Notice of Pre-AIA or AIA Status
The present application, filed on or after March 16, 2013, is being examined under the first inventor to file provisions of the AIA .
Priority
Acknowledgment is made of Applicant's claim for foreign priority under 35 U.S.C. 119 (a)-(d).
Information Disclosure Statement
The information disclosure statement (IDS) submitted on April 28, 2025 is in compliance with the provisions of 37 CFR 1.97. Accordingly, the information disclosure statement is being considered by the examiner.
The information disclosure statement (IDS) submitted on July 17, 2025 is in compliance with the provisions of 37 CFR 1.97. Accordingly, the information disclosure statement is being considered by the examiner.
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.
This application currently names joint inventors. In considering patentability of the claims the examiner presumes that the subject matter of the various claims was commonly owned as of the effective filing date of the claimed invention(s) absent any evidence to the contrary. Applicant is advised of the obligation under 37 CFR 1.56 to point out the inventor and effective filing dates of each claim that was not commonly owned as of the effective filing date of the later invention in order for the examiner to consider the applicability of 35 U.S.C. 102(b)(2)(C) for any potential 35 U.S.C. 102(a)(2) prior art against the later invention.
Claims 1-4, 8-14, and 16-18 are rejected under 35 U.S.C. 103 as being unpatentable over Bo (“Modeling Relevance Ranking under the Pre-training and Fine-tuning Paradigm, August 12, 2021), and further in view of Gu (PG Pub. No. 2022/0382975 A1).
Regarding Claim 1, Bo discloses a computer-implemented method for training a machine-learning algorithm (MLA) to rank in-use digital documents at an online search platform, the MLA comprising a plurality of transformer blocks, and a given transformer block of the plurality of transformer blocks comprising:
a transformer encoder block (see Bo, Abstract, where to verify the effectiveness of Pre-Rank, we showed … implementations by using BERT … as the underlying ranking model [it is the position of the Examiner that BERT (Bi-directional Encoder Representations for Transformers] inherently comprises a transformer encoder block, a multi-head attention layer, and a feed-forward neural network layer; see also Devlin, et. al. “Bert: Pre-training of deep bidirectional transformers for language understanding”, cited by reference in citation [13]);
a multi-head attention layer configured to determine dependencies between inputs to the given transformer block (see Bo, Abstract, where to verify the effectiveness of Pre-Rank, we showed … implementations by using BERT … as the underlying ranking model [it is the position of the Examiner that BERT (Bi-directional Encoder Representations for Transformers] inherently comprises a transformer encoder block, a multi-head attention layer, and a feed-forward neural network layer; see also Devlin, et. al. “Bert: Pre-training of deep bidirectional transformers for language understanding”, cited by reference in citation [13]); and
a feed-forward neural network layer configured to: (i) receive outputs of the multi-head attention layer; and (ii) process the outputs of the multi-head attention layer in parallel (see Bo, Abstract, where to verify the effectiveness of Pre-Rank, we showed … implementations by using BERT … as the underlying ranking model [it is the position of the Examiner that BERT (Bi-directional Encoder Representations for Transformers] inherently comprises a transformer encoder block, a multi-head attention layer, and a feed-forward neural network layer; see also Devlin, et. al. “Bert: Pre-training of deep bidirectional transformers for language understanding”, cited by reference in citation [13]);
the method being executable by at least one processor, the method comprising: receiving, by the at least one processor, training data associated with a given user, the training data including (i) a plurality of past queries having been submitted by the given user to the online search platform; (ii) respective sets of past digital documents generated, by the online search platform, in response to submitting thereto each one of the plurality of past queries, and a given past digital document including a respective past user interaction parameter indicative of whether the given user has interacted with the given past digital document (see Bo, Section 1, paragraph 5, where in the Pre-Rank implementation with BERT, we start with BERT-BASE and continue the pre-training with binary cross-entropy loss on the click log; see also Section 3, paragraph 2, where we used click log to generate supervision signals; see also Fig. 1a, where Queries and Documents are fed into the encoder to generate a click prediction which is compared with actual click data);
during a first training phase: generating, by the at least one processor, using the training data, a first set of training digital objects, a given training digital object of the first set of training digital objects including: (i) a respective past query from the plurality of past queries; and (ii) a predetermined number of past digital documents responsive to the respective past query without including respective past user interaction parameters of the predetermined number of past digital documents (see Bo, Section 1, paragraph 5, where in the Pre-Rank implementation with BERT, we start with BERT-BASE and continue the pre-training with binary cross-entropy loss on the click log; see also Section 3, paragraph 2, where we used click log to generate supervision signals; see also Fig. 1a, where Queries and Documents are fed into the encoder to generate a click prediction which is compared with actual click data);
applying, by the at least one processor, to the given training digital object of the first set of training digital object a tokenizer to generate a respective one of a first plurality of tokens (see Bo, Section 4.1.1, where as for the encoder layer Epre in BERT, the input is a concatenation of query tokens and the clicked document tokens);
feeding, by the at least one processor, the first plurality of tokens to the plurality of transformer blocks, thereby pre-training the MLA for determining, for the given training digital object of the first set of training digital objects if the given user has interacted with each one of the predetermined number of past digital documents (see Bo, Section 4.1.1, where as for the encoder layer Epre in BERT, the input is a concatenation of query tokens and the clicked document tokens);
during a second training phase, following the first training phase: generating, by the at least one processor, using the training data, a second set of training digital objects, a given training digital object of the second set of training digital including: (i) the respective past query from the plurality of past queries; and (ii) a number of past digital documents responsive to the respective past query with which the given user has interacted with including the respective past user interaction parameters of the number of past digital documents (see Bo, Section 1, paragraph 5, where in the Pre-Rank implementation with BERT, we start with BERT-BASE and continue the pre-training with binary cross-entropy loss on the click log; see also Section 3, paragraph 2, where we used click log to generate supervision signals; see also Fig. 1a, where Queries and Documents are fed into the encoder to generate a click prediction which is compared with actual click data);
applying, by the at least one processor, to the given training digital object of the second set of training digital object the tokenizer to generate a respective one of a second plurality of tokens (see Bo, Section 4.1.1, where as for the encoder layer Epre in BERT, the input is a concatenation of query tokens and the clicked document tokens); and
feeding, by the at least one processor, the second plurality of tokens to the plurality of transformer blocks (see Bo, Section 4.1.1, where as for the encoder layer Epre in BERT, the input is a concatenation of query tokens and the clicked document tokens), thereby finetuning the MLA (see Gu, paragraph [0004], where the present disclosure includes a framework that learns from unlabeled documents (e.g., during a pre-training stage) and is fine-tuned for specific downstream applications and/or tasks) to determine, for a given in-use digital document, a likelihood parameter of the given user interacting with the given in-use digital document (see Bo, Section 1, paragraph 5, where in the Pre-Rank implementation with BERT, we start with BERT-BASE and continue the pre-training with binary cross-entropy loss on the click log; see also Section 3, paragraph 2, where we used click log to generate supervision signals; see also Fig. 1a, where Queries and Documents are fed into the encoder to generate a click prediction which is compared with actual click data).
Bo does not explicitly disclose:
the pre-training comprising determining preliminary weights for the plurality of transformer blocks of the MLA; and
the finetuning comprising adjusting the preliminary weights to determine adjusted weights of the plurality of transformer blocks of the MLA for further use in personalized ranking of in-use digital documents for the given user.
Gu discloses:
the pre-training comprising determining preliminary weights for the plurality of transformer blocks of the MLA (see Gu, paragraph [0093], where pre-training the machine learning model 110 includes applying sample-dependent attention weights); and
the finetuning comprising adjusting the preliminary weights to determine adjusted weights of the plurality of transformer blocks of the MLA for further use in personalized ranking of in-use digital documents for the given user (see Gu, paragraph [0062], where at block 302 the process 300 involves accessing training data that includes a set of documents for training and a set of documents for validation; at block 304, the process involves pre-training a machine learning model 110 … at block 306, the process involves fine-tuning the machine learning model 110 using the set of documents for validation).
Both Bo and Gu discloses training BERT based models. Therefore it would have been obvious to one of ordinary skill in the art before the effective filing date of the invention to combine Bo with Gu as it amounts to combining prior art elements according to known techniques to yield predictable results (see MPEP 2143(I)(A)).
Regarding Claim 2, Bo in view of Gu discloses the method of Claim 1, wherein the past digital documents associated with the given training digital objects of the first set of training digital objects have been randomly selected from a respective set of digital documents responsive to the respective past query (see Bo, Section 5.1.3, where for the click prediction, the positive instances are derived from the click data of ORCAS, and the negative instances are randomly selected from the top 100 documents of the corresponding query if users did not click on them).
Regarding Claim 3, Bo in view of Gu discloses the method of Claim 1, wherein the respective past user interaction parameter associated with the given past digital document has been determined based on past click data of the given user (see Bo, Section 1, paragraph 5, where in the Pre-Rank implementation with BERT, we start with BERT-BASE and continue the pre-training with binary cross-entropy loss on the click log; see also Section 3, paragraph 2, where we used click log to generate supervision signals; see also Fig. 1a, where Queries and Documents are fed into the encoder to generate a click prediction which is compared with actual click data).
Regarding Claim 4, Bo in view of Gu discloses the method of Claim 3, wherein the click data includes data of at least one click of the given user on the given past digital document made in response to submitting the respective past query to the online search platform (see Bo, Section 1, paragraph 5, where in the Pre-Rank implementation with BERT, we start with BERT-BASE and continue the pre-training with binary cross-entropy loss on the click log; see also Section 3, paragraph 2, where we used click log to generate supervision signals; see also Fig. 1a, where Queries and Documents are fed into the encoder to generate a click prediction which is compared with actual click data).
Regarding Claim 8, Bo in view of Gu discloses the method of Claim 1, wherein the number of past digital documents responsive to the respective past query with which the given user has interacted are all the past digital documents in a respective set of digital documents responsive to the respective past query that the user has interacted with (see Bo, Section 5.1.3, where for the click prediction, the positive instances are derived from the click data of ORCAS, and the negative instances are randomly selected from the top 100 documents of the corresponding query if users did not click on them).
Regarding Claim 9, Bo in view of Gu discloses the method of Claim 1, wherein a first total number of members in the first set of training digital objects and a second total number of members in the second set of training digital objects are the same (see Bo, Section 5.1.3, where for the click prediction, the positive instances are derived from the click data of ORCAS, and the negative instances are randomly selected from the top 100 documents of the corresponding query if users did not click on them).
Regarding Claim 10, Bo in view of Gu discloses the method of Claim 1, wherein a first total number of members in the first set of training digital objects and a second total number of members in the second set of training digital objects are pre-determined (see Bo, Section 5.1.3, where for the click prediction, the positive instances are derived from the click data of ORCAS, and the negative instances are randomly selected from the top 100 documents of the corresponding query if users did not click on them).
Regarding Claim 1, Bo discloses a system for training a machine-learning algorithm (MLA) to rank in-use digital documents at an online search platform, the MLA comprising a plurality of transformer blocks, and a given transformer block of the plurality of transformer blocks comprising:
a transformer encoder block (see Bo, Abstract, where to verify the effectiveness of Pre-Rank, we showed … implementations by using BERT … as the underlying ranking model [it is the position of the Examiner that BERT (Bi-directional Encoder Representations for Transformers] inherently comprises a transformer encoder block, a multi-head attention layer, and a feed-forward neural network layer; see also Devlin, et. al. “Bert: Pre-training of deep bidirectional transformers for language understanding”, cited by reference in citation [13]);
a multi-head attention layer configured to determine dependencies between inputs to the given transformer block (see Bo, Abstract, where to verify the effectiveness of Pre-Rank, we showed … implementations by using BERT … as the underlying ranking model [it is the position of the Examiner that BERT (Bi-directional Encoder Representations for Transformers] inherently comprises a transformer encoder block, a multi-head attention layer, and a feed-forward neural network layer; see also Devlin, et. al. “Bert: Pre-training of deep bidirectional transformers for language understanding”, cited by reference in citation [13]); and
a feed-forward neural network layer configured to: (i) receive outputs of the multi-head attention layer; and (ii) process the outputs of the multi-head attention layer in parallel (see Bo, Abstract, where to verify the effectiveness of Pre-Rank, we showed … implementations by using BERT … as the underlying ranking model [it is the position of the Examiner that BERT (Bi-directional Encoder Representations for Transformers] inherently comprises a transformer encoder block, a multi-head attention layer, and a feed-forward neural network layer; see also Devlin, et. al. “Bert: Pre-training of deep bidirectional transformers for language understanding”, cited by reference in citation [13]);
the method being executable by at least one processor, the method comprising: receiving, by the at least one processor, training data associated with a given user, the training data including (i) a plurality of past queries having been submitted by the given user to the online search platform; (ii) respective sets of past digital documents generated, by the online search platform, in response to submitting thereto each one of the plurality of past queries, and a given past digital document including a respective past user interaction parameter indicative of whether the given user has interacted with the given past digital document (see Bo, Section 1, paragraph 5, where in the Pre-Rank implementation with BERT, we start with BERT-BASE and continue the pre-training with binary cross-entropy loss on the click log; see also Section 3, paragraph 2, where we used click log to generate supervision signals; see also Fig. 1a, where Queries and Documents are fed into the encoder to generate a click prediction which is compared with actual click data);
during a first training phase: generating, by the at least one processor, using the training data, a first set of training digital objects, a given training digital object of the first set of training digital objects including: (i) a respective past query from the plurality of past queries; and (ii) a predetermined number of past digital documents responsive to the respective past query without including respective past user interaction parameters of the predetermined number of past digital documents (see Bo, Section 1, paragraph 5, where in the Pre-Rank implementation with BERT, we start with BERT-BASE and continue the pre-training with binary cross-entropy loss on the click log; see also Section 3, paragraph 2, where we used click log to generate supervision signals; see also Fig. 1a, where Queries and Documents are fed into the encoder to generate a click prediction which is compared with actual click data);
applying, by the at least one processor, to the given training digital object of the first set of training digital object a tokenizer to generate a respective one of a first plurality of tokens (see Bo, Section 4.1.1, where as for the encoder layer Epre in BERT, the input is a concatenation of query tokens and the clicked document tokens);
feeding, by the at least one processor, the first plurality of tokens to the plurality of transformer blocks, thereby pre-training the MLA for determining, for the given training digital object of the first set of training digital objects if the given user has interacted with each one of the predetermined number of past digital documents (see Bo, Section 4.1.1, where as for the encoder layer Epre in BERT, the input is a concatenation of query tokens and the clicked document tokens);
during a second training phase, following the first training phase: generating, by the at least one processor, using the training data, a second set of training digital objects, a given training digital object of the second set of training digital including: (i) the respective past query from the plurality of past queries; and (ii) a number of past digital documents responsive to the respective past query with which the given user has interacted with including the respective past user interaction parameters of the number of past digital documents (see Bo, Section 1, paragraph 5, where in the Pre-Rank implementation with BERT, we start with BERT-BASE and continue the pre-training with binary cross-entropy loss on the click log; see also Section 3, paragraph 2, where we used click log to generate supervision signals; see also Fig. 1a, where Queries and Documents are fed into the encoder to generate a click prediction which is compared with actual click data);
applying, by the at least one processor, to the given training digital object of the second set of training digital object the tokenizer to generate a respective one of a second plurality of tokens (see Bo, Section 4.1.1, where as for the encoder layer Epre in BERT, the input is a concatenation of query tokens and the clicked document tokens); and
feeding, by the at least one processor, the second plurality of tokens to the plurality of transformer blocks (see Bo, Section 4.1.1, where as for the encoder layer Epre in BERT, the input is a concatenation of query tokens and the clicked document tokens), thereby finetuning the MLA (see Gu, paragraph [0004], where the present disclosure includes a framework that learns from unlabeled documents (e.g., during a pre-training stage) and is fine-tuned for specific downstream applications and/or tasks) to determine, for a given in-use digital document, a likelihood parameter of the given user interacting with the given in-use digital document (see Bo, Section 1, paragraph 5, where in the Pre-Rank implementation with BERT, we start with BERT-BASE and continue the pre-training with binary cross-entropy loss on the click log; see also Section 3, paragraph 2, where we used click log to generate supervision signals; see also Fig. 1a, where Queries and Documents are fed into the encoder to generate a click prediction which is compared with actual click data).
Bo does not explicitly disclose:
the pre-training comprising determining preliminary weights for the plurality of transformer blocks of the MLA; and
the finetuning comprising adjusting the preliminary weights to determine adjusted weights of the plurality of transformer blocks of the MLA for further use in personalized ranking of in-use digital documents for the given user.
Gu discloses:
the pre-training comprising determining preliminary weights for the plurality of transformer blocks of the MLA (see Gu, paragraph [0093], where pre-training the machine learning model 110 includes applying sample-dependent attention weights); and
the finetuning comprising adjusting the preliminary weights to determine adjusted weights of the plurality of transformer blocks of the MLA for further use in personalized ranking of in-use digital documents for the given user (see Gu, paragraph [0062], where at block 302 the process 300 involves accessing training data that includes a set of documents for training and a set of documents for validation; at block 304, the process involves pre-training a machine learning model 110 … at block 306, the process involves fine-tuning the machine learning model 110 using the set of documents for validation).
Both Bo and Gu discloses training BERT based models. Therefore it would have been obvious to one of ordinary skill in the art before the effective filing date of the invention to combine Bo with Gu as it amounts to combining prior art elements according to known techniques to yield predictable results (see MPEP 2143(I)(A)).
Regarding Claim 12, Bo in view of Gu discloses the system of Claim 11, wherein the at least one processor further causes the system to select the past digital documents associated with the given training digital objects of the first set of training digital objects from a respective set of digital documents responsive to the respective past query randomly (see Bo, Section 5.1.3, where for the click prediction, the positive instances are derived from the click data of ORCAS, and the negative instances are randomly selected from the top 100 documents of the corresponding query if users did not click on them).
Regarding Claim 13, Bo in view of Gu discloses the system of Claim 11, wherein the at least one processor further causes the system to determine the respective past user interaction parameter associated with the given past digital document based on past click data of the given user (see Bo, Section 1, paragraph 5, where in the Pre-Rank implementation with BERT, we start with BERT-BASE and continue the pre-training with binary cross-entropy loss on the click log; see also Section 3, paragraph 2, where we used click log to generate supervision signals; see also Fig. 1a, where Queries and Documents are fed into the encoder to generate a click prediction which is compared with actual click data).
Regarding Claim 14, Bo in view of Gu discloses the system of Claim 13, wherein the click data includes data of at least one click of the given user on the given past digital document made in response to submitting the respective past query to the online search platform (see Bo, Section 1, paragraph 5, where in the Pre-Rank implementation with BERT, we start with BERT-BASE and continue the pre-training with binary cross-entropy loss on the click log; see also Section 3, paragraph 2, where we used click log to generate supervision signals; see also Fig. 1a, where Queries and Documents are fed into the encoder to generate a click prediction which is compared with actual click data).
Regarding Claim 16, Bo in view of Gu discloses the system of Claim 11, wherein the number of past digital documents responsive to the respective past query with which the given user has interacted are all the past digital documents in a respective set of digital documents responsive to the respective past query that the user has interacted with (see Bo, Section 5.1.3, where for the click prediction, the positive instances are derived from the click data of ORCAS, and the negative instances are randomly selected from the top 100 documents of the corresponding query if users did not click on them).
Regarding Claim 17, Bo in view of Gu discloses the system of Claim 11, wherein a first total number of members in the first set of training digital objects and a second total number of members in the second set of training digital objects are the same (see Bo, Section 5.1.3, where for the click prediction, the positive instances are derived from the click data of ORCAS, and the negative instances are randomly selected from the top 100 documents of the corresponding query if users did not click on them).
Regarding Claim 18, Bo in view of Gu discloses the system of Claim 11, wherein a first total number of members in the first set of training digital objects and a second total number of members in the second set of training digital objects are pre-determined (see Bo, Section 5.1.3, where for the click prediction, the positive instances are derived from the click data of ORCAS, and the negative instances are randomly selected from the top 100 documents of the corresponding query if users did not click on them).
Claims 5, 6, and 15 are rejected under 35 U.S.C. 103 as being unpatentable over Bo and Gu as applied to Claims 1-4, 8-14, and 16-18 above, and further in view of Vorobev (PG Pub. No. 2017/0140053 A1).
Regarding Claim 5, Bo in view of Gu discloses the method of Claim 1, further comprising:
Bo does not disclose:
receiving, by the at least one processor, an in-use query;
retrieving, by the at least one processor, a set of in-use digital documents responsive to the in-use query;
applying, by the at least one processor, the MLA to each one of the set of in-use digital documents to generate respective likelihood parameters of the given user interacting therewith; and
using, by the at least one processor, the respective likelihood parameters for ranking each one of the set of in-use digital documents.
Vorobev discloses:
receiving, by the at least one processor, an in-use query (see Vorobev, Claim 1a, where the method comprises receiving, by the search engine server, a search query from an electronic device associated with a user);
retrieving, by the at least one processor, a set of in-use digital documents responsive to the in-use query (see Vorobev, Claim 1b, where the method comprises selecting, by a ranking algorithm of the search engine server, at least one relevant web resource to the search query);
applying, by the at least one processor, the MLA to each one of the set of in-use digital documents to generate respective likelihood parameters of the given user interacting therewith (see Vorobev, Claim 1d, where the method comprises applying a first machine learned algorithm to determine, for each of the candidate web resources of the plurality of candidate web resources, a predicted relevancy parameter, the predicted relevancy parameter being based, at least in part, on a respective web-resource-inherent data, the predicted relevancy parameter being indicative of a predicted relevancy of the respective candidate web resource to the search query); and
using, by the at least one processor, the respective likelihood parameters for ranking each one of the set of in-use digital documents (see Vorobev, Claim 1e, where the method comprises applying a second machine learning algorithm to determine, for each of the plurality of candidate web resources, an exploration score based at least in part on the respective predicted relevancy parameter, and inputting the determined exploration score of the plurality of candidate resources into a bandit-based ranking algorithm for i. ranking the plurality of candidate web resources).
Bo is directed to generating a BERT-based language model to ranking search results based on click logs (see Bo, Abstract). Vorobev is directed to using such a model to rank search results in response to a user query (see Vorobev, Abstract). Accordingly, it would have been obvious to one of ordinary skill in the art before the effective filing date of the invention to combine Bo with Vorobev as it amounts to combining prior art elements according to known techniques to yield predictable results (see MPEP 2143(I)(A)).
Regarding Claim 6, Bo in view of Gu and Vorobev discloses the method of Claim 5, wherein:
Bo dose not disclose using the respective likelihood parameters comprises feeding, by the at least one processor, the respective likelihood parameters as an input to an other MLA, the other MLA having been configured to rank the set of in-use digital documents based at least on the respective likelihood values of the given user interacting therewith. Vorobev discloses using the respective likelihood parameters comprises feeding, by the at least one processor, the respective likelihood parameters as an input to an other MLA, the other MLA having been configured to rank the set of in-use digital documents based at least on the respective likelihood values of the given user interacting therewith (see Vorobev, Claim 1e, where the method comprises applying a second machine learning algorithm to determine, for each of the plurality of candidate web resources, an exploration score based at least in part on the respective predicted relevancy parameter, and inputting the determined exploration score of the plurality of candidate resources into a bandit-based ranking algorithm for i. ranking the plurality of candidate web resources).
Bo is directed to generating a BERT-based language model to ranking search results based on click logs (see Bo, Abstract). Vorobev is directed to using such a model to rank search results in response to a user query (see Vorobev, Abstract). Accordingly, it would have been obvious to one of ordinary skill in the art before the effective filing date of the invention to combine Bo with Vorobev as it amounts to combining prior art elements according to known techniques to yield predictable results (see MPEP 2143(I)(A)).
Regarding Claim 5, Bo in view of Gu discloses the method of Claim 1, further comprising:
Bo does not disclose:
receiving an in-use query;
retrieving a set of in-use digital documents responsive to the in-use query;
applying the MLA to each one of the set of in-use digital documents to generate respective likelihood parameters of the given user interacting therewith; and
using the respective likelihood parameters for ranking each one of the set of in-use digital documents.
Vorobev discloses:
receiving an in-use query (see Vorobev, Claim 1a, where the method comprises receiving, by the search engine server, a search query from an electronic device associated with a user);
retrieving a set of in-use digital documents responsive to the in-use query (see Vorobev, Claim 1b, where the method comprises selecting, by a ranking algorithm of the search engine server, at least one relevant web resource to the search query);
applying the MLA to each one of the set of in-use digital documents to generate respective likelihood parameters of the given user interacting therewith (see Vorobev, Claim 1d, where the method comprises applying a first machine learned algorithm to determine, for each of the candidate web resources of the plurality of candidate web resources, a predicted relevancy parameter, the predicted relevancy parameter being based, at least in part, on a respective web-resource-inherent data, the predicted relevancy parameter being indicative of a predicted relevancy of the respective candidate web resource to the search query); and
using the respective likelihood parameters for ranking each one of the set of in-use digital documents (see Vorobev, Claim 1e, where the method comprises applying a second machine learning algorithm to determine, for each of the plurality of candidate web resources, an exploration score based at least in part on the respective predicted relevancy parameter, and inputting the determined exploration score of the plurality of candidate resources into a bandit-based ranking algorithm for i. ranking the plurality of candidate web resources).
Bo is directed to generating a BERT-based language model to ranking search results based on click logs (see Bo, Abstract). Vorobev is directed to using such a model to rank search results in response to a user query (see Vorobev, Abstract). Accordingly, it would have been obvious to one of ordinary skill in the art before the effective filing date of the invention to combine Bo with Vorobev as it amounts to combining prior art elements according to known techniques to yield predictable results (see MPEP 2143(I)(A)).
Bo is directed to generating a BERT-based language model to ranking search results based on click logs (see Bo, Abstract). Vorobev is directed to using such a model to rank search results in response to a user query (see Vorobev, Abstract). Accordingly, it would have been obvious to one of ordinary skill in the art before the effective filing date of the invention to combine Bo with Vorobev as it amounts to combining prior art elements according to known techniques to yield predictable results (see MPEP 2143(I)(A)).
Claim 7 is rejected under 35 U.S.C. 103 as being unpatentable over Bo, Gu, and Vorobev as applied to Claims 5, 6, and 15, above, and further in view of Zhu (PG Pub. No. 2019/0236132 A1).
Regarding Claim 7, Bo in view of Gu and Vorobev discloses the method of Claim 6, wherein:
Bo does not disclose the other MLA is an ensemble of CatBoost decision trees. Zhu discloses the other MLA is an ensemble of CatBoost decision trees (see Zhu, paragraph [0059], where text ranking algorithms 320 can include an ensemble of tree-based algorithms such as CatBoost).
Bo is directed to generating a BERT-based language model to ranking search results based on click logs, a form of relevance ranking (see Bo, Abstract). Zhu is also directed to relevance ranking (see Zhu, paragraph [0015]). Therefore, it would have been obvious to one of ordinary skill in the art before the effective filing date of the invention to combine Bo with Zhu as it amounts to combining prior art elements according to known techniques to yield predictable results (see MPEP 2143(I)(A)).
Conclusion
Any inquiry concerning this communication or earlier communications from the examiner should be directed to FARHAD AGHARAHIMI whose telephone number is (571)272-9864. The examiner can normally be reached M-F 9am - 5pm ET.
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, Apu Mofiz can be reached at 571-272-4080. 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.
/FARHAD AGHARAHIMI/Examiner, Art Unit 2161
/APU M MOFIZ/Supervisory Patent Examiner, Art Unit 2161
Prosecution Insights