DETAILED ACTION
Notice of Pre-AIA or AIA Status
1. 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
2. Applicant's arguments received 01/20/2026 have been considered but are moot in view of the new ground(s) of rejection. Detailed response is given in sections 3-4 as set forth below in this Office Action.
Regarding the claim eligibility, Applicant argues that (REMARKS, p.2):
PNG
media_image1.png
843
643
media_image1.png
Greyscale
Examiner respectfully disagrees. Applicant is advised that, according to MPEP 2106 and the 2019 Revised Patent Subject Matter Eligibility Guidance (2019 PEG), the Office determines claim eligibility under 35 U.S.C. § 101 using the Alice framework. The analysis under Step 2A - Prong 1 evaluates whether the claim recites a judicial exception. Step 2A - Prong 2 asks if the claim recites additional elements that integrate the judicial exception into a practical application, and, if necessary, Step 2B further analyzes whether or not the claim provides an Inventive Concept. That is, the claim needs to be analyzed limitation by limitation, and/or element by element, following the MPEP/2019 PEG guidelines.
In the instant case, focusing on what the inventors have invented exactly and giving the broadest reasonable interpretation (BRI) to the claims, Examiner asserts that the pending claims 1-20 are directed to an abstract idea of multi-stage classification of queries using AI model but without reciting any additional element that amounts to “significantly more” than the judicial exception. Specifically, when applying the two-step Alice test to analyze the subject matter eligibility, the Examiner separates the claim limitations which constitute an abstract idea (e.g., the bolded portion of representative claim 8) from “additional elements” that might be sufficient to amount to significantly more than the judicial exceptions to integrate the abstract idea into a practical application and/or to add an "inventive concept”. Under the BRI to the claim, the Examiner asserts that each or the combination of the limitations in the bolded portion of representative claim 8 encompasses mental processes of data manipulation/ evaluation/judgment and mathematical concepts and/or calculations, all of which can be performed in the human mind and/or with the aid of pen and paper. As such, the bolded portion of instant claim 8 falls within a combination of the “Mathematical Concepts” and “Mental Process” Groupings of Abstract Ideas defined by the 2019 PEG. Further, the Examiner maintains that none of the claimed additional elements is qualified for meaningful limitations to integrate the identified judicial exception into a practical application or provides an inventive concept when the claim is considered as a whole (see detailed analyses set forth in section 4 below).
It is held that simply setting forth advantages (i.e. benefits) of use without providing any rational/evidence to how/why the claimed elements amount to significantly more than the judicial exception could be treated as mere instructions to apply the judicial exception on a computer component (MPEP 2106.05(f)), but not qualified for an improvement (i.e. enhancement) in the functioning of a computer or an improvement to another technology or technical field. The key is to show that the claim goes beyond just performing a calculation and provides a practical application or significant improvement through the use of that calculation. See MPEP 2106.04(d)(I) and 2106.05(a).
Regarding Applicant’s arguments that “the technology stores a record of the relevant regulatory compliances respected by the models and/or AI agents that receive the query, enabling compliance auditing and enforcement” (REMARKS, p.2), the Examiner considers that the claimed step of “storing an audit data entry …” is recited at a high level of generality which is nothing more than conventional technology of storing audit/reference data entries in commonly known storage media on a general-purpose computer. The claim does not recite or provide, for example, such details as how the audit data are acquired, where the audit data entry is stored, and what structure the recited audit data structure applies exactly to organize and store the audit data, as opposed to conventional methods of storing audit/reference data entries (see Applicant’s Spec., para. 0078, 0120). Accordingly, the newly added limitations to the audit data do not integrate the identified judicial exception into a practical application or reflect a qualified improvement. Instead, the claim would monopolize the judicial exception across a wide range of applications.
The rest of the Applicant’s arguments regarding the claim eligibility are reliant upon the issues discussed above or have been fully addressed by the analyses as set forth in section 4 below in this Office Action, thus are deemed unpersuasive as well.
Claim Rejections - 35 USC § 101
3. The following is a quotation of the appropriate paragraphs of 35 U.S.C. 101 that form the basis for the rejections under this section made in this Office action:
35 U.S.C. 101 reads as follows:
Whoever invents or discovers any new and useful process, machine, manufacture, or composition of matter, or any new and useful improvement thereof, may obtain a patent therefor, subject to the conditions and requirements of this title.
4. Claims 1-20 are rejected under 35 U.S.C. 101 because the claimed invention is directed to a judicial exception (i.e., a law of nature, a natural phenomenon, or an abstract idea) without significantly more.
Under the 2019 PEG (now been incorporated into MPEP 2106), the revised procedure for determining whether a claim is "directed to" a judicial exception requires a two-prong inquiry into whether the claim recites: (1) any judicial exceptions, including certain groupings of abstract ideas (i.e., mathematical concepts, certain methods of organizing human interactions such as a fundamental economic practice, or mental processes); and (2) additional elements that integrate the judicial exception into a practical application (see MPEP § 2106.05(a)-(c), (e)-(h)).
Only if a claim (1) recites a judicial exception and (2) does not integrate that exception into a practical application, do we then look to whether the claim: (3) adds a specific limitation beyond the judicial exception that is not "well-understood, routine, conventional" in the field (see MPEP § 2106.0S(d)); or (4) simply appends well-understood, routine, conventional activities previously known to the industry, specified at a high level of generality, to the judicial exception.
Claims 1-20 are directed to an abstract idea of multi-stage classification of queries using AI model.
Specifically, representative claim 8 recites:
A computer-implemented method for routing queries by performing multi-stage classification of the queries using a hierarchical cascade of artificial intelligence (Al) models, the computer-implemented method comprising:
S1: obtaining an output generation request that comprises an input for generation of an output using one or more Al agents of a plurality of Al agents;
S2: transmitting a representation of the input to a first Al model set that is configured to generate (a) a first classification of the input and (b) an associated first score;
S3: in response to a determination that the first score fails to satisfy a first threshold stored in a threshold data structure, transmitting the representation of the input to a second Al model set, wherein the second Al model set is configured to generate (a) a second classification of the input and (b) an associated second score, using a larger number of Al model parameters than the first Al model set;
S4: in response to a determination that the second score satisfies a second threshold stored in the threshold data structure, aggregating the first classification and the second classification by:
S4.I associating a weight to each of the first classification and the second classification using a first historical performance metric value set and a second historical performance metric value set, respectively, and
S4.II determining a composite classification output that ranks the first classification and the second classification in accordance with respective weights of the first classification and the second classification;
S5: selecting a subset of the plurality of Al agents in accordance with the composite classification output, wherein each Al agent in the subset of the plurality of Al agents is configured to execute a computer-executable task set using the representation of the input;
S6: storing an audit data entry, wherein the audit data entry comprises an audit data structure comprising the composite classification output, a list of identifiers corresponding to the subset of the plurality of AI agents, and a list of model compliances corresponding to the subset of the plurality of AI agents, wherein each model compliance corresponds to a compliance of a model of the subset of the plurality of AI with one or more regulatory constraints;
S7: determining, based on the audit data structure, that the subset of the plurality of AI agents complies with one or more compliance requirements associated with the output generation request; and
S8: transmitting the representation of the input to the subset of the plurality of Al agents to execute respective computer-executable task sets.
The claim limitations in the abstract idea have been highlighted in bold above; the remaining limitations are “additional elements”.
The highlighted portion of the claim constitutes an abstract idea under the 2019 Revised Patent Subject Matter Eligibility Guidance and the additional elements are NOT sufficient to amount to significantly more than the judicial exceptions, as analyzed below:
Step
Analysis
1. Statutory Category ?
Yes.
Method
2A - Prong 1: Judicial Exception Recited?
Yes.
See bolded portion listed above.
Under its broadest reasonable interpretation (BRI), the limitation S2 reads on a process of performing data manipulation, evaluation, and judgment to generate outputs using AI models which encompasses mathematical concepts and/or calculations, namely a series of calculations leading to one or more numerical results or answers, that can be performed in the human mind and/or with pen and paper. The claim also does not provide any details about how the AI models operate or how the outputs (i.e., a first classification of the input and (b) an associated first score) are made. In view of the USPTO’s July 17, 2024 Subject Matter Eligibility Examples), a "prediction using a machine learning model" is considered an "abstract idea" if the claim focuses solely on the concept of making predictions using a generic machine learning algorithm, without any specific technical improvements or applications that go beyond the basic idea of using a computer to analyze data and generate predictions.
The limitation S3 comprises two parts: determining that the first score fails to satisfy a first threshold stored in a threshold data structure (S3.I), and transmitting the representation of the input to a second Al model set, wherein the second Al model set is configured to generate (a) a second classification of the input and (b) an associated second score, using a larger number of Al model parameters than the first Al model set (S3.II). Under the BRI, the limitation S3.I reads on a data analysis process that can be performed by the human mind using mental steps/critical thinking and/or with pen and paper. Similar to S2, the limitation S3.II encompasses mathematical concepts and/or calculations, namely a series of calculations leading to one or more numerical results or answers that can be performed in the human mind and/or with pen and paper.
Under the BRI, the limitation S4 encompasses mental processes and mathematical concepts and/or calculations. Both of the mental processes and the mathematical concepts and/or calculations can be performed in the human mind and/or with pen and paper.
Under the BRI, the limitation S5 reads on a mental process that can be performed by the human mind using mental steps/critical thinking and/or with pen and paper, while the “wherein” clause encompasses an abstract idea of using the AI agent to generate output. See discussion of S2 above.
Under the BRI, the limitation S7 reads on a mental process that can be performed by the human mind using mental steps/critical thinking and/or with pen and paper based on observation, evaluation, judgment, and opinion.
Further, the computer related elements are recited at a high level of generality. Under the BRI, they may be interpretated as software, hardware or combinations thereof. An element directed to functional descriptive material, including computer programs, per se, amounts to no more than mere instructions to apply the exception using a generic computer as a tool commonly known in the art.
Nothing in the claimed limitations preclude the mental processes and math calculation from practically being performed in the mind and/or with pen/paper.
As such, the bolded portion of instant claim 8 falls within a combination of the “Mathematical Concepts” and “Mental Process” Groupings of Abstract Ideas defined by the 2019 PEG.
2A - Prong 2: Integrated into a Practical Application?
No.
The claim as a whole does not integrate the abstract idea into a practical application.
Under the BRI, the steps S1 and S6 read on merely processes of gathering the data/information necessary for performing the identified abstract. According to MPEP 2106.05(g)(3): … that were described as mere data gathering in conjunction with a law of nature or abstract idea. As such, it represents an extra-solution activity to the judicial exception.
The newly added limitations including “wherein the audit data entry comprises an audit data structure comprising ... , wherein each model compliance corresponds to a compliance of a model of the subset of the plurality of AI with one or more regulatory constraints”, under the BRI, encompasses merely data characterization which can be viewed as nothing more than an attempt to generally link the use of the judicial exception to the relevant technological environment or field of use. The claim does not provide such details as, for example, how the audit data are acquired, where the audit data entry is stored, and what structure the recited audit data structure applies exactly to organize and store the audit data, so that these additional elements effect a meaningful way beyond generally linking the use of the judicial exception to a particular technological environment. Accordingly, the limitations to the audit data do not
integrate the identified judicial exception into a practical application or reflect a qualified improvement. Instead, the claim would monopolize the judicial exception across a wide range of applications.
The limitation S8 reads on an insignificant post-solution activity, which is not qualified for meaningful limitations to integrate the identified judicial exception into a practical application.
In general, the claim as a whole does not meet any of the following criteria to integrate the abstract idea into a practical application:
An additional element reflects an improvement in the functioning of a computer, or an improvement to other technology or technical field;
an additional element that applies or uses a judicial exception to effect a particular treatment or prophylaxis for a disease or medical condition;
an additional element implements a judicial exception with, or uses a judicial exception in conjunction with, a particular machine or manufacture that is integral to the claim;
an additional element effects a transformation or reduction of a particular article to a different state or thing; and
an additional element applies or uses the judicial exception in some other meaningful way beyond generally linking the use of the judicial exception to a particular technological environment, such that the claim as a whole is more than a drafting effort designed to monopolize the exception.
Various considerations are used to determine whether the additional elements are sufficient to integrate the abstract idea into a practical application. However, in all of these respects, the claim fails to recite additional elements which might possibly integrate the claim into a particular practical application. Instead, based on the above considerations, the claim would tend to monopolize the algorithm across a wide range of applications.
2B: Claim provides an Inventive Concept?
No.
See analysis given in 2A - Prong 2 above.
The claim does not recite any additional element that amounts to “significantly more” or an “inventive concept”.
The claim is therefore ineligible under 35 USC 101.
Independent claims 1 and 15 are treated as ineligible subject matter under 35 U.S.C. § 101 for the same reasons as for claim 8 set forth above.
The dependent claims 2-7, 9-14 and 16-20 inherit attributes of the independent claim 1, 8 or 15 on which they depend, but do not add anything which would render the claimed invention a patent eligible application of the abstract idea. These claims merely extend (or narrow) the abstract idea which do not amount for "significant more" because they merely add details to the algorithm which forms the abstract idea as discussed above.
In particular, claim 6 recites: wherein each computer-executable task set executed by each Al agent causes the Al agent to perform one or more of: invocation of a software application, retrieval of data from a database, or a return of a response to the computing device. With weight given to all the additional elements of claim 6 and focusing on the extent to which (or how) an additional element imposes meaningful limits on the claim, it is deemed that claim 6 is not qualified for meaningful limitations to integrate the identified judicial exception into a practical application because it only generally attaches one or more extra-solution activities to the identified judicial exception. See MPEP 2106.04(d): “A claim that integrates a judicial exception into a practical application will apply, rely on, or use the judicial exception in a manner that imposes a meaningful limit on the judicial exception, such that the claim is more than a drafting effort designed to monopolize the judicial exception. …”.
Claim 7 recites: wherein the first historical performance metric value set includes at least one of: model accuracy, average response latency, or a model reliability score determined using previous classifications performed by the first language model set. Under the BRI, the additional elements recited in claim 7 encompass merely data characterization which can be viewed as nothing more than an attempt to generally link the use of the judicial exception to the relevant technological environment.
Claim 13 recites: causing training of one or more of: the first or second Al model set using a labeled corpus that identifies historical inputs, historical classifications, and historical computer-executable task sets stored over a pre-defined time window. Under its BRI, this limitation encompasses processes of binning/clustering/labelling a corpus of existing training data and training the first or second Al model set using the labelled training data. In light of the USPTO’s July 2024 Subject Matter Eligibility Examples (e.g., Example 47, claim 2), discretizing continuous training data to generate input data by processes including binning or clustering continuous data may be practically performed in the human mind using observation, evaluation, judgment, and opinion, while said training is recited at a high level of generality which may involve optimizing the AI models using a series of mathematical calculations to iteratively adjust the algorithms and/or parameter values of the AI models, therefore encompasses mathematical concepts.
Claim 14 recites: generating a cryptographic hash representing the composite classification output and the subset of the plurality of Al agents; and storing the cryptographic hash on a distributed ledger. Under its BRI, the step of generating a cryptographic hash representing the composite classification output and the subset of the plurality of Al agents encompasses mathematical concepts, and the step of storing the cryptographic hash on a distributed ledge reads on merely an insignificant post-solution activity which does not amount the abstract idea to be significantly more.
Claim 17 recites: wherein the first Al model set is trained to generate the first classification by mapping the input to one or more predefined intents based on a degree of similarity between (a) a vector representation of the input and (b) a vector representation of each of the one or more predefined intents, and wherein the second Al model set is trained on a domain-specific training dataset associated with a predefined subject matter. Under the BRI, the limitation “wherein the first Al model set is trained to generate the first classification by mapping the input to one or more predefined intents based on a degree of similarity between (a) a vector representation of the input and (b) a vector representation of each of the one or more predefined intents” encompasses mathematical concepts and/or calculations that are implemented by the first Al model set, thus it is a part or an extension of the identified abstract idea but does not amount for "significantly more" because it merely adds details to the abstract idea. As to the claimed training data for the second Al model set, under its BRI, “a domain-specific training dataset associated with a predefined subject matter” is considered merely a way of binning/labelling the training data which is also treated as a part or an extension of the identified abstract idea.
Regarding claim 18, the “language model” is recited at a high level of generality, which can be viewed as nothing more than an attempt to generally link the use of the judicial exception to a relevant technological environment.
Claim 20 recites: wherein the system is further caused to: display a representation of the subset of the plurality of Al agents on a user interface of a computing device, wherein the representation is displayed in accordance with a display attribute set determined based on the first score and the second score. This limitation encompasses an insignificant post-solution activity which does not amount the abstract idea to be significantly more as it merely displays the results of the abstract idea, while a user interface of a computing device is "well-understood, routine, conventional" in the field.
Examiner’s Note
5. While there are related references that discuss orchestrating task execution among Al agentic models responsive to a received query by using a hierarchical model cascade to classify queries into agent domains, the prior art of record does not specifically provide teachings for the claimed limitations including:
transmit a vector representation of the query to a first language model set that is configured to generate (a) a first classification of the query that maps the query to a first subset of the plurality of Al agents and (b) an associated first confidence score;
compare the first confidence score to a first threshold value stored in a threshold data structure accessed by the computing device, the first threshold value determined using a first historical performance metric value set of the first language model set;
in response to a determination that the first confidence score fails to satisfy the first threshold value, transmit the vector representation of the query to a second language model set, wherein the second language model set is configured to generate (a) a second classification of the query that maps the query to a second subset of the plurality of Al agents and (b) an associated second confidence score, using a larger number of model parameters than the first language model set;
compare the second confidence score to a second threshold value stored in the threshold data structure accessed by the computing device, the second threshold value determined using a second historical performance metric value set of the second language model set;
in response to a determination that the second confidence score satisfies the second threshold value, aggregate the first classification and the second classification by:
assigning a weight to each of the first classification and the second classification using the first historical performance metric value set and the second historical performance metric value set, respectively, and
generating a composite classification that ranks the first classification and the second classification in accordance with respective weights of the first classification and the second classification;
select a routing Al agent set from the first and second subsets of the plurality of Al agents, wherein each Al agent in the routing Al agent set is configured to autonomously execute a computer-executable task set using the vector representation of the query.
It is these limitations listed above, as they are claimed in the combination recited in independent claim 1, that would make the pending claims 1-7 of the present application distinguish over the prior art of record.
Regarding claims 8-14, the prior art of record does not specifically provide teachings for the claimed limitations including:
transmitting a representation of the input to a first Al model set that is configured to generate (a) a first classification of the input and (b) an associated first score;
in response to a determination that the first score fails to satisfy a first threshold stored in a threshold data structure, transmitting the representation of the input to a second Al model set, wherein the second Al model set is configured to generate (a) a second classification of the input and (b) an associated second score, using a larger number of Al model parameters than the first Al model set;
in response to a determination that the second score satisfies a second threshold stored in the threshold data structure, aggregating the first classification and the second classification by:
associating a weight to each of the first classification and the second classification using a first historical performance metric value set and a second historical performance metric value set, respectively, and
determining a composite classification output that ranks the first classification and the second classification in accordance with respective weights of the first classification and the second classification;
selecting a subset of the plurality of Al agents in accordance with the composite classification output, wherein each Al agent in the subset of the plurality of Al agents is configured to execute a computer-executable task set using the representation of the input.
It is these limitations listed above, as they are claimed in the combination recited in independent claim 8, that would make the pending claims 8-14 of the present application distinguish over the prior art of record.
Regarding claims 15-20, the prior art of record does not specifically provide teachings for the claimed limitations including:
transmit a representation of the input to a first Al model set that is configured to generate (a) a first classification of the input and (b) an associated first score;
in response to a determination that the first score fails to satisfy a first threshold stored in a threshold data structure, transmit the representation of the input to a second Al model set, wherein the second Al model set is configured to generate (a) a second classification of the input and (b)an associated second score, using a larger number of Al model parameters than the first Al model set;
in response to a determination that the second score satisfies a second threshold stored in the threshold data structure, determine a composite classification output that ranks the first classification and the second classification based on a first historical performance metric value set and a second historical performance metric value set; and
select a subset of the plurality of Al agents in accordance with the composite classification output, wherein each Al agent in the subset of the plurality of Al agents is configured to execute a computer-executable task set using the representation of the input.
It is these limitations listed above, as they are claimed in the combination recited in independent claim 15, that would make the pending claims 15-20 of the present application distinguish over the prior art of record.
Conclusion
6. 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 extension fee 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 date of this final action.
Citation of Relevant Prior Art
7. The prior art made of record and not relied upon is considered pertinent to applicant's disclosure:
BHATHENA et al. (US 20240282296 A1) -- METHOD AND SYSTEM FOR CONDITIONAL HIERARCHICAL DOMAIN ROUTING AND INTENT CLASSIFICATION FOR VIRTUAL ASSISTANT QUERIES
Wei et al. (CN 115905471 A) -- Business Object Model Generating Method, Electronic Device, Computer Readable Storage Medium
Liu et al. (CN 115811401 A) -- Supervision Method, Device And System
Lee et al. (US 20220377107 A1) -- SYSTEM AND METHOD FOR DETECTING PHISHING-DOMAINS IN A SET OF DOMAIN NAME SYSTEM (DNS) RECORDS
SABHARWAL et al. (US 20220300716 A1) -- SYSTEM AND METHOD FOR DESIGNING ARTIFICIAL INTELLIGENCE (AI) BASED HIERARCHICAL MULTI-CONVERSATION SYSTEM
Contact Information
8. Any inquiry concerning this communication or earlier communications from the examiner should be directed to XIUQIN SUN whose telephone number is (571)272-2280. The examiner can normally be reached 9:30am-6:00pm.
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, Shelby A. Turner can be reached on (571) 272-6334. 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.
/X.S/Examiner, Art Unit 2857
/SHELBY A TURNER/Supervisory Patent Examiner, Art Unit 2857