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 .
Introduction
This office action is in response to communications filed 12/22/2024. Claims 1-20 are pending and likewise have been examined.
Information Disclosure Statement
The information disclosure statement (IDS) submitted on 12/26/2024 is in compliance with the provisions of 37 CFR 1.97. Accordingly, the information disclosure statement is being considered by the examiner.
Claim Rejections - 35 USC § 102
The following is a quotation of the appropriate paragraphs of 35 U.S.C. 102 that form the basis for the rejections under this section made in this Office action:
A person shall be entitled to a patent unless –
(a)(2) the claimed invention was described in a patent issued under section 151, or in an application for patent published or deemed published under section 122(b), in which the patent or application, as the case may be, names another inventor and was effectively filed before the effective filing date of the claimed invention.
Claim(s) 1, 8-9 and 16 is/are rejected under 35 U.S.C. 102(a)(2) as being anticipated by Sheward et al. (US 20240394563 A1).
Regarding Claim 1:
Sheward teaches a method comprising: inputting a first prompt to a large language model to cause the large language model to output a list of keywords based on a context window(Para [0007], Ln 1-10, receiving one or more documents from one or more information sources; receiving an indication of one or more exemplary entities associated with the topic from a user; automatically, using one or more artificial intelligence models: extracting a plurality of textual entities from the one or more documents);
inputting a second prompt to the large language model to cause the large language model to output an adjacency matrix for keywords in the list of keywords that indicates which of the keywords in the list of keywords are related in the context window(Para [0007], Ln 1-25, automatically, using one or more artificial intelligence models: extracting a plurality of textual entities from the one or more documents, determining a quality level of each of the plurality of extracted textual entities based on the one or more exemplary entities received from the user, wherein a quality level of a textual entity indicates a degree of similarity between the textual entity and each exemplary entity of the one or more exemplary entities, identifying a plurality of high-quality textual entities from the plurality of textual entities based on the respective quality levels of the plurality of extracted textual entities. See Fig 1B. Para [0054], Ln 1-11, One or more artificial intelligence models may be leveraged in the determination of the quality levels.);
determining a graph including nodes corresponding to respective keywords in the list of keywords and edges corresponding to relationships between keywords indicated by the adjacency matrix(Para [0007], Ln 1-25, categorizing each the plurality of high-quality entities according to one or more sub-topics associated with the topic, and determining connection information indicating relationships between the one or more sub-topics based on documents of the one or more documents from which each of the plurality of high-quality entities originated; and generating, based on the one or more sub-topics and the connection information, a first knowledge graph for the topic that represents the one or more sub-topics in the one or more documents and the relationships between said sub-topics);
and storing, displaying, or transmitting the graph(Para [0021], Ln 1-3, In some embodiments of the method, providing a graphical representation of the first knowledge graph to the user using a graphical user interface).
Regarding Claim 8:
Sheward teaches the method of claim 1, in which the first prompt causes the list of keywords to be ordered based on relevance to a topic occurring in one or more documents included in the context window(Para [0049], Ln 1-10, Optionally, at step 304, the user may explicitly indicate the topic of interest with which the exemplary entities are associated. Additionally, the user can indicate a level of importance of each exemplary entity, for example by ranking each exemplary entity with a numerical score that quantifies the relevance or closeness of said exemplary entity to the topic of interest. Para [0036], Ln 1-5, Exemplary entities 104 may be words, phrases, or concepts associated with a topic (or topics) of interest to user).
Regarding Claim 9:
Sheward teaches a system comprising: a memory, and a processor, wherein the memory includes instructions executable by the processor to cause the system to(Para [0046], Ln 1-8, instructions, processors, memory):
input a first prompt to a large language model to cause the large language model to output a list of keywords based on a context window(Para [0007], Ln 1-10, receiving one or more documents from one or more information sources; receiving an indication of one or more exemplary entities associated with the topic from a user; automatically, using one or more artificial intelligence models: extracting a plurality of textual entities from the one or more documents);
input a second prompt to the large language model to cause the large language model to output an adjacency matrix for keywords in the list of keywords that indicates which of the keywords in the list of keywords are related in the context window(Para [0007], Ln 1-25, automatically, using one or more artificial intelligence models: extracting a plurality of textual entities from the one or more documents, determining a quality level of each of the plurality of extracted textual entities based on the one or more exemplary entities received from the user, wherein a quality level of a textual entity indicates a degree of similarity between the textual entity and each exemplary entity of the one or more exemplary entities, identifying a plurality of high-quality textual entities from the plurality of textual entities based on the respective quality levels of the plurality of extracted textual entities. See Fig 1B. Para [0054], Ln 1-11, One or more artificial intelligence models may be leveraged in the determination of the quality levels);
and determine a graph including nodes corresponding to respective keywords in the list of keywords and edges corresponding to relationships between keywords indicated by the adjacency matrix(Para [0007], Ln 1-25, categorizing each the plurality of high-quality entities according to one or more sub-topics associated with the topic, and determining connection information indicating relationships between the one or more sub-topics based on documents of the one or more documents from which each of the plurality of high-quality entities originated; and generating, based on the one or more sub-topics and the connection information, a first knowledge graph for the topic that represents the one or more sub-topics in the one or more documents and the relationships between said sub-topics).
Regarding Claim 16:
Claim 16 contains similar limitations as Claim 8 and is therefore rejected for the same reasons.
Claim(s) 17 is/are rejected under 35 U.S.C. 102(a)(2) as being anticipated by Hoang et al. (US 12217007 B2)(first cited Hoang reference of two, hereinafter Hoang).
Regarding Claim 17:
Hoang teaches a method comprising: accessing a graph including nodes corresponding to respective keywords in a list of keywords and edges corresponding to relationships between keywords in a context window(Col 3, Ln 35-45, The second model (e.g., a semantic decoder) is a neural network that takes as input a list of triples (e.g., subject, predicate, object) where the subject and object are entities and the predicate is a relation. The semantic decoder outputs a fragment of text describing the knowledge representing the triples. Abstract, Ln 1-5, Embodiments are provided for unsupervised learning of domain specific knowledge graph from textual data and language generation from knowledge graph);
selecting a set of distinct tuples of keywords corresponding to different paths through the graph, wherein each tuple of keywords includes the respective keywords corresponding to nodes of a path in the graph that passes through up to M nodes, where M is an integer(Col 3, Ln 35-45, The second model (e.g., a semantic decoder) is a neural network that takes as input a list of triples (e.g., subject, predicate, object) where the subject and object are entities and the predicate is a relation. The semantic decoder outputs a fragment of text describing the knowledge representing the triples);
inputting prompts, which each include the keywords in a tuple of keywords from the set of distinct tuples of keywords, to a large language model to cause the large language model to output texts(Col 3, Ln 35-45, The second model (e.g., a semantic decoder) is a neural network that takes as input a list of triples (e.g., subject, predicate, object) where the subject and object are entities and the predicate is a relation. The semantic decoder outputs a fragment of text describing the knowledge representing the triples);
and storing, displaying, or transmitting the texts(Col 1, Ln 30-35, An embodiment includes a computer system. The computer system includes a processor, a computer-readable memory).
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.
Claim(s) 2 and 10 is/are rejected under 35 U.S.C. 103 as being unpatentable over Sheward as applied to claim 1 above, and further in view of Dognin et al. (US 20240070404 A1)(first cited Dognin of two).
Regarding Claim 2:
Sheward teaches the method of claim 1, but does not teach comprising: selecting a path in the graph that passes through up to M nodes, where M is an integer; inputting a third prompt, which includes the respective keywords for nodes in the selected path in the graph, to the large language model to cause the large language model to output a text; and storing, displaying, or transmitting the text.
In the same field of knowledge graph construction, Dognin teaches comprising: selecting a path in the graph that passes through up to M nodes, where M is an integer; inputting a third prompt, which includes the respective keywords for nodes in the selected path in the graph, to the large language model to cause the large language model to output a text(Para [0034], Ln 1-31, The input for T2G is shown at 1005. The model generates a graph from the input text by simultaneously extracting relevant nodes and linking them coherently. Examples of these nodes are given as 1009, 1011, 1013, 1015, 1017, and 1019. Some of these nodes are clearly extracted from the input text (1011, 1013, 1015, 1017, 1019), whereas a node like 1009 is generated by the model (since this is a generative model). This graph is created by capturing the facts from the input sentences. Then, this graph was used as input to generate text. For the G2T task, another model starts from the generated graph and generates semantically relevant text from it. Para [0036], Ln 1-7, Consider a directed graph G with vertices…and edges…. vertices are given by……{a, b, c, d}.);
and storing, displaying, or transmitting the text(Para [0086], Ln 1-5, In still another aspect, an exemplary apparatus includes a memory; and at least one processor, coupled to the memory, and operative to perform any one, some, or all of the method steps just described).
It would have been obvious for one skilled in the art, at the effective time of filling, to modify Sheward, with the knowledge graph system of Dognin, as it can help improve performance of both graph-to-text and text-to-graph tasks(Para [0013], Ln 1-3, Para [0021], Ln 1-2).
Regarding Claim 10:
Claim 10 contains similar limitations as Claim 2 and is therefore rejected for the same reasons.
Claim(s) 4 and 12 is/are rejected under 35 U.S.C. 103 as being unpatentable over Sheward as applied to claim 1 above, and further in view of Hoang.
Regarding Claim 4:
Sheward teaches the method of claim 1, but does not teach comprising: selecting a set of distinct tuples of keywords corresponding to different paths through the graph, wherein each tuple of keywords includes the respective keywords corresponding to nodes of a path in the graph that passes through up to M nodes, where M is an integer; inputting prompts, which each include the keywords in a tuple of keywords from the set of distinct tuples of keywords, to the large language model to cause the large language model to output texts; and storing, displaying, or transmitting the texts.
In the same field of Knowledge Graph Construction, Hoang teaches comprising: selecting a set of distinct tuples of keywords corresponding to different paths through the graph, wherein each tuple of keywords includes the respective keywords corresponding to nodes of a path in the graph that passes through up to M nodes, where M is an integer(Col 3, Ln 35-45, The second model (e.g., a semantic decoder) is a neural network that takes as input a list of triples (e.g., subject, predicate, object) where the subject and object are entities and the predicate is a relation. The semantic decoder outputs a fragment of text describing the knowledge representing the triples);
inputting prompts, which each include the keywords in a tuple of keywords from the set of distinct tuples of keywords, to the large language model to cause the large language model to output texts(Col 3, Ln 35-45, The second model (e.g., a semantic decoder) is a neural network that takes as input a list of triples (e.g., subject, predicate, object) where the subject and object are entities and the predicate is a relation. The semantic decoder outputs a fragment of text describing the knowledge representing the triples);
and storing, displaying, or transmitting the texts(Col 1, Ln 30-35, An embodiment includes a computer system. The computer system includes a processor, a computer-readable memory).
It would have been obvious for one skilled in the art, at the effective time of filling to modify Sheward, with the Knowledge graph system of Hoang, as it can help improve accuracy of the model(Col 1, Ln 40-49, Col 1, Ln 63-67).
Regarding Claim 12:
Claim 12 contains similar limitations as Claim 4 and is therefore rejected for the same reasons.
Claim(s) 7 and 15 is/are rejected under 35 U.S.C. 103 as being unpatentable over the combination of Sheward and Hoang as applied to claim 4 above, and further in view of Liu et al. (US 20220180065 A1).
Regarding Claim 7:
The combination of Sheward and Hoang teaches the method of claim 4, but does not teach in which the context window includes one or more documents that are materials for an educational course, and the texts are test questions on a topic covered in the educational course.
In the same field of knowledge graphs, Liu teaches in which the context window includes one or more documents that are materials for an educational course, and the texts are test questions on a topic covered in the educational course(Para [0106], Ln 1-12, In certain embodiments, the function is to provide answers to customer questions, and the knowledge graph is constructed using customer question and answer documents. The knowledge graph includes question entities and answer entities, the edges represent a suitable answer to a question. When a customer purchases a product in an e-commerce platform, the function module 160 is configured to query the knowledge graph using the question, obtain answers tailored to the question, and provide the answer to the customer. In certain embodiments, the knowledge graph include sub-graphs corresponding to different type of products, such that the answers provided to the customer can be more accurate).
It would have been obvious for one skilled in the art, at the effective time of filling, to modify the combination of Sheward and Hoang with the subgraph structure of Liu as it can improve question answering accuracy(Para [0106], Ln 1-12).
Regarding Claim 15:
Claim 15 contains similar limitations as Claim 7 and is therefore rejected for the same reasons.
Claim(s) 20 is/are rejected under 35 U.S.C. 103 as being unpatentable over Hoang as applied to claim 17 above, and further in view of Liu.
Regarding Claim 20:
Hoang teaches the method of claim 17, but does not teach in which the context window includes one or more documents that are materials for an educational course, and the texts are test questions on a topic covered in the educational course.
In the same field of knowledge graphs, Liu teaches in which the context window includes one or more documents that are materials for an educational course, and the texts are test questions on a topic covered in the educational course(Para [0106], Ln 1-12, In certain embodiments, the function is to provide answers to customer questions, and the knowledge graph is constructed using customer question and answer documents. The knowledge graph includes question entities and answer entities, the edges represent a suitable answer to a question. When a customer purchases a product in an e-commerce platform, the function module 160 is configured to query the knowledge graph using the question, obtain answers tailored to the question, and provide the answer to the customer. In certain embodiments, the knowledge graph include sub-graphs corresponding to different type of products, such that the answers provided to the customer can be more accurate).
It would have been obvious for one skilled in the art, at the effective time of filling, to modify Hoang with the subgraph structure of Liu as it can improve question answering accuracy(Para [0106], Ln 1-12).
Allowable Subject Matter
Claims 3, 5-6, 11, 13-14 and 18-19 objected to as being dependent upon a rejected base claim, but would be allowable if rewritten in independent form including all of the limitations of the base claim and any intervening claims.
The following is a statement of reasons for the indication of allowable subject matter:
Regarding Claim 3:
The combination of Sheward and Dognin teaches the method of claim 2, but does not teach in which selecting the path in the graph comprises: starting from a node corresponding to a respective keyword with a highest relevance to a topic from amongst the keywords in the list of keywords; and traversing edges of the graph to adjacent nodes until reaching an Mth distinct node or reaching a node with no edges that connect to a node that has not been reached by the path.
Claim 11 contains similar limitations as Claim 3 and therefore also contains allowable subject matter.
Regarding Claim 5:
The combination of Sheward and Hoang teaches the method of claim 4, but does not teach in which selecting the set of distinct tuples of keywords corresponding to different paths through the graph comprises: determining embedding vectors for tuples of keywords corresponding to paths in the graph, wherein an embedding vector for one of the tuples of keywords is determined as a linear combination of embedding vectors for keywords in the one of the tuples of keywords in an embedding vector space of the large language model;
In the same field of Knowledge Graph Construction, Hoang teaches in which selecting the set of distinct tuples of keywords corresponding to different paths through the graph comprises: determining embedding vectors for tuples of keywords corresponding to paths in the graph, wherein an embedding vector for one of the tuples of keywords is determined as a linear combination of embedding vectors for keywords in the one of the tuples of keywords in an embedding vector space of the large language model(See Fig 6B. Col 15, Ln 32-43, entity relation matching network component 628 may turn/convert embeddings of the input texts, candidate relations and candidate entities (e.g., the relation embeddings 622, sentence embeddings 624, and entity embeddings 626) into distributions over relations/triple);
It would have been obvious for one skilled in the art, at the effective time of filling to modify the combination of Sheward and Hoang, with the Knowledge graph system of Hoang, as it can help improve accuracy of the model(Col 1, Ln 40-49, Col 1, Ln 63-67).
The combination of Sheward and Hoang and applying a clustering algorithm to the embedding vectors for the tuples of keywords to select tuples of keywords from among the tuples of keywords corresponding to paths in the graph.
Hoang Thanh et al. (US 20230297855 A1)(second cited Hoang reference, hereinafter Hoang Thanh) teaches knowledge graph relation embedding clustering, but not in relation to selection for conversion to text as in claim 4.
The prior art of record alone or in combination does not teach the above limitations.
Claim 13 contains similar limitations as Claim 5 and therefore also contains allowable subject matter.
Regarding Claim 6:
The combination of Sheward and Hoang teaches the method of claim 4, but does not teach in which selecting the set of distinct tuples of keywords corresponding to different paths through the graph comprises: removing tuples of keywords that are from the set of distinct tuples of keywords that are subsets of other tuples of keywords in the set of distinct tuples of keywords.
The prior art of record alone or in combination does not teach the above limitations.
Claim 14 contains similar limitations as Claim 6 and therefore also contains allowable subject matter.
Regarding Claim 18:
Hoang teaches the method of claim 17, in which selecting the set of distinct tuples of keywords corresponding to different paths through the graph comprises: determining embedding vectors for tuples of keywords corresponding to paths in the graph, wherein an embedding vector for one of the tuples of keywords is determined as a linear combination of embedding vectors for keywords in the one of the tuples of keywords in an embedding vector space of the large language model(See Fig 6B. Col 15, Ln 32-43, entity relation matching network component 628 may turn/convert embeddings of the input texts, candidate relations and candidate entities (e.g., the relation embeddings 622, sentence embeddings 624, and entity embeddings 626) into distributions over relations/triple);
Hoang does not teach and applying a clustering algorithm to the embedding vectors for the tuples of keywords to select tuples of keywords from among the tuples of keywords corresponding to paths in the graph.
Hoang Thanh(US 20230297855 A1) teaches knowledge graph relation embedding clustering, but not in relation to selection for conversion to text as in claim 17.
The prior art of record alone or in combination does not teach the above limitations.
Regarding Claim 19:
Hoang teaches the method of claim 17, but does not teach in which selecting the set of distinct tuples of keywords corresponding to different paths through the graph comprises: removing tuples of keywords that are from the set of distinct tuples of keywords that are subsets of other tuples of keywords in the set of distinct tuples of keywords.
The prior art of record alone or in combination does not teach the above limitations.
Conclusion
The prior art made of record and not relied upon is considered pertinent to applicant's disclosure.
Kuo et al. (US 20220147835 A1)
Recognition of entity triples in text and construction of knowledge graph, using machine learning.
Song (US 20210374333 A1)
Graph to text generation using machine learning models and graph triple relations.
Dognin(US 11829726 B2)
Bidirectional text to graph and graph to text generation machine learning model.
Chen et al. “AutoKG: Efficient Automated Knowledge Graph Generation for Language Models”
Automated Knowledge graph generation using unsupervised learning.
Guo et al. “CycleGT: Unsupervised Graph-to-Text and Text-to-Graph Generation via Cycle Training”
Graph-to-Text and Text-to-Graph machine learning model trained in unsupervised manner.
Any inquiry concerning this communication or earlier communications from the examiner should be directed to ALEXANDER G MARLOW whose telephone number is (571)272-4536. The examiner can normally be reached Monday - Thursday 10:00 am - 8:00 pm EST.
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, Richmond Dorvil can be reached at (571)272-7602. 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.
/ALEXANDER G MARLOW/ Assistant Examiner, Art Unit 2658
/RICHEMOND DORVIL/ Supervisory Patent Examiner, Art Unit 2658
Prosecution Insights