge connection” does not meet the requirements of the improvements analysis.
The applicant asserts generically that “the claims are applied in a meaningful way beyond generically linking the use of the judicial exception to a particular technological environment” [remarks, page 3]. This argument is not persuasive because the applicant has not clearly identified which additional elements explicitly recited in the claims, if any, actually serve to integrate the abstract ideas recited in the claim into a practical application.
The applicant asserts generically that “when the external data source is used, above additional elements in combination with the knowledge graph construction steps provide a computer implemented method for automatically constructing the knowledge graph” and that the claims therefore “amount to significantly more than the judicial exception” [remarks, page 4]. This argument is not persuasive because, again, the applicant has not clearly identified which additional elements explicitly recited in the claims, if any, lead the claims to amount to significantly more than the judicial exceptions. The applicant’s reference to “above additional elements” in their remarks is indeterminate; the only additional elements previously discussed in the applicant’s remarks appear to be (1) the fact that the knowledge graph construction method is performed by a computer, and (2) the external data source, but neither of these additional elements amount to significantly more than a judicial exception (see the first bullet point above for the former, and MPEP 2106.05(f) and/or MPEP 2106.05(h) for the latter).
The complete 101 analysis, with minor updates in view of the applicant’s amendments, is given below.
Regarding the rejections under 35 USC 103, the applicant’s remarks have been fully considered. The applicant asserts that Steps 4-6 of the independent claim are “not disclosed by Zhang and Wei” [remarks, page 7] without clearly indicating how the language that is actually used in these claim limitations patentably distinguishes them from the prior art made for record. While the applicant includes purported summaries of the prior art made of record [remarks, pages 5-6] (whose accuracy the examiner does not concede), and some remarks regarding purported points of contrast between the references and the invention [remarks, pages 6-7] (whose persuasiveness the examiner also does not concede), these remarks are generic in the sense that they are not focused on the language that is actually used in the claims. In other words, the applicant's assertions fail to comply with 37 CFR 1.111(b) because they amount to a general allegation that the claims define a patentable invention without specifically pointing out how the language of the claims patentably distinguishes them from the references. The examiner maintains that the pending claims (including the limitations regarding Steps 4-6) are not patentably distinguishable from the prior art made of record for the reasons indicated in the previous Office action and below. The complete prior art mapping, with minor updates in view of the applicant’s amendments, is given below.
Claim Objections
Claim(s) 1-6 is/are objected to because of the following informalities:
Claim 1 Step 1 recites obtaining prompt data by storing information relevant to a knowledge graph topic as character strings [emphasis added] but the underlined phrase remains syntactically ambiguous: it is unclear whether it is the information relevant to a knowledge graph topic that is to be stored “as character strings” or if it is the prompt data that is to be obtained “as character strings”. The specification suggests that it is the relevant topic information that is to be stored “as character strings” [specification, 0044], and for the purpose of compact prosecution, the claim is interpreted broadly as encompassing at least this interpretation. Alternative language that is syntactically unambiguous is advised (e.g., “Step 1: obtaining prompt data by storing information relevant to a knowledge graph topic is stored as character strings and is used for constructing a knowledge graph”). Dependent claims 2-5 inherit the objection.
Claim 1 Step 6 recites completing construction of a knowledge graph [emphasis added] but this should be “completing construction of the knowledge graph” for proper antecedent basis (because Claim 1 Step 1 already introduces a knowledge graph; cf. “wherein the information relevant to the knowledge graph topic is used for constructing a knowledge graph”). Dependent claims 2-5 inherit the objection.
Appropriate correction is required.
Claim Rejections - 35 USC 112(b)
The following is a quotation of 35 USC 112(b):
(b) CONCLUSION.—The specification shall conclude with one or more claims particularly pointing out and distinctly claiming the subject matter which the inventor or a joint inventor regards as the invention.
The following is a quotation of 35 USC 112 (pre-AIA ), second paragraph:
The specification shall conclude with one or more claims particularly pointing out and distinctly claiming the subject matter which the applicant regards as his invention.
Claim(s) 4 is/are rejected under 35 USC 112(b) or 35 USC 112 (pre-AIA ), second paragraph, as being indefinite for failing to particularly point out and distinctly claim the subject matter which the inventor or a joint inventor (or for applications subject to pre-AIA 35 USC 112, the applicant), regards as the invention.
Claim 4 continues to recite an expected function of the large language model agent of cognition/annotation/reasoning/association [emphasis added] but it is not clear what the intended scope of the “expected function” of a large language model agent is. Expectations regarding the function of a model can vary from person to person. MPEP 2173.05(b)(IV) indicates that, in the presence of subjective claim terminology, “[s]ome objective standard must be provided in order to allow the public to determine the scope of the claim. A claim term that requires the exercise of subjective judgment without restriction may render the claim indefinite”. The specification provides no objective guidance as to the interpretation of the “expected function” of these large language model agents. For the purpose of compact prosecution, the claim is interpreted broadly so that the “expected function” of the large language model agent of cognition, for example, encompasses any form of cognition, and so forth.
Claim Rejections - 35 USC 101
35 USC 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.
Claim(s) 1-6 is/are rejected under 35 USC 101 because the claimed invention(s) is/are directed to abstract ideas without significantly more.
Claim 1
Step 1. The claim and its dependents 2-6 fall under the statutory category of methods. An analysis of step 2 for each of these claims follows.
Step 2A Prong 1. The claim recites the following abstract ideas:
An [automatic] knowledge graph construction method based on prior knowledge and knowledge connection, comprising: (This recites a mental process that can be performed in the human mind or by a human using pen and paper. A human mind, perhaps assisted with pen and paper, can construct a knowledge graph based on prior knowledge and connections between knowledge. See MPEP 2106.04(a)(2)(III).)
and Step 6: configuring the effective data related to the knowledge graph topic and pre- defined input prompts as an input layer of a large language model automatic agent framework, (This recites a mental process that can be performed in the human mind or by a human using pen and paper. A human mind can regard data as input for a machine learning model. See MPEP 2106.04(a)(2)(III).)
obtaining entity-relation-entity triples [through multi-round feedback of the large language model automatic agent framework,] (This recites a mental process that can be performed in the human mind or by a human using pen and paper. A human mind can extract/obtain entity-relation-entity triples from textual data. See MPEP 2106.04(a)(2)(III).)
and completing construction of a knowledge graph; (This recites a mental process that can be performed in the human mind or by a human using pen and paper. A human mind can complete construction of a knowledge graph. See MPEP 2106.04(a)(2)(III).)
wherein Step 5 is as follows: Step 5.1: configuring the article paragraphs as extraction data and the prior knowledge as reference data; (This recites a mental process that can be performed in the human mind or by a human using pen and paper. A human mind can regard data as being of a particular type. See MPEP 2106.04(a)(2)(III).)
Step 5.2: configuring the pre-defined contrastive prompt text as prompt instructions (This recites a mental process that can be performed in the human mind or by a human using pen and paper. A human mind can regard data as being prompt instructions. See MPEP 2106.04(a)(2)(III).)
Step 5.3: based on the prompt instructions, extracting extended information of prior knowledge from the article paragraphs [by the knowledge-connecting large language model] based on the prior knowledge; (This recites a mental process that can be performed in the human mind or by a human using pen and paper. A human mind can extract information from paragraphs based on prompt instructions and their prior knowledge. See MPEP 2106.04(a)(2)(III).)
and Step 5.4: obtaining the effective data related to the knowledge graph topic by integrating the prior knowledge with the extended information. (This recites a mental process that can be performed in the human mind or by a human using pen and paper. A human mind can integrate their prior knowledge with new information. See MPEP 2106.04(a)(2)(III).)
Step 2A Prong 2. The claim recites the following additional elements which, considered individually and as an ordered combination, do not integrate the abstract idea into a practical application:
[An] automatic [knowledge graph construction method] (This recites merely applying (or equivalent) an abstract idea, or implementing an abstract idea on a computer, or using a computer as a tool to perform an abstract idea. See MPEP 2106.05(f).)
Step 1: obtaining prompt data by storing information relevant to a knowledge graph topic as character strings, (This recites insignificant extra-solution activity. See MPEP 2106.05(g).)
wherein the information relevant to the knowledge graph topic is used for constructing a knowledge graph; (This recites a general link between an abstract idea and a particular field of use or technological environment. See MPEP 2106.05(h).)
Step 2: retrieving and saving article paragraphs from an external data source based on the prompt data; (This recites insignificant extra-solution activity. See MPEP 2106.05(g).)
Step 3: respectively injecting prompt templates for four large language model agents of annotation, reasoning, cognition, and association; (This recites insignificant extra-solution activity. See MPEP 2106.05(g).)
Step 4: obtaining prior knowledge by inputting the injected prompt templates, the article paragraphs, and specific task requirements into the four large language model agents; (This recites merely recites obtaining generic output from a generic machine learning model based on certain input. In other words, this recites merely applying (or equivalent) an abstract idea, or implementing an abstract idea on a computer, or using a computer as a tool to perform an abstract idea. See MPEP 2106.05(f).)
Step 5: obtaining effective data related to the knowledge graph topic by inputting the article paragraphs, the prior knowledge, and a pre-defined contrasting prompt text into a knowledge-connecting large language model; (This recites merely recites obtaining generic output from a generic machine learning model based on certain input. In other words, this recites merely applying (or equivalent) an abstract idea, or implementing an abstract idea on a computer, or using a computer as a tool to perform an abstract idea. See MPEP 2106.05(f).)
[obtaining entity-relation-entity triples] through multi-round feedback of the large language model automatic agent framework, (This recites merely applying (or equivalent) an abstract idea, or implementing an abstract idea on a computer, or using a computer as a tool to perform an abstract idea. See MPEP 2106.05(f).)
and inputting to the knowledge-connecting large language model; (This recites insignificant extra-solution activity. See MPEP 2106.05(g).)
[extracting extended information of prior knowledge from the article paragraphs] by the knowledge-connecting large language model (This recites merely applying (or equivalent) an abstract idea, or implementing an abstract idea on a computer, or using a computer as a tool to perform an abstract idea. See MPEP 2106.05(f).)
Step 2B. The claim recites the following additional elements which, considered individually and as an ordered combination, do not amount to significantly more than the abstract idea:
[An] automatic [knowledge graph construction method] (This recites merely applying (or equivalent) an abstract idea, or implementing an abstract idea on a computer, or using a computer as a tool to perform an abstract idea. See MPEP 2106.05(f).)
Step 1: obtaining prompt data by storing information relevant to a knowledge graph topic as character strings, (This insignificant extra-solution activity is well-understood, routine, conventional as it is mere data transfer and storage. See MPEP 2106.05(d)(II), “Receiving or transmitting data over a network” and/or “Storing and retrieving information in memory” and/or “Electronic recordkeeping” and/or “Storing and retrieving information in memory”.)
wherein the information relevant to the knowledge graph topic is used for constructing a knowledge graph; (This recites a general link between an abstract idea and a particular field of use or technological environment. See MPEP 2106.05(h).)
Step 2: retrieving and saving article paragraphs from an external data source based on the prompt data; (This insignificant extra-solution activity is well-understood, routine, conventional as it is mere data transfer and storage. See MPEP 2106.05(d)(II), “Receiving or transmitting data over a network” and/or “Storing and retrieving information in memory” and/or “Electronic recordkeeping” and/or “Storing and retrieving information in memory”.)
Step 3: respectively injecting prompt templates for four large language model agents of annotation, reasoning, cognition, and association; (This insignificant extra-solution activity is well-understood, routine, conventional as it is mere data transfer. See MPEP 2106.05(d)(II), “Receiving or transmitting data over a network” and/or “Storing and retrieving information in memory”.)
Step 4: obtaining prior knowledge by inputting the injected prompt templates, the article paragraphs, and specific task requirements into the four large language model agents; (This recites merely recites obtaining generic output from a generic machine learning model based on certain input. In other words, this recites merely applying (or equivalent) an abstract idea, or implementing an abstract idea on a computer, or using a computer as a tool to perform an abstract idea. See MPEP 2106.05(f).)
Step 5: obtaining effective data related to the knowledge graph topic by inputting the article paragraphs, the prior knowledge, and a pre-defined contrasting prompt text into a knowledge-connecting large language model; (This recites merely recites obtaining generic output from a generic machine learning model based on certain input. In other words, this recites merely applying (or equivalent) an abstract idea, or implementing an abstract idea on a computer, or using a computer as a tool to perform an abstract idea. See MPEP 2106.05(f).)
[obtaining entity-relation-entity triples] through multi-round feedback of the large language model automatic agent framework, (This recites merely applying (or equivalent) an abstract idea, or implementing an abstract idea on a computer, or using a computer as a tool to perform an abstract idea. See MPEP 2106.05(f).)
and inputting to the knowledge-connecting large language model; (This insignificant extra-solution activity is well-understood, routine, conventional as it is mere data transfer. See MPEP 2106.05(d)(II), “Receiving or transmitting data over a network” and/or “Storing and retrieving information in memory”.)
[extracting extended information of prior knowledge from the article paragraphs] by the knowledge-connecting large language model (This recites merely applying (or equivalent) an abstract idea, or implementing an abstract idea on a computer, or using a computer as a tool to perform an abstract idea. See MPEP 2106.05(f).)
Claim 2
Step 2A Prong 1. The claim recites the following abstract ideas:
The abstract idea(s) in the parent claim(s).
[The automatic knowledge graph construction method based on prior knowledge and knowledge connection according to claim 1, wherein Step 1 is as follows:] Step 1.1: converting all original corpus data in different formats, comprising text, voice, and PDF images from actual scenarios, into a unified character text and performing preliminary integration to obtain original character data; (This recites a mental process that can be performed in the human mind or by a human using pen and paper. A human being using pen and paper can convert text, voice, and PDF data into text. See MPEP 2106.04(a)(2)(III).)
and Step 1.2: performing data cleaning on the original character data to eliminate blank data and redundant data, and obtaining the prompt data. (This recites a mental process that can be performed in the human mind or by a human using pen and paper. A human being can manually or mentally perform data cleaning. See MPEP 2106.04(a)(2)(III).)
Step 2A Prong 2. The claim recites the following additional elements which, considered individually and as an ordered combination, do not integrate the abstract idea into a practical application:
The additional element(s) in the parent claim(s).
Step 2B. The claim recites the following additional elements which, considered individually and as an ordered combination, do not amount to significantly more than the abstract idea:
The additional element(s) in the parent claim(s).
Claim 3
Step 2A Prong 1. The claim recites the following abstract ideas:
The abstract idea(s) in the parent claim(s).
[The automatic knowledge graph construction method based on prior knowledge and knowledge connection according to claim 1, wherein Step 2 is as follows:] Step 2.1: configuring the prompt data as an input to a retrieval framework; (This recites a mental process that can be performed in the human mind or by a human using pen and paper. A human mind can regard certain data as being input. See MPEP 2106.04(a)(2)(III).)
according to a maximum likelihood estimation search algorithm; (This recites a mathematical concept. Maximum likelihood estimation is a mathematical concept. See MPEP 2106.04(a)(2)(I).)
Step 2A Prong 2. The claim recites the following additional elements which, considered individually and as an ordered combination, do not integrate the abstract idea into a practical application:
The additional element(s) in the parent claim(s).
Step 2.2: retrieving a paragraph text with a highest similarity to the prompt data from the external data source (This recites insignificant extra-solution activity. See MPEP 2106.05(g).)
and Step 2.3: saving the retrieved paragraph text in the character string format. (This recites insignificant extra-solution activity. See MPEP 2106.05(g).)
in a character string format. (This recites data of a particular type or source, merely linking an abstract idea to a particular field of use. See MPEP 2106.05(h).)
Step 2B. The claim recites the following additional elements which, considered individually and as an ordered combination, do not amount to significantly more than the abstract idea:
The additional element(s) in the parent claim(s).
Step 2.2: retrieving a paragraph text with a highest similarity to the prompt data from the external data source (This insignificant extra-solution activity is well-understood, routine, conventional as it is mere data transfer. See MPEP 2106.05(d)(II), “Receiving or transmitting data over a network” and/or “Storing and retrieving information in memory”.)
and Step 2.3: saving the retrieved paragraph text in the character string format. (This insignificant extra-solution activity is well-understood, routine, conventional as it is mere data storage. See MPEP 2106.05(d)(II), “Electronic recordkeeping” and/or “Storing and retrieving information in memory”.)
in a character string format. (This recites data of a particular type or source, merely linking an abstract idea to a particular field of use. See MPEP 2106.05(h).)
Claim 4
Step 2A Prong 1. The claim recites the following abstract ideas:
The abstract idea(s) in the parent claim(s).
an expected function of the large language model agent of cognition; (This recites a mental process that can be performed in the human mind or by a human using pen and paper. A human mind can perform cognition. See MPEP 2106.04(a)(2)(III).)
an expected function of the large language model agent of annotation; (This recites a mental process that can be performed in the human mind or by a human using pen and paper. A human mind can perform annotation. See MPEP 2106.04(a)(2)(III).)
an expected function of the large language model agent of reasoning; (This recites a mental process that can be performed in the human mind or by a human using pen and paper. A human mind can perform reasoning. See MPEP 2106.04(a)(2)(III).)
an expected function of the large language model agent of association. (This recites a mental process that can be performed in the human mind or by a human using pen and paper. A human mind can perform association. See MPEP 2106.04(a)(2)(III).)
Step 2A Prong 2. The claim recites the following additional elements which, considered individually and as an ordered combination, do not integrate the abstract idea into a practical application:
The additional element(s) in the parent claim(s).
[The automatic knowledge graph construction method based on prior knowledge and knowledge connection according to claim 1, wherein Step 3 is as follows:] Step 3.1: injecting a cognitive prompt template according to [an expected function of the cognitive agent;] Step 3.2: injecting an annotation prompt template according to [an expected function of the annotation agent;] Step 3.3: injecting a reasoning prompt template according to [an expected function of the reasoning agent;] and Step 3.4: injecting an associative prompt template according to [an expected function of the associative agent.] (This recites insignificant extra-solution activity. See MPEP 2106.05(g).)
Step 2B. The claim recites the following additional elements which, considered individually and as an ordered combination, do not amount to significantly more than the abstract idea:
The additional element(s) in the parent claim(s).
[The automatic knowledge graph construction method based on prior knowledge and knowledge connection according to claim 1, wherein Step 3 is as follows:] Step 3.1: injecting a cognitive prompt template according to [an expected function of the cognitive agent;] Step 3.2: injecting an annotation prompt template according to [an expected function of the annotation agent;] Step 3.3: injecting a reasoning prompt template according to [an expected function of the reasoning agent;] and Step 3.4: injecting an associative prompt template according to [an expected function of the associative agent.] (This insignificant extra-solution activity is well-understood, routine, conventional as it is mere data transfer. See MPEP 2106.05(d)(II), “Receiving or transmitting data over a network” and/or “Storing and retrieving information in memory”.)
Claim 5
Step 2A Prong 1. The claim recites the following abstract ideas:
The abstract idea(s) in the parent claim(s).
and Step 4.4: generating prior knowledge related to the knowledge graph topic [by the four large language model agents] from four aspects based on the prompt templates, the article paragraphs, and the pre-defined generated data format from Step 4.1 to Step 4.3. (This recites a mental process that can be performed in the human mind or by a human using pen and paper. A human mind can produce/generate prior knowledge related to a certain topic. See MPEP 2106.04(a)(2)(III).)
Step 2A Prong 2. The claim recites the following additional elements which, considered individually and as an ordered combination, do not integrate the abstract idea into a practical application:
The additional element(s) in the parent claim(s).
[The automatic knowledge graph construction method based on prior knowledge and knowledge connection according to claim 1, wherein Step 4 is as follows:] Step 4.1: inputting the prompt templates as data correlation requirements to the four large language model agents; (This recites insignificant extra-solution activity. See MPEP 2106.05(g).)
Step 4.2: inputting the article paragraphs as an extraction corpus to the four large language model agents; (This recites insignificant extra-solution activity. See MPEP 2106.05(g).)
Step 4.3: inputting a pre-defined generated data format as the specific task requirements to the four large language model agents; (This recites insignificant extra-solution activity. See MPEP 2106.05(g).)
by the four large language model agents (This recites merely applying (or equivalent) an abstract idea, or implementing an abstract idea on a computer, or using a computer as a tool to perform an abstract idea. See MPEP 2106.05(f).)
Step 2B. The claim recites the following additional elements which, considered individually and as an ordered combination, do not amount to significantly more than the abstract idea:
The additional element(s) in the parent claim(s).
[The automatic knowledge graph construction method based on prior knowledge and knowledge connection according to claim 1, wherein Step 4 is as follows:] Step 4.1: inputting the prompt templates as data correlation requirements to the four large language model agents; (This insignificant extra-solution activity is well-understood, routine, conventional as it is mere data transfer. See MPEP 2106.05(d)(II), “Receiving or transmitting data over a network” and/or “Storing and retrieving information in memory”.)
Step 4.2: inputting the article paragraphs as an extraction corpus to the four large language model agents; (This insignificant extra-solution activity is well-understood, routine, conventional as it is mere data transfer. See MPEP 2106.05(d)(II), “Receiving or transmitting data over a network” and/or “Storing and retrieving information in memory”.)
Step 4.3: inputting a pre-defined generated data format as the specific task requirements to the four large language model agents; (This insignificant extra-solution activity is well-understood, routine, conventional as it is mere data transfer. See MPEP 2106.05(d)(II), “Receiving or transmitting data over a network” and/or “Storing and retrieving information in memory”.)
by the four large language model agents (This recites merely applying (or equivalent) an abstract idea, or implementing an abstract idea on a computer, or using a computer as a tool to perform an abstract idea. See MPEP 2106.05(f).)
Claim 6
Step 2A Prong 1. The claim recites the following abstract ideas:
The abstract idea(s) in the parent claim(s).
[The automatic knowledge graph construction method based on prior knowledge and knowledge connection according to claim 1, wherein Step 6 is as follows:] Step 6.1: setting agent types of two large language models in the large language model automatic agent framework as “customer” and “knowledge graph construction expert”; (This recites a mental process that can be performed in the human mind or by a human using pen and paper. A human mind can label models. See MPEP 2106.04(a)(2)(III).)
and Step 6.3: according to the multi-round feedback, obtaining the entity-relation-entity triples comprised in a local knowledge graph obtained by each round of the multi-round feedback, and finally obtaining a complete knowledge graph according to all local knowledge graphs. (This recites a mental process that can be performed in the human mind or by a human using pen and paper. A human mind can obtain entity-relation-entity triples over multiple rounds, and can complete knowledge graphs. See MPEP 2106.04(a)(2)(III).)
Step 2A Prong 2. The claim recites the following additional elements which, considered individually and as an ordered combination, do not integrate the abstract idea into a practical application:
The additional element(s) in the parent claim(s).
Step 6.2: transmitting the effective data related to the knowledge graph topic to the "customer" as context, and transmitting an input prompt to the "knowledge graph construction expert" as an instruction; (This recites insignificant extra-solution activity. See MPEP 2106.05(g).)
Step 2B. The claim recites the following additional elements which, considered individually and as an ordered combination, do not amount to significantly more than the abstract idea:
The additional element(s) in the parent claim(s).
Step 6.2: transmitting the effective data related to the knowledge graph topic to the "customer" as context, and transmitting an input prompt to the "knowledge graph construction expert" as an instruction; (This insignificant extra-solution activity is well-understood, routine, conventional as it is mere data transfer. See MPEP 2106.05(d)(II), “Receiving or transmitting data over a network” and/or “Storing and retrieving information in memory”.)
Claim Rejections - 35 USC 103
The following is a quotation of 35 USC 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.
This application currently names joint inventors. In considering patentability of the claims the examiner presumes that the subject matter of the various claims was commonly owned as of the effective filing date of the claimed invention(s) absent any evidence to the contrary. Applicant is advised of the obligation under 37 CFR 1.56 to point out the inventor and effective filing dates of each claim that was not commonly owned as of the effective filing date of the later invention in order for the examiner to consider the applicability of 35 USC 102(b)(2)(C) for any potential 35 USC 102(a)(2) prior art against the later invention.
Claim(s) 1 and 4-6 is/are rejected under 35 USC 103 as being unpatentable over Yichong ZHANG et al. (Traditional Chinese Medicine Knowledge Graph Construction Based on Large Language Models, published 2024-04-07; hereafter “Zhang”) and Yanbin WEI et al. (KICGPT: Large Language Model with Knowledge in Context for Knowledge Graph Completion, published 2024-02-23; hereafter “Wei”).
Claim 1
Zhang discloses:
An automatic knowledge graph construction method based on prior knowledge and knowledge connection, comprising: Step 1: obtaining prompt data by storing information relevant to a knowledge graph topic as character strings, wherein the information relevant to a knowledge graph is used for constructing a knowledge graph; Step 2: retrieving and saving article paragraphs from an external data source based on the prompt data; ([Zhang, abstract and section 3.1]: Zhang discusses “the use of large language models in constructing a knowledge graph for Traditional Chinese Medicine (TCM)” [Zhang, abstract]. The method “utilize[s] thematic web crawling techniques to obtain TCM-related text data from specified websites and save them in .txt format” [Zhang, section 3.1 paragraph beginning “In the process”]. In other words, TCM maps to the “knowledge graph topic” of the claim, the knowledge graph that is constructed maps to the “knowledge graph” of the claim, and the TCM-related text data maps to the “information relevant to a knowledge graph topic”, the “article paragraphs”, and the “prompt data” of the claim. The fact that the articles are stored as .txt files means that they are stored “as character strings” as required by the claim (cf. 112(b) rejections). The web and/or the specified websites map to the “external data source” of the claim.)
Step 3: respectively injecting prompt templates for [four] large language model agents of annotation, [reasoning,] cognition, and association; ([Zhang, sections 3.2-3]: Zhang discloses using LLMs to perform three tasks: named entity recognition [Zhang, sections 3.2.1 and 3.3.2, figure 3, and table A.1], self-validation [Zhang, sections 3.2.3 and 3.3.3, and figure 4], and entity relationship extraction [Zhang, sections 3.2.4 and 3.3.4, and table A.2]. These three tasks respectively fall under the broadest reasonable interpretation of the “annotation” (named entity recognition being an act of “annotating” named entities), “cognition” (self-validation being a form of both “cognition”), and “association” (establishing relationships between entities being an act of “association”) tasks recited in the claim. The examiner notes that this is merely one mapping out of several which are consistent with the broadest reasonable interpretation of the claim as recited (e.g., any of the three tasks could also be viewed as an act of “reasoning”). Only for the sake of concreteness is it the one adopted herein. Zhang further discloses providing prompts to LLMs to perform these knowledge extraction tasks [Zhang, section 3.2 first paragraph]. These prompts include a “few-shot demonstration” [Zhang, section 3.2 first paragraph] which maps to the “prompt templates” of the claim. In other words, the act of providing the few-shot demonstrations to the LLMs maps to the “injecting” step of the claim.)
Step 4: obtaining prior knowledge by inputting the injected prompt templates, the article paragraphs, and specific task requirements into the [four] large language model agents; ([Zhang, sections 3.2-3]: As noted above, Zhang discloses performing knowledge extraction tasks by providing prompts to LLMs [Zhang, section 3.2]. These prompts contain a “task description”, a “few-shot demonstration” and an “input sentence” [Zhang, section 3.2 first paragraph]. The task description maps to the “specific task requirements” of the claim. The few-shot demonstration maps to the “[injected] prompt templates” of the claim. The input sentences from the TCM-related text data map to the “article paragraphs” of the claim. In other words, the act of providing these prompts into LLMs maps to the step of “inputting the injected prompt templates, article paragraphs, and specific task requirements” as recited by the claim. Any knowledge obtained from the LLMs in this way maps to the “prior knowledge” of the claim.)
Step 5: obtaining effective data related to the knowledge graph topic by inputting the article paragraphs, the prior knowledge, and a pre-defined contrasting prompt text into a knowledge-connecting large language model; ([Zhang, section 3.2.4 and 3.3.4 and table A.2]: As noted above, Zhang discloses using an LLM to perform entity relationship extraction [Zhang, sections 3.2.4 and 3.3.4; see also, table A.2]. The LLM performing entity relationship extraction maps to the “knowledge-connecting large language model” of the claim. The input sentence together with the associated question maps to the “pre-defined contrasting prompt text” of the claim. The examiner notes that this is merely one of several possible mappings that are consistent with the broadest reasonable interpretation of the claim as recited (e.g., the question about the input sentence alone could be mapped to the “pre-defined contrasting prompt text” instead). The entity relationship extraction also involves inputting the “article paragraphs” as described above. It also involves inputting “prior knowledge” as mapped above; for example, to extract relationships between entities, the model needs to be provided entities which are recognized by the named entity recognition task, and those entity names are part of the “prior knowledge” as mapped above.)
wherein Step 5 is as follows: Step 5.1: configuring the article paragraphs as extraction data and the prior knowledge as reference data; Step 5.2: configuring the pre-defined contrastive prompt text as prompt instructions and inputting to the knowledge-connecting large language model; ([Zhang, sections 3.2-3]: As noted above, Zhang discloses an LLM performing entity relationship extraction [Zhang, sections 3.2.4 and 3.3.4] which maps to the “knowledge-connecting large language model” of the claim. The input sentence maps to the “article paragraphs” and to the “extraction data” of the claim. It (or the question about the input sentence) maps to the “pre-defined contrasting prompt text” and the “prompt instructions” of the claim.)
Step 5.3: based on the prompt instructions, extracting extended information of prior knowledge from the article paragraphs by the knowledge-connecting large language model based on the prior knowledge; ([Zhang, sections 3.3.4]: Zhang discloses that the relationships extracted by the entity relationship extraction task “can be formalized as relationship triplets <e1, r, e2>, where e1 and e2 are entities, and r belongs to the target relationship set” [Zhang, section 3.3.4]. The relationship r maps to the “extended information” of the claim. It is “based on the prompt instructions” with the “prompt instructions” being as mapped above, it is “of prior knowledge” since it relates to the entities identified by the named entity recognition task, and it is “from the article paragraphs” because it is extracted from an input sentence.)
and Step 5.4: obtaining the effective data related to the knowledge graph topic by integrating the prior knowledge with the extended information. ([Zhang, section 3.3.4]: As noted above, the relationships identified by the entity relationship extraction task map to the “effective data” of the claim. Since these relationships are formalized as triples <e1, r, e2> [Zhang, section 3.3.4], this can be viewed as “integrating prior knowledge with the extended information” as recited by the claim, since the entities e1 and e2 are part of the “prior knowledge” as mapped above and the relationship r is the “extended information” as mapped above.)
Zhang does not distinctly disclose:
four [large language model agents of annotation,] reasoning, [cognition, and association; … the] four [large language model agents;]
and Step 6: configuring the effective data related to the knowledge graph topic and pre- defined input prompts as an input layer of a large language model automatic agent framework,
obtaining entity-relation-entity triples through multi-round feedback of the large language model automatic agent framework,
and completing construction of the knowledge graph;
Wei is in the field of knowledge graphs. More precisely, it discloses a framework for “Knowledge Graph Completion (KGC)” called “KICGPT” which “integrates a large language model” [Wei, abstract]. In the combination, the knowledge graph produced by Zhang is provided to the KGC method of Wei as input. Then Zhang in view of Wei discloses:
four [large language model agents of annotation,] reasoning, [cognition, and association; … the] four [large language model agents;] ([Wei, abstract and section 3.1]: KICGPT can be mapped to the fourth “large language model” of the claim because it performs a “link prediction” task [Wei, section 3.1] which falls under the broadest reasonable interpretation of “reasoning” as recited by the claim.)
and Step 6: configuring the effective data related to the knowledge graph topic and pre- defined input prompts as an input layer of a large language model automatic agent framework, ([Wei, abstract and figure 1]: The KICGPT framework disclosed by Wei maps to the “large language model automatic agent framework” of the claim. As noted above, the knowledge graph produced by Zhang is provided to KICGPT as input; this means that the “effective data related to the knowledge graph topic” is “configur[ed]… as an input layer” of the framework. The prompts provided to the LLMs in the KICGPT framework [Wei, figure 1] map to the “pre-defined input prompts” of the claim.)
obtaining entity-relation-entity triples through multi-round feedback of the large language model agent framework, ([Wei, section 3.2]: The KICGPT framework iterates over “each query triple (h, r, ?)” and provides, for each such query, an ordered list R_{KICGPT} of entities which complete the query [Wei, section 3.2 first paragraph]. The fact that the method iterates over query triples means that its output maps to the “multi-round feedback of the large language model agent framework” of the claim. For each query (h, r, ?) and each entity t in the ordered list R_{KICGPT} produced by KICGPT, the resulting triple (h, r, t) maps to one of the “entity-relation-entity triples” of the claim.)
and completing construction of a knowledge graph; ([Wei, abstract]: As noted above, the method of Zhang performs “Knowledge Graph Completion (KGC)” [Wei, abstract].)
Before the effective filing date of the invention, it would have been obvious to a person of ordinary skill in the art to combine the method of knowledge graph construction of Zhang with the method of knowledge graph completion of Wei because it “alleviates the long-tail problem without incurring additional training overhead” and because “[e]mpirical results on benchmark datasets demonstrate the effectiveness of KICGPT with smaller training overhead and no finetuning” [Wei, abstract], so the combination would be both effective and efficient overall.
Claim 4
Zhang in view of Wei discloses the elements of the parent claim(s). It also discloses:
[The automatic knowledge graph construction method based on prior knowledge and knowledge connection according to claim 1, wherein Step 3 is as follows:] Step 3.1: injecting a cognitive prompt template according to an expected function of the large language model agent of cognition; ([Zhang, sections 3.2-3]: As noted under the parent claim, the LLM performing the self-validation task maps to the “large language model agent of cognition” of the claim. The “prompt template” provided to this LLM as mapped under the parent claim maps to the “cognitive prompt template” of the claim.)
Step 3.2: injecting an annotation prompt template according to an expected function of the large language model agent of annotation; ([Zhang, sections 3.2-3]: As noted under the parent claim, the LLM performing the named entity extraction task maps to the “large language model agent of annotation” of the claim. The “prompt template” provided to this LLM as mapped under the parent claim maps to the “annotation prompt template” of the claim.)
Step 3.3: injecting a reasoning prompt template according to an expected function of the large language model agent of reasoning; ([Wei, figure 1]: As noted under the parent claim, KICGPT maps to the “large language model agent of reasoning” of the claim. Wei discloses the use of a “Unified Prompt Template” [Wei, figure 1] which maps to the “reasoning prompt template” of the claim.)
and Step 3.4: injecting an associative prompt template according to an expected function of the large language model agent of association. ([Zhang, sections 3.2-3]: As noted under the parent claim, the LLM performing the entity relationship extraction task maps to the “large language model agent of association” of the claim. The “prompt template” provided to this LLM as mapped under the parent claim maps to the “associative prompt template” of the claim.)
The same motivation to combine applies.
Claim 5
Zhang in view of Wei discloses the elements of the parent claim(s). It also discloses:
[The automatic knowledge graph construction method based on prior knowledge and knowledge connection according to claim 1, wherein Step 4 is as follows:] Step 4.1: inputting the prompt template as data correlation requirements to the four large language model agents; ([Zhang, sections 3.2-3]: As noted under the parent claim, the few-shot demonstration maps to the “prompt template[s]” of the claim. It also maps to the “data correlation requirements” of the claim since it describes the requirements regarding how data is to be correlated.)
Step 4.2: inputting the article paragraphs as an extraction corpus to the four large language model agents; ([Zhang, sections 3.2-3]: As noted under the parent claim, the input sentence from the TCM-related text data maps to the “article paragraphs” of the claim. It also maps to the “extraction corpus” of the claim since it is a corpus of data from which data is to be extracted.)
Step 4.3: inputting a pre-defined generated data format as the specific task requirements to the four large language model agents; ([Zhang, sections 3.2-3]: As noted under the parent claim, the task description maps to the “specific task requirements” of the claim. It also maps to the “pre-defined generated data format” of the claim since it describes how the format of the output data to be generated by the models.)
and Step 4.4: generating prior knowledge related to the knowledge graph topic by the four large language model agents from four aspects based on the prompt templates, the article paragraphs, and the pre-defined generate data format from Step 4.1 to Step 4.3. ([Zhang, sections 3.2-3]: As noted under the parent claim, Zhang in view of Wei discloses generating “prior knowledge related to the [knowledge] graph topic” using the four large language model agents.)
The same motivation to combine applies.
Claim 6
Zhang in view of Wei discloses the elements of the parent claim(s). It also discloses:
[The automatic knowledge graph construction method based on prior knowledge and knowledge connection according to claim 1, wherein Step 6 is as follows:] Step 6.1: setting agent types of two large language models in the large language model automatic agent framework as “customer” and “knowledge graph construction expert”; ([Wei, section 3.2]: Wei discloses that the KICGPT framework has two components [Wei, section 3.2]. The triple-based KGC retriever maps to the “customer” of the claim and the LLM to the “knowledge graph construction expert” of the claim.)
Step 6.2: transmitting the effective data related to the knowledge graph topic to the “customer” as context, ([Zhang, section 3.3.4; Wei, abstract]: As noted under the parent claim, in the combination, the knowledge graph produced by Zhang is passed on to the KGC method of Wei. This means, in particular, that the “effective data” as mapped under the parent claim is “transmit[ed]… to the ‘customer’ as context” as recited by the claim.)
and transmitting an input prompt to the “knowledge graph construction expert” as the instruction; ([Wei, figure 1]: Wei discloses the use of a “Unified Prompt Template” [Wei, figure 1] in the LLM. This unified prompt template maps to the “input prompt” and the “instruction” of the claim.)
and Step 6.3: according to the multi-round feedback, obtaining the entity-relation-entity triples comprised in a local knowledge graph obtained by each round of the multi-round feedback, ([Wei, section 3.2]: As explained under the parent claim, Wei discloses “multi-round feedback” (one “round” for each query triple (h, r, ?)) and obtains the “entity-relation-entity triples” (h, r, t) for each t in the ordered list R_{KICGPT} output by the framework. Any such triple falls under the broadest reasonable interpretation of a “local knowledge graph” (since it can be regarded as a knowledge graph with two nodes and one edge between them).)
and finally obtaining a complete knowledge graph according to all local knowledge graphs. ([Wei, abstract]: The completed knowledge graph obtained by the KGC method of Wei maps to the “complete knowledge graph” of the claim. This complete knowledge graph is based on the triples produced by the system, so it is obtained “according to all local knowledge graphs” as recited by the claim, with the “local knowledge graphs” being as mapped above.)
The same motivation to combine applies.
Claim(s) 2 is/are rejected under 35 USC 103 as being unpatentable over Zhang in view of Wei, further in view of Sekar KRISHNAN (US20230096474A1, published 2023-03-30; hereafter “Krishnan”).
Claim 2
Zhang in view of Wei discloses the elements of the parent claim(s). It also discloses:
[The automatic knowledge graph construction method based on prior knowledge and knowledge connection according to claim 1, wherein Step 1 is as follows:] Step 1.1: converting all original corpus data in different formats, comprising text, [voice, and PDF images] from actual scenarios, into a unified character text and performing preliminary integration to obtain original character data; ([Zhang, section 3.1]: As noted above, Zhang discloses “web crawling techniques to obtain TCM-related text data from specified websites and save them in .txt format” [Zhang, section 3.1 paragraph beginning “In the process”]. The data on the websites is the ”original corpus data” and the data is of “different formats” (namely, semi-structured and unstructured [Zhang, section 3.1 paragraph beginning “In the process”]). The data being saved in .txt format means that the data is converted “into a unified character text” and the data thus obtained is the “original character data” of the claim.)
and Step 1.2: performing data cleaning on the original character data to eliminate [blank data] and redundant data, and obtaining the prompt data. ([Zhang, section 3.1]: Zhang discloses conducting a “meticulous data cleaning process” by identifying and removing “possible HTML tags, URL links, special characters, garbled or inconsistently encoded data, and recurring data” [Zhang, section 3.1 last paragraph]. Removing recurring data maps to “eliminat[ing]… redundant data” as recited by the claim. The cleaned data is the “prompt data” of the claim.)
The same motivation to combine applies.
Zhang in view of Wei does not distinctly disclose:
[data in different formats, comprising text,] voice, and PDF images
[eliminate] blank data
Krishnan is in the field of data analysis. Moreover, Zhang in view of Wei and Krishnan discloses:
[data in different formats, comprising text,] voice, and PDF images ([Krishnan, 0029]: Krishnan discloses “converting the electronic file to a text file if the electronic file is not already in a text format” [Krishnan, 0029]. It specifically mentions that “if the data category is audio, the text converter 111 may convert the audio file into text by sending the audio file into an automated speech recognition (ASR) engine” as well as “if the unstructured data has a category of PDF, the text converter may convert the PDF file into text by sending the PDF file to a PDF to text converter engine” [Krishnan, 0029]. The audio data files of Krishnan map to the “voice” data of the claim, and the PDF data files of Krishnan map to the “PDF images” of the claim.)
[eliminate] blank data ([Krishnan, 0019]: Krishnan discloses a text cleaner 112 which “may process the text file from the text converter 111… to remove noise” where noise includes “unwanted information presented in the extracted text, such as… white spaces” [Krishnan, 0019]. Removing white spaces maps to “eliminat[ing] blank data” as recited by the claim.)
Before the effective filing date of the invention, it would have been obvious to a person of ordinary skill in the art to combine the knowledge graph construction method of Zhang in view of Wei with the data conversion and cleaning processes described in Krishnan because conversion would ensure unity of input data format, and cleaning would increase the “accuracy of the machine learning engines” [Krishnan, 0019], thereby resulting in a more robust system overall.
Claim(s) 3 is/are rejected under 35 USC 103 as being unpatentable over Zhang in view of Wei, further in view of David BODOFF et al. (A Unified Maximum Likelihood Approach to Document Retrieval, published 2001; hereafter “Bodoff”).
Claim 3
Zhang in view of Wei discloses the elements of the parent claim(s). It also discloses:
[The automatic knowledge graph construction method based on prior knowledge and knowledge connection according to claim 1, wherein Step 2 is as follows:] Step 2.1: configuring the prompt data as an input to a retrieval framework; ([Zhang, sections 3.2-3]: As noted under the parent claim, Zhang discloses providing the data to LLMs for knowledge extraction. The knowledge extraction tasks as described in Zhang map to the “retrieval framework” of the claim.)
Step 2.2: retrieving a paragraph text with a highest similarity to the prompt data from the external data source [according to a maximum likelihood estimation search algorithm;] and Step 2.3: saving the retrieved paragraph text in a character string format. ([Zhang, section 3.1]: As noted above, Zhang discloses “web crawling techniques to obtain TCM-related text data from specified websites and save them in .txt format” [Zhang, section 3.1 paragraph beginning “In the process”]. As noted under the parent claim, the web and/or the specified websites map to the “external data source” of the claim. One of the data files thus obtained is a “paragraph text with a highest similarity to the prompt data” with the “prompt data” as mapped above, so Zhang discloses “retrieving a paragraph text with a highest similarity” as recited by the claim. Zhang does not distinctly disclose the use of a maximum likelihood search algorithm for identifying this paragraph, but this is disclosed in the combination as proposed below. The fact that the data is saved in .txt format means that the “retrieved paragraph” is saved “in a character string format” as recited by the claim)
The same motivation to combine applies.
While Zhang discloses similarity computations [Zhang, section 3.2.2], Zhang in view of Wei might not distinctly disclose:
[retrieving a paragraph text with a highest similarity to the prompt data from the external data source] according to a maximum likelihood estimation search algorithm;
Bodoff is in the field of data analysis. Moreover, Zhang in view of Wei and Bodoff discloses:
[retrieving a paragraph text with a highest similarity to the prompt data from the external data source] according to a maximum likelihood estimation search algorithm; ([Bodoff, abstract]: Bodoff discloses a method of “information retrieval (IR)… using a maximum likelihood framework” [Bodoff, abstract]. The IR method of Bodoff maps to the “maximum likelihood estimation search algorithm” of the claim.)
Before the effective filing date of the invention, it would have been obvious to a person of ordinary skill in the art to combine the knowledge graph construction method of Zhang in view of Wei with the information retrieval method of Bodoff because it uses “all the available data… to simultaneously estimate both documents and queries in proportions that are optimal in a maximum likelihood sense” and “[t]he resulting algorithm is directly applicable to many approaches to IR” [Bodoff, abstract], thereby resulting in a more effective system.
Conclusion
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 Shishir AGRAWAL whose telephone number is +1 703-756-1183. The examiner can normally be reached Monday through Thursday, 08:30-14:30 Pacific Time.
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, Alexey SHMATOV can be reached on +1 571-270-3428. The fax phone number for the organization where this application or proceeding is assigned is +1 571-273-8300.
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 +1 866-217-9197 (toll-free). If you would like assistance from a USPTO Customer Service Representative, call +1 800-786-9199 (IN USA OR CANADA) or +1 571-272-1000.
/S.A./Examiner, Art Unit 2123
/ALEXEY SHMATOV/Supervisory Patent Examiner, Art Unit 2123
Prosecution Insights