DETAILED ACTION
This communication is responsive to Amendment filed 10/21/2025.
Claims 1-2 and 21-38 are pending in this application. In the Amendment, claims 1-2 are amended, claims 3-20 are cancelled and claims 21-38 are new. This action is made Final.
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 .
Response to Arguments
Applicant’s arguments with respect to claims amended 10/21/2025 have been considered but are moot because the new ground of rejection does not rely on any reference applied in the prior rejection of record for any teaching or matter specifically challenged in the argument.
Claim Rejections - 35 USC § 112
The following is a quotation of the first paragraph of 35 U.S.C. 112(a):
(a) IN GENERAL.—The specification shall contain a written description of the invention, and of the manner and process of making and using it, in such full, clear, concise, and exact terms as to enable any person skilled in the art to which it pertains, or with which it is most nearly connected, to make and use the same, and shall set forth the best mode contemplated by the inventor or joint inventor of carrying out the invention.
The following is a quotation of the first paragraph of pre-AIA 35 U.S.C. 112:
The specification shall contain a written description of the invention, and of the manner and process of making and using it, in such full, clear, concise, and exact terms as to enable any person skilled in the art to which it pertains, or with which it is most nearly connected, to make and use the same, and shall set forth the best mode contemplated by the inventor of carrying out his invention.
Claim 1 is rejected under 35 U.S.C. 112(a) or 35 U.S.C. 112 (pre-AIA ), first paragraph, as failing to comply with the written description requirement. The claim(s) contains subject matter which was not described in the specification in such a way as to reasonably convey to one skilled in the relevant art that the inventor or a joint inventor, or for applications subject to pre-AIA 35 U.S.C. 112, the inventor(s), at the time the application was filed, had possession of the claimed invention. Claim 1 recites “a multi-cluster subset of intents”. Although Applicant’s specification describes intent clusters (Applicant’s specification para.67-68), there is not mention of multi-clusters.
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, 21-22 and 25-27 are rejected under 35 U.S.C. 103 as being unpatentable over Sachindran et al. (“Sachindran”, US 2024/0086489) in view of Galitsky (US 11,580,144) and further in view of Yue et al. (“Yue”, US 2021/0192134).
As per claim 1, Sachindran teaches a system for generating cluster-based outputs from dual-pathway models, the system comprising:
one or more processors (Sachindran, para.38, Fig.1, user system 110 includes processors; para.106, Fig.6, processing device 602) and one or more non-transitory media comprising instructions recorded thereon that, when executed by the one or more processors, cause operations (Sachindran, para.40, Fig.1, data store 116; para.113, Fig.6, medium 642) comprising:
receiving user interaction data of a user interaction of a user with a user interface (Sachindran, Fig.5, operation 502, para.12, 17-19, 60, 75, 95, user typing first few characters of search in input box, context data);
after using a first model with the user interaction data to obtain a first output of the first model (Sachindran, Fig.4A-B, para.60, 68, 83-84, 89, search suggestions 408/458 based on first partial search input), invoking, based on the first output of the first model satisfying a pre-fetch threshold, a second model (Sachindran, para.62, partial search input and context data provided to predictive model) to determine a plurality of intent clusters corresponding to the user interaction of the user (Sachindran, para.18-20, partial search term with context data used to determine intended entity type clusters), wherein each intent cluster of the plurality of intent clusters groups semantically-related intents (Sachindran, para. 20-21, 25, 44, 46, 68, 83-84, 89, intent by entity type grouped into ranked clusters) and comprises a ranked set of intents (Sachindran, para.21, 36, 46, ranking model ranks entity type clusters corresponding to intent);
determining that a multi-cluster subset of intents, comprising a highest ranked intent from each intent cluster of the plurality of intent clusters, satisfies a confidence threshold (Sachindran, para.25, 60, 68-71, highest ranked subset matching intent); and
based on the multi-cluster subset of intents satisfying the confidence threshold, generating for display, on the user interface, a set of responses that correspond to the multi-cluster subset of intents, respectively (Sachindran, Fig.4A-B, para.60, 68-71, 83-84, 89, subset of suggestions generated in rank order based on first partial search input).
Although Sachindran teaches a threshold query length (Sachindran, para.62, 75-76, 96), the system does not teach invoking, based on the first output of the first deterministic model failing to satisfy an intent number threshold corresponding to a required number of intents, a second non-deterministic model. Galitsky teaches a method of generating outputs from a query that includes invoking, based on the first output of the first model failing to satisfy an intent number threshold corresponding to a required number of intents, a second model (Galitsky, col.2, lines 19-33, 53-64; col.17, line 53-col.18, line 10, when threshold number of matches not satisfied then further search performed on second index). It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to include Galitsky’s teaching with Sachindran’s system in order to locate sufficient responses.
Furthermore, the system of Sachindran and Galitsky does not explicitly teach generating outputs from deterministic/non-deterministic models. Yue teaches a system of intent completion wherein deterministic/non-deterministic models are used for intent completion (Yue, para.47, 51-52, 56, 65, 73). It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to include Yue’s teaching with the system of Sachindran and Galitsky to enable rapid customization and personalization (Yue, para.47).
As per claim 2, Sachindran teaches a method for generating cluster-based outputs from dual-pathway models, the method comprising:
receiving user data of a user interacting with a user interface (Sachindran, Fig.5, operation 502, para.12, 17-19, 60, 75, 95, user typing first few characters of search in input box, context data);
using a first deterministic model with the user data to obtain a first output of the first model (Sachindran, Fig.4A-B, para.60, 68, 83-84, 89, search suggestions 408/458 based on first partial search input);
invoking, based on the first output of the first model satisfying a pre-fetch thresholds, a second model (Sachindran, para.62, partial search input and context data provided to predictive model) to determine a plurality of intent clusters corresponding to the user data of the user (Sachindran, para.18-20, partial search term with context data used to determine intended entity type clusters), wherein each intent cluster of the plurality of intent clusters comprise a ranked set of intents (Sachindran, para. 20-21, 25, 44, 46, 68, 83-84, 89, intent by entity type grouped into ranked clusters, para.21, 36, 46, ranking model ranks entity type clusters corresponding to intent);
determining that a first intent subset comprising a highest ranked intent from each of two or more of the plurality of intent clusters, is greater than or equal to confidence threshold (Sachindran, para.25, 60, 68-71, highest ranked subset matching intent); and
based on determining that the first intent subset is greater than or equal to the confidence threshold, generating for display, on the user interface, a set of responses that correspond to the first intent subset, respectively (Sachindran, Fig.4A-B, para.60, 68, 83-84, 89, subset of suggestions generated in rank order based on first partial search input).
Although Sachindran teaches a threshold query length (Sachindran, para.62, 75-76, 96), the method does not teach invoking, based on the first output of the first deterministic model failing to satisfy one or more thresholds, a second non-deterministic model and determining a subset is greater than or equal to an intent number threshold corresponding to a required number of intents. Galitsky teaches a method of generating outputs from a query that includes invoking, based on the first output of the first model failing to satisfy one or more thresholds, a second model and determining a subset is greater than or equal to an intent number threshold corresponding to a required number of intents (Galitsky, col.2, lines 19-33, 53-64; col.17, line 53-col.18, line 10, when threshold number of matches not satisfied then further search performed on second index). It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to include Galitsky’s teaching with Sachindran’s method in order to locate sufficient responses.
Furthermore, the method of Sachindran and Galitsky does not explicitly teach generating outputs from deterministic/non-deterministic models. Yue teaches a method of intent completion wherein deterministic/non-deterministic models are used for intent completion (Yue, para.47, 51-52, 56, 65, 73). It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to include Yue’s teaching with the method of Sachindran and Galitsky to enable rapid customization and personalization (Yue, para.47).
As per claim 21, the method of Sachindran, Galitsky and Yue teaches the method of claim 2, wherein invoking the second non-deterministic model comprises, after using the first deterministic model with the user data to obtain an initial set of suggested inputs to a search input field of the user interface, invoking, based on the initial set of suggested inputs to the search input field failing to satisfy the one or more thresholds, the second non-deterministic model to obtain a plurality of suggested inputs to the search input field (Sachindran, Fig.4A-B, para.60, 68-71, 83-84, 89, subset of suggestions generated 408/458; Galitsky, col.2, lines 19-33, 53-64; col.17, line 53-col.18, line 10).
As per claim 22, the method of Sachindran, Galitsky and Yue teaches the method of claim 2, wherein invoking the second non-deterministic model comprises, after using the first deterministic model with the user data to obtain the first output of the first deterministic model, invoking, based on the first output of the first deterministic model failing to satisfy the intent number threshold corresponding to the required number of intents, the second non-deterministic model to determine the plurality of intent clusters corresponding to the user data of the user (Sachindran, para.18-20, partial search term with context data used to determine intended entity type clusters, Fig.4A-B, para.60, 68-71, 83-84, 89, subset of suggestions generated 408/458; Galitsky, col.2, lines 19-33, 53-64; col.17, line 53-col.18, line 10).
As per claim 25, the method of Sachindran, Galitsky and Yue teaches the method of claim 2, wherein invoking the second non-deterministic model comprises, after using the first deterministic model with the user data to obtain the first output of the first deterministic model, invoking, based on the first output of the first deterministic model failing to satisfy the one or more thresholds (Sachindran, para.18-20, Fig.4A-B, para.60, 68-71, 83-84, 89; Galitsky, col.2, lines 19-33, 53-64; col.17, line 53-col.18, line 10), a non-deterministic semantic autocomplete model to determine the plurality of intent clusters corresponding to the user data of the user, the non-deterministic semantic autocomplete model being configured to generate outputs based on predicted confidences (Sachindran, para.14, 64, 66, 70-71, predictive model 134 generates output results based on confidence value).
As per claim 26, the method of Sachindran, Galitsky and Yue teaches the method of claim 2, further comprising: determining that an earlier intent subset, comprising the highest ranked intent from each intent cluster of the plurality of intent clusters, is greater than a second intent number threshold (Yue, para.47, 51-52, 56, 61, 65, 73); and based on determining that the earlier intent subset is greater than the second intent number threshold, generating the first intent subset by filtering one or more intents from the earlier intent subset such that the first intent subset is greater than or equal to the intent number threshold and less than the second intent number threshold (Yue, para.47, 51-52, 56, 61, 65, 73, filter intent; Sachindran, para. 20-21, 25, 44, 46, 68, 83-84, 89, intent by entity type grouped into ranked clusters, para.21, 36, 46, ranking model ranks entity type clusters corresponding to intent; Galitsky, col.2, lines 19-33, 53-64; col.17, line 53-col.18, line 10).
As per claim 27, the method of Sachindran, Galitsky and Yue teaches the method of claim 2, further comprising: determining that an earlier intent subset, comprising the highest ranked intent from each intent cluster of the plurality of intent clusters without lower ranked intents from each intent cluster of the plurality of intent clusters, is less than the intent number threshold (Yue, para.47, 51-52, 56, 61, 65, 73); and based on determining that the earlier intent subset is less than the intent number threshold, generating the first intent subset to include the earlier intent subset and one or more of the lower ranked intents from the plurality of intent clusters (Yue, para.47, 51-52, 56, 61, 65, 73, filter intent; Sachindran, para. 20-21, 25, 44, 46, 68, 83-84, 89, intent by entity type grouped into ranked clusters, para.21, 36, 46, ranking model ranks entity type clusters corresponding to intent; Galitsky, col.2, lines 19-33, 53-64; col.17, line 53-col.18, line 10).
Claim 23 is rejected under 35 U.S.C. 103 as being unpatentable over Sachindran et al. (“Sachindran”, US 2024/0086489), Galitsky (US 11,580,144) and Yue et al. (“Yue”, US 2021/0192134) in view of Galimovich (US 2020/0401638).
As per claim 23, the method of Sachindran, Galitsky and Yue teaches the method of claim 2, wherein invoking the second non-deterministic model comprises, after using a deterministic word graph (Yue, para.47, 51-52, 56, 65, 73) of the first deterministic model with the user data to obtain the first output of the first deterministic model, invoking, based on the first output of the first deterministic model failing to satisfy the one or more thresholds, the second non-deterministic model to determine the plurality of intent clusters corresponding to the user data of the user (Sachindran, para.18-20, partial search term with context data used to determine intended entity type clusters, Fig.4A-B, para.60, 68-71, 83-84, 89, subset of suggestions generated 408/458; Galitsky, col.2, lines 19-33, 53-64; col.17, line 53-col.18, line 10).
However, the method of Sachindran, Galitsky and Yue does not explicitly teach the deterministic word graph being configured to generate outputs based on respective popularities of combinations of different text characters. Galimovich teaches a method of generating query completion suggestions wherein a first model comprises a deterministic word graph that generates outputs based on respective popularities of combinations of different text characters (Galimovich, para.92, 114-115, 125, popularity of determinative terms). It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to include Galimovich‘s teaching with the method of Sachindran, Galitsky and Yue in order to narrow the search query to common suggestions.
Claim 24 is rejected under 35 U.S.C. 103 as being unpatentable over Sachindran et al. (“Sachindran”, US 2024/0086489), Galitsky (US 11,580,144) and Yue et al. (“Yue”, US 2021/0192134) in view of Hamilton et al. (US 2024/0104091).
As per claim 24, the method of Sachindran, Galitsky and Yue teaches the method of claim 2, wherein invoking the second non-deterministic model comprises, after using a directed acyclic graph of the first deterministic model (Yue, para.47, 51-52, 56, 65, 73) with the user data to obtain the first output of the first deterministic model, invoking, based on the first output of the first deterministic model failing to satisfy the one or more thresholds, the second non- deterministic model to determine the plurality of intent clusters corresponding to the user data of the user (Sachindran, para.18-20, partial search term with context data used to determine intended entity type clusters, Fig.4A-B, para.60, 68-71, 83-84, 89, subset of suggestions generated 408/458; Galitsky, col.2, lines 19-33, 53-64; col.17, line 53-col.18, line 10).
However, the method of Sachindran, Galitsky and Yue does not explicitly teach using a directed acyclic graph with an initial vertex and a set of final vertices such that paths from the initial vertex to final vertices represent suffixes of a string based on the user data. Hamilton teaches a method of generating autocomplete predictions wherein the first model generates a directed acyclic graph with an initial vertex and a set of final vertices such that paths from the initial vertex to final vertices represent suffixes of a string based on the user data (Hamilton, para.74, general search result trie data object is a tree data structure with leaf node results). It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to include Hamilton‘s teaching with the method of Sachindran, Galitsky and Yue in order to narrow the search query to general suggestions.
Claim 28 is rejected under 35 U.S.C. 103 as being unpatentable over Sachindran et al. (“Sachindran”, US 2024/0086489), Galitsky (US 11,580,144) and Yue et al. (“Yue”, US 2021/0192134) in view of Mao et al. (“Mao”, US 2023/0092702).
As per claim 28, the method of Sachindran, Galitsky and Yue teaches the method of claim 2, however does not teach wherein the second non-deterministic model is trained to: minimize cosine distances between a first set of historical user inputs, wherein the first set of historical user inputs corresponds to a single intent; and maximize cosine distances between a second set of historical user inputs, wherein the second set of historical user inputs corresponds to different intents.
Mao teaches a method selecting suggested responses wherein a second model is trained to: minimize cosine distances between a first set of historical user inputs, wherein the first set of historical user inputs corresponds to a single intent; and maximize cosine distances between a second set of historical user inputs, wherein the second set of historical user inputs corresponds to different intents (Mao, para.25, 36, 40, 75, 87, 94, 102, 104, 107-108, 111, 122, 134, 140 clustering process 1000 assigns input into semantically similar clusters using cosine distances). It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to include Mao‘s teaching with the method of Sachindran, Galitsky and Yue in order to refine the search results to similar intentions.
Claims 29, 32-33 and 37 are rejected under 35 U.S.C. 103 as being unpatentable over Sachindran et al. (“Sachindran”, US 2024/0086489) in view of Galitsky (US 11,580,144).
As per claim 29, Sachindran teaches the one or more non-transitory computer-readable media comprising instructions that, when executed by one or more processors (Sachindran, para.38, Fig.1, user system 110 includes processors; para.106, Fig.6, processing device 602), cause operations comprising:
receiving user data of a user interacting with a user interface (Sachindran, Fig.5, operation 502, para.12, 17-19, 60, 75, 95, user typing first few characters of search in input box, context data);
using a first model with the user data to obtain a first output of the first model (Sachindran, Fig.4A-B, para.60, 68, 83-84, 89, search suggestions 408/458 based on first partial search input);
invoking, based on the first output of the first model satisfying one or more thresholds, a second model to obtain a first intent subset (Sachindran, para.62, partial search input and context data provided to predictive model based on threshold); and
generating one or more responses that correspond to one or more intents of the first intent subset (Sachindran, Fig.4A-B, para.60, 68-71, 83-84, 89, subset of suggestions generated in rank order based on first partial search input).
Although Sachindran teaches a threshold query length (Sachindran, para.62, 75-76, 96), the medium does not teach invoking, based on the first output of the first model failing to satisfy one or more thresholds, a second model. Galitsky teaches a method of generating outputs from a query that includes invoking, based on the first output of the first model failing to satisfy one or more thresholds, a second model (Galitsky, col.2, lines 19-33, 53-64; col.17, line 53-col.18, line 10, when threshold number of matches not satisfied then further search performed on second index). It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to include Galitsky’s teaching with Sachindran’s medium in order to locate sufficient responses.
As per claim 32, the medium of Sachindran and Galitsky teaches the one or more non-transitory computer-readable media of claim 29, wherein invoking the second model comprises, after using the first model with the user data to obtain an initial set of suggested inputs to a search input field of the user interface, invoking, based on the initial set of suggested inputs to the search input field failing to satisfy the one or more thresholds, the second model to obtain a plurality of suggested inputs to the search input field (Sachindran, Fig.4A-B, para.60, 68-71, 83-84, 89, subset of suggestions generated 408/458; Galitsky, col.2, lines 19-33, 53-64; col.17, line 53-col.18, line 10).
As per claim 33, the medium of Sachindran and Galitsky teaches the one or more non-transitory computer-readable media of claim 29, wherein invoking the second model comprises invoking, based on the first output of the first model failing to satisfy an intent number threshold corresponding to a required number of intents, the second model to obtain the first intent subset (Sachindran, para.18-20, partial search term with context data used to determine intended entity type clusters, Fig.4A-B, para.60, 68-71, 83-84, 89, subset of suggestions generated 408/458; Galitsky, col.2, lines 19-33, 53-64; col.17, line 53-col.18, line 10).
As per claim 37, the medium of Sachindran and Galitsky teaches the one or more non-transitory computer-readable media of claim 29, wherein invoking the second model comprises invoking, based on the first output of the first model failing to satisfy the one or more thresholds (Sachindran, para.18-20, Fig.4A-B, para.60, 68-71, 83-84, 89; Galitsky, col.2, lines 19-33, 53-64; col.17, line 53-col.18, line 10), a semantic autocomplete model to obtain the first intent subset, the semantic autocomplete model being configured to generate outputs based on predicted confidences (Sachindran, para.14, 64, 66, 70-71, predictive model 134 generates output results based on confidence value).
Claims 30-31 and 34 are rejected under 35 U.S.C. 103 as being unpatentable over Sachindran et al. (“Sachindran”, US 2024/0086489) and Galitsky (US 11,580,144) in view of Yue et al. (“Yue”, US 2021/0192134).
As per claim 30, the medium of Sachindran and Galitsky teaches the one or more non-transitory computer-readable media of claim 29, wherein invoking the second model comprises, after using the first model with the user data to obtain the first output of the first model, invoking, based on the first output of the first model failing to satisfy the one or more thresholds, a second model to obtain the first intent subset (Galitsky, col.2, lines 19-33, 53-64; col.17, line 53-col.18, line 10).
However, the medium of Sachindran and Galitsky does not explicitly teach generating outputs from a non-deterministic model. Yue teaches a medium of intent completion wherein deterministic/non-deterministic models are used for intent completion (Yue, para.47, 51-52, 56, 65, 73). It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to include Yue’s teaching with the medium of Sachindran and Galitsky to enable rapid customization and personalization (Yue, para.47).
As per claim 31, the medium of Sachindran and Galitsky teaches the one or more non-transitory computer-readable media of claim 29, wherein invoking the second model comprises, after using a first model with the user data to obtain the first output of the first model, invoking, based on the first output of the first model failing to satisfy the one or more thresholds, a second model to obtain the first intent subset (Galitsky, col.2, lines 19-33, 53-64; col.17, line 53-col.18, line 10).
However, the medium of Sachindran and Galitsky does not explicitly teach generating outputs from deterministic/non-deterministic models. Yue teaches a medium of intent completion wherein deterministic/non-deterministic models are used for intent completion (Yue, para.47, 51-52, 56, 65, 73). It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to include Yue’s teaching with the medium of Sachindran and Galitsky to enable rapid customization and personalization (Yue, para.47).
As per claim 34, the medium of Sachindran and Galitsky teaches the one or more non-transitory computer-readable media of claim 29, wherein invoking the second model comprises invoking, based on the first output of the first model failing to satisfy an intent number threshold, the second model to obtain the first intent subset (Galitsky, col.2, lines 19-33, 53-64; col.17, line 53-col.18, line 10).
However, the medium of Sachindran and Galitsky does not teach a threshold derived from one or more interface display parameters associated with the user interface. Yue teaches a medium of intent completion wherein intent thresholds are based on the display (Yue, para.56, threshold to fit screen). It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to include Yue’s teaching with the medium of Sachindran and Galitsky to maintain simplicity.
Claim 35 is rejected under 35 U.S.C. 103 as being unpatentable over Sachindran et al. (“Sachindran”, US 2024/0086489) and Galitsky (US 11,580,144) in view of Yue et al. (“Yue”, US 2021/0192134) and further in view of Galimovich (US 2020/0401638).
As per claim 35, the medium of Sachindran and Galitsky teaches the one or more non-transitory computer-readable media of claim 29, wherein invoking the second model comprises, using the first model with the user data to obtain the first output of the first model, invoking, based on the first output of the first model failing to satisfy the one or more thresholds, the second model to obtain the first intent subset (Sachindran, para.18-20, partial search term with context data used to determine intended entity type clusters, Fig.4A-B, para.60, 68-71, 83-84, 89, subset of suggestions generated 408/458; Galitsky, col.2, lines 19-33, 53-64; col.17, line 53-col.18, line 10).
However, the medium of Sachindran and Galitsky does not explicitly teach generating outputs using a deterministic word graph. Yue teaches a medium of intent completion using a deterministic word graph for intent completion (Yue, para.47, 51-52, 56, 65, 73). It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to include Yue’s teaching with the medium of Sachindran and Galitsky to enable rapid customization and personalization (Yue, para.47).
Furthermore, the medium of Sachindran, Galitsky and Yue does not explicitly teach the deterministic word graph being configured to generate outputs based on respective popularities of combinations of different text characters. Galimovich teaches a medium of generating query completion suggestions wherein a first model comprises a deterministic word graph that generates outputs based on respective popularities of combinations of different text characters (Galimovich, para.92, 114-115, 125, popularity of determinative terms). It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to include Galimovich‘s teaching with the medium of Sachindran, Galitsky and Yue in order to narrow the search query to common suggestions.
Claim 36 is rejected under 35 U.S.C. 103 as being unpatentable over Sachindran et al. (“Sachindran”, US 2024/0086489) and Galitsky (US 11,580,144) in view of Yue et al. (“Yue”, US 2021/0192134) and further in view of Hamilton et al. (US 2024/0104091).
As per claim 36, the medium of Sachindran and Galitsky teaches the one or more non-transitory computer-readable media of claim 29, wherein invoking the second model comprises, using the first model with the user data to obtain the first output of the first model, invoking, based on the first output of the first model failing to satisfy the one or more thresholds, the second model to obtain the first intent subset.
However, the medium of Sachindran and Galitsky does not explicitly teach generating outputs using a directed acyclic graph. Yue teaches a medium of intent completion using a directed acyclic graph for intent completion (Yue, para.47, 51-52, 56, 65, 73). It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to include Yue’s teaching with the medium of Sachindran and Galitsky to enable rapid customization and personalization (Yue, para.47).
However, the medium of Sachindran, Galitsky and Yue does not explicitly teach using a directed acyclic graph with an initial vertex and a set of final vertices such that paths from the initial vertex to final vertices represent suffixes of a string based on the user data. Hamilton teaches a medium of generating autocomplete predictions wherein the first model generates a directed acyclic graph with an initial vertex and a set of final vertices such that paths from the initial vertex to final vertices represent suffixes of a string based on the user data (Hamilton, para.74, general search result trie data object is a tree data structure with leaf node results). It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to include Hamilton‘s teaching with the medium of Sachindran, Galitsky and Yue in order to narrow the search query to general suggestions.
Claim 38 is rejected under 35 U.S.C. 103 as being unpatentable over Sachindran et al. (“Sachindran”, US 2024/0086489) and Galitsky (US 11,580,144) in view of Mao et al. (“Mao”, US 2023/0092702).
As per claim 38, the medium of Sachindran and Galitsky teaches the one or more non-transitory computer-readable media of claim 29, however does not teach wherein the second model is trained to: minimize cosine distances between a first set of historical user inputs, wherein the first set of historical user inputs corresponds to a single intent; and maximize cosine distances cosine distances between a second set of historical user inputs, wherein the second set of historical user inputs corresponds to different intents.
Mao teaches a medium selecting suggested responses wherein a second model is trained to: minimize cosine distances between a first set of historical user inputs, wherein the first set of historical user inputs corresponds to a single intent; and maximize cosine distances cosine distances between a second set of historical user inputs, wherein the second set of historical user inputs corresponds to different intents (Mao, para.25, 36, 40, 75, 87, 94, 102, 104, 107-108, 111, 122, 134, 140 clustering process 1000 assigns input into semantically similar clusters using cosine distances). It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to include Mao‘s teaching with the medium of Sachindran and Galitsky in order to refine the search results to similar intentions.
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.
Inquiries
Any inquiry concerning this communication or earlier communications from the examiner should be directed to SAJEDA MUHEBBULLAH whose telephone number is (571)272-4065. The examiner can normally be reached Mon-Tue/Thur-Fri 10am-8pm.
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, William L Bashore can be reached at 571-272-4088. 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.
/S.M./
Sajeda MuhebbullahExaminer, Art Unit 2174
/WILLIAM L BASHORE/ Supervisory Patent Examiner, Art Unit 2174