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 .
This action is in response to amendments filed January 6th, 2026. The status of the claims is as follows. Claims 1, 5 and 6 are amended. Claims 1-7 are currently pending.
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.
The factual inquiries for establishing a background for determining obviousness under 35 U.S.C. 103 are summarized as follows:
1. Determining the scope and contents of the prior art.
2. Ascertaining the differences between the prior art and the claims at issue.
3. Resolving the level of ordinary skill in the pertinent art.
4. Considering objective evidence present in the application indicating obviousness or nonobviousness.
Claims 1-3, 6 and 7 are rejected under 35 U.S.C. 103 as being unpatentable over Rosset et al. (US11615149B2, hereinafter “Rosset”) in view of Chapelle et al. (“Multi-task learning for boosting with application to web search ranking” [2010], hereinafter “Chapelle”).
Regarding Claim 1,
Rosset discloses A ranking function generating apparatus comprising: a memory; and a processor configured to execute (Rosset [Abstract]; “Described herein is a mechanism for utilizing a neural network to identify and rank search results. A machine learning model is trained by converting training data comprising query-document entries into query term-document entries. The query term-document entries are utilized to train the machine learning model. A set of query terms are identified. The query terms can be derived from a query history. The trained machine learning model is used to calculate document ranking scores for the query terms and the resultant scores are stored in a pre-calculated term-document index. A query to search the document index is broken down into its constituent terms and an aggregate document ranking score is calculated from a weighted sum of the document ranking scores corresponding to the individual query terms. Because the term-document index can be pre-calculated, it can be downloaded to provide deep learning search capabilities in a computationally limited environment.”
Rosset [Column 16 Line 50]; “The example of the machine 900 includes at least one processor 902 (e.g., a central processing unit (CPU), a graphics processing unit (GPU), advanced processing unit (APU), or combinations thereof), one or more memories such as a main memory 904, a static memory 906, or other types of memory, which communicate with each other via link 908. Link 908 may be a bus or other type of connection channel. The machine 900 may include further optional aspects such as a graphics display unit 910 comprising any type of display. The machine 900 may also include other optional aspects such as an alphanumeric input device 912 (e.g., a keyboard, touch screen, and so forth), a user interface (UI) navigation device 914 (e.g., a mouse, trackball, touch device, and so forth), a storage unit 916 (e.g., disk drive or other storage device(s)), a signal generation device 918 (e.g., a speaker), sensor(s) 921 (e.g., global positioning sensor, accelerometer(s), microphone(s), camera(s), and so forth), output controller 928 (e.g., wired or wireless connection to connect and/or communicate with one or more other devices such as a universal serial bus (USB), near field communication (NFC), infrared (IR), serial/parallel bus, etc.), and a network interface device 920 (e.g., wired and/or wireless) to connect to and/or communicate over one or more networks 926”)
Producing training data including at least a first search log related to a first item included in a search result of a search query, a second search log related to a second item included in the search result, and respective domains of the first search log and the second search log (Rosset [Column 5 Line 1]; “The machine-learning algorithms utilize the training data 304 to find correlations among features of the data 302 that affect the outcome or assessment 314. In some example embodiments, the training data 304 includes labeled data, which identifies a correct outcome for the input data. In the context of search queries, training data comprises query-document “pairs” 306. Such query-document pairs often contain an example search query, a document that is relevant to the query (referred to as Doc+) and a document that is not relevant to the query (referred to as Doc−). The Doc+ is a positive example for the search query and the Doc− is a negative example for the search query. In the context of this disclosure, a query-document pair will refer to different combinations of a query and example documents. For example, a query-document pair can comprise a query and an associated Doc+. A query-document pair can also comprise a query and an associated Doc−. Finally, a query-document pair can comprise a query, an associated Doc+, and an associated Doc−“ wherein the plurality of query-document pairs is interpreted as first and second query pair items associated with first and second search query logs with respective document domains of the search logs represented in each pairing
Rosset [Abstract]; “A machine learning model is trained by converting training data comprising query-document entries into query term-document entries. The query term-document entries are utilized to train the machine learning model. A set of query terms are identified. The query terms can be derived from a query history. The trained machine learning model is used to calculate document ranking scores for the query terms and the resultant scores are stored in a pre-calculated term-document index. A query to search the document index is broken down into its constituent terms and an aggregate document ranking score is calculated from a weighted sum of the document ranking scores corresponding to the individual query terms. Because the term-document index can be pre-calculated, it can be downloaded to provide deep learning search capabilities in a computationally limited environment.” wherein the production of training data including query term-document entries including query terms derived from a query history reads on a unit producing training data including search logs related to items in the search result with each of the query terms having respective document-type domains of their unique constituent query items)
and learning, using the training data, parameters of a neural network that implements ranking functions for a plurality of domains through … learning (Rosset [Column 5 Line 20]; “The machine learning model is trained using the training data 304 via a training process 308. Training processes 308 are known in the art and adjust weights and other parameters of the machine learning model such that a loss function, which expresses errors that the model makes, is minimized with respect to the training data. The training process presents each of the query-document pairs to the machine learning model, evaluates the output assessment, compares it to the correct output, and adjusts the weights and other parameters of the model based on the comparison. The process is repeated for the training data set 304 so that the loss function is minimized. The result of the training is the trained machine-learning model 312.”
Rosset [Column 6 Line 56]; “Given a collection (sometimes referred to as a corpus) of documents, C, and a vocabulary of query terms, V, the trained machine learning model can be used to precompute each of the ranking score contribution values, ϕ.sub.t,d, for all terms t∈V and documents d∈C. From these ranking contribution values, the query independent machine learning model for query, q, and document, d, can be calculated as a weighted sum of the ranking contribution values, as given in equation (5). In addition, the pre-computed ranking contribution values can be stored in an inverted index to perform retrieval from the full collection using the learned relevance function, Φ.” wherein ranking functions are implemented)
wherein the neural network comprises a plurality of output layers, respective output layers of the plurality of output layers correspond to respective domains of the plurality of domains, (Rosset [Column 5 Line 1]; “The machine-learning algorithms utilize the training data 304 to find correlations among features of the data 302 that affect the outcome or assessment 314. In some example embodiments, the training data 304 includes labeled data, which identifies a correct outcome for the input data. In the context of search queries, training data comprises query-document “pairs” 306. Such query-document pairs often contain an example search query, a document that is relevant to the query (referred to as Doc+) and a document that is not relevant to the query (referred to as Doc−). The Doc+ is a positive example for the search query and the Doc− is a negative example for the search query. In the context of this disclosure, a query-document pair will refer to different combinations of a query and example documents. For example, a query-document pair can comprise a query and an associated Doc+. A query-document pair can also comprise a query and an associated Doc−. Finally, a query-document pair can comprise a query, an associated Doc+, and an associated Doc−“ wherein the plurality of query-document pairs is interpreted as first and second query pair items associated with first and second search query logs with respective document domains of the search logs represented in each pairing
Rosset [Column 4 Line 1]; “Deep learning models in search systems compare an input query 202 to a candidate document 206 in order to produce a ranking score 218 for the candidate document 206 with respect to the input query 202. The deep learning models typically comprise an embedding layer 204, 208 which produce word embeddings for the input query 202 and/or candidate document 206. After the embedding layers are one or more neural network and/or encoding layers 210, 214. The neural network and/or encoding layers can comprise various types of layers, depending on the particular model architecture. These are the layers which, in a trained model, use and produce correlations of features of the input query 202 and/or candidate document 206 so that a ranking score 218 can be produced that accounts for the correlations. The model has one or more output layers 216 that produce the ranking score 218. Various output layers 216 can be used such as softmax, summation, sigmoid, and so forth, as known in the art.“)
the respective domains of the plurality of domains are distinct among one another (Rosset [Column 5 Line 1]; “The machine-learning algorithms utilize the training data 304 to find correlations among features of the data 302 that affect the outcome or assessment 314. In some example embodiments, the training data 304 includes labeled data, which identifies a correct outcome for the input data. In the context of search queries, training data comprises query-document “pairs” 306. Such query-document pairs often contain an example search query, a document that is relevant to the query (referred to as Doc+) and a document that is not relevant to the query (referred to as Doc−). The Doc+ is a positive example for the search query and the Doc− is a negative example for the search query. In the context of this disclosure, a query-document pair will refer to different combinations of a query and example documents. For example, a query-document pair can comprise a query and an associated Doc+. A query-document pair can also comprise a query and an associated Doc−. Finally, a query-document pair can comprise a query, an associated Doc+, and an associated Doc−“ wherein the plurality of query-document pairs is interpreted as first and second query pair items associated with first and second search query logs with respective distinct document domains (Relevant documents labeled as Doc+, irrelevant documents labeled as Doc-) of the search logs represented in each pairing)
an output layer of the plurality of output layers outputs a ranking score of a corresponding domain of the plurality of domains, (Rosset [Figure 2];
PNG
media_image1.png
570
701
media_image1.png
Greyscale
Rosset [Column 4 Line 12]; “These are the layers which, in a trained model, use and produce correlations of features of the input query 202 and/or candidate document 206 so that a ranking score 218 can be produced that accounts for the correlations. The model has one or more output layers 216 that produce the ranking score 218. Various output layers 216 can be used such as softmax, summation, sigmoid, and so forth, as known in the art” wherein the output layers derived from the respective query-document correlations thus reads on the ranking score correspondent to a domain of the plurality of domains (+Doc, -Doc))
and the neural network after learning performs generating, based on an input search query, one or more domain-specific ranking scores that respectively rank an item in a search result (Rosset [Column 5 Line 33]; “When the machine-learning model 312 is used to perform an assessment, new data 310 (e.g., a new query and a documents to be evaluated with respect to the query) is provided as an input to the trained machine-learning model 312, and the trained machine-learning model 312 generates the assessment 314 as output. In this case the assessment 314 would be a ranking score for the document.” wherein the generation of an assessment of the trained machine-learning model comprising ranking scores for the document associated with a plurality of documents and a query thus reads on one or more domain-specific ranking scores that respectively rank an item in a search result”)
Rosset fails to explicitly disclose but Chapelle discloses Multi-task learning regarding each of the domains as a task (Chapelle [Introduction]; “Multi-task learning algorithms [2] aim to improve the performance of several learning tasks through shared models. Previous work focussed primarily on neural networks, k-nearest neighbors [2] and support vector machines [6]. In this paper, we introduce a novel multi-task learning algorithm for gradient boosting. This is motivated by our interest in web search ranking: gradient boosted decision trees are indeed among the state-of-the-art algorithms for large scale web-search ranking”
Chapelle [Introduction];“On the other hand, a large fraction of queries are region-insensitive. Thus, it seems worthwhile to treat the different countries as tasks that are not completely independent of one another as they share some commonalities, yet, differ enough that one cannot naıvely combine their training data sets.” wherein the search log queries associated with different country domains by which multi-task learning is applied with each country being considered a different task is performed)
It would have been obvious to modify Rosset’s method of producing search log training data queries for training neural network parameters to rank search queries by using Chapelle’s multi-task learning with a plurality of domain tasks. One would have been motivated to do so because multi-task learning with a joint model “enables implicit data sharing and regularization” (Chapelle [Abstract]) thus allowing the model to consider commonalities and differences between queries given their domains.
Regarding Claim 2,
The combination of Rosset/Chapelle teaches the method of Claim 1 (and thus the rejection of Claim 1 is incorporated). The combination already discloses wherein the neural network includes a plurality of output layers that output scholar values representing ranks of items in the plurality of individual domains (Rosset [Column 4 Line 11]; “These are the layers which, in a trained model, use and produce correlations of features of the input query 202 and/or candidate document 206 so that a ranking score 218 can be produced that accounts for the correlations. The model has one or more output layers 216 that produce the ranking score 218. Various output layers 216 can be used such as softmax, summation, sigmoid, and so forth, as known in the art.”)
and the learning unit learns the parameters so as to minimize a value of a loss function defined using a difference between a first output value from the neural network for the domains included in the training data and the first item and a second output value from the neural network for the domains and the second item and using the first search log and the second search log (Rosset [Column 5 Line 20]; “The machine learning model is trained using the training data 304 via a training process 308. Training processes 308 are known in the art and adjust weights and other parameters of the machine learning model such that a loss function, which expresses errors that the model makes, is minimized with respect to the training data. The training process presents each of the query-document pairs to the machine learning model, evaluates the output assessment, compares it to the correct output, and adjusts the weights and other parameters of the model based on the comparison. The process is repeated for the training data set 304 so that the loss function is minimized. The result of the training is the trained machine-learning model 312.” wherein a loss function minimized
Rosset [Column 6 Equation 4];
PNG
media_image2.png
356
497
media_image2.png
Greyscale
Wherein the loss function is defined in part with a difference between neural network output values and their corresponding respective item queries, wherein the queries are determined through the first and second search logs)
Regarding Claim 3,
The combination of Rosset/Chapelle teaches the method of Claim 2 (and thus the rejection of Claim 2 is incorporated). The combination already discloses wherein the training data includes a feature value of the first item and a feature value of the second item, wherein the first output value is an output value from the output layer corresponding to the domains included in the training data, among a plurality of output values output by inputting the feature value of the first item to the neural network, and wherein the second output value is an output value from the output layer corresponding to the domains included in the training data, among a plurality of output values output by inputting the feature value of the second item to the neural network (Rosset [Column 4 Line 1]; “Deep learning models in search systems compare an input query 202 to a candidate document 206 in order to produce a ranking score 218 for the candidate document 206 with respect to the input query 202. The deep learning models typically comprise an embedding layer 204, 208 which produce word embeddings for the input query 202 and/or candidate document 206. After the embedding layers are one or more neural network and/or encoding layers 210, 214. The neural network and/or encoding layers can comprise various types of layers, depending on the particular model architecture. These are the layers which, in a trained model, use and produce correlations of features of the input query 202 and/or candidate document 206 so that a ranking score 218 can be produced that accounts for the correlations. The model has one or more output layers 216 that produce the ranking score 218. Various output layers 216 can be used such as softmax, summation, sigmoid, and so forth, as known in the art.” wherein the training data comprising of query terms and their respective features reads on feature values of first and second items; wherein the various output layers to produce a plurality or ranking scores for each query document-type domain comprising query items reads on first and second output values corresponding to domains of the training data obtained by inputting the feature values of the input queries into the neural network)
Claim 6 recites the exact method performed by the apparatus of Claim 1. Thus, Claim 6 is rejected for reasons set forth in the rejection of Claim 1.
Claim 7 recites a non-transitory computer-readable recording medium having computer-readable instructions to execute the exact method performed by the apparatus of Claim 1. Thus, Claim 7 is rejected for reasons set forth in the rejection of Claim 1.
Claim 4 is rejected under 35 U.S.C. 103 as being unpatentable over Rosset et al. (US11615149B2, hereinafter “Rosset”) in view of Chapelle et al. (“Multi-task learning for boosting with application to web search ranking” [2010], hereinafter “Chapelle”) in view of Dwork et al. (US20030037074, hereinafter “Dwork”)
Regarding Claim 4,
The combination of Rosset/Chapelle teaches the method of Claim 2 (and thus the rejection of Claim 2 is incorporated). The combination already discloses to calculate a value determined from the first search log and the second search log to calculate a value of the loss function and a gradient of the loss function related to the parameters (Rosset [Column 6 Equation 5];
PNG
media_image3.png
242
479
media_image3.png
Greyscale
Wherein the corresponding query term used in the calculation of the loss function and its gradient is determined from the weighting factor for I terms (including first and second log terms)
Rosset [Column 6 Line 56]; “Given a collection (sometimes referred to as a corpus) of documents, C, and a vocabulary of query terms, V, the trained machine learning model can be used to precompute each of the ranking score contribution values, ϕt,d, for all terms t∈V and documents d∈C. From these ranking contribution values, the query independent machine learning model for query, q, and document, d, can be calculated as a weighted sum of the ranking contribution values, as given in equation “ wherein the ranking contribution values are determined from terms t∈V. Thus, the calculation of ranking contribution values to calculate values of the loss functions and the gradient of loss function that are then used to learn the parameters reads on calculation of a value based on the first and second search logs (weighting factor for the ith ranking score contribution factor) to calculate a value of the loss function and the gradient of the loss function for the learning of parameters)
and wherein the learning unit uses the value of the loss function and the gradient of the loss function related to the parameters to learn the parameters (Rosset [Column 5 Line 20]; “The machine learning model is trained using the training data 304 via a training process 308. Training processes 308 are known in the art and adjust weights and other parameters of the machine learning model such that a loss function, which expresses errors that the model makes, is minimized with respect to the training data. The training process presents each of the query-document pairs to the machine learning model, evaluates the output assessment, compares it to the correct output, and adjusts the weights and other parameters of the model based on the comparison. The process is repeated for the training data set 304 so that the loss function is minimized. The result of the training is the trained machine-learning model 312.” wherein the learning unit uses the value of the loss function
Rosset [Column 5 Line 47];
PNG
media_image4.png
154
490
media_image4.png
Greyscale
Wherein equation 1 reads on the loss function utilizing the gradient of the loss function for updating parameters)
The combination of Rosset/Chapelle does not explicitly disclose but Dwork discloses wherein the learning unit calculates, from the difference, a probability that the first item is ranked higher than the second item in one of the domains (Dwork [103]; “A general method for obtaining an initial aggregation of partial lists is proposed, using Markov chains. The states of each Markov chain correspond to the n candidates to be ranked, and the states' transition probabilities depend in some particular way on the given (partial) lists. The stationary probability distribution of the Markov chain is used to sort the n candidates to produce the final ranking. There are several motivations for using Markov chains:
Handling partial lists and top d lists: Rather than require every pair of pages (candidates) i and j to be compared by every search engine (voter), the available comparisons between i and j are used to determine the transition probability between i and j, and exploit the connectivity of the chain to (transitively) “infer” comparison outcomes between pairs that were not explicitly ranked by any of the search engines.” wherein state transitional probabilities derived from difference comparisons between candidates representing the ranking relationship between item queries in the list read on probabilities of first items being ranked higher than second items (transitional probabilities between items) being calculated)
It would have been obvious to modify Rosset/Chapelle’s method of producing search log training data queries for multi-task learning of neural network parameters to replace the search queries of Rosset/Chapelle with the Markov chain probability-based search query rankings of Dwork. One would have been motivated to do so because “the intuition is that Markov chains provide a more holistic viewpoint of comparing all n candidates against each other—significantly more meaningful than ad hoc and local inferences like “if a majority prefer A to B and a majority prefer B to C, then A should be better than C” (Dwork [0104]).
Claim 5 is rejected under 35 U.S.C. 103 as being unpatentable over Rosset et al. (US11615149B2, hereinafter “Rosset”) in view of Chapelle et al. (“Multi-task learning for boosting with application to web search ranking” [2010], hereinafter “Chapelle”) and further in view of Duzhik et al. (US11681713B2, hereinafter “Duzhik”).
Regarding Claim 5,
The combination of Rosset/Chapelle teaches the method of Claim 1 (and thus the rejection of Claim 1 is incorporated). The combination fails to explicitly disclose but Duzhik discloses wherein each search log of the first and second search logs is information representing a number of times a user behavior of a predetermined type was performed with respect to the item included in the search result of the search query, and wherein the plurality of domains comprises a domain of the predetermined type of the user behavior corresponding to the said each search log of the first and second search logs (Duzhik [Column 16 Line 54]; “The aggregator 420 may retrieve, from the user interaction log 218 of the search log database 212 of the search engine server 210, an indication of a plurality of user-interaction parameters 410. The indication of the plurality of user-interaction parameters 410 includes a plurality of respective sets of user-interaction parameters 412, each respective set of user-interaction parameters 412 corresponding to a respective set of search results 408, where a given search result of the respective set of search results 408 is associated with one or more user-interaction parameters of the respective set of user-interaction parameters 412. Generally, each user-interaction parameter of each of the respective set of user-interaction parameters 412 may be indicative of user behavior of one or more users after having submitted the respective search query 404 on the search engine server 210, and clicked on one or more search results in the respective set of search results 408 during a search session on the search engine server 210, as an example via one of the first client device 110, the second client device 120, the third client device 130, and the fourth client device 14”
Duzhik [Column 13 Line 7]; “Non-limiting examples of user-interaction parameters tracked by the analytics server 220 include: Loss/Win: was the document clicked in response to the search query or not. Dwell time: time a user spends on a document before returning to the SERP. Long/short click: was the user interaction with the document long or short, compared to the user-interaction with other documents on the SERP. Click-through rate (CTR): Number of clicks on an element divided by the number of times the element is shown (impressions” wherein the number of clicks on an element in the user-interaction parameters associated with search logs reads on the number of times a user behavior of a predetermined types was performed with respect to a query element; wherein the domain of the user interaction search log is the indication of a plurality of user-interaction parameters thus read on as user behavior corresponding to a search log)
It would have been obvious to modify Rosset/Chapelle’s method of producing search log training data queries for multi-task learning of neural network parameters to be performed with search logs representative of instances of user behaviors. One would have been motivated to do so for the purpose of “generating an additional ranking feature—based on the knowledge (for example, of past user interactions with documents shown to past users associated with the previously-seen search query) associated with the previously-seen search query, which additional ranking feature may then be used by ranking algorithms” (Duzhik [Column 3 Line 12]) thus allowing models to take into consideration human interactions in search query learning.
Response to Arguments
The Examiner acknowledges the Applicant’s amendments to Claims 1, 5 and 6.
Applicant’s arguments filed January 6th, 2026, traversing the rejection of claims 1-7 under 35 U.S.C. § 101 have been fully considered, and are fully persuasive.
Applicant’s arguments filed January 6th, 2026, traversing the rejection of claims 1, 5 and 6 under 35 U.S.C. § 103 have been fully considered, but are not fully persuasive.
Examiner notes that Rosset discloses “The model has one or more output layers 216 that produce the ranking score 218. Various output layers 216 can be used such as softmax, summation, sigmoid, and so forth, as known in the art.” (Rosset []). Rosset also discloses the output layers comprising a softmax/summation/sigmoid activation function of correlated features derived from their respective query-document inputs, thus reading on the output layers outputting ranking scores of a corresponding domain (associated with the query-document input (+Doc, -Doc)) of the plurality of domains (+Doc, -Doc). The neural network’s generation of an assessment to gauge the performance of the trained machine learning model comprising inputted query-document pairs to generate a ranking score respective to the inputted query-document pairs (and the associated specific +Doc/-Doc domain) thus reads on generated domain-specific ranking scores correspondent to the search query. The query-document pairs being of +Doc or -Doc domains reads on distinct respective domains.
The rejection of Claim 1 under 35 U.S.C. § 103 has been maintained. Similarly, the rejection of Claims 5 and 6 under 35 U.S.C. § 103 have been maintained.
The rejection of Claims 2-4, 7 under 35 U.S.C. § 103, which depend directly or indirectly from Claim 1, have been maintained.
Conclusion
Applicant’s amendment necessitated the new ground(s) of rejection presented in this Office action. Accordingly, THIS ACTION IS MADE FINAL. See MPEP § 706.07(a). 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 JONATHAN J KIM whose telephone number is (571)272-0523. The examiner can normally be reached 8-6.
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, Matt El can be reached on (571) 270-3264. 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.
/JONATHAN J KIM/Examiner, Art Unit 2141
/MATTHEW ELL/Supervisory Patent Examiner, Art Unit 2141
Prosecution Insights