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 .
Priority
Acknowledgment is made of applicant's claim for foreign priority based on an application filed in the Republic of India on 28 November 2024. It is noted, however, that applicant has not filed a certified copy of the India application (IN20241093076) as required by 37 CFR 1.55.
A Note on Intended Use
The Examiner notes there are multiple elements in the claims that will be interpreted as intended use. A recitation directed to the manner in which a claimed apparatus is intended to be used does not distinguish the claimed apparatus from the prior art, if the prior art has the capability to so perform, see MPEP 2114 (II) and Ex parte Masham, 2 USPQ2d 1647 (Bd. Pat. App. & Inter. 1987). “Language that suggest or makes optional but does not require steps to be performed does not limit a claim to a particular structure, nor limits the scope of a claim or claim limitation”, see MPEP 2111.04. The Examiner notes the recited prior art has the capability to perform the limitations indicated as intended use.
An incomplete list of the limitations that could be interpreted as intended use is as follows:
Claims 3, 12, and 20 recite “to enhance [document prioritization based on temporal proximity]”.
Claim Rejections - 35 USC § 112
The following is a quotation of 35 U.S.C. 112(b):
(b) CONCLUSION.—The specification shall conclude with one or more claims particularly pointing out and distinctly claiming the subject matter which the inventor or a joint inventor regards as the invention.
The following is a quotation of 35 U.S.C. 112 (pre-AIA ), second paragraph:
The specification shall conclude with one or more claims particularly pointing out and distinctly claiming the subject matter which the applicant regards as his invention.
Claims 6 and 15 are rejected under 35 U.S.C. 112(b) or 35 U.S.C. 112 (pre-AIA ), second paragraph, as being indefinite for failing to particularly point out and distinctly claim the subject matter which the inventor or a joint inventor (or for applications subject to pre-AIA 35 U.S.C. 112, the applicant), regards as the invention.
The claims recite “wherein aggregating the semantic encoding and the temporal encoding comprises: providing the semantic encoding and the temporal encoding to a neural network that aggregates the semantic encoding using weights for the semantic encoding and the temporal encoding”. It is unclear what is meant by “aggregating the semantic encoding” within the context of the claim and claimed invention, as this pertains to aggregating two encodings, not one. For purposes of examination, the interpretation “aggregates the semantic encoding and temporal encoding using weights for the semantic encoding and the temporal encoding” has been taken.
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 claims are directed to a judicial exception (i.e., an abstract idea) without significantly more.
Independent claims 1, 10, and 18 recite generating a final unified representation for each document of a set of documents, wherein generating the final unified representation comprises encoding, using a semantic embedding vector, a document’s core semantic features in a semantic encoding, mapping a time domain for at least the document into a dimensional vector space to encode temporal information into a temporal encoding, and aggregating the semantic encoding and temporal encoding to generate the final unified representation; and generating a query embedding. These encompass mathematical relationships, mathematical formulas or equations, and mathematical calculations, which fall under the “Mathematical Concepts” grouping of abstract ideas.
The independent claims further recite comparing a query embedding to the final unified representation for each document of the set of documents, and identifying one or more documents of the set of documents based on the comparing. These encompass mathematical formulas or equations and mathematical calculations, which fall under the “Mathematical Concepts” grouping of abstract ideas. Additionally, these encompass an evaluation, observation, and/or judgment, which falls under the “Mental Processes” grouping of abstract ideas.
Dependent claims 2, 11, and 19 recite ranking one or more documents based on temporal relevance. This encompasses mathematical relationships, mathematical formulas or equations, and mathematical calculations, which fall under the “Mathematical Concepts” grouping of abstract ideas. These also encompass an evaluation, observation, and/or judgment, which falls under the “Mental Processes” grouping of abstract ideas.
Dependent claims 3, 12, and 20 recite wherein ranking the one or more documents based on temporal relevance comprises applying a scoring method to enhance document prioritization based on temporal proximity. Dependent claims 4 and 13 similarly recite wherein mapping the time domain into the dimensional vector space to encode the temporal information into the temporal encoding comprises applying a theorem/formula to parameterize the temporal information. Dependent claims 7 and 16 similarly recite applying a particular similarity calculation to the query embedding and the final unified representation to generate a score for the document. These encompass mathematical relationships, mathematical formulas or equations, and mathematical calculations, which fall under the “Mathematical Concepts” grouping of abstract ideas.
Because the claims recite limitations that fall under “Mental Processes” and “Mathematical Concepts” groupings of abstract ideas (but for the recitation of generic computer components), accordingly, the claims recite an abstract idea.
The judicial exception is not integrated into a practical application of the idea. More particularly, the claims do not recite additional elements that amount to significantly more than the judicial exception. The claimed computing elements are recited at a high level of generality and recited so generically that they represent no more than mere instructions to apply the judicial exception on a computer (see MPEP 2106.05(f)). These limitations can also be viewed as nothing more than an attempt to generally link the use of the judicial exception to the technological environment of a computer (see MPEP 2106.05(h)).
Furthermore, the additional elements recite insignificant extra-solution activities, including accessing a set of documents (independent claims 1, 10, and 18) and providing one or more identified documents for downstream use (independent claims 1, 10, and 18; dependent claims 8-9 and 17); providing (ranked) documents (dependent claims 2, 11, and 19); providing the document to a computing component (for generating the semantic embedding vector) (dependent claims 5 and 14); providing the semantic encoding and temporal encoding to a computing component that performs the aggregation (dependent claims 6 and 15); providing the one or more identified documents to a remote computing device for display (dependent claims 8 and 17); and inputting the documents to receive an output, and providing the output to a remote computing device for display (dependent claim 9).
The claims further recite insignificant field-of-use limitations, describing the context rather than a particular manner of achieving the result. These include performing an iterative process for each document (i.e., repeat the abstract steps) (independent claims 1, 10, and 18); where the query embedding comprises a time-aware embedding for a query (independent claims 1, 10, and 18); that the provided documents is based on the ranking (dependent claims 2, 11, and 19); the scoring method is Gaussian (dependent claims 3, 12, and 20); the theorem/formula for parameterizing the temporal information is Bochner’s theorem (dependent claims 4 and 13); the document is provided to an encoder component of a transformer-based network for generating the semantic embedding vector (dependent claims 5 and 14); the semantic encoding and temporal encoding is provided to a neural network for performing the aggregation (dependent claims 6 and 15); the similarity scoring method is a cosine similarity (dependent claims 7 and 16); that the type of downstream use comprises providing the one or more identified documents to a remote computing device for display (also an insignificant extra-solution activity) (dependent claims 8 and 17); and that the type of downstream use comprises providing the one or more identified documents as input to a large language model (LLM) with the query, receiving a natural language output from the LLM, and providing the natural language output to a remote computing device for display (also insignificant extra-solution activities) (dependent claim 9), as well as the claimed element that the natural language output is provided for output.
Accordingly, these additional elements do not integrate the abstract idea into a practical application, because they do not impose any meaningful limits on practicing the abstract idea.
The claims do not include any additional elements that are sufficient to amount to significantly more than the judicial exception. As discussed above with respect to the integration of the abstract idea into a practical application, the additional elements reciting the use of various computing software and hardware components amount to no more than mere instructions to apply the exception using generic computer components. Mere instructions to apply an exception using generic computer components cannot provide an inventive concept.
The additional elements recite insignificant extra-solution activities that are well-understood, routine, and conventional activities within the computing realm. See MPEP § 2106.05(d)(II) (“Electronic recordkeeping” and “Storing and retrieving information in memory” with respect to “accessing a set of documents”; “Receiving or transmitting data over a network” with respect to “providing one or more identified documents for downstream use”, providing ranked documents, providing documents/information to various computing components, including for display (see MPEP § 2106.05(d)(II) with respect to “Presenting offers and gathering statistics”).
Even as an ordered combination, the claimed elements do not add anything that is not already present when the steps are considered separately. The claims recite a series of abstract steps at a high level of generality by which the claimed steps are performed. See, e.g., Affinity Labs of Texas LLC v. DirecTV., 838 F.3d 1266 (Fed. Cir. 2016) at p. 7-8 (“At that level of generality, the claims do no more than describe a desired function or outcome, without providing any limiting detail that confines the claim to a particular solution to an identified problem. The purely functional nature of the claim confirms that it is directed to an abstract idea, not to a concrete embodiment of that idea”); and Elec. Power Grp., LLC v. Alstom S.A., 830 F.3d 1350 (Fed. Cir. 2016), slip op. 12 (“[The] essentially result-focused, functional character of claim language has been a frequent feature of claims held ineligible under § 101”).
Even with the inclusion of slightly narrowing steps, e.g., specifying the type of mathematical formula utilized in calculations, such limitations do not move the claims outside the realm of abstract ideas. See, e.g., SAP America, Inc. v. InvestPic, LLC, 890 F.3d 1016, 126 USPQ2d 1638 (Fed. Cir. 2018) at p. 12) (finding that the claimed limitations attempting to narrow the claimed statistical methods to bootstrap, jackknife, and cross-validation were all particular methods of resampling, thus doing no more than simply providing further narrowing of what were still mathematical operations, and added nothing outside the abstract realm).
Furthermore, the rest of the claimed steps are merely invoked in a manner to attempt to limit the claims to a particular technological environment—namely, implementation via computers. Similarly, attempting to narrow the type of computing elements involved, e.g., encoder component of a transformer-based network, neural network, etc., do nothing more than recite mere instructions to apply the judicial exception to a computer, while attempting to narrow it to a particular field-of-use (e.g., encoder component of a transformer-based network, neural network, etc.), describing the context rather than a particular manner of achieving the result.
In other words, at this level of generality, the claims do no more than describe a desired function or outcome, and without providing any limiting detail that confines the claims to a particular solution to an identified problem. The purely functional nature of the claims confirm that they are directed to an abstract idea, not to a concrete embodiment of the idea.
A desired goal (i.e., result or effect), absent of structural or procedural means for achieving that goal, is an abstract idea. In this case, the claims are directed to an abstract idea for failing to describe how—by what particular process or structure—the goal is accomplished. Even with the additional elements, the claimed limitations fail to restrict how the goal is accomplished.
Thus, for at least the aforementioned reasons, the claims are rejected under 35 U.S.C. 101 for being directed to a judicial exception (i.e., an abstract idea) without significantly more.
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.
Claims 1-2, 5-11, and 14-19 are rejected under 35 U.S.C. 103 as being unpatentable over Vinodkumar (“Vinodkumar”) (US 2025/0291768 A1), in view of Yang et al. (“Yang”) (US 12,210,576 B1).
Regarding claim 1: Vinodkumar teaches A computer-implemented method comprising:
accessing a set of documents (Vinodkumar, [0014] and [0019], where a log file may be transmitted to object-based system 102 and stored at log file data store 118 for further access and processing, e.g., generating a data structure using various data objects such as a vector embedding object. A log file may be stored as a document object);
generating a final unified representation for each document of the set of documents (Vinodkumar, [FIG. 4, items 420 and 430] and [0044-0045], where temporal embeddings are added to vector embeddings, resulting in a temporally-enhanced data/vector embeddings), wherein generating the final unified representation comprises performing an iterative process for each document (Vinodkumar, [0014] and [0044-0045], where the system receives a log file stored as a document object, and generates a vector embedding object for the document object, where the vector embedding object may have other embeddings added to it, e.g., temporal embedding, positional embedding, etc. This process is performed for each log file; see, e.g., Vinodkumar, [0062] and [0069], where each log file is accessed and processed to generate a data object, the original data objects being used to convert the data objects into vector embeddings), and wherein the iterative process comprises:
encoding … a document's core semantic features in a semantic encoding (Vinodkumar, [0014], where a vector embedding object of the document object may be generated, where the vector embedding object converts characters and other data from the document object into numbers that capture the characters’ meaning and relationships. See, e.g., Vinodkumar, [0039], where a vector embedding may correspond with a representation of the data/document objects as vectors, where the embedding stores semantic relationships or similarities between objects to help define the inherent structure and meaning of the data),
mapping a time domain for at least the document into a dimensional vector space to encode temporal information into a temporal encoding (Vinodkumar, [FIG. 4, items 420 and 430], [0014] and [0043], where the vector embedding object may be used to generate a temporal embedding of the vector embedding object, where the temporal embedding includes additional numbers to capture the temporal/time component of the data), and
aggregating the semantic encoding and temporal encoding to generate the final unified representation (Vinodkumar, [FIG. 4, items 420 and 430] and [0044-0045], where temporal embeddings are added to vector embeddings, e.g., using sine/cosine functions to add the temporal embeddings to the vector embeddings, resulting in a temporally-enhanced data/vector embeddings. See, e.g., Vinodkumar, [0048], where temporally-enhanced data results from this, and passed to an LLM so that the LLM can use the data to provide responses to search queries);
comparing the query … to the final unified representation for each document of the set of documents; identifying one or more documents of the set of documents based on the comparing; and providing the one or more identified documents for downstream use (Vinodkumar, [0080-0082], where a user submits a question/query to analytics portal interface 550 with a time component, e.g., “last two months”. The question/query is parsed and provided to vector database 530. A portion, e.g., top number, of relevant vectors based on a query/question is retrieved, e.g., using cosine similarity (implying a “comparing” step, as the retrieved vectors are “based on” the query/question). The top number of relevant vectors are then passed into machine learning model 570 to generate the most accurate response back to the user (i.e., “downstream use”)).
Vinodkumar does not appear to explicitly teach [generating the semantic encoding] using a semantic embedding vector; generating a query embedding, where the query embedding comprises a time-aware embedding for a query; [and comparing the query] embedding [to the final unified representation for each document in the set of documents].
Yang teaches [generating the semantic encoding] using a semantic embedding vector (Yang, [11:27-40], where items are tokenized into token embeddings via a language model, where the token embeddings represent each token of the input text, and the token embeddings are passed through a feed-forward layer (e.g., feed forward layer 206) to reduce dimensionality prior to further processing);
generating a query embedding, where the query embedding comprises a time-aware embedding for a query (Yang, [FIG. 4] and [10:61-67]-[12:1-3], where the system determines the seasonal relevance of an item/query (see Yang, [11:11-15], where the steps disclosed in [FIG. 4] may be applied to a query). A seasonal relevance vector may be predicted and subsequently outputted. Feature data may be computed using the seasonal relevance vector, e.g., for each relevant time period (i.e., “time-aware embedding for a query”), where the feature data may be used as an input into a search algorithm); [and]
[comparing the query] embedding [to the final unified representation for each document in the set of documents] (Vinodkumar, [0080-0082], where a user submits a question/query to analytics portal interface 550 with a time component, e.g., “last two months”. The question/query is parsed and provided to vector database 530. A portion, e.g., top number, of relevant vectors based on a query/question is retrieved, e.g., using cosine similarity (implying a “comparing” step, as the retrieved vectors are “based on” the query/question). See Yang, [10:61-67]-[12:1-3] above with respect to the query being in the form of a “query embedding”, as claimed, and is used as input to a search algorithm to retrieve and/or rank content).
It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to have combined the teachings of Vinodkumar and Yang (hereinafter “Vinodkumar as modified”) with the motivation of (1) reducing dimensionality which increases the efficiency of processing (due to less volume of data being dealt with); and (2) converting the query into a query embedding to approximate and simplify the distance calculations by converting the query into an embedding1, and for enable temporal-based query retrieval such that potentially more relevant content is provided to users (Yang, [Background]).
Regarding claim 2: Vinodkumar as modified teaches The computer-implemented method of claim 1, further comprising:
ranking the one or more documents based on temporal relevance, wherein the documents are provided based on the ranking (Yang, [11:11-15] and [11:67]-[12:1-3], where the search algorithm retrieves and/or ranks content based at least in part on the feature that represents the seasonal relevance (i.e., “temporal relevance”) of the first item/query. See also, e.g., Yang, [4:44-56], where the query result pertains to relevant documents).
Regarding claim 5: Vinodkumar as modified teaches The computer-implemented method of claim 1, wherein encoding the document's core semantic features in a semantic encoding comprises:
providing the document to encoder component of a transformer-based network that generates the semantic embedding vector (Yang, [3:1-9], where when input data is passed into a transformer machine learning model, such as a (transformer) neural network (Yang, [2:11-67]), attention weights are calculated between every token simultaneously. The attention unit produces embeddings for every token in context that contain information not only about the token itself, but also a weighted combination of other relevant tokens weighted by the attention weights. See Yang, [9:5-16] and [11:27-40], where dense representation of words such as for an input item/text (i.e., “document; see Yang, [4:44-56]), may have been generated using a pre-trained language model (i.e., “semantic embedding vector”)).
Although Yang does not appear to explicitly state that the pre-trained language model is the disclosed transformer machine learning model, one of ordinary skill in the art would have found it obvious to have modified Yang by incorporating the transformer machine learning model into the learning model for generating the embeddings with the motivation of improving the contextual understanding and computational speed of generating those initial embeddings.
It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to have combined the teachings of Vinodkumar and Yang with the motivation of quickly generating an initial embedding for a document.
Regarding claim 6: Vinodkumar as modified teaches The computer-implemented method of claim 1, wherein aggregating the semantic encoding and the temporal encoding comprises:
providing the semantic encoding and the temporal encoding to a neural network that aggregates the semantic encoding using weights for the semantic encoding and the temporal encoding (Yang, [3:1-9], where when input data is passed into a transformer machine learning model, such as a (transformer) neural network (Yang, [2:11-67]), attention weights are calculated between every token simultaneously. The attention unit produces embeddings for every token in context that contain information not only about the token itself, but also a weighted combination of other relevant tokens weighted by the attention weights. See Vinodkumar in claim 1 above with respect to the “semantic encoding” and “temporal encoding”).
It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to have combined the teachings of Vinodkumar and Yang with the motivation of automatically detecting patterns in complex, e.g., seasonal data and improving over time by retraining models as more and more data becomes available (see, e.g., Yang, [1:54-67]), e.g., as neural networks have a proven track record of semantically modeling text for downstream tasks (Yang, [9:5-16]).
Regarding claim 7: Vinodkumar as modified teaches The computer-implemented method of claim 1, wherein comparing the query embedding to the final unified representation comprises:
applying cosine similarity to the query embedding and the final unified representation to generate a score for the document (Vinodkumar, [0081-0082], where the portion of the vectors outputted and provided to the machine learning model 570 may comprise a top number of relevant vectors (for the log files/documents, i.e., “for the document”) based on a query/question using a cosine similarity algorithm (i.e., “generate a score”). See Yang, [FIG. 4] and [10:61-67]-[12:1-3] in claim 1 above with respect to the “query embedding” being used for retrieving and/or ranking relevant content).
Regarding claim 8: Vinodkumar as modified teaches The computer-implemented method of claim 1, wherein providing the one or more identified documents for downstream use comprises:
providing the one or more identified documents to a remote computing device for display (Vinodkumar, [0049] and [0094], where the user may access a display, e.g., via computer system 700 that includes a display 712 that includes a user interface module (GUI) (Vinodkumar, [0100-0101]), that receives search queries. The system may return (i.e., to the display) relevant data in response to the search query. See Vinodkumar, [0051], where the user operates user device 130, which may send/receive (and display) information from the system. See Yang, [4:44-56] and [11:67]-[12:1-3], where the retrieved and ranked content pertains to items/documents).
It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to have combined the teachings of Vinodkumar and Yang with the motivation of allowing users to view/browse the original data.
Regarding claim 9: Vinodkumar as modified teaches The computer-implemented method of claim 1, wherein providing the one or more identified documents for downstream use comprises:
providing the one or more identified documents as input to a large language model (LLM) with the query; receiving a natural language output from the LLM; and providing the natural language output to a remote computing device for display (Vinodkumar, [0082], where the cosine similarity may be used to retrieve the top number of relevant vectors, which are then passed into machine learning model 570 (such as, e.g., a large language model (Vinodkumar, [0078])), to generate the most accurate response back to the user. The response can include question-answer (e.g., as seen in Vinodkumar, [FIG. 5]), summary, reports, etc., e.g., a response comprising “yes, there was an anomaly detected with your us-west2 data center on Dec. 25, 2023 at 10:47 AM” (i.e., “natural language output”). See Vinodkumar, [0049] and [0094], where the user may access a display, e.g., via computer system 700 that includes a display 712 that includes a user interface module (GUI) (Vinodkumar, [0100-0101]), that receives search queries. The system may return (i.e., to the display) relevant data in response to the search query).
Regarding claim 10: Claim 10 recites substantially the same claim limitations as claim 1, and is rejected for the same reasons.
Note that Vinodkumar teaches A system comprising: one or more processors; [and] a memory coupled to the one or more processors, the memory storing a plurality of instructions executable by the one or more processors, the plurality of instructions comprising instructions that when executed by the one or more processors perform a method comprising [the claimed steps] (Vinodkumar, [FIG. 7] and [0097-0099], where the disclosed system for implementing the described steps includes a bus 702, one or more hardware processors 704 coupled with bus 702 for processing information, and main memory 706 which is coupled to bus 702 for storing information and instructions to be executed by processor 704, where such instructions render the system 700 to perform the disclosed operations).
Regarding claim 11: Claim 11 recites substantially the same claim limitations as claim 2, and is rejected for the same reasons.
Regarding claim 14: Claim 14 recites substantially the same claim limitations as claim 5, and is rejected for the same reasons.
Regarding claim 15: Claim 15 recites substantially the same claim limitations as claim 6, and is rejected for the same reasons.
Regarding claim 16: Claim 16 recites substantially the same claim limitations as claim 7, and is rejected for the same reasons.
Regarding claim 17: Claim 17 recites substantially the same claim limitations as claim 8, and is rejected for the same reasons.
Regarding claim 18: Claim 18 recites substantially the same claim limitations as claim 1, and is rejected for the same reasons.
Note that Vinodkumar teaches A non-transitory computer-readable memory storing a plurality of instructions executable by one or more processors, the plurality of instructions comprising instructions that when executed by the one or more processors cause the one or more processors to perform a method comprising [the claimed steps] (Vinodkumar, [FIG. 6] and [0084-0086], where the disclosed steps may be implemented as a computing component that includes a hardware processor 602 and machine-readable storage medium 604, where the machine-readable storage medium may be a non-transitory storage medium, and stores executable instructions 606-614, which pertain to the instructions for carrying out the disclosed steps (see, e.g., Vinodkumar, [0087-0096])).
Regarding claim 19: Claim 19 recites substantially the same claim limitations as claim 2, and is rejected for the same reasons.
Claims 3, 12, and 20 are rejected under 35 U.S.C. 103 as being unpatentable over Vinodkumar (“Vinodkumar”) (US 2025/0291768 A1), in view of Yang et al. (“Yang”) (US 12,210,576 B1), in further view of Shughrue et al. (“Shughrue”) (US 2024/0331010 A1).
Regarding claim 3: Vinodkumar as modified teaches The computer-implemented method of claim 2, but does not appear to explicitly teach wherein ranking the one or more documents based on temporal relevance comprises: applying a Gaussian scoring method to enhance document prioritization based on temporal proximity.
Shughrue teaches wherein ranking the one or more documents based on temporal relevance comprises: applying a Gaussian scoring method to enhance document prioritization based on temporal proximity (Shughrue, [0072], where the system can fit a Gaussian distribution to the specific place based on the set of check-ins associated with the specific place over a period of time. See Shughrue, [0096], where given a user and a set of candidate places, the recommendation component uses a Gaussian weight function to compute the user-specific relevance score for each place and takes into account historical interactions between users and each place using a Gaussian weight function. See Vinodkumar and Yang in claim 2 above with respect to the ranking pertaining to “document prioritization based on temporal proximity”).
It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to have combined the teachings of Vinodkumar as modified and Shughrue with the motivation of performing such calculations efficiently (i.e., the Gaussian scoring being computationally efficient and simple).2
The Examiner notes that “to enhance [document prioritization based on temporal proximity]” has been considered as an intended use/result, and is not afforded patentable weight.
The Examiner notes that “A claim containing a ‘recitation with respect to the manner in which a claimed apparatus is intended to be employed does not differentiate the claimed apparatus from a prior art apparatus’ if the prior art apparatus teaches all the structural limitations of the claim.” Ex parte Masham, 2 USPQ2d 1647 (Bd. Pat. App. & Inter. 1987); see also MPEP § 2114. The recited prior art has the capability to perform these intended use limitations, and therefore, the prior art meets the claimed limitations. See MPEP § 2111.02; see also In re Schreiber, 128 F.3d 1473, 1477, 44 USPQ2d 1429, 1431 (Fed.Cir. 1997).
Because the combination of Huang as modified disclose all the claimed features, the claimed invention does not distinguish over the prior art since the combination of Vinodkumar as modified would confer the same intended use/result as claimed.
Regarding claim 12: Claim 12 recites substantially the same claim limitations as claim 3, and is rejected for the same reasons.
Regarding claim 20: Claim 20 recites substantially the same claim limitations as claim 3, and is rejected for the same reasons.
Claims 4 and 13 are rejected under 35 U.S.C. 103 as being unpatentable over Vinodkumar (“Vinodkumar”) (US 2025/0291768 A1), in view of Yang et al. (“Yang”) (US 12,210,576 B1), in further view of Xu et al. (“Xu”) (US 2021/0133846 A1).
Regarding claim 4: Vinodkumar as modified teaches The computer-implemented method of claim 1, but does not appear to explicitly teach wherein mapping the time domain into the dimensional vector space to encode the temporal information into the temporal encoding comprises: applying Bochner's theorem to parameterize the temporal information.
Xu teaches applying Bochner's theorem to parameterize the temporal information (Xu, [0041], where time embeddings may be generated using harmonic analysis such as applying a Bochner time embedding method).
It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to have combined the teachings of Vinodkumar as modified and Xu with the motivation of transforming the temporal information into an effective/efficient representation for certain machine learning types, e.g., functional time representation learning when using with self-attention in continuous-time event sequence prediction.3
Regarding claim 13: Claim 13 recites substantially the same claim limitations as claim 4, and is rejected for the same reasons.
Conclusion
The prior art made of record and not relied upon is considered pertinent to applicant’s disclosure. See the enclosed 892 form.
Chen et al. (US 2022/0253435 A1) is cited to show why one of ordinary skill in the art would have found it obvious to have utilized a query embedding for comparing against document vectors as claimed (Chen et al., [0129]).
Brown et al. (US 2016/0217701 A1) is cited to show why one of ordinary skill in the art would have found it obvious to have utilized Gaussian scoring (Brown et al., [0093] and throughout).
Xu et al. (“Self-attention with Functional Time Representation Learning”) is cited to show why one of ordinary skill in the art would have found it obvious to have utilized Bochner's theorem for parameterizing temporal information (Xu et al., [Abstract] and [Conclusion]).
The prior art should be considered to define the claims over the art of record.
Any inquiry concerning this communication or earlier communications from the examiner should be directed to IRENE BAKER whose telephone number is (408)918-7601. The examiner can normally be reached M-F 8-5PM PT.
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, Boris Gorney can be reached at (571) 270-5626. 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.
/IRENE BAKER/Primary Examiner, Art Unit 2154
8 June 2026
1 Chen et al. US 2022/0253435 A1 at [0129] (“…a k-NN search with a product quantizer performs an exhaustive search, such that a product quantizer still requires comparing a query vector to every vector in the database. The benefit is that it approximates and simplifies the distance calculations”).
2 See Brown et al. US 2016/0217701 A1 at [0093] (“…the present disclosure features algorithms that are computationally efficient and simple to implement for real-time learning analyses”), describing Gaussian techniques throughout.
3 Xu et al. “Self-attention with Functional Time Representation Learning”. See [Abstract] and [Conclusion].
Prosecution Insights