DETAILED ACTION
This rejection is in response to Request for Continued filed on 04/23/2026.
Claims 1-20 are currently pending and have been examined.
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 .
Continued Examination Under 37 CFR 1.114
A request for continued examination under 37 CFR 1.114, including the fee set forth in 37 CFR 1.17(e), was filed in this application after final rejection. Since this application is eligible for continued examination under 37 CFR 1.114, and the fee set forth in 37 CFR 1.17(e) has been timely paid, the finality of the previous Office action has been withdrawn pursuant to 37 CFR 1.114. Applicant's submission filed on 04/23/2026 has been entered.
Response to Arguments
Applicant's arguments filed 4/23/2026 have been fully considered but they are not persuasive.
With respect to applicant’s arguments on page 7-11 of remarks filed 04/23/2026 that the claim amendments are patent eligible because the amended claims are not directed to an abstract idea because the claims recite a specific machine learning implementation for generating a recommendation and natural language explanation, Examiner respectfully disagrees.
One of the enumerated groupings of abstract ideas is defined as certain methods of organizing human activity that includes fundamental economic principles or practices (including hedging, insurance, mitigating risk); commercial or legal interactions (including agreements in the form of contracts; legal obligations; advertising, marketing or sales activities or behaviors; business relations); managing personal behavior or relationships or interactions between people (including social activities, teaching, and following rules or instructions). Another enumerated grouping of abstract ideas is defined as mathematical concepts grouping is defined as mathematical relationships, mathematical formulas or equations, and mathematical calculations. See MPEP § 2106.04(a)(2).
The claim limitations regarding machine learning are not interpreted as being directed towards the abstract idea. However, the other limitations amount to certain methods of organizing human activity as it relates to sales activities and commercial interactions because the claims include providing recommendation rating scores and natural language explanations of the recommendations based on receiving a query and item data that are initialized as tokens and used to generate recommendation rating score and the natural language explanation. The above-recited limitations also amount to mathematical concepts such as the knowledge graphs, the embedding vector using a matrix of parameters, and recommendation rating scores.
With respect to applicant’s arguments on page 11-13 of remarks filed 04/23/2026 that the additional elements integrate the judicial exception into a practical application because the claims are directed to a specific machine learning based recommendation architecture that processes data to generate a recommendation rating score and natural language explanation which imposes a meaningful limit on the judicial exception and improves how the computer system outputs recommendations, Examiner respectfully disagrees.
In addition, a specific way of achieving a result is not a stand-alone consideration in Step 2A Prong Two. However, the specificity of the claim limitations is relevant to the evaluation of several considerations including the use of a particular machine, particular transformation and whether the limitations are mere instructions to apply an exception. See MPEP §§ 2106.04(d)(I), 2106.05(b), 2106.05(c), and 2106.05(f).
If it is asserted that the invention improves upon conventional functioning of a computer, or upon conventional technology or technological processes, a technical explanation as to how to implement the invention should be present in the specification. That is, the disclosure must provide sufficient details such that one of ordinary skill in the art would recognize the claimed invention as providing an improvement. The specification need not explicitly set forth the improvement, but it must describe the invention such that the improvement would be apparent to one of ordinary skill in the art. Conversely, if the specification explicitly sets forth an improvement but in a conclusory manner (i.e., a bare assertion of an improvement without the detail necessary to be apparent to a person of ordinary skill in the art), the examiner should not determine the claim improves technology. An indication that the claimed invention provides an improvement can include a discussion in the specification that identifies a technical problem and explains the details of an unconventional technical solution expressed in the claim, or identifies technical improvements realized by the claim over the prior art. MPEP § 2106.05(a).
The courts have also identified limitations that did not integrate a judicial exception into a practical application:
Merely reciting the words "apply it" (or an equivalent) with the judicial exception, or merely including instructions to implement an abstract idea on a computer, or merely using a computer as a tool to perform an abstract idea, as discussed in MPEP §§ 2106.04(d)(I) and 2106.05(f);
Adding insignificant extra-solution activity to the judicial exception, as discussed in MPEP §§ 2106.04(d)(I) and 2106.05(g); and
Generally linking the use of a judicial exception to a particular technological environment or field of use, as discussed in MPEP §§ 2106.04(d)(I) and 2106.05(h).
A specific way of achieving a result (e.g. using machine learning to generate recommendation) is not a stand-alone consideration in Step 2A Prong Two. It is unclear to a person of ordinary skill in the art how using machine learning to generate recommendations improves the computer system. Applicant’s specification fails to describe how the computer system is improved and merely recites in paragraph [0131] that process, methods, or algorithms can be stored and instructions executable by a computer in many forms and in paragraph [0131] that the machine learning model is used to input vectors and output recommendations and natural language explanation. Therefore, the additional elements (e.g. using a transformer-based encoder, a machine learning model that uses a graph-aware encoder and an auto- regressive decoder, and a display) when considered individually and in combination do not integrate the judicial exception into a practical application because the claim limitations merely use a computer as a tool to perform the abstract idea.
With respect to applicant’s arguments on page 13-14 of remarks filed 04/23/2026 that the prior art does not teach amended claim features, Examiner respectfully disagrees.
Applicant’s arguments with respect to claim amendments have been considered but are moot because the new ground of rejection does not rely on any reference applied in the prior rejection of record for any teaching or matter specifically challenged in the argument.
Claim Rejections - 35 USC § 101
35 U.S.C. 101 reads as follows:
Whoever invents or discovers any new and useful process, machine, manufacture, or composition of matter, or any new and useful improvement thereof, may obtain a patent therefor, subject to the conditions and requirements of this title.
Claims 1-20 are rejected under 35 U.S.C. 101 because the claimed invention is directed to a judicial exception (an abstract idea) without significantly more.
Under Step 1 of the Subject Matter Eligibility Test, it must be considered whether the claims are directed to one of the four statutory classes of invention. See MPEP § 2106. In the instant case, claims 1-11 are directed to a method, claims 12-19 is directed to a system, and claim 20 is directed to an apparatus ( which falls within one of the four statutory categories of invention (process/apparatus). Accordingly, the claims will be further analyzed under revised step 2:
Under step 2A (prong 1) of the Subject Matter Eligibility Test, it must be considered whether the claims recite a judicial exception if so, then determine in Prong Two if the recited judicial exception is integrated into a practical application of that exception. If the claim recites a judicial exception (i.e., an abstract idea), the claim requires further analysis in Prong Two. One of the enumerated groupings of abstract ideas is defined as certain methods of organizing human activity that includes fundamental economic principles or practices (including hedging, insurance, mitigating risk); commercial or legal interactions (including agreements in the form of contracts; legal obligations; advertising, marketing or sales activities or behaviors; business relations); managing personal behavior or relationships or interactions between people (including social activities, teaching, and following rules or instructions). The mathematical concepts grouping is defined as mathematical relationships, mathematical formulas or equations, and mathematical calculations. See MPEP § 2106.04(a)(2).
Regarding representative independent claim 1, recites the abstract idea of:
a method for generating a recommendation and a natural language explanation of the recommendation, the method comprising:
receiving query data corresponding to a query;
receiving item data corresponding to an item;
initializing the query data as at least one natural language query token;
initializing the item data as at least one natural language item token;
generating a knowledge graph for the item based on the at least one natural language item token;
generating,…, a knowledge graph string by flattening the knowledge graph for the item;
mapping at least one token associated with the knowledge graph string and the at least one natural language query token to an embedding vector using a matrix of parameters;
generating,…, a recommendation rating score and a natural language explanation of the recommendation based on the query data, the item data, and the embedding vector,… to generate the recommendation rating score and the natural language explanation of the recommendation,… encodes a topological structure of the knowledge graph for the item; and
outputting, to a user…, the recommendation rating score and the natural language explanation of the recommendation.
The above-recited limitations amount to certain methods of organizing human activity as it relates to sales activities and commercial interactions because the claims include providing recommendation rating scores and natural language explanations of the recommendations based on receiving a query and item data that are initialized as tokens and used to generate recommendation rating score and the natural language explanation. The above-recited limitations also amount to mathematical concepts such as the knowledge graphs, the embedding vector using a matrix of parameters, and recommendation rating scores. Accordingly, the claim recites an abstract idea. See MPEP § 2106.
The Step 2A (prong 2) of the Subject Matter Eligibility Test, is the next step in the eligibility analyses and looks at whether the abstract idea is integrated into a practical application. This requires an additional element or combination of additional elements in the claims to apply, rely on, or use the judicial exception in a manner that imposes a meaningful limit on the judicial exception, such that the claim is more than a drafting effort designed to monopolize the exception. See MPEP § 2106.
In this instance, the claims recite the additional elements such as:
using a transformer-based encoder…by a machine learning model…, wherein the machine learning model uses a graph-aware encoder and an auto- regressive decoder…, wherein the graph-aware encoder…;… at a display …(Claims 1, 12, and 20);
A system for providing a recommendation and a natural language explanation of the recommendation, the system comprising: a processor; and a memory including instructions that, when execute by the processor, cause the processor to:… (Claim 12);
An apparatus for providing a recommendation and a natural language explanation of the recommendation, the apparatus comprising: a processor; and a memory including instructions that, when executed by the processor, cause the processor to: (Claim 20).
However, these elements do not amount to an improvement in the functioning of a computer or any other technology or technical field, apply the judicial exception with, or by use of, a particular machine, or apply or use the judicial exception in some other meaningful way beyond generally linking the use of the judicial exception to a particular technological environment, such that the claim as a whole is more than a drafting effort designed to monopolize the exception.
Independent claims and dependent claims also fail to recite elements which amount to an improvement in the functioning of a computer or any other technology or technical field, apply the judicial exception with, or by use of, a particular machine, or apply or use the judicial exception in some other meaningful way beyond generally linking the use of the judicial exception to a particular technological environment, such that the claim as a whole is more than a drafting effort designed to monopolize the exception. For example, independent claims and dependent claims are directed to the abstract idea itself and do not amount to an integration according to any one of the considerations above.
Step 2B is the next step in the eligibility analyses and evaluates whether the claims recite additional elements that amount to an inventive concept (i.e., “significantly more”) than the recited judicial exception. According to Office procedure, revised Step 2A overlaps with Step 2B, and thus, many of the considerations need not be re-evaluated in Step 2B because the answer will be the same. See MPEP § 2106.
In Step 2A, several additional elements were identified as additional limitations:
using a transformer-based encoder…by a machine learning model…, wherein the machine learning model uses a graph-aware encoder and an auto- regressive decoder…, wherein the graph-aware encoder…;… at a display …(Claims 1, 12, and 20);
A system for providing a recommendation and a natural language explanation of the recommendation, the system comprising: a processor; and a memory including instructions that, when execute by the processor, cause the processor to:… (Claim 12);
An apparatus for providing a recommendation and a natural language explanation of the recommendation, the apparatus comprising: a processor; and a memory including instructions that, when executed by the processor, cause the processor to: (Claim 20).
These additional limitations, including the limitations in the independent claims and dependent claims, do not amount to an inventive concept because the recitations above do not amount to an improvement in the functioning of a computer or any other technology or technical field, apply the judicial exception with, or by use of, a particular machine, or apply or use the judicial exception in some other meaningful way beyond generally linking the use of the judicial exception to a particular technological environment, such that the claim as a whole is more than a drafting effort designed to monopolize the exception. In addition, they were already analyzed under Step 2A and did not amount to a practical application of the abstract idea.
For these reasons, the claims are rejected under 35 U.S.C. 101.
Claim Rejections - 35 USC § 103
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.
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) 1-20 are rejected under 35 U.S.C. 103 as being unpatentable over Boteanu et al. (US Pub. No. 20210232633 A1, hereinafter “Boteanu”) in view of Pesaranghader et al. (US Pub. No. 20240193667 A1, hereinafter “Pesaranghader”) in further view of Alomrani et al. (US Pub. No. 20240119294 A1).
Regarding claims 1, 12, and 20
Boteanu discloses a method for providing a recommendation and a natural language explanation of the recommendation, the method comprising (Boteanu, [0064]: recommendations with annotations and descriptors based on searches using phrases; [0066]: semantic language):
receiving query data corresponding to a query; receiving item data corresponding to an item; initializing the query data as at least one natural language query token; initializing the item data as at least one natural language item token (Boteanu, [0082]: query is received and defined as words, letters, or phrases and descriptors associated with the item are retrieved; [0015]: passing the query and providing subsequent query in combination with descriptors may describe activities, audiences, interests; [0016]: descriptors are semantically related to query as words);
generating a knowledge graph for the item based on the at least one natural language item token; (Boteanu, FIG. 6C, [0078]: provide a knowledge graph used to find variations of descriptors and knowledge graph represents entities using multiple equivalent names, which may be variations of descriptors; [0079]: knowledge graph includes each entity (e.g. the word television) includes alternative names (e.g. tv, LCD, or display) where an aggregate confidence is assigned to each word to the entity such as a unique identifier (e.g. identification code or value); [0015]: descriptors may describe activities, audiences, interests; [0016]: descriptors are semantically related to query as words; [0070]: training vectors may be a transformation of one-dimensional training vectors to form a multi-dimensional representation of words; [0071]: using a neural network with a hidden layer and multiple related words which may be trained to recognize the single word from multiple related words when one-dimensional representation of a word may be prepared for the transformation to a multi-dimensional representation of words);
mapping at least one token associated with the knowledge graph string and the at least one natural language query token to an embedding vector using a matrix of parameters (Boteanu, [0072]: word-to-word comparison may be performed, and may be extended to multiple words in a window of words. A training vector for a word, as illustrated in reference number 510, may be first converted into a feature representation using a feature matrix; [0015]: descriptors may describe activities, audiences, interests; [0016]: descriptors are semantically related to query as words);
generating, by a machine learning model, a recommendation rating score and a natural language explanation based on the query data, the item data, the embedding vector, wherein the machine learning model uses … to generate the recommendation rating score and the natural language explanation of the recommendation,… the knowledge graph for the item (Boteanu, [0066]: training neural network with words from a review from item purchases and combined with words from query; [0071]: using neural network, training vector maybe trained; [0073]: input a plurality of words and relationships into the trained neural network to output level of closeness between the words in a review to the query and output a words that bears relationship to both words inputted; [0035]: train one or more neural networks with each of the new review or feedback for a purchased item and for the used query; [0039]: terms in the review or feedback are within a predetermined number of words from each other; [0074]: the closest determined terms, by semantic relationships, may have the least distance as calculated; [0078]: knowledge graph for items);
and outputting, to a user at a display, … and the natural language explanation of the recommendation (Boteanu, [0063]: query suggestions for words searched; [0064]: recommendations are related to the search query; [0073]: a trained neural network will be able to identify relationships based on the numerical values associated with words and predict a word that is bears a relationship to both words in a review to a query and provide a level of closeness between words; FIG. 4, [0057]: the machine learning using the previously described neural network(s) may then use the learning that the word SHOES bear semantic similarity to HIKING and RUNNING; Fig. 4, [0056]: interface is presented; FIG. 4, [0058]: query assist includes selectable links titled with the query and descriptors or portions of descriptors; FIG.4, [0059]: display prominent items on user interface that are semantically the closest to the word).
Boteanu does not teach:
generating, using a transformer-based encoder, a knowledge graph string by flattening the knowledge graph for the item;
…wherein the machine learning model uses a graph-aware encoder and an auto-regressive decoder…, wherein the graph-aware encoder encodes a topological structure of the .. graph…;
outputting, to a user…, the recommendation rating score…
However, Pesaranghader teaches:
generating, using a transformer-based encoder, a knowledge graph string by flattening the knowledge graph for the item (Pesaranghader, [0202]: for KGE-based recommendations, the query embedding is aligned with the KGE space. This alignment is facilitated by a pre-trained autoencoder, which translates NLP embeddings into their equivalent KG embeddings (Step 3). Once the aligned query embedding is generated; [0205]: recommendation based on products; [0184]: a pre-trained natural language processing (NLP) embedding model configuration to encode; [0144]: use knowledge graph to encode; [0017]: aligning the NLP embedding representation to a knowledge graph embedding);
outputting, to a user…, the recommendation rating score…(Pesaranghader, [0187]: provide calculated rankings to the user).
It would have been obvious to one of ordinary skill in the art at the time the invention was made to have modified the neural network and output to a user at a display of Boteanu with generate knowledge graph string using a transformer-based encoder and output recommendation rating scores to a user as taught by Pesaranghader because the results of such a modification would be predictable. Specifically, Boteanu would continue to teach the neural network and output to a user at a display except that now generate knowledge graph string using a transformer-based encoder and output recommendation rating scores to a user are taught according to the teachings of Pesaranghader in order to improve recommendation accuracy. This is a predictable result of the combination. (Pesaranghader, [0005-0007]).
However, Alomrani teaches
…wherein the machine learning model uses a graph-aware encoder and an auto-regressive decoder…, wherein the graph-aware encoder encodes a topological structure of the …graph… (Alomrani, [0036] Graph Neural Networks (GNNs) are neural network models that learn to encode the structural information of a graph into vector representations for each node i.e. node embeddings. GNNs output node embeddings that capture the topological structure of each node's neighborhood through a series of neighborhood aggregation layers; [0041]: Encoder-Decoder Architecture; [0024]: decoder comprises multi-layer perception neural networks).
It would have been obvious to one of ordinary skill in the art at the time the invention was made to have modified the machine learning model of Boteanu and Pesaranghader with a graph-aware encoder and an auto-regressive decoder as taught by Alomrani because the results of such a modification would be predictable. Specifically, Boteanu and Pesaranghader would continue to teach the machine learning model except that now a graph-aware encoder and an auto-regressive decoder is taught according to the teachings of Alomrani in order to improve the system. This is a predictable result of the combination. (Alomrani, [0009-0016]).
Regarding claims 12 and 20
Claim 12 is substantially similar to claim 1, however claim 12 additionally recites a system comprising: a processor; and a memory including instructions that, when execute by the processor, cause the processor to...(Boteanu, [0031]: processor and memory). Claim 20 is substantially similar to claim 1 but additionally recites an apparatus comprising: a processor; and a memory including instructions that, when executed by the processor, cause the processor to: … generate a star-shaped knowledge graph (Boteanu, [0031]: processor and memory; FIG. 6C, [0079]: knowledge graph 610 appears star shaped).
Regarding claims 2 and 13
The combination of Boteanu, Pesaranghader, and Alomrani teaches the method of claim 1, wherein the query data includes purchase history data (Boteanu, [0061]: queries based on past purchases of related items).
Regarding claims 3 and 14
The combination of Boteanu, Pesaranghader, and Alomrani teaches the method of claim 2, wherein the purchase history data includes a string representation of previously purchased items associated with at least one of the user and at least one other user (Boteanu, [0080]: behavioral data from past purchase and interaction logs related to a unique identifier; [0016]: previously purchased items include a sentence or statement about the item from prior feedback or reviews; [0014]: feedback is from other sources including reviews about the product; [0025]: feedback or reviews left by users or prior purchasers of associated items or products to the query).
Regarding claim 4 and 15
The combination of Boteanu, Pesaranghader, and Alomrani teaches the method of claim 1, wherein the query data includes customer requirement data (Boteanu, [0024]: requiring a user to filter the reviews to determine if items are relevant to the latent interests of the user; FIG. 5, [0065]: reviews with matches to queries).
Regarding claims 5 and 16
The combination of Boteanu, Pesaranghader, and Alomrani teaches the method of claim 4, wherein the customer requirement data is represented via a tokenization of extracted keywords associated with the query (Boteanu, [0024]: requiring a user to filter the reviews to determine if items are relevant to the latent interests of the user; FIG. 5, [0065]: reviews with matches to queries; FIG. 5, [0068]: parse and extract reviews).
Regarding claims 6 and 17
The combination of Boteanu, Pesaranghader, and Alomrani teaches the method of claim 1, wherein the knowledge graph for the item includes denotation tokens (Boteanu, FIG. 6C, [0078]: knowledge graph with variations of descriptors).
Regarding claims 7 and 18
The combination of Boteanu, Pesaranghader, and Alomrani teaches the method of claim 6, wherein the denotation tokens include at least a head token (Boteanu, FIG. 6C, [0078]: knowledge graph with primary term).
Regarding claims 8 and 19
The combination of Boteanu, Pesaranghader, and Alomrani teaches the method of claim 7, wherein the head token includes a topic of the knowledge graph for the item (Boteanu, FIG. 6C, [0078]: knowledge graph with primary term includes television).
Regarding claim 9
The combination of Boteanu, Pesaranghader, and Alomrani teaches the method of claim 1, wherein the knowledge graph for the item includes a star-shaped knowledge graph (Boteanu, FIG. 6C, [0079]: knowledge graph 610 appears star shaped).
Regarding claim 10
The combination of Boteanu, Pesaranghader, and Alomrani teaches the method of claim 9, wherein a center node of the knowledge graph for the item includes an item entity associated with the item (Boteanu, FIG. 6C, [0079]: knowledge graph in center includes an entity such as television).
Regarding claim 11
The combination of Boteanu, Pesaranghader, and Alomrani teaches the method of claim 1, wherein the parameters include randomly initialized parameters (Boteanu, [0071]: correct initial random values in the hidden layer to an accurate representation of a multi-dimensional vector for the word's relation to other words; [0072]: word-to-word comparison may be performed, and may be extended to multiple words in a window of words. A training vector for a word, as illustrated in reference number 510, may be first converted into a feature representation using a feature matrix).
Conclusion
The prior art made of record and not relied upon is considered pertinent to applicant's disclosure is cited as Mohanty et al. (US Pub. No. 20250200281 A1) related to AI models that facilitate building and searching the knowledge graph, Sardina et al. (US Pub. No. 20250021871 A1) related to a framework that provides multiple graph fields in knowledge graphs that embed nodes in different dimensions, and non-patent literature, Knowledge Graphs and Pretrained Language Models Enhanced Representation Learning for Conversational Recommender Systems, related to utilizing natural language interactions and dialogue history to infer user preferences and provide accurate recommendations.
Any inquiry concerning this communication or earlier communications from the examiner should be directed to LATASHA DEVI RAMPHAL whose telephone number is (571)272-2644. The examiner can normally be reached 11 AM - 7:30 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, Jeffrey A. Smith can be reached at 5712726763. 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.
/LATASHA D RAMPHAL/Examiner, Art Unit 3688
/Jeffrey A. Smith/Supervisory Patent Examiner, Art Unit 3688
Prosecution Insights