CTNF 18/454,155 CTNF 101938 DETAILED ACTION Notice of Pre-AIA or AIA Status 07-03-aia AIA 15-10-aia The present application, filed on or after March 16, 2013, is being examined under the first inventor to file provisions of the AIA. Information Disclosure Statement 06-52 The information disclosure statement (IDS) submitted on 2024/05/15 & 2023/08/23. The submission 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 § 101 07-04-01 AIA 07-04 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. Claim 1-7 rejected under 35 U.S.C. 101 because the claimed invention is directed to a judicial exception (an abstract idea) without reciting significantly more. Regarding independent claims 1, 6, and 7 Step 1 -- whether the claim falls within any statutory category. See MPEP 2106.03 Independent claim 1 is drawn to a non-transitory computer-readable recording medium claim. Independent claim 6 is drawn to an information processing apparatus comprising a memory and a processor coupled to the memory. Independent claim 7 is drawn to a prediction method implemented by a computer. Therefore, claims 1, 6, and 7 each fall under at least one of the four categories of statutory subject matter: claim 1 as a manufacture, claim 6 as a machine, and claim 7 as a process. Step 2A Prong 1 – whether the claim recites a judicial exception. See MPEP 2106.04, subsection II. Regarding independent claim 1 , the claim is directed to a non-transitory computer-readable recording medium storing a prediction program that uses knowledge graph embedding, for causing a computer to execute prediction processing. The limitation of “determining whether or not graph data to be predicted is data that includes a node link that indicates a relationship between nodes not included in training data used for training of the knowledge graph embedding” recites an abstract idea because it describes evaluating graph data and determining whether the graph data includes a relationship/link not included in training data. This limitation falls within the mental processes grouping of abstract ideas, i.e., concepts performed in the human mind including observation, evaluation, judgment, and opinion. See MPEP § 2106.04(a)(2), subsection III . The limitation of “specifying… graph data similar to the graph data to be predicted from the training data based on a result of embedding prediction for a label of a node included in the graph data to be predicted” recites an abstract idea because it describes identifying or selecting similar graph data based on embedding prediction. Under the broadest reasonable interpretation in light of the specification, the embedding prediction/similarity operation involves vector-based prediction or similarity processing and therefore falls within the mathematical concepts grouping of abstract ideas, including mathematical relationships and mathematical calculations. See MPEP § 2106.04(a)(2), subsection I . The limitation of “determining a prediction result for the graph data to be predicted based on the specified similar graph data” recites an abstract idea because it describes determining a result from selected similar graph data, which is an evaluation or judgment based on information. This limitation falls within the mental processes grouping of abstract ideas. See MPEP § 2106.04(a)(2), subsection III . Accordingly, independent claim 1 recites an abstract idea under Step 2A Prong 1 because the claim includes limitations falling within the mathematical concepts and mental processes groupings of abstract ideas. Therefore, the analysis should proceed to Step 2A Prong 2 . Independent claim 6 is directed to an information processing apparatus comprising a memory and a processor configured to perform prediction processing that uses knowledge graph embedding, and recites limitations that are the same as those of independent claim 1. As discussed above with respect to independent claim 1, the limitations of determining whether graph data includes a node link not included in training data, specifying similar graph data based on embedding prediction, and determining a prediction result recite abstract ideas, including mental processes (e.g., evaluation, judgment, and decision-making) and mathematical concepts (e.g., embedding-based similarity or prediction operations). See MPEP § 2106.04(a)(2), subsections I and III . Therefore, independent claim 6 is directed to an abstract idea for the same reasons as independent claim 1. Independent claim 7 is directed to a prediction method implemented by a computer and recites limitations that are the same as those of independent claim 1. As discussed above with respect to independent claim 1, the limitations of determining whether graph data includes a node link not included in training data, specifying similar graph data based on embedding prediction, and determining a prediction result recite abstract ideas, including mental processes and mathematical concepts . See MPEP § 2106.04(a)(2), subsections I and III . Therefore, independent claim 7 is directed to an abstract idea for the same reasons as independent claim 1. Step 2A Prong 2 -- whether the claim as a whole integrates the recited judicial exception into a practical application of the exception or whether the claim is “directed to” the judicial exception. This evaluation is performed by (1) identifying whether there are any additional elements recited in the claim beyond the judicial exception, and (2) evaluating those additional elements individually and in combination to determine whether the claim as a whole integrates the exception into a practical application. See MPEP 2106.04(d). Regarding independent claim 1 , this claim recites additional elements of “a non-transitory computer-readable recording medium,” “storing a prediction program,” “for causing a computer to execute processing,” and “uses knowledge graph embedding.” These limitations amount to no more than generally linking the use of the judicial exception to a particular technological environment or field of use, namely implementation on a computer using knowledge graph embedding. See MPEP § 2106.05(h). The recited computer-readable recording medium and computer are generic computer components that merely serve as tools for performing the abstract idea. The claim does not recite a particular improvement to the functioning of the computer or to knowledge graph embedding technology itself. See MPEP § 2106.05(a). The claim also does not recite a particular machine that is integral to the claimed process, a transformation of a particular article, or any other meaningful limitation that applies the exception in a manner beyond merely implementing the abstract idea on a computer. See MPEP §§ 2106.05(b), 2106.05(c), and 2106.05(e). Accordingly, the additional elements, individually and in combination, do not integrate the recited judicial exception into a practical application. Therefore, independent claim 1 is directed to the abstract idea under Step 2A Prong 2. Regarding independent claim 6 , this claim recites additional elements of: “a memory” , and “a processor coupled to the memory, the processor being configured to perform prediction processing that uses knowledge graph embedding.” These limitations amount to no more than generally linking the use of the judicial exception to a particular technological environment, namely implementation on a computer system. The recited memory and processor are generic computer components that merely serve as tools for performing the abstract idea. The limitation “processor configured to perform prediction processing using knowledge graph embedding” constitutes no more than instructions to apply the abstract idea using a generic computer, without reciting any improvement to the functioning of the computer or to knowledge graph embedding technology itself. See MPEP § 2106.05(f) . Further, the recitation of knowledge graph embedding merely limits the abstract idea to a particular technological field, which is insufficient to integrate the exception into a practical application. See MPEP § 2106.05(h) . The claim does not recite: an improvement to computer functionality ( MPEP § 2106.05(a) ), a particular machine integral to the claim ( MPEP § 2106.05(b) ), a transformation of an article ( MPEP § 2106.05(c) ), or any other meaningful limitation beyond generally linking the abstract idea to a computer ( MPEP § 2106.05(e) ). Accordingly, the additional elements, individually and in combination, do not integrate the judicial exception into a practical application. Therefore, independent claim 6 is directed to the abstract idea under Step 2A Prong 2. Regarding independent claim 7 , this claim recites the additional element of: “a prediction method implemented by a computer.” This limitation amounts to no more than generally linking the use of the judicial exception to a particular technological environment, namely implementation on a computer. The recited computer is a generic computer performing generic functions and merely acts as a tool to execute the abstract idea. See MPEP § 2106.05(f) . Further, limiting the abstract idea to execution on a computer constitutes a field-of-use limitation and does not integrate the exception into a practical application. See MPEP § 2106.05(h) . The claim does not recite any improvement to the functioning of the computer or to another technology ( MPEP § 2106.05(a) ), nor does it include any additional elements that impose a meaningful limit on the abstract idea beyond generally linking it to a computer environment ( MPEP § 2106.05(e) ). Accordingly, the additional elements, individually and in combination, do not integrate the judicial exception into a practical application. Therefore, independent claim 7 is directed to the abstract idea under Step 2A Prong 2. Step 2B -- whether the claim amounts to significantly more than the judicial exception. See MPEP § 2106.05. Regarding independent claim 1 , the claim recites additional elements of: a non-transitory computer-readable recording medium, a prediction program, and a computer executing processing using knowledge graph embedding. These additional elements, individually and in combination, do not amount to significantly more than the judicial exception. The recited computer-readable medium and program execution constitute generic computer implementation , which is well-understood, routine, and conventional in the field. See MPEP § 2106.05(d) . Further, the use of knowledge graph embedding merely represents the application of a known mathematical technique or algorithm to the abstract idea, without improving the functioning of the computer or the underlying technology. This amounts to no more than instructions to apply the abstract idea using a computer. See MPEP § 2106.05(f) . Additionally, these elements merely limit the abstract idea to a particular technological environment, which does not provide an inventive concept. See MPEP § 2106.05(h) . Accordingly, the claim does not include any additional elements that amount to significantly more than the judicial exception. Regarding independent claim 6 , the claim recites additional elements of: a memory, and a processor coupled to the memory. These elements are generic computer components performing generic computer functions and are well-understood, routine, and conventional in the field. See MPEP § 2106.05(d) . The limitation that the processor is configured to perform prediction processing using knowledge graph embedding amounts to no more than instructions to apply the abstract idea using a computer. See MPEP § 2106.05(f) . Accordingly, the claim does not include any additional elements that amount to significantly more than the judicial exception. Regarding independent claim 7 , the claim recites the additional element of: a prediction method implemented by a computer. This limitation merely recites the use of a generic computer to perform the abstract idea and is well-understood, routine, and conventional. See MPEP § 2106.05(d) . The claim does not include any additional elements that impose a meaningful limitation on the abstract idea or provide an inventive concept. Accordingly, the claim does not amount to significantly more than the judicial exception. Regarding dependent claims 2-5 Regarding dependent claims 2–5 , these claims merely narrow the previously identified abstract idea limitations recited in independent claim 1. For the reasons described above with respect to independent claim 1, the judicial exceptions recited in these dependent claims are not meaningfully integrated into a practical application, nor do they amount to significantly more than the abstract ideas. The additional limitations introduced in claims 2–5, including calculating similarity using embedding vectors, defining similarity as a distance, selecting graph data based on minimum distance or total distance, and determining a prediction result based on a label of a specific node, merely further define the abstract idea using mathematical relationships and evaluative processes. These limitations constitute mathematical concepts and mental processes , including organizing, comparing, and selecting data based on calculated similarity metrics, which are practically capable of being performed in the human mind or with the assistance of pen and paper. Accordingly, dependent claims 2–5 also recite abstract ideas that do not integrate into a practical application and do not amount to significantly more than the judicial exception. Therefore, claims 2–5 are rejected under 35 U.S.C. § 101 . Step 1 -- whether the claim falls within any statutory category. See MPEP § 2106.03. Dependent claims 2–5 are drawn to a non-transitory computer-readable recording medium according to claim 1. Therefore, each of claims 2–5 falls under at least one of the four categories of statutory subject matter, namely a manufacture . Step 2A Prong 1 – whether the claim recites a judicial exception. See MPEP 2106.04, subsection II. Regarding claim 2 , this claim recites the limitations of “calculating similarity between the label of the node and each piece of graph data” by performing link prediction using an embedding vector, and “specifying ... graph data most similar from among the plurality of pieces of graph data.” These limitations are directed toward the abstract idea of mathematical concepts , because calculating similarity using embedding vectors and link prediction involves mathematical relationships and calculations. These limitations also involve organizing, comparing, and selecting information based on calculated similarity, which may also be characterized as a mental process involving evaluation and judgment. See MPEP § 2106.04(a)(2), subsection I and subsection III . Regarding claim 3 , this claim recites the limitations that “the similarity between the label of the node and the graph data is a distance,” and that graph data having “a smallest distance” is specified using distances between labels and graph data. These limitations are directed toward the abstract idea of mathematical concepts , because determining distance and selecting the smallest distance are mathematical calculations or mathematical relationships. These limitations also involve comparing and selecting information, which may be practically performed as an evaluative process. Regarding claim 4 , this claim recites the limitation of specifying graph data having “a smallest total value of the distances” by using “a total value of the distances” from respective labels of nodes. This limitation is directed toward the abstract idea of mathematical concepts , because it expressly involves determining a total value of distances and selecting the smallest total value. This is a mathematical calculation and comparison. Regarding claim 5 , this claim recites determining, as a prediction result, “a value that corresponds to a label of a specific node included in the similar graph data.” This limitation is directed toward the abstract idea of a mental process , because it involves evaluating selected graph data and determining a corresponding value based on a node label. To the extent the value is determined through the similarity and embedding-based framework of claim 1, this limitation also relates to mathematical concepts . Conclusion for Step 2A Prong 1 Dependent claims 2–5 further narrow the abstract idea recited in independent claim 1 by adding limitations involving similarity calculation, distance calculation, smallest-distance selection, total-distance calculation, and determining a prediction value based on a node label. Accordingly, claims 2–5 recite judicial exceptions, namely mathematical concepts and mental processes , under MPEP § 2106.04(a)(2) . Step 2A Prong 1 result for claims 2–5: YES -- the claims recite a judicial exception. Step 2A Prong 2 -- whether the claim as a whole integrates the recited judicial exception into a practical application of the exception or whether the claim is “directed to” the judicial exception. This evaluation is performed by (1) identifying whether there are any additional elements recited in the claim beyond the judicial exception, and (2) evaluating those additional elements individually and in combination to determine whether the claim as a whole integrates the exception into a practical application. See MPEP 2106.04(d). Regarding dependent claims 2–5 , these claims recite additional limitations including: calculating similarity using embedding vectors and link prediction, defining similarity as a distance, selecting graph data based on smallest distance or smallest total distance, and determining a prediction result corresponding to a node label. These limitations amount to no more than further defining the abstract idea using mathematical calculations and evaluative processes. The recited limitations constitute mathematical operations and data evaluation steps , which merely refine the abstract idea and do not improve the functioning of a computer or any other technology. See MPEP § 2106.05(a) . Further, these limitations amount to no more than instructions to apply the abstract idea using mathematical techniques and data processing, without imposing meaningful limits on the abstract idea. See MPEP § 2106.05(f) . Additionally, the recited limitations merely limit the abstract idea to a particular technological environment involving knowledge graph embedding and similarity calculations, which constitutes a field-of-use limitation. See MPEP § 2106.05(h) . The claims do not recite a particular machine integral to the claim ( MPEP § 2106.05(b) ), do not transform an article ( MPEP § 2106.05(c) ), and do not include any other meaningful limitations that integrate the judicial exception into a practical application ( MPEP § 2106.05(e) ). Accordingly, the additional elements, individually and in combination, do not integrate the judicial exception into a practical application. Step 2B -- whether the claim amounts to significantly more than the judicial exception. See MPEP § 2106.05. Regarding dependent claims 2–5 , the claims recite additional limitations including: calculating similarity using embedding vectors and link prediction, defining similarity as a distance, selecting graph data based on smallest distance or smallest total distance, and determining a prediction result corresponding to a node label. These additional elements, individually and in combination, do not amount to significantly more than the judicial exception. The recited similarity calculations, distance calculations, aggregation of distances, and selection of minimum values constitute mathematical operations and data processing techniques that are well-understood, routine, and conventional in the field. See MPEP § 2106.05(d) . Further, these limitations are expressed in result-oriented functional language , such as “calculating,” “specifying,” and “determining,” without reciting how the results are achieved. Such limitations amount to no more than instructions to apply the abstract idea. See MPEP § 2106.05(f) . Additionally, the recited limitations merely refine the abstract idea by applying known mathematical techniques and selecting data based on calculated metrics, and do not provide any technological improvement or inventive concept beyond the abstract idea. Accordingly, the claims do not include any additional elements that amount to significantly more than the judicial exception. Claim Rejections - 35 USC § 103 07-06 AIA 15-10-15 In the event the determination of the status of the application as subject to AIA 35 U.S.C. 102 and 103 (or as subject to pre-AIA 35 U.S.C. 102 and 103) is incorrect, any correction of the statutory basis (i.e., changing from AIA to pre-AIA) for the rejection will not be considered a new ground of rejection if the prior art relied upon, and the rationale supporting the rejection, would be the same under either status. 07-20-aia AIA 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. 07-103 AIA The text of those sections of Title 35, U.S. Code not included in this action can be found in a prior Office action. 07-23-aia AIA The factual inquiries for establishing a background for determining obviousness under 35 U.S.C. 103 are summarized as follows: 1. Determining the scope and contents of the prior art. 2. Ascertaining the differences between the prior art and the claims at issue. 3. Resolving the level of ordinary skill in the pertinent art. 4. Considering objective evidence present in the application indicating obviousness or nonobviousness. 07-21-aia AIA Claim s 1, 2, 6 and 7 are rejected under 35 U.S.C. 103 as being unpatentable over Albooyeh et al. (Albooyeh) , Non-Patent Literature, “Out-of-Sample Representation Learning for Knowledge Graphs”, published on November 2020 and cited in the IDS filed on 5/15/24, and relied upon at pages 2658-2659, in view of Zhang et al. (Zhang), Non-Patent Literature, “Reasoning Through Memorization: Nearest Neighbor Knowledge Graph Embeddings,” published on January 14, 2022, and relied upon at pages 1-6 . As to independent Claim 1 , Albooyeh teaches a non-transitory computer-readable recording medium storing a prediction program that uses knowledge graph embedding, for causing a computer to execute processing comprising ( Albooyeh , p. 2658, right column, Section 2 “Background and Notation,” para. 4, “KG embedding models map entities and relations to hidden representations known as embeddings and define a function φ from the embeddings of the entities and the relation in a triple to a score…” and Albooyeh , p. 2658, right column, Section 2, para. 5, “for each triple (v, r, u) in the batch, the embeddings for v, r and u are looked up and the score for the triple is computed according to φ. Then the embeddings and the parameters of φ are updated…”): determining whether or not graph data to be predicted is data that includes a node link that indicates a relationship between nodes not included in training data used for training of the knowledge graph embedding ( Albooyeh, p. 2659, left column, Section 3 “Out-of-Sample KG Reasoning”, Definition 1, para. 1 , “Out-of-sample reasoning for KGs is the problem of training a model on a KG G with entities V and relations R such that at test time, the model can be used for making predictions about any out-of-sample entity v ∉ V”; and Albooyeh, p. 2659, right column, Section 3, para. 2 , “According to the definition, G v is observed only at the test time and so during training, the model does not observe any triples involving v”); and determining a prediction result for the graph data to be predicted ( Albooyeh, p. 2659, right column, Section 3 “Out-of-Sample KG Reasoning”, Definition 1, para. 1 , “at test time, the model can be used for making predictions about any out-of-sample entity v ∉ V”; and Albooyeh, p. 2659, right column, Section 3, para. 2 “a function from G v and the in-sample entity and relation embeddings to an embedding for v that can be used to make further predictions about v”). Albooyeh teaches computing an embedding for an out-of-sample entity based on relationships with in-sample entities, for example, by learning embeddings for in-sample entities and relations and using a function from those embeddings to produce an embedding for the out-of-sample entity that can be used for prediction. However, Albooyeh does not teach specifying, in a case where it is determined that the graph data to be predicted is the data that includes the node link not included in the training data, graph data similar to the graph data to be predicted from training data based on a result of embedding prediction for a label of a node included in the graph data to be predicted , and determining a prediction result for the graph data to be predicted based on the specified similar graph data . In the same field of endeavor, Zhang teaches specifying graph data similar to the graph data to be predicted from training data by retrieving nearest neighbors based on similarity in embedding space ( Zhang, p. 1, left column, Abstract, para. 1 , “We compute the nearest neighbors based on the distance in the entity embedding space from the knowledge store”; Zhang, p. 1, right column to p. 2, left column, Introduction, para. beginning “Inspired by this…” , “we construct a knowledge store of entities and retrieve nearest neighbors according to distance in the entity embedding space”; Zhang, p.3, right column, §2.5, para. 1 , “we use the representation of [MASK] output as the predict anchor entity embedding to find the nearest neighbor in the knowledge store”, and Zhang, p.3, left column, §2.5, para. 1 , “we can obtain the probability distribution over neighbors based on a softmax of k-nearest neighbors”). Albooyeh and Zhang are analogous to the claimed invention as both are from the same field of endeavor of knowledge graph embedding and reasoning for predicting relationships between entities. Therefore, it would have been obvious to one of ordinary skill in the art, before the effective filing date of the claimed invention, to combine the out-of-sample representation learning framework of Albooyeh with the nearest neighbor retrieval and memorization mechanism of Zhang. The motivation to combine is as recited by Zhang ( Zhang, p. 1, left column, Abstract, para. 1 : “we compute the nearest neighbors based on the distance in the entity embedding space from the knowledge store… [which] can allow rare or emerging entities to be memorized explicitly… [and] improve inductive and transductive link prediction results”), such that one would enhance the out-of-sample prediction capability of Albooyeh by incorporating similarity-based nearest neighbor retrieval in embedding space, since Albooyeh recognizes the challenge of learning representations for unseen entities and making predictions based on limited relational information ( Albooyeh, p.2658, left column, Introduction, para. 4, “The main challenge of out-of-sample representation learning for non-attributed KGs is that an entity representation must be learned using only the relations the entity participates in…” ; and Albooyeh, p. 2659, right column, Section 3, para. 2, “According to the definition, G v is observed only at the test time and so during training, the model does not observe any triples involving v.”), and Zhang provides a known technique to address this problem by retrieving similar entities based on embedding distance to improve prediction accuracy for rare or unseen entities. As to dependent Claim 2, Albooyeh does not teach but Zhang further teaches wherein the specifying includes calculating similarity between the label of the node and each piece of graph data and specifying graph data most similar from among a plurality of pieces of graph data by performing link prediction using embedding vectors, by computing distances between embeddings and retrieving nearest neighbors from a knowledge store ( Zhang, p. 3, right column, §2.5, para. 1 , “the model queries the knowledge store with the last hidden output vector h [MASK] obtain the distribution of p kNN according to a distance function d(., .)”; Zhang, p.3, right column, §2.5, Eq. (4) , “d(e i ; e j ) = |e i ; e j | 2 ”; Zhang, p.4, left column, §2.5, para. 1 , “we can obtain the probability distribution over neighbors based on a softmax of k–nearest neighbors”; and Zhang, p.4, left column, §2.5, para. 1 “For each entity retrieved by the model, we choose only one nearest embedding in the knowledge store to represent the entity,” wherein similarity is explicitly calculated via embedding distance and the most similar graph data is selected as the nearest neighbor based on that distance). Albooyeh and Zhang are analogous to the claimed invention as both are from the same field of endeavor of knowledge graph embedding and reasoning for predicting relationships between entities. Therefore, it would have been obvious to one of ordinary skill in the art, before the effective filing date of the claimed invention, to modify the system of Albooyeh to include “wherein the specifying includes calculating similarity between the label of the node and each piece of graph data and specifying graph data most similar from among a plurality of pieces of graph data by performing link prediction using embedding vectors, by computing distances between embeddings and retrieving nearest neighbors from a knowledge store” as taught by Zhang for the purpose of providing a known technique to address this problem by retrieving similar entities based on embedding distance to improve prediction accuracy for rare or unseen entities. Claims 6 and 7 are apparatus and method claims, respectively. Claims 6 and 7 contain similar limitations of claim 1. Therefore, claim 6 and 7 are rejected under the same rationale . 07-21-aia AIA Claim 3 is rejected under 35 U.S.C. 103 as being unpatentable over Albooyeh et al. (Albooyeh) , Non-Patent Literature, “Out-of-Sample Representation Learning for Knowledge Graphs”, published on November 2020 and cited in the IDS filed on 5/15/24, and relied upon at pages 2658-2659 and Zhang et al. (Zhang), Non-Patent Literature, “Reasoning Through Memorization: Nearest Neighbor Knowledge Graph Embeddings,” published on January 14, 2022, and relied upon at pages 1-6 as discussed in claims 1, 2, 6 and 7 above, and further in view of Bordes et al. (Bordes), Non-Patent Literature, “Translating Embeddings for Modeling Multi-relational Data”, published in 2013, cited in the IDS filed on 8/23/23, and relied upon at page 3 . As to dependent Claim 3, Albooyeh does not explicitly teach the similarity between the label of the node and the graph data is a distance between the label of the node and the graph data, the label of the node included in the graph data to be predicted is the same label as the label of the node that has already been used for training included in the training data, and the specifying includes in a case where there is a plurality of labels of the nodes included in the graph data to be predicted, specifying graph data that has a smallest distance from among the plurality of pieces of graph data by using a distance between each label of the plurality of nodes and each of the plurality of pieces of graph data. In the same field of endeavor, Zhang teaches in Zhang, p.6, Section 2.5, Eq.(6) , “d(h i , h j )=|h i , h j | 2 ”, wherein the similarity between entities is determined based on the distance between their corresponding embedding representations), and further teaches selecting the nearest neighbor based on the smallest distance in the embedding space ( Zhang, p.6, Section 2.5, para. 2 , “use KNN to choose the nearest embedding in the knowledge store,” and “find the nearest neighbor in the knowledge store,” wherein the most similar graph data is identified as that having the smallest distance. Additionally, Bordes further teaches modeling similarity using a distance function in embedding space ( Bordes, p. 3, Section 2, para. 1 , “the energy of a triplet is equal to d (h+ℓ, t) for some dissimilarity measure d,” wherein the relationship between entities is evaluated based on a distance metric such as L1 or L2 norm), thereby reinforcing that similarity between graph elements is expressed as a distance and that selecting the most similar data corresponds to choosing the smallest distance. Albooyeh, Zhang , and Bordes are analogous to the claimed invention, as all originate from the same field of endeavor of knowledge graph embedding and reasoning for predicting relationships between entities using learned vector representations. Therefore, it would have been obvious to one of ordinary skill in the art, before the effective filing date of the claimed invention, to modify the teaching of Albooyeh to include calculating similarity between the label of the node and each piece of graph data, by performing link prediction with a plurality of pieces of graph data included in the training data by using an embedding vector of a label of a node that has already been used for training for the label of the node included in the graph data to be predicted; and specifying, as graph data similar to the graph data to be predicted, graph data most similar from among the plurality of pieces of graph data, as taught by Zhang and Bordes. The motivation to combine Albooyeh, Zhang, and Bordes is supported by the shared goal of all three references to improve knowledge graph completion and reasoning performance using embedding-based techniques, as Albooyeh identifies the challenge of predicting relationships for out-of- sample entities ( Albooyeh, p. 2658, left column, Introduction, para. 4, “The main challenge of out-of-sample representation learning for non-attributed KGs is that an entity representation must be learned using only the relations the entity participates in…”), Zhang addresses this challenge by retrieving nearest neighbors in embedding space to enhance reasoning for rare or emerging entities ( Zhang, p. 1, left column, Abstract, para. 1 , “Our approach can allow rare or emerging entities to be memorized explicitly rather than implicitly in model parameters”); and Zhang, p. 3, right column, Section 2.5 “Memorized Inference”, para. 1 , “we use the representation of [MASK] output as the predict anchor entity embedding to find the nearest neighbor in the knowledge store”), and Bordes provides a well-established framework for modeling relationships and optimizing predictions via minimizing embedding distances ( Bordes, p. 3, left column, Section 2 “Transaction-based model”, para. 3 , “The loss function (1) favors lower values of the energy for training triplets than for corrupted triplets…”) . 07-21-aia AIA Claim 4 is rejected under 35 U.S.C. 103 as being unpatentable over Albooyeh et al. (Albooyeh) , Non-Patent Literature, “Out-of-Sample Representation Learning for Knowledge Graphs”, published on November 2020 and cited in the IDS filed on 5/15/24, and relied upon at pages 2658-2659 and Zhang et al. (Zhang), Non-Patent Literature, “Reasoning Through Memorization: Nearest Neighbor Knowledge Graph Embeddings,” published on January 14, 2022, and relied upon at pages 1-6 and Bordes et al. (Bordes), Non-Patent Literature, “Translating Embeddings for Modeling Multi-relational Data”, published in 2013, cited in the IDS filed on 8/23/23, and relied upon at page 3 as discussed in claims 3, and further in view of Ferré, Sebastien (Ferré) , Non-Patent Literature, “Link Prediction in Knowledge Graphs with Concepts of Nearest Neighbours,”, published in June 2019, and relied upon at pages 9-10 . As to dependent Claim 4, Albooyeh in view of Zhang and Bordes as applied in Claim 3 does not explicitly teach the specifying includes specifying, from among the plurality of pieces of graph data, graph data that has a smallest total value of the distances by using a total value of the distances from the respective labels of the plurality of nodes for each of the plurality of pieces of graph data. In the same field of endeavor, Ferré teaches this limitation ( Ferré, p. 10, Section 5 “Link Prediction”, para. 3 , “The Dempster’s rule is then used to combine the evidence from all clusters…”, and equation defining belief PNG media_image1.png 46 334 media_image1.png Greyscale , wherein multiple distances d l from different neighbour clusters are aggregated to compute a global belief score for each candidate entity; and “we can rank the entities by decreasing belief,” wherein selecting the highest belief corresponds to selecting the candidate with optimal (i.e., effectively minimal aggregated distance / maximal similarity)). The combination of Albooyeh, Zhang, Bordes, and Ferré teaches computing distances in embedding space ( Zhang, p. 4, left column, Section 2.5, Eq. (6), Euclidean distance between embeddings), representing entities via embeddings ( Bordes, p. 3, Section 2 “Translation-based model” ), and further aggregating multiple distance-based evidences across neighbors to compute a global score and selecting the best candidate (Ferré as cited above). Therefore, it would have been obvious to one of ordinary skill in the art, before the effective filing date of the claimed invention, to combine the embedding-based similarity computation of Zhang and Bordes with the multi-neighbor aggregation and selection mechanism of Ferré in order to improve prediction accuracy by leveraging multiple related nodes rather than a single nearest neighbor, thereby yielding more robust and reliable graph data selection. The motivation to combine Albooyeh, Zhang, Bordes, and Ferré is to as recited by Ferré ( Ferré, p. 9-10, Section 5, “Link Prediction”, para. 2-4 ), describing combining evidence from multiple nearest neighbours to compute a global belief score and rank candidate entities, such that by aggregating distances across multiple related nodes rather than replying on a single similarity measure, the prediction process becomes more robust, accurate, and less sensitive to noise or sparse graph structures, thereby improving the reliability of selecting the most appropriate graph data for prediction . 07-21-aia AIA Claim 5 is rejected under 35 U.S.C. 103 as being unpatentable over Albooyeh et al. (Albooyeh) , Non-Patent Literature, “Out-of-Sample Representation Learning for Knowledge Graphs”, published on November 2020 and cited in the IDS filed on 5/15/24, and relied upon at pages 2658-2659 and Zhang et al. (Zhang), Non-Patent Literature, “Reasoning Through Memorization: Nearest Neighbor Knowledge Graph Embeddings,” published on January 14, 2022, and relied upon at pages 1-6 as discussed in claims 1, 2, 6 and 7 above, and further in view of Costabello et al. (Costabello), US 10,157,226 B1, issued on December 18, 2018 . As to dependent Claim 5, Albooyeh and Zhang as applied in Claims 1 do not explicitly teach the determining o the prediction result includes determining, as a prediction result for the graph data to be predicted, a value that corresponds to a label of a specific node included in the similar graph data is determined. Costabello teaches in Costabello, p. 14, col. 3, lines 1–4 , “the prediction platform may score the candidate responses based on the revised knowledge graph embeddings, and may identify a particular candidate response, that best answers the query, based on scoring the candidate responses”; Costabello, p. 16, col. 7, lines 57-62 , “the prediction engine may score the candidate drugs based on the revised knowledge graph embeddings… and determine values associated with the candidate statements”; and Costabello, p. 16, col. 7, line 63 – col. 8, line 1 , “the prediction engine may then utilize the values to calculate the probability estimates for the candidate drugs… and provide the scored candidate drugs…”, wherein scoring candidate entities and selecting the best candidate corresponds to determining a prediction result as a value associated with a label of a specific node. Albooyeh, Zhang, and Costabello are analogous to the claimed invention as they are from the same field of endeavor of knowledge graph embedding and link prediction. Therefore, it would have been obvious to one of ordinary skill in the art, before the effective filing date of the claimed invention, to combine Albooyeh’s out-of-sample representation learning framework with Zhang’s nearest-neighbor retrieval and Costabello’s scoring and selection mechanism. The motivation to combine is to as recited by Costabello ( Costabello, p. 14, col. 3, lines 3-5 , “identify a particular candidate response, that best answers the query, based on scoring the candidate responses”) such that to enable accurate prediction by selecting the highest-scoring candidate entity from among possible nodes, thereby improving prediction interpretability and effectiveness in knowledge graph completion tasks. Conclusion Any inquiry concerning this communication or earlier communications from the examiner should be directed to HUNG VAN LE whose telephone number is (571)270-0164. The examiner can normally be reached 8 a.m. - 5 p.m.. 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, Cesar Paula can be reached at (571) 272-4128. 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. /HUNG VAN LE/Examiner, Art Unit 2145 /CESAR B PAULA/Supervisory Patent Examiner, Art Unit 2145 Application/Control Number: 18/454,155 Page 2 Art Unit: 2145 Application/Control Number: 18/454,155 Page 3 Art Unit: 2145 Application/Control Number: 18/454,155 Page 4 Art Unit: 2145 Application/Control Number: 18/454,155 Page 5 Art Unit: 2145 Application/Control Number: 18/454,155 Page 6 Art Unit: 2145 Application/Control Number: 18/454,155 Page 7 Art Unit: 2145 Application/Control Number: 18/454,155 Page 8 Art Unit: 2145 Application/Control Number: 18/454,155 Page 9 Art Unit: 2145 Application/Control Number: 18/454,155 Page 10 Art Unit: 2145 Application/Control Number: 18/454,155 Page 11 Art Unit: 2145 Application/Control Number: 18/454,155 Page 12 Art Unit: 2145 Application/Control Number: 18/454,155 Page 13 Art Unit: 2145 Application/Control Number: 18/454,155 Page 14 Art Unit: 2145 Application/Control Number: 18/454,155 Page 15 Art Unit: 2145 Application/Control Number: 18/454,155 Page 16 Art Unit: 2145 Application/Control Number: 18/454,155 Page 17 Art Unit: 2145 Application/Control Number: 18/454,155 Page 18 Art Unit: 2145 Application/Control Number: 18/454,155 Page 19 Art Unit: 2145
Prosecution Insights