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 .
Status of Claims
This final office action is responsive to Applicant’s submission filed 08/12/2025. Currently, claims 1-18 are pending. Claims 1, 8 and 12 have been amended. No newly added and/or cancelled claims.
Allowable Subject Matter
Claims 1-18 are allowed over prior art.
The following is a statement of reasons for the indication of allowable subject matter:
None of the cited and/or relevant prior art, teaches or suggests the combination:
“generating a training set including one or more examples comprising natural language descriptions of items of the item catalog and values of one or more attributes for the item from the one or more templates and the item catalog, each example corresponding to the item and including a plurality of tokens in positions, with values of one or more tokens determined from values of one or more attributes of the item;
training a corpus model to receive a natural language description of the item and to output one or more embeddings in a vector space for one or more tokens in the natural language description of the item by:
applying the corpus model to each example of the training set and backpropagating one or more error terms based on a difference between a predicted token generated by the corpus model for a position of an example and a token included at the position of the example until one or more loss functions satisfy one or more criteria;
obtaining selection training data from prior searches for items obtained by the online concierge system, the selection training data comprising a plurality of selection examples, a selection example including a query term and a plurality of pairs that each include an item identifier and an affinity score between an item corresponding to the item identifier and the query term;
training a model comprising the corpus model and a mapping layer that receives an embedding from the corpus model and outputs a predicted similarity of the embedding to item identifier embeddings generated by the corpus model for each item of the item catalog, wherein the corpus model is further configured to receive item identifiers as tokens and to generate token embeddings for the item identifiers in a same shared latent vector space as embeddings generated for query attributes, wherein training the model comprises:
applying the model to each selection example of the training set and backpropagating one or more error terms based on a difference between a predicted similarity between the embedding from the corpus model and an item embedding and the affinity score between the query term of the selection example and the item embedding until one or more loss functions satisfy one or more criteria; and
applying the trained model to the one or more attributes in the query to generate, in natural language, item identifiers that fit the query, wherein the one or more attributes and the item identifiers are embedded by the corpus model into the shared latent vector space, wherein the trained model generates predicted similarities between the query and each item identifier of the item catalog in a single iteration by operating on embeddings in the shared latent vector space, without retrieving or separately encoding the item entries from the database; and
retrieving an item based on the item identifiers generated by the trained model to bypass searching for the one or more attributes in the database that stores the data entries, thereby improving the latency of response in reducing the query by the online concierge system”,
as recited in claim 1.
Claims 8 and 12 recite similar limitations as set forth in claim 1, and therefore are patentable over prior art.
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 1-18 are rejected under 35 U.S.C. 101 because the claimed invention is directed to a judicial exception (i.e., abstract idea) without significantly more.
The claims recite method and apparatus for searching an online catalog.
Exemplary claim 1, recites in part,
“generating a training set including one or more examples comprising natural language descriptions of items of the item catalog and values of one or more attributes for the item from the one or more templates and the item catalog, each example corresponding to the item and including a plurality of tokens in positions, with values of one or more tokens determined from values of one or more attributes of the item; (creating model input data)
training a corpus model to receive a natural language description of the item and to output one or more embeddings in a vector space for one or more tokens in the natural language description of the item by:
applying the corpus model to each example of the training set and backpropagating one or more error terms based on a difference between a predicted token generated by the corpus model for a position of an example and a token included at the position of the example until one or more loss functions satisfy one or more criteria; (training model using input data and mathematical calculations)
training a model comprising the corpus model and a mapping layer that receives an embedding from the corpus model and outputs a predicted similarity of the embedding to item identifier embeddings generated by the corpus model for each item of the item catalog, wherein the corpus model is further configured to receive item identifiers as tokens and to generate token embeddings for the item identifiers in a same shared latent vector space as embeddings generated for query attributes, wherein training the model comprises:
applying the model to each selection example of the training set and backpropagating one or more error terms based on a difference between a predicted similarity between the embedding from the corpus model and an item embedding and the affinity score between the query term of the selection example and the item embedding until one or more loss functions satisfy one or more criteria. (training model using input data and mathematical calculations)
The limitations describe the steps of, 1) creating model input data, 2) training a corpus model using the input data and backpropagation (mathematical calculations), and 3) training a model using input data and backpropagation (mathematical calculations). The steps describe training a model using mathematical calculations on generated input data.
The above limitations, under their broadest reasonable interpretation, encompass "Mathematical Concepts" (mathematical calculations) enumerated in MPEP 2106.04(a)(2)(I). If a claim limitation, under its broadest reasonable interpretation, covers mathematical calculations, then it falls within the “Mathematical Concepts” grouping of abstract ideas. Accordingly, the claim recites an abstract idea.
The judicial exception is not integrated into a practical application. The claim recites the additional elements of an “online concierge system”, a “database” and a “computer-readable storage medium” in the form of computer elements to perform the limitations encompassing the abstract ideas identified above. The recited computer elements represent using a computer as a tool to perform the judicial exception (MPEP 2106.05 (f)(2)).
In addition, the claim recites the additional elements of “obtaining an item catalog for one or more warehouses…”, “storing data entries corresponding to the items in the catalogue…”, “creating one or more templates for natural language descriptions of attributes for each item of the item catalog…”, “obtaining selection training data from prior searches for items obtained by the online concierge system…”, “storing parameters comprising the trained model…”, and “retrieving an item based on the item identifiers generated by the trained model to bypass searching for the one or more attributes…”. The recited steps describe collecting data (obtaining item catalog and selection training data), processing data (generating model input data), and storing data (trained model) which amounts to insignificant pre-solution and post-solution activities (MPEP 2106.05(g)).
Further, the additional elements of “receiving a query”, “applying a trained model to generate, in natural language, item identifiers that fit the query”, and “retrieving data” are recited at a high level of generality that amounts to retrieving and formatting data (MPEP 2106.05(g)).
When considered both individually and as a whole, the additional elements do not integrate the abstract idea into a practical application.
The recitation of additional elements is acknowledged as identified above. The discussion with respect to the practical application is equally applicable to consideration of whether the claims amount to significantly more. The steps of “obtaining an item catalog for one or more warehouses…”, “creating one or more templates for natural language descriptions of attributes for each item of the item catalog…”, “obtaining selection training data from prior searches for items obtained by the online concierge system…”, “storing parameters comprising the trained model…”, “receiving a query”, “applying a trained model to generate, in natural language, item identifiers that fit the query” and “retrieving the results from database”, while amounting to extra-solution activities, also amounts to appending with well-understood, routine, conventional activity, specified at a high level of generality. “Receiving/transmitting information over a network” and “storing and retrieving data” have been recognized by the courts as a well-understood, routine, and conventional function (MPEP 2106.05(d)).
In addition, the recited “online concierge system” and “computer-readable storage medium” represents computer elements used in performing the limitations encompassing the abstract ideas identified above. The computer elements recited represent using a computer as a tool to perform the judicial exception (MPEP 2106.05 (f)(2)).
Therefore, there are no meaningful recitations, considered in combination, that transform the judicial exception into a patent eligible application such that the claim amounts to significantly more than the judicial exception itself.
Accordingly, claim 1 is directed to a judicial exception (i.e., abstract idea) without significantly more.
Claims 8 and 12 recite similar limitations as set forth in claim 1, and therefore are rejected based on the same rationale.
Dependent claims 2-7, 9-11 and 13-18 recite limitations directed to the abstract idea, and does not integrate the abstract idea into a practical application nor amounts to significantly more.
Response to Arguments
101 Rejection
Applicant's arguments filed 01/28/2026 with respect to the rejection of claims 1-18 under 35 U.S.C. §101 have been fully considered but they are not persuasive.
In response to Applicant’s arguments, Examiner respectfully disagrees.
As discussed under section 101 above, the claimed invention is directed to a judicial exception (i.e., abstract idea) without significantly more.
The claimed invention uses a machine-learning based retrieval bypass to improve latency of a database. The machine-learning solution comprising a corpus model and a model (including the corpus model and mapping layer). The corpus model and model apply backpropagating algorithms (mathematical calculations) to training data comprising prior searches and natural language descriptions template. A search query is transformed into “item identifiers” that fits the query, and the required result is retrieved from a database using the “item identifiers”.
In Intellectual Ventures I LLC v. Erie Indemnity Co., the Court held that “[w]hile limiting the index to XML tags certainly narrows the scope of the claims, in this instance, it is simply akin to limiting an abstract idea to one field of use or adding token post-solution components that do not convert the otherwise ineligible concept into an inventive concept. See Bilski v. Kappos, 561 U.S. 593, 612 (2010). Similarly, the metafiles associated with these tags do not transform the claim into something beyond a conventional computer practice for facilitating searches.”
Similar to Intellectual Ventures I, the present application uses “item identifiers” instead of one or more associated item attributes, to retrieve catalogue items within a database. Limiting a search query to “item identifiers” that fits a query, instead of using one or more item attributes in a search query, simply limits the database indexing to “item identifiers” and does not transform the claim into something beyond a conventional computer practice for facilitating searches.
Applicant’s filed specification describes that “a token corresponds to a word” and “a token embedding is an embedding for the word”. The system determines a token embedding for each token in a natural language description for an item. A positional embedding identifies a position of a token within a natural language description for an item. A positional embedding allows the system to identify an order in which tokens, and their corresponding token embeddings, occur in a natural language description for an item. Paragraphs 0067, 0068.
In Versata Dev. Group, Inc. v. SAP Am., Inc., it was argued that the claims recite “a specific approach to determining the price of a product on a computer, using hierarchies so as to enable the desired benefit for the computing environment: fewer software tables and searches, leading to improvements in computer performance and ease of maintenance.” However, the Court explained that the claims are not directed to improving computer performance. The claims are directed to using a computer to improve the performance of price determination – not the performance of a computer.
Similar to Versata, the claimed invention provides for eliminating using one or more item attributes to search a database (larger data size) to achieve the solution quickly. The use of tokens (item identifiers) and token embeddings (item identifiers within natural language item descriptions) improves the performance of searching a database. However, the claimed invention does not improve the functioning of the computer, computer technology or technical field.
Accordingly, the claimed invention(s) is/are directed to a judicial exception (i.e., abstract idea) without significantly more.
Conclusion
THIS ACTION IS MADE FINAL. Applicant is reminded of the extension of time policy as set forth in 37 CFR 1.136(a).
A shortened statutory period for reply to this final action is set to expire THREE MONTHS from the mailing date of this action. In the event a first reply is filed within TWO MONTHS of the mailing date of this final action and the advisory action is not mailed until after the end of the THREE-MONTH shortened statutory period, then the shortened statutory period will expire on the date the advisory action is mailed, and any nonprovisional extension fee (37 CFR 1.17(a)) pursuant to 37 CFR 1.136(a) will be calculated from the mailing date of the advisory action. In no event, however, will the statutory period for reply expire later than SIX MONTHS from the mailing date of this final action.
Any inquiry concerning this communication or earlier communications from the examiner should be directed to OLUSEGUN GOYEA whose telephone number is (571)270-5402. The examiner can normally be reached M-F: 9am-5pm 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, FAHD OBEID can be reached at 5712703324. 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.
/OLUSEGUN GOYEA/Primary Examiner, Art Unit 3627
Prosecution Insights