DETAILED ACTION
Introduction
This office action is in response to applicant’s request for continued examination filed 6/17/2026. Claims 1-20 are currently pending and have been examined. Applicant’s IDS have been considered. There is no claim to foreign priority.
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 .
Continued Examination Under 37 CFR 1.114
A request for continued examination under 37 CFR 1.114, including the fee set forth in 37 CFR 1.17(e), was filed in this application after final rejection. Since this application is eligible for continued examination under 37 CFR 1.114, and the fee set forth in 37 CFR 1.17(e) has been timely paid, the finality of the previous Office action has been withdrawn pursuant to 37 CFR 1.114. Applicant's submission filed on 6/17/2026 has been entered.
Response to Arguments
Applicant’s arguments, see remarks, filed 6/17/26, with respect to the rejection(s) of claim(s) 1-20 under 35 USC 103 have been fully considered and are not persuasive.
Applicant’ argues, “The portions of Gadde cited in the Office Action are directed to entity-level data augmentation, such as generating artificial utterances to augment training data. The cited portions of Gadde do not teach or suggest: resolving a value span corresponding to a token into a standard format via a first machine learning model; generating a processed natural language utterance by combining the utterance with the class label, the standard format, and database schema information, and utilizing a second machine learning model configured to predict a meaning representation by encoding the processed natural language utterance and decoding the encoded processed natural language utterance based on a grammar-based generative process.”
The Examiner notes, the above argued limitations are addressed below with respect to the previously cited prior art. The Examiner further notes, Gadde explicitly teaches wherein the first machine learning model is configured to resolve a value span corresponding to the token into a standard format for the named entity category (paragraphs [0135-0158]-for example, his currency, and his other classes of entities and combinations thereof, as a named entity category, and the list of values to cover for a given entity, as a value span, which is resolved via the chatbot system, including the machine learning model, identified and resolved into his standard format, see format for currency, etc.). The Examiner notes the importance of resolving the value span corresponding to the input token, see the rejection below regarding tokenization of the input for machine learning model processing, due to the fact that the value of an entity may comprise a word or a string of words and/or other variations, and this aspect of the classification is resolved directly by converting any corresponding value spans into a standard format, which is attributed to a corresponding class label. The additional limitations, which involve a model processing relevant data, to generate a result is clearly articulated below, wherein the utterance, the standard format, and database schema information, input into a machine learning model, utilizes these relevant elements to generate a response (see rejection below). Therefore, the applicant’s corresponding arguments and arguments to claims that depend on the above arguments are deemed non-persuasive.
The claim objections to claims 7 and 15 have been withdrawn.
Claim Rejections - 35 USC § 103
In the event the determination of the status of the application as subject to AIA 35 U.S.C. 102 and 103 (or as subject to pre-AIA 35 U.S.C. 102 and 103) is incorrect, any correction of the statutory basis (i.e., changing from AIA to pre-AIA ) for the rejection will not be considered a new ground of rejection if the prior art relied upon, and the rationale supporting the rejection, would be the same under either status.
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 1-20 are rejected under 35 U.S.C. 103 as being unpatentable over Gadde et al. (Gadde, US 2021/0390951) in view of Quamar et al. (Quamar, Natural Language Interfaces to Data) in view of Fujimoto et al. (Fujimoto, Us 2020/0402509), and further in view of Naganathan et al. (Naganathan, US 11,550,786).
As per claim 1, Gadde teaches a method comprising:
accessing a natural language utterance comprising a plurality of tokens (Fig. 2, his input utterance, paragraph [0088-0090, 0100]-as his utterance and natural language utterance, and units in the utterance are tokenized);
using a first machine learning model to predict a class label for a token of the plurality of tokens, wherein the class label represents that the token corresponds to a named entity category of a plurality of named entity categories, (ibid, paragraphs [0091, 0030, 0046, 0116, 0117]-his named entity recognizer, and corresponding predicted entities thereof, his predicted entities to include date/time, entities, see his prediction model discussion, ML model), and wherein the first machine learning model is configured to resolve a value span corresponding to the token into a standard format for the named entity category (paragraphs [0135-0158]-for example, his currency, and his other classes of entities and combinations thereof, as a named entity category, and the list of values to cover for a given entity, as a value span, which is resolved via the chatbot system, including the machine learning model, identified and resolved into his standard format, see format for currency, etc.);
processing the natural language utterance to generate a processed natural language utterance, wherein processing the natural language utterance comprises combining the natural language utterance with the class label, the standard format, and database [schema] information (paragraph [0161-0165])-his generated artificial utterance with respect to the template, including class labels, which include the currency, standard format for currencies, as described above, and his entity and database configuration mapping);
providing the processed natural language utterance to a second machine learning model;]
[using the second machine learning model] to predict a meaning representation for the processed natural language utterance, wherein the meaning representation comprises a value associated with the class label and an operator (ibid, paragraphs [0116, 0117, 0159], Fig. 4 items 425 and 410-his predicted intent, template and artificial utterance, the generated template and utterance as included a meaning, as a logical form representation of the utterance, and including a value associated with the class label (date_time) and an operator (on|at),
[and wherein the second machine learning model is configured to predict the meaning representation by encoding the processed natural language utterance and decoding the encoded processed natural language utterance based on a grammar-based generative process;]
[detecting that the value matches a predetermined value type or that the operator matches a predetermined operator; and
in response to detecting that the value matches the predetermined value type or that the operator matches the predetermined operator, modifying at least one of the value and the operator and generating an executable statement for the meaning representation, wherein the executable statement for the meaning representation comprises a modified version of at least one of the value and the operator].
Gadde lacks explicitly teaching that which Quamar teaches, processing the natural language utterance to generate a processed natural language utterance, wherein processing the natural language utterance comprises combining the natural language utterance with the class label, [the standard format], and database schema information (pages 36-44, section 3.2, Fig. 3.4, page 37, see his concatenation discussion, including the utterance text query tokens, consisting of entity mappings and schema elements, his encoder as the first machine learning model);
providing the processed natural language utterance to a second machine learning model (ibid-Fig. 3.4-his passed information to his encoder/decoder, his encoder/decoder as the second machine learning model);
using the second machine learning model to predict a meaning representation for the processed natural language utterance, wherein the meaning representation comprises a value associated with the class label and an operator (ibid, see also Fig. 3.4-his output from his decoder, see also SQL query discussion), and wherein the second machine learning model is configured to predict the meaning representation by encoding the processed natural language utterance and decoding the encoded processed natural language utterance based on a grammar-based generative process (ibid-pages 36-44, section 3.2-his natural language input representation, natural language query tokens, encoded, and decoded, page 38, “where the model learns to generate the output” as a sequence of tokens, and corresponding “grammar-based rules are applied to sequentially generate” the output).
Thus, it would have been obvious to one of ordinary skill in the linguistics art, before the effective filing date of the invention, as all the claimed elements were known in the prior art and one skilled in the art could have combined the elements as claimed by known methods (computer implemented techniques and algorithms combining processes and steps in natural language processing), in view of the teachings of Gadde and Quamar to combine the prior art element of a user natural language user request/utterance with date/time entities, that are class labeled, with resolved values covering a span, converted into a logical and standardized form, using predictive machine learning techniques, as taught by Gadde with the enhanced RATSQL, including a concatenation of utterance text, entities and schema to be input into a second machine learning model to generate a meaning representation as taught by Quamar as each element performs the same function as it does separately, as the combination would yield predictable results, KSR International Co. v. Teleflex Inc., 550 US. -- 82 USPQ2nd 1385 (2007), wherein the predictable result would be generating a logical form or structured query, such as SQL, for a database (ibid, Quamar, see also Quamar abstract NLID discussion).
Gadde lacks explicitly teaching that which Fujimoto teaches,
detecting that the value matches a predetermined value type or that the operator matches a predetermined operator (Fujimoto, paragraph [0027-0041]-as his value type that is associated with a temporal ambiguity, time information types, from his time extraction unit, and corresponding entity categories, wherein the categories are associated with words in his sentence, including date and time, see his ISO8601, his representation of dates and times, his predetermined value types including, “time of the year, month, day, hour minute and second” as ambiguous, when matched, they are processed to disambiguate or resolve the ambiguities, see also paragraphs [0037-0039, 0025, 0033]);
in response to detecting that the value matches the predetermined value type or that the operator matches the predetermined operator (ibid-based on the input sentence and matching values), modifying at least one of the value and the operator; and generating an executable statement for the meaning representation (ibid-the time value information, is modified, based on “nearest past” or “nearest future time”), wherein the executable statement for the meaning representation comprises a modified version of at least one of the value and the operator (ibid-the new time value, generated, and included in a execution of the task command, paragraphs [0045)).
Thus, it would have been obvious to one of ordinary skill in the linguistics art, before the effective filing date of the invention, as all the claimed elements were known in the prior art and one skilled in the art could have combined the elements as claimed by known methods (computer implemented techniques and algorithms combining processes and steps in natural language processing), in view of the teachings of Gadde and Fujimoto to combine the prior art element of a user natural language user request/utterance with date/time entities converted into a logical form, using predictive machine learning techniques, as taught by Gadde with resolving the ambiguities of the date/time entities by modifying the executable command for the task as taught by Fujimoto as each element performs the same function as it does separately, as the combination would yield predictable results, KSR International Co. v. Teleflex Inc., 550 US. -- 82 USPQ2nd 1385 (2007), wherein the predictable result would be resolving date/time, based ambiguities in user utterances, wherein the natural language used for execution and having an updated instruction includes matching values and/or operands (ibid, Fujimoto, paragraph [0004], see Naganathan-abstract).
Gadde with Fujimoto lack explicitly teaching that which Naganathan teaches, in response to detecting that the value matches the predetermined value type or that the operator matches the predetermined operator (C.11 lines 23-34-his “over” operator, matching his predetermined “greater than” operator); wherein the executable statement for the meaning representation comprises a modified version of at least one of the value and the operator (ibid, C.11 lines 23-34, C.2 lines 44-53, as the executable and modified structured query statement generated).
Thus, it would have been obvious to one of ordinary skill in the linguistics art, before the effective filing date of the invention, as all the claimed elements were known in the prior art and one skilled in the art could have combined the elements as claimed by known methods (computer implemented techniques and algorithms combining processes and steps in natural language processing), in view of the teachings of Gadde and Fujimoto to combine the prior art element of a user natural language user request/utterance with date/time entities converted into a logical form, using predictive machine learning techniques, as taught by Gadde with resolving the ambiguities of the date/time entities by modifying the executable command for the task as taught by Fujimoto with detecting an operator matches a predetermined operator as taught by Naganathan as each element performs the same function as it does separately, as the combination would yield predictable results, KSR International Co. v. Teleflex Inc., 550 US. -- 82 USPQ2nd 1385 (2007), wherein the predictable result would be resolving date/time, based ambiguities in user utterances, wherein the natural language used for execution and having an updated instruction includes matching values and/or operands (ibid, Fujimoto, paragraph [0004], see Naganathan-abstract).
As per claims 2, 10 and 18, Gadde further makes obvious the method of claim 1, wherein the plurality of entity categories comprises at least one of a date category that is representative of a date entity in an utterance, a time category that is representative of a time entity in an utterance, and a datetime category that is representative of a date entity and a time entity in an utterance (ibid-see above entity categories, paragraphs [0030, 0091, 0116]-see date, time and datetime categories).
As per claims 3 and 11, Gadde with Quamar with Fujimoto with Naganathan further makes obvious the method of claim 1, wherein the predetermined value type corresponds to an entity category of the plurality of entity categories (ibid—Fujimoto, paragraph [0027-0041]-as his value type that is associated with a temporal ambiguity, time information types, from his time extraction unit, and corresponding entity categories, wherein the categories are associated with words in his sentence, including date and time, see his ISO8601, his representation of dates and times, his predetermined value types including, “time of the year, month, day, hour minute and second” as ambiguous, and not known, the Examiner notes, Fujimoto is similarly motivated and combined with Gadde and Naganathan, as seen in claim 1).
Thus, it would have been obvious to one of ordinary skill in the linguistics art, before the effective filing date of the invention, as all the claimed elements were known in the prior art and one skilled in the art could have combined the elements as claimed by known methods (computer implemented techniques and algorithms combining processes and steps in natural language processing), in view of the teachings of Gadde and Fujimoto to combine the prior art element of a user natural language user request/utterance with date/time entities converted into a logical form, using predictive machine learning techniques, as taught by Gadde with the predetermined value types for ambiguities associated with date/time as taught by Fujimoto as each element performs the same function as it does separately, as the combination would yield predictable results, KSR International Co. v. Teleflex Inc., 550 US. -- 82 USPQ2nd 1385 (2007), wherein the predictable result would be resolving date/time, based ambiguities in user utterances, having predetermined value types associated with entity categories, Naganathan value types associated with entity categories (C.5 lines 62-64-his date/time entities as tagged), in order to select and resolve the date/time values, wherein the natural language used for execution and having an updated instruction includes matching values and/or operands (ibid, Fujimoto, paragraph [0004], see Naganathan-abstract).
As per claims 4 and 12, Gadde with Quamar with Fujimoto with Naganathan further makes obvious the method of claim 1, wherein the predetermined operator corresponds to an operator for selecting a duration of time (Naganathan, C.5 lined 57-64, as operators as defining range(s) of time, see also C.11 lines 23-34-his “over” operator, matching his predetermined “greater than” operator, and “between” operator, as defining range(s) of time).
As per claims 5, 13 and 19, Gadde with Quamar with Fujimoto with Naganathan further makes obvious the method of claim 1, further comprising:
detecting that the value matches the predetermined value type (ibid-see claim 1, corresponding and similar limitation); and
modifying the value, wherein modifying the value comprises modifying a date based associated with the natural language utterance based on preference information included in database schema information (ibid-see claim 1, Fujimoto modification discussion, wherein the time value information, is modified, based on “nearest past” or “nearest future time”, which is preference information included in database schema information, paragraphs [0022-0044]-his time units, as including database schema information, Figs. 1 and 2, the new time value, generated, and included in a execution of the task command, paragraphs [0045], as similarly motivated and combined as seen in claim 1).
As per claims 6, 14 and 20, Gadde with Quamar with Fujimoto with Naganathan further makes obvious the method of claim 1, further comprising:
detecting that the operator matches the predetermined operator (ibid-see claim 1, corresponding and similar limitation); and
modifying the operator, wherein modifying the operator comprises replacing the operator with another operator (Naganathan, C.11 lines 23-34-his “over” operator, matching and replaced by his predetermined “greater than” operator, as similarly motivated and combined).
As per claims 7 and 15, Gadde with Quamar with Fujimoto with Naganathan further makes obvious the method of claim 1, wherein the second machine learning model is a trained machine learning model that was trained with training data comprised of a plurality of training meaning representations, [wherein each training meaning representation of the plurality of training meaning representations comprises an operator] (ibid-see claim 1, Quamar, decoder as the second machine learning model, text-to-SQL training including operators-as similarly combined and motivated, Gadde, machine learning discussion, paragraphs [0091, 0102, 0111, 0115, 0116-0123, 0159 ]-his training algorithms, machine learning, using all labeled data, and prediction based intent/meaning classification, wherein the labeled data includes an operator, as seen in paragraph [0159, 0161]-Fig. 4 items 425 and 410).
Gadde with Quamar lack explicitly teaching that which Naganathan teaches, wherein each training meaning representation of the plurality of training meaning representations comprises an operator (C.18 line 49-C.19 line 6-his intent/meaning and corresponding training based on the query and operand, C.7 line 57-C.8 line 11-for each query, corresponding operator is created, thus intent/meaning training comprises the operator in training).
Thus, it would have been obvious to one of ordinary skill in the linguistics art, before the effective filing date of the invention, as all the claimed elements were known in the prior art and one skilled in the art could have combined the elements as claimed by known methods (computer implemented techniques and algorithms combining processes and steps in natural language processing), in view of the teachings of Gadde and Fujimoto to combine the prior art element of a user natural language user request/utterance with date/time entities converted into a logical form, using predictive machine learning techniques, as taught by Gadde with resolving the ambiguities of the date/time entities by modifying the executable command for the task as taught by Fujimoto with training data with respect to meaning representations which include an operator as taught by Naganathan as each element performs the same function as it does separately, as the combination would yield predictable results, KSR International Co. v. Teleflex Inc., 550 US. -- 82 USPQ2nd 1385 (2007), wherein the predictable result would be resolving date/time, based ambiguities in user utterances, wherein the natural language used for execution and having an updated instruction includes matching values and/or operands (ibid, Fujimoto, paragraph [0004], see Naganathan-abstract, and previously cited sections).
As per claims 8 and 16, Gadde with Quamar with Fujimoto with Naganathan further makes obvious the method of claim 1, further comprising: executing a query on a computing system based on the executable statement (ibid-Gadde, paragraph [0028-0031, 0041]-his user requests/statements, and corresponding action by the digital assistant, to carry out the user’s intent).
As per claim 9, claim 9 sets forth limitations similar to claim 1 and is thus rejected under similar reasons and rationale, wherein the system is deemed to embody the system, such that Gadde with Quamar with Fujimoto with Naganathan make obvious a system comprising: one or more processors (Gadde, paragraphs [0013, 0014]-see his processors and non-transitory computer readable storage medium and instructions); and one or more computer-readable media storing instructions which, when executed by the one or more processors, cause the system to perform operations comprising (ibid): accessing a natural language utterance comprising a plurality of tokens (ibid-see claim 1, corresponding and similar limitation); using a first machine learning model to predict a class label for a token of the plurality of tokens, wherein the class label represents that the token corresponds to a named entity category of a plurality of named entity categories (ibid), and wherein the first machine learning model is configured to resolve a value span corresponding to the token into a standard format for the named entity category (ibid); processing the natural language utterance to generate a processed natural language utterance, wherein processing the natural language utterance comprises combining the natural language utterance with the class label, the standard format, and database schema information (ibid); providing the processed natural language utterance to a second machine learning model (ibid); using the second machine learning model to predict a meaning representation for the processed natural language utterance (ibid), wherein the meaning representation comprises a value associated with the class label and an operator (ibid), and wherein the second machine learning model is configured to predict the meaning representation by encoding the processed natural language utterance and decoding the encoded processed natural language utterance based on a grammar-based generative process (ibid); detecting that the value matches a predetermined value type or that the operator matches a predetermined operator (ibid); in response to detecting that the value matches the predetermined value type or that the operator matches the predetermined operator, modifying at least one of the value and the operator; and generating an executable statement for the meaning representation, wherein the executable statement for the meaning representation comprises a modified version of at least one of the value and the operator (ibid).
As per claim 17, claim 17 sets forth limitations similar to claim 1 and is thus rejected under similar reasons and rationale, wherein the non-transitory computer-readable media storing instructions is deemed to embody the method, such that Gadde with Quamar with Fujimoto with Naganathan make obvious a one or more non-transitory computer-readable media storing instructions which, when executed by one or more processors, cause a system to perform a method comprising (Gadde, paragraphs [0013, 0014]-see his processors and non-transitory computer readable storage medium and instructions): accessing a natural language utterance comprising a plurality of tokens (ibid-see claim 1, corresponding and similar limitation);
using a first machine learning model to predict a class label for a token of the plurality of tokens, wherein the class label represents that the token corresponds to a named entity category of a plurality of named entity categories (ibid), and wherein the first machine learning model is configured to resolve a value span corresponding to the token into a standard format for the named entity category (ibid); processing the natural language utterance to generate a processed natural language utterance, wherein processing the natural language utterance comprises combining the natural language utterance with the class label, the standard format, and database schema information (ibid); providing the processed natural language utterance to a second machine learning model (ibid); using the second machine learning model to predict a meaning representation for the processed natural language utterance, wherein the meaning representation comprises a value associated with the class label and an operator (ibid); and wherein the second machine learning model is configured to predict the meaning representation by encoding the processed natural language utterance and decoding the encoded processed natural language utterance based on a grammar-based generative process (ibid); detecting that the value matches a predetermined value type or that the operator matches a predetermined operator (ibid); in response to detecting that the value matches the predetermined value type or that the operator matches the predetermined operator, modifying at least one of the value and the operator (ibid); and generating an executable statement for the meaning representation, wherein the executable statement for the meaning representation comprises a modified version of at least one of the value and the operator (ibid).
Conclusion
The prior art made of record and not relied upon is considered pertinent to applicant's disclosure (See PTO-892).
Poirel et al. (Poirel, US 2019/0325061) teaches resolving a value span corresponding to a token into a standard format for a named entity.
Any inquiry concerning this communication or earlier communications from the examiner should be directed to LAMONT M SPOONER whose telephone number is (571)272-7613. The examiner can normally be reached 8:00 AM -5:00 PM.
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, Daniel Washburn can be reached at (571)272-5551. 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.
/LAMONT M SPOONER/ Primary Examiner, Art Unit 2657
9/18/2026
Prosecution Insights