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 .
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 U.S.C. 102 and 103) is incorrect, any correction of the statutory basis 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.
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and the fee set forth in 37 CFR 1.17(e) has been timely paid, the finality of the previous Office action has been withdrawn pursuant to 37 CFR 1.114.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 7/23/2026 has been entered.
Status of Claims
Claims 1-2, 5-9, 12-16, and 19-20 remain pending, and are rejected.
Claims 3-4, 10-11, and 17-18 have been cancelled.
Response to Arguments
Applicant’s arguments filed on 7/23/2026 with respect to the rejection under 35 U.S.C. 101 have been fully considered, but are not persuasive for at least the following rationale:
Applicant’s arguments filed on 7/23/2026 with respect to the rejection under 35 U.S.C. 101 for claims directed to a judicial exception are not persuasive.
Notably, on pages 11-13 of the Applicant’s Remarks, comparisons are drawn to Core Wireless for improved user interfaces for electronic devices, faster and easier than conventional navigation approaches, and presenting UI elements without navigating away from the current view of the interface. Applicant argues that the present claims restrain how suggested queries and recommended item are arranged and revealed to the user, presenting UI elements without navigating away from the current view of the interface, citing specification paragraph [0076] in support, which discloses how conventional interfaces “may simply provide suggestions for auto-completion”, and where the present claims arranged the suggested queries so that the user can “select an additional query to get a new list of recommended items”.
Examiner respectfully disagrees. In Core Wireless, the specification specifically discussed the technical problems of the interface, such as devices with small screens (technical specification) where data and functionality needs to be divided into many layers and views, and the navigation of an interface, such as having to scroll around and switch views to find the right data/functionality. The claims contained precise language delimiting a limited set of information to the user, rather than using conventional user interface methods to merely display data. The present claims and specification do not have such technical problems/solutions as in Core Wireless. The cited paragraph of the Applicant’s specification merely discloses selecting the additional query to get a new list of recommended items, which merely discusses selecting a suggestion to display a new list of items, and is not any technical changes to the interface. Furthermore, the claims merely recite displaying data of the abstract idea in a particular location among the abstract data, rather than reciting any particular user interface functionality. It was not simply the ability to display information without navigating away from an interface, but the particular method of delimiting data of applications based on technical specifications of the device, without having to launch the applications (while the applications are un-launched). The claims merely present an abstract algorithm for providing query suggestions, and providing results of the suggested query.
In view of the above, the rejection under 35 U.S.C. 101 has been maintained below.
Applicant’s arguments filed on 7/23/2026 with respect to the rejection under 35 U.S.C. 103 have been fully considered, but moot in light of new grounds of rejection. Applicant’s amendments necessitated new grounds of rejection. However, on pages 13-14 of the Applicant’s Remarks, it is argued that the cited references fail to teach or suggest the features of “generating or expanding a panel UI elements below the suggested query and presenting a second list of recommended items within the panel UI element, the second list of recommended items being displayed between the suggested query and a next suggested query… without navigating away from the user interface”.
Examiner respectfully disagrees. Bai discloses allowing the user to preview search results from a particular query suggestion by generating a preview window (expanding a panel UI element) on the same interface by the query suggestion (Bai: [0035]). Figure 8, #805 of Bai discloses an image of the expanded window that displays over the selected suggested query and in between the other suggested queries. Furthermore, the exact location of the expanded panel is an obvious matter of design choice in light of the system already disclosed in the combination of references. Such modification would not have otherwise affected the invention of the combination of Wan, Boteanu, and Bai, and would have merely represented one of numerous placements of the expanded panel. Notably, Applicant has also failed to persuasively demonstrate the criticality of the expanded panel being between the suggested query and a next suggested query.
Claim Objections
Regarding Claims 7 and 14, with Claim 7 as representative: Claim 7 recites wherein the machine-learned language model has a transformer architecture including one or more attention layers.
However, claim 1 already recites wherein the machine-learned language model is configured as a transformer architecture including one or more attention layers…
The transformer architecture including one or more attention layers reads as another architecture in addition to what is recited in claim 1. It appears that the claim has been amended into the independent claim, and the dependent claim was not cancelled. Appropriate correction is required.
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 therefore, subject to the conditions and requirements of this title.
Claims 1-2, 5-9, 12-16, and 19-20 are rejected under 35 U.S.C. 101 because the claims are directed to a judicial exception without significantly more.
Step 1:
Claims 1-2 and 5-7 are directed to a method, which is a process. Claims 8-9 and 12-14 are directed to a non-transitory computer-readable storage medium, which is an article of manufacture. Claim 15-16 and 19-20 are directed to a system, which is an apparatus. Therefore, claims 1-2, 5-9, 12-16, and 19-20 are directed to one of the four statutory categories of invention.
Step 2A (Prong 1):
Taking claim 15 as representative, claim 15 sets forth the following limitations reciting the abstract idea of suggesting queries for a user based on an input query or the recommended items:
receiving, an input search query;
generating a list of recommended items as a response to the input query, wherein the list of recommended items retrieved are related to the input query;
presenting the list of recommended items to the user;
generating a prompt for input, the prompt specify at least content of the input query or the list of recommended items, and a request to formulate suggested queries related to the input query or the list of recommended items;
providing the prompt to a model;
receiving a response generated by executing the model on the prompt;
parsing the response from the model to extract a set of suggested queries;
presenting the set of suggested queries adjacent to the list of recommended items as a user is entering the input query, wherein the one or more suggested queries are natural-language based queries;
the second list of recommended items being displayed between the suggested query, so that the second list of recommended items is presented without navigating away, while the next suggested query and at least one other suggested query remained displayed, wherein the second list of recommended items is related to the suggested query and comprises a response to the suggested query.
The recited limitations above set forth the process for suggesting queries for a user based on an input query or recommended items. These limitations amount to certain methods of organizing human activity, including commercial or legal transactions (e.g. agreements in the form of contracts, advertising, marketing or sales activities or behaviors, etc.). The claims are directed to determining queries to suggest based on an input query and the recommended items from that query (see specification [0002] disclosing the searching for a list of items from a provider), which is an advertising and marketing activity. The limitations also amount to mental processes, including observation and evaluation. The claims are directed to using an input query and recommended items in an algorithm (analysis) to determine suggested queries, which are activities of observation and evaluation. Such concepts have been identified by the courts as abstract ideas (see: MPEP 2106.04(a)(2)).
Step 2A (Prong 2):
Examiner acknowledges that representative claim 15 recites additional elements, such as:
a computer processor;
a non-transitory computer-readable storage medium;
a client device;
a search element on a user interface;
machine-learned language model;
wherein the machine-learned model is configured to be a transformer architecture including one or more attention layers, the transformer architecture coupled to receive a set of input tokens and generate a set of output tokens;
responsive to receiving an indication that the user clicked or hovered over a suggested query, generating or expanding a panel UI element below the suggested query and presenting a second list of recommended items with the panel UI element.
training parameters of the machine-learned language model, further comprising:
obtaining training data including the input query and the suggested query that the user interacted with;
generating estimated outputs by applying the parameters of the machine-learned language model to the input query;
computing a loss function based on the estimated outputs and the suggested query;
updating the parameters of the machine-learned language model based on error terms obtained from the loss function.
Taken individually and as a whole, representative claim 15 does not integrate the recited judicial exception into a practical application of the exception. The additional elements do no more than generally link the use of a judicial exception to a particular technological environment or field of use.
Furthermore, this is also because the claim fails to (i) reflect an improvement in the functioning of a computer, or an improvement to other technology or technical field, (ii) implement a judicial exception with a particular machine, (iii) effect a transformation or reduction of a particular article to a different state or thing, or (iv) apply the judicial exception in some other meaningful way beyond generally linking the use of the judicial exception to a particular technological environment.
While the claims recite a computer processor and non-transitory computer-readable storage medium, these elements are recited with a very high level of generalization. The specification does not provide any particular disclosure to the processor except in paragraph [0085], which merely describes the processor as comprising one or more processors or processing units that perform the steps of instructions. The non-transitory computer-readable storage medium is also disclosed without much description, except that it stores information (specification: [0086]). As such, it is evident that these elements are generic computing components that merely execute the abstract idea, such that it is performed on a computer. The client device is disclosed in specification paragraph [0014] as any personal or mobile computing device, such as a smartphone, tablet, laptop computer, or desktop computer. As such, it is clear the client device merely represents the user within a computing environment, and the user interface is merely functions as any user interface to provide a general link to a computing environment. Paragraph [0036] discloses the machine-learned language model as any of a transformer-based architecture, LSTM, Markov networks, BART, GAN, diffusion models, and the like. Furthermore, the claims do not recite any of the underlying technology of the machine learned models and merely receives an input to provide an output. As such, it is evident that the machine learned models are any generic machine learning model that is merely applied to the abstract idea to provide an output of data. The training parameters, obtaining training data, generating estimated outputs by applying parameters, computing a loss function based on estimated outputs, and updating the parameters based on error terms obtained from the loss function also do not recite any particular changes of machine learning. These limitations merely flesh out generic machine learning processes, but these steps merely represent general functions of any machine learning model, and do not change or improve any underlying technology of machine learning.
In view of the above, under Step 2A (Prong 2), representative claim 15 does not integrate the recited exception into a practical application (see: MPEP 2106.04(d)).
Step 2B:
Returning to representative claim 15, taken individually or as a whole, the additional elements of claim 15 do not provide an inventive concept (i.e. whether the additional elements amount to significantly more than the exception itself). As noted above, the additional elements recited in claim 15 are recited in a generic manner with a high level of generality and only serve to implement the abstract idea on a generic computing device. The claims result only in an improved abstract idea itself and do not reflect improvements to the functioning of a computer or another technology or technical field. As discussed above with respect to the integration of the abstract idea into a practical application, the additional elements used to perform the claimed process ultimately amount to no more than the mere instructions to apply the exception using a generic computer and/or no more than a general link to a technological environment.
Even when considered as an ordered combination, the additional elements of claim 15 do not add anything further than when they are considered individually.
In view of the above, claim 15 does not provide an inventive concept under step 2B, and is ineligible for patenting.
Regarding Claim 1 (method): Claim 1 recites at least substantially similar concepts and elements as recited in claim 15 such that similar analysis of the claims would be readily apparent to one of ordinary skill in the art. As such, claims 1 is rejected under at least similar rationale as provided above regarding claim 15.
Regarding Claim 8 (non-transitory computer-readable storage medium): Claim 8 recites at least substantially similar concepts and elements as recited in claim 15 such that similar analysis of the claims would be readily apparent to one of ordinary skill in the art. As such, claims 8 is rejected under at least similar rationale as provided above regarding claim 15.
Dependent claims 2, 5-7, 9, 12-14, 16, and 19-20 recite further complexity to the judicial exception (abstract idea) of claim 15, such as by further defining the algorithm of suggesting queries for a user based on an input query or the recommended items, and do not recite any further additional elements. Thus, each of claims 2, 5-7, 9, 12-14, 16, and 19-20 are held to recite a judicial exception under Step 2A (Prong 1) for at least similar reasons as discussed above.
Under prong 2 of step 2A, the additional elements of dependent claims 2, 5-7, 9, 12-14, 16, and 19-20 also do not integrate the abstract idea into a practical application, considered both individually or as a whole. More specifically, dependent claims 2, 5-7, 9, 12-14, 16, and 19-20 rely on at least similar elements as recited in claim 15. Further additional elements are also acknowledged (e.g., an application programming interface (claim 6); a transformer architecture (claim 7)); however, the additional elements of claims 2, 5-7, 9, 12-14, 16, and 19-20 are recited only at a high level of generality (i.e. as generic computing hardware) such that they amount to nothing more than the mere instructions to implement or apply the abstract idea on generic computing hardware (or, merely uses a computer as a tool to perform an abstract idea). Further, the additional elements do no more than generally link the use of a judicial exception to a particular technological environment or field of use (such as the Internet or computing networks).
Secondly, this is also because the claims fails to (i) reflect an improvement in the functioning of a computer, or an improvement to other technology or technical field, (ii) implement the judicial exception with, or use the judicial exception in conjunction with, a particular machine or manufacture that is integral to the claim, (iii) effect a transformation or reduction of a particular article to a different state or thing, or (iv) applies or uses the judicial exception in some other meaningful way beyond generally linking the use of the judicial exception to a particular technological environment.
Taken individually and as a whole, dependent claims 2, 5-7, 9, 12-14, 16, and 19-20 do not integrate the recited judicial exception into a practical application of the exception under step 2A (prong 2).
Lastly, under step 2B, claims 2, 5-7, 9, 12-14, 16, and 19-20 also fail to result in “significantly more” than the abstract idea under step 2B. The dependent claims recite additional functions that describe the abstract idea and use the computing device to implement the abstract idea, while failing to provide an improvement to the functioning of a computer, another technology, or technical field. The dependent claims fail to confer eligibility under step 2B because the claims merely apply the exception on generic computing hardware and generally link the exception to a technological environment.
Even when viewed as an ordered combination (as a whole), the additional elements of the dependent claims do not add anything further than when they are considered individually.
Taken individually or as an ordered combination, the dependent claims simply convey the abstract idea itself applied on a generic computer and are held to be ineligible under Steps 2B for at least similar rationale as discussed above regarding claim 15. Thus, dependent claims 2, 5-7, 9, 12-14, 16, and 19-20 do not add “significantly more” to the abstract idea.
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, 6-9, 13-16, and 20 are rejected under 35 U.S.C. 103 as being unpatentable by Wan (US 20210097063 A1) in view of Boteanu (US 11,036,801 B1), and in further view of Bai (US 20170147680 A1), and in even further view of Volkovs (US 20220058489 A1).
Regarding Claim 1: Wan discloses a method comprising:
receiving, from a client device, an input search query via a search element on a user interface generated on the client device; (Wan: [0037] – “Search component 136 includes an interface that enables users and/or automatic processes to initiate searches of entity data store 132 or session data store 134 or content system 150 and to retrieve results of those searches. Thus, search component 136 may provide a user interface to allow users of entity system 110 to search entity data store”; Wan: [0084] – “a target query is identified in the session data using a first temporal constraint. An example of an approach for identifying a target query is described above in connection with training data generation component 144, 200. In an embodiment, block 322 includes identifying the target query by determining a query of the at least three search queries that has a most recent timestamp data”).
generating a list of recommended items as a response to the input query, wherein the list of recommended items retrieved are related to the input query; (Wan: [0039] – “Notification component 140 generates and delivers electronic content, such as search results, search recommendations, and notifications, to user accounts of users of entity management system 130. Examples of electronic notifications include synchronous or asynchronous messages, alerts, news feed items, recommendations, listings of search results”
presenting the list of recommended items to the user interface of the client device, wherein the one or more suggested queries are natural-language based queries; (Wan: [0039] – “delivers electronic content, such as search results, search recommendations, and notifications, to user accounts of users of entity management system 130. Examples of electronic notifications include synchronous or asynchronous messages, alerts, news feed items, recommendations, listings of search results”; Wan: Fig. 4C, #448 displaying natural language AI in the system for determining queries).
generating a prompt for input to a machine-learned language model, the prompt specify at least content of the input query or the list of recommended items, and a request to formulate suggested queries related to the input query or the list of recommended items; (Wan: [0079] – “in response to a search query, the learned model produced by block 306 is used to generate a related query. An example of a mechanism that can be used to generate the related query is shown in FIG. 3F, described below. In an embodiment, block 308 includes using the learned model to generate at least one recommended query that is semantically related to a new query, in response to the new query”).
providing the prompt to a model serving system for execution by the machine-learned language model; (Wan: [0079] – “in response to a search query, the learned model produced by block 306 is used to generate a related query”).
receiving, from the model serving system, a response generated by executing the model machine-learned language model on the prompt; (Wan: [0079] – “in response to a search query, the learned model produced by block 306 is used to generate a related query. An example of a mechanism that can be used to generate the related query is shown in FIG. 3F, described below. In an embodiment, block 308 includes using the learned model to generate at least one recommended query that is semantically related to a new query”; Wan: [0080] – “generating at least one recommended query using the learned model and a vocabulary of words extracted from the at least one search log based on frequency of occurrence of the words in the at least one search log. In an embodiment, block 308 includes generating the at least one recommended query by the learned model iteratively selecting words from the vocabulary according to a probability”).
parsing the response from the model serving system to extract a set of suggested queries; (Wan: [0080]- “generating at least one recommended query using the learned model and a vocabulary of words extracted from the at least one search log based on frequency of occurrence of the words in the at least one search log. In an embodiment, block 308 includes generating the at least one recommended query by the learned model iteratively selecting words from the vocabulary according to a probability”).
training parameters of the machine-learned language model, further comprising: obtaining training data including the input query and the suggested query that the user interacted with; (Wan: [0024] – “The training sequence includes at least one context query, a source query, and a target query arranged in a temporal order. The training sequence is used to train a model using a machine learning-based process. Through the machine learning-based process, the model learns a mapping between the at least one context query and the source query, on the one hand, and the target query, on the other hand. The resulting learned model then can be used to generate related search suggestions that may better represent a user's intent than the query entered by the user, even for previously unseen queries”).
Wan does not explicitly teach a method comprising:
wherein the machine-learned language model is configured as a transformer architecture including one or more attention layers, the transformer architecture coupled to receive a set of input tokens and generate a set of output tokens;
presenting the set of suggested queries on the user interface of the client device as a user of the client device is entering the input query in the search element;
presenting one or more suggested queries adjacent to the list of recommended items;
responsive to receiving an indication that the user clicked or hovered over a suggested query, generating or expanding a panel UI element below the suggested query and presenting a second list of recommended items within the panel UI element, the second list of recommended items being displayed between the suggested query and a next suggested query, so that the second list of recommended items is presented without navigating away from the user interface on which the one or more suggested queries is displayed, and while the next suggested query and at least one other suggested query remains displayed on the user interface, wherein the second list of recommended items is related to the suggested query and comprises a response to the suggested query;
generating estimated outputs by applying the parameters of the machine-learned language model to the input query;
computing a loss function based on the estimated outputs and the suggested query;
updating the parameters of the machine-learned language model based on error terms obtained from the loss function.
Notably, however, Wan does disclose displaying recommended queries adjacent to the search input box and generating a notification in response to the recommended queries, which includes search results (Wan: [0081]; [0039]).
To that accord, Boteanu does teach a method comprising:
presenting the set of suggested queries on the user interface of the client device as a user of the client device is entering the input query in the search element; (Boteanu: col. 18, ln. 58-col. 19, ln. 3 – “auto-completion in the search query is more relevant using query assists, so that when an query is provided for “headphones,” then the search bar suggests “headphones . . . for airplane travel” or plainly, provides an selectable link of “for running” as an available result or for selection. In response to a query for “backpacks,” for instance, the search bar suggests “for parents” or “for hiking” based on these terms appearing in reviews left by customers who previously purchased select items using the query in searches prior to the current search provided. The present method and system supplements other methods of query suggestion, for example based on frequency of usage”; Boteanu: Fig. 4, #420 displaying suggested queries as the input query is being entered).
presenting one or more suggested queries adjacent to the list of recommended items; (Boteanu: col. 16, ln. 58-col 17, ln. 16 – “responsive to a new query that may be determined as associated with a stored query, a query assist menu 420 may be provided. Query assist menu 420 at least modifies the first interface. The query assist menu 420 may also be provided over the popular products of the first interface. In the query assist menu 420, there is area 420A provided for one or more query assists—e.g., “Shoes for Running ‘Sneakerun's sneaker is the lightest running . . . ’ 420B, “Shoes for Hiking” 420C, and “Shoes for Cold Weather Climbing” 420D. The query assist includes selectable links titled with the query and descriptors or portions of descriptors, as explained with respect to FIGS. 2 and 3. In an alternate implementation, an area 416A may be provided in interface 402 with the query assist for the query. On selection of one of the selectable links, the interface is modified (or only area 404 is modified) to display items of the listed items that are associated to a descriptor underlying the selected selectable link. Furthermore, the modification to the interface may be to rearrange or rank the displayed items as most relevant to the selected selectable link. One of ordinary skill would understand that example 400 provides results 410-414 in interface 402 as an example only, but that if the query is not submitted, then the query assist menu 420 may be provided over any prior listing of items or results to invite a selection from the user entering the query”; Boteanu: Fig. 4, #420,418,410 displaying the suggested queries over the search results).
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 invention of Wan disclosing the system of determining recommended queries using a machine learning model with the presenting suggested queries as the user is entering the input query as taught by Boteanu. One of ordinary skill in the art would have been motivated to do so in order to reduce latency issues of requesting additional pages from the server (Boteanu: col. 1, ln. 15-27).
Wan in view of Boteanu does not explicitly teach a method comprising:
wherein the machine-learned language model is configured as a transformer architecture including one or more attention layers, the transformer architecture coupled to receive a set of input tokens and generate a set of output tokens;
responsive to receiving an indication that the user clicked or hovered over a suggested query, generating or expanding a panel UI element below the suggested query and presenting a second list of recommended items within the panel UI element, the second list of recommended items being displayed between the suggested query and a next suggested query, so that the second list of recommended items is presented without navigating away from the user interface on which the one or more suggested queries is displayed, and while the next suggested query and at least one other suggested query remains displayed on the user interface, wherein the second list of recommended items is related to the suggested query and comprises a response to the suggested query;
generating estimated outputs by applying the parameters of the machine-learned language model to the input query;
computing a loss function based on the estimated outputs and the suggested query;
updating the parameters of the machine-learned language model based on error terms obtained from the loss function.
Notably, however, Wan does disclose displaying recommended queries adjacent to the search input box and generating a notification in response to the recommended queries, which includes search results (Wan: [0081]; [0039]).
To that accord, Bai does teach a method comprising:
responsive to receiving an indication that the user clicked or hovered over a suggested query, generating or expanding a panel UI element below the suggested query and presenting a second list of recommended items within the panel UI element, the second list of recommended items being displayed between the suggested query and a next suggested query, so that the second list of recommended items is presented without navigating away from the user interface on which the one or more suggested queries is displayed, and while the next suggested query and at least one other suggested query remains displayed on the user interface, wherein the second list of recommended items is related to the suggested query and comprises a response to the suggested query; (Bai: [0035] – “The suggestion engine 170 may further allow the user to preview the search results 165 that would be generated by the search engine 160 from a particular query suggestion 163 if the user were to select the user interface element corresponding to the query suggestion 163. For example, if the user hovers a finger above a user interface element, or selects the user interface element for a predetermined duration or with a predetermined force, rather than replace the GUI 175 with search results 165 corresponding to the query suggestion 163 corresponding to the user interface element, the suggestion engine 170 may generate a preview window that displays some or all of the search results 165 to the user in the GUI 175 or other user interface”; Bai: Fig. 8, #805 displaying an image of the expanded window that displays over the selected suggested query and in between the other suggested queries ).
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 invention of Wan in view of Boteanu disclosing the system of determining recommended queries using a machine learning model with the clicking of a suggested query to generate a second list of recommended items responsive to the suggested query as taught by Bai. One of ordinary skill in the art would have been motivated to do so in order to display the results without replacing the contents of the GUI (Bai: [0048]).
Wan in view of Boteanu and Bai does not explicitly teach a method comprising:
wherein the machine-learned language model is configured as a transformer architecture including one or more attention layers, the transformer architecture coupled to receive a set of input tokens and generate a set of output tokens;
generating estimated outputs by applying the parameters of the machine-learned language model to the input query;
computing a loss function based on the estimated outputs and the suggested query;
updating the parameters of the machine-learned language model based on error terms obtained from the loss function.
Notably, however, Wan does disclose machine-learning based processes between a context query and a source query (Wan: [0024]).
To that accord, Volkovs does teach a method comprising:
wherein the machine-learned language model is configured as a transformer architecture including one or more attention layers, the transformer architecture coupled to receive a set of input tokens and generate a set of output tokens; (Volkovs: [0032] – “Each encoder in the encoders 230 comprises multiple neural network layers that transform the input data into abstract feature vectors, which are used as input for subsequent neural network layers”; Volkovs: [0020] – “the recommendation system 130 may generate recommendations for the online system 110 by using a trained deep neural network. The deep neural network may be a two-headed attention fused deep neural network”).
generating estimated outputs by applying the parameters of the machine-learned language model to the input query; (Volkovs: [0047] – “The training data may be split into three data sets, namely, a training dataset for learning the set of parameters, a validation dataset for an unbiased estimate of the model performance, and a test dataset for evaluating final performance”).
computing a loss function based on the estimated outputs and the suggested query; (Volkovs: [0054] – “The recommendation system 130 determines a loss function that indicates a difference between the estimated outputs 760 and actual labels for the plurality of training instances”).
updating the parameters of the machine-learned language model based on error terms obtained from the loss function. (Volkovs: [0054] – “The recommendation system 130 determines a loss function that indicates a difference between the estimated outputs 760 and actual labels for the plurality of training instances. During the backpropagation step, the recommendation system 130 repeatedly updates the set of parameters for the prediction model by backpropagating error terms obtained from the loss function”).
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 invention of Wan in view of Boteanu disclosing the system of determining recommended queries using a machine learning model with the machine-learned model with a transformer architecture including attention layers, generating estimated outputs, computing a loss function, and updating parameters based on error terms as taught by Volkovs. One of ordinary skill in the art would have been motivated to do so in order to generate more effective recommendations by leveraging more personalized information (Volkovs: [0004]).
Regarding Claim 2: Wan in view of Boteanu, Bai, and Volkovs discloses the limitations of claim 1 above.
Wan does not explicitly teach wherein the set of suggested queries are placed below the list of recommended items or above the list of recommended item on the user interface. Notably, however, Wan does disclose displaying recommended queries adjacent to the search input box (Wan: [0081]).
To that accord, Boteanu does teach presenting one or more suggested queries adjacent to the list of recommended items; (Boteanu: col. 16, ln. 58-col 17, ln. 16 – “responsive to a new query that may be determined as associated with a stored query, a query assist menu 420 may be provided. Query assist menu 420 at least modifies the first interface. The query assist menu 420 may also be provided over the popular products of the first interface. In the query assist menu 420, there is area 420A provided for one or more query assists—e.g., “Shoes for Running ‘Sneakerun's sneaker is the lightest running . . . ’ 420B, “Shoes for Hiking” 420C, and “Shoes for Cold Weather Climbing” 420D. The query assist includes selectable links titled with the query and descriptors or portions of descriptors, as explained with respect to FIGS. 2 and 3. In an alternate implementation, an area 416A may be provided in interface 402 with the query assist for the query. On selection of one of the selectable links, the interface is modified (or only area 404 is modified) to display items of the listed items that are associated to a descriptor underlying the selected selectable link. Furthermore, the modification to the interface may be to rearrange or rank the displayed items as most relevant to the selected selectable link. One of ordinary skill would understand that example 400 provides results 410-414 in interface 402 as an example only, but that if the query is not submitted, then the query assist menu 420 may be provided over any prior listing of items or results to invite a selection from the user entering the query”; Boteanu: Fig. 4, #420,418,410 displaying the suggested queries over the search results).
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 invention of Wan disclosing the system of determining recommended queries using a machine learning model with the presenting suggested queries below or above the list of recommended items as taught by Boteanu. One of ordinary skill in the art would have been motivated to do so in order to reduce latency issues of requesting additional pages from the server (Boteanu: col. 1, ln. 15-27).
Regarding Claim 6: Wan in view of Boteanu, Bai, and Volkovs discloses the limitations of claim 1 above.
Wan further discloses wherein providing the prompt comprises making an application programming interface (API) call to an API of the model serving system. (Wan: [0028] – “Event interface 112 may be implemented as a user interface operable by one or more end users of entity management system 130 and/or as an application program interface (API) through which other components and/or systems may interact with entity management system”).
Regarding Claim 7: Wan in view of Boteanu, Bai, and Volkovs discloses the limitations of claim 1 above.
Wan in view of Boteanu and Bai does not explicitly teach wherein the machine-learned language model has a transformer architecture including one or more attention layers. Notably, however, Wan does disclose a machine learning model to determine the recommended queries, including various attention layers (Wan: [0100]; see also: [0090]; [0098]). Wan does not explicitly disclose a transformer architecture.
To that accord, Volkovs does teach wherein the machine-learned language model has a transformer architecture including one or more attention layers. (Volkovs: [0020] – “the recommendation system 130 may generate recommendations for the online system 110 by using a trained deep neural network. The deep neural network may be a two-headed attention fused deep neural network”).
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 invention of Wan in view of Boteanu and Bai disclosing the system of determining recommended queries using a machine learning model with the machine-learned model with a transformer architecture including attention layers as taught by Volkovs. One of ordinary skill in the art would have been motivated to do so in order to leverage attention mechanisms to assign weights of different modalities (Volkovs: [0005]).
Regarding Claims 8 and 15: Claims 8 and 15 recite substantially similar limitations as claim 1. Therefore, claims 8 and 15 are rejected under the same rationale as claim 1 above.
Regarding Claims 9 and 16: Claims 9 and 16 recite substantially similar limitations as claim 2. Therefore, claims 9 and 16 are rejected under the same rationale as claim 2 above.
Regarding Claims 13 and 20: Claims 13 and 20 recite substantially similar limitations as claim 6. Therefore, claims 13 and 20 are rejected under the same rationale as claim 6 above.
Regarding Claim 14: Claim 14 recites substantially similar limitations as claim 7. Therefore, claim 14 is rejected under the same rationale as claim 7 above.
Claims 5, 12, and 19 are rejected under 35 U.S.C. 103 as being unpatentable by the combination of Wan (US 20210097063 A1), Boteanu (US 11,036,801 B1), Bai (US 20170147680 A1), and Volkovs (US 20220058489 A1), in view of Yoon (US 11,748,413 B1).
Regarding Claim 5: The combination of Wan, Boteanu, Bai, and Volkovs discloses the limitations of claim 1 above.
The combination does not explicitly teach a method comprising:
obtaining a second set of suggested queries related to the input query;
generating a dropdown element below the search interface presenting the second set of suggested queries to the user.
Notably, however, Wan does disclose generating a set of query recommendations (Wan: [0040]).
To that accord, Yoon does teach a method comprising:
obtaining a second set of suggested queries related to the input query; (Yoon: claim 2 – “generating code effective to cause a second natural language query suggestion associated with the third search string to be displayed on the display of the user, wherein the second natural language query suggestion is displayed as a second suggested search string for querying the online item catalog”).
generating a dropdown element below the search interface presenting the second set of suggested queries to the user. (Yoon: col. 6, ln. 38-50 – “Query reformulator 142 may search the query index 140 to determine if a natural language query corresponds to the string data generated by query reformulator 142 from various combinations of keywords 144 and/or the input search query. If a natural language query stored in query index 140 corresponds to the input string data, query reformulator 142 may determine whether or not greater than a threshold number of search results (e.g., greater than a threshold number of items) is associated with the natural language query. If so, the natural language query may be sent by query reformulator 142 as one of the suggested queries 150 (e.g., for display in a drop down menu (or output as an audio suggestion and/or output in another manner) on an interface of computing device”).
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 invention of the combination of Wan, Boteanu, Bai, and Volkovs disclosing the system of determining recommended queries using a machine learning model with the generating of a UI element below the suggested query with the second set of suggested queries and displaying a dropdown element as taught by Yoon. One of ordinary skill in the art would have been motivated to do so in order to provide more accurate queries representative of what the user intends (Yoon: col. 1, ln. 17-31).
Regarding Claims 12 and 19: Claims 12 and 19 recite substantially similar limitations as claim 5. Therefore, claims 12 and 19 are rejected under the same rationale as claim 5 above.
Conclusion
The prior art made of record and not relied upon is considered pertinent to applicant's disclosure.
Braundmeier (US 20200065412 A1) discloses [0024] – “The query prediction system 100 uses a neural network 116 according to the present disclosure to predict a user's query. For example, historical query data, such as historical usage patterns for the user, is processed using the neural network 116. In one example, given a large dataset, wherein ad hoc queries are often performed, the present disclosure is able to reduce the amount time to generate a result from a user query on that large data set by using predicted queries based on analyzed historical usage patterns”.
Yang (US 12,210,576 B1) discloses col. 9, ln. 27-40 – “The subsequent feed-forward layer 206 may be used to modify the FastText embeddings (e.g., by reducing the dimensionality) for input into the self-attention layer(s) 210. The embeddings output by feed-forward layer 206 may be provided along with positional encodings 208a, . . . , 208k (e.g., positional embeddings for each token) to the self-attention layer(s) 210. Self-attention layer(s) 210 may be used to extract the relationship of words in the product title before reducing them into a single embedding (e.g., via pooling layer 212) for the feed forward prediction layer 214. For example, the self-attention layers 210 may output per-token embeddings and the pooling layer 212 may apply average pooling (or another type of pooling) to generate a sentence-level embedding representing the entire title”.
Any inquiry concerning this communication or earlier communications from the examiner should be directed to TIMOTHY J KANG whose telephone number is (571)272-8069. The examiner can normally be reached Monday - Friday: 8:30am - 7:00pm 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, Maria-Teresa Thein can be reached at 571-272-6764. 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.
/T.J.K./Examiner, Art Unit 3689
/VICTORIA E. FRUNZI/Primary Examiner, Art Unit 3689 9/15/2026
Prosecution Insights