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 Amendment
The amendment filed 01/02/2026 has been entered. Claims 1-20 are presented for examination.
Information Disclosure Statement
The information disclosure statements (IDS) submitted 10/16/2025 is being considered by the examiner.
Claim Rejections - 35 USC § 103
The following is a quotation of 35 U.S.C. 103 which forms the basis for all obviousness rejections set forth in this Office action:
A patent for a claimed invention may not be obtained, notwithstanding that the claimed invention is not identically disclosed as set forth in section 102, if the differences between the claimed invention and the prior art are such that the claimed invention as a whole would have been obvious before the effective filing date of the claimed invention to a person having ordinary skill in the art to which the claimed invention pertains. Patentability shall not be negated by the manner in which the invention was made.
Claim(s) 1, 5-8, 12-15 and 19-20 are rejected under 35 U.S.C. 103 as being unpatentable over Le United States Patent Application Publication US 2022/0094647, in view of Polleri United States Patent Application Publication US 2021/0081819, and further in view of Howard United States Patent Application Publication 2020/0057964.
Le and Polleri were cited in the previous office action.
Regarding claim 1, Le discloses a computer-implemented method (Le, Abstract Line 1-3 Methods and systems are described for generating dynamic interface options using machine learning models; Le, para [0053], computer interpreted as device with processing circuitry that runs and stores applications) comprising:
receiving, by one or more computer processors, data and associated metadata corresponding to a machine learning task from a user (Le, para [0052], processors for sending and receiving commands; Le, para [0037], with regards to fig. 3, System 300 may receive user action data based on user actions (interpreted as data corresponding to a machine learning task from the user) with user interfaces during a device session. The user action data (e.g., data 304) may include metadata (interpreted as associated metadata));
determining, by one or more computer processors, a task context and a problem domain based on the received data and the associated metadata (Le, para [0041], transfer learning allows system 300 to deal with current scenarios (e.g., detecting user intent) by leveraging the already existing labeled data of some related task or domain. System 300 may store knowledge gained through other tasks and apply it to the current task. This is equivalent to determining a task context and a problem domain based on the received data and the associated metadata);
selecting from a set of pre-complied models, by one or more computer processors, pre-compiled models (Le, para [0020], the system may include different supervised and unsupervised machine learning models and human devised rules that may reflect accumulated domain (interpreted as problem domain) expertise. The system may include deep learning models that may include neural factorization machines, deep and wide, and multimodal models (interpreted as pre-compiled models));
generating, by one or more computer processors, two multimodal model combinations with the selected pre-compiled models (Le, para [0048], First model 310, second model 312, and third model 308 may receive inputs and generate outputs that are processed by fourth model 314. Fourth model 314 may then generate a final classification 318. Fourth model 314 may include ensemble prediction. For example, ensemble methods use multiple learning algorithms to obtain better predictive performance than could be obtained from any of the constituent learning algorithms alone);
executing, by one or more computer processors, two or more multimodal model combinations with the received data and the associated metadata (Le, para [0042], with regards to fig. 3, textual image and metadata 304 serves as an input to multimodal 310. System 300 processes this information in first model 310 which is equivalent to executing the one or more multimodal model combinations with the received data and the associated metadata);
displaying, by one or more computer processors, output results of the executed one or more multimodal model combinations to the user (Le, para [0075], with regards to fig. 5, At step 516, process 500 generates the dynamic conversational response during the conversational interaction. For example, the system may generate, at the user interface, the dynamic conversational response during the conversational interaction); and
determining, by one or more computer processors, whether a level of error associated with the results is acceptable to the user based on a response from the user (Le, para [0058], update its configurations (e.g., weights, biases, or other parameters) based on the assessment of its prediction. In further cases, uses backpropagation of error. ‘Reflective of the magnitude’ is interpreted as determining a level).
Le does not disclose based on the task context and the problem domain, identifying, by one or more computer processors, the machine learning task, evaluating, by one or more computer processors, a match between the problem domain and a set of pre-compiled models, selecting pre-compiled models, by one or more computer processors, that match the problem domain.
However, Polleri discloses:
based on the task context and the problem domain, identifying, by one or more computer processors, the machine learning task (Polleri, [0213], historical data interpreted as context and problem to be solved description interpreted as problem domain; Polleri, para [0214-215], transcribes the inputs to be text fragments. Correlates machine learning models with text fragments represents matching);
evaluating, by one or more computer processors, a match between the problem domain and a set of pre-compiled models (Polleri, para [0216], analyzes metadata of the frameworks of the machine learning models with the text fragments, where the text fragments are generated from the problem to be solved);
selecting pre-compiled models, by one or more computer processors, that match the problem domain (Polleri, para [0217], performs a selection based on the match);
It would have been obvious to one of ordinary skill in the art before the time of the effective filing date of the claimed invention to have included the concept of based on the task context and the problem domain, identifying, by one or more computer processors, the machine learning task, evaluating, by one or more computer processors, a match between the problem domain and a set of pre-compiled models, selecting pre-compiled models, by one or more computer processors, that match the problem domain as suggested in Polleri into Le since both of these systems are addressing the need of generating new machine learning model. By incorporating the teaching of Polleri into Le’s system would improve the integrity of Le’s system by allow for adaptation at run-time due to changes in data (Polleri, para [0008]).
The combination of Le and Polleri does not teaches executing, by one or more computer processors, the two or more multimodal model combinations with the received data and the associated metadata to identify a new model architecture for the machine learning task, wherein the two or more multimodal model combinations are evaluated to identify a multimodal model combination that is more accurate at performing the machine learning task as compared to other multimodal model combinations in the two or more multimodal model combinations, and the new model architecture comprises at least in part the identified multimodal model combination, and displaying, by one or more computer processors, output results of the new model architecture.
However, Howard teaches executing, by one or more computer processors, the two or more multimodal model combinations with the received data and the associated metadata to identify a new model architecture for the machine learning task, wherein the two or more multimodal model combinations are evaluated to identify a multimodal model combination that is more accurate at performing the machine learning task as compared to other multimodal model combinations in the two or more multimodal model combinations, and the new model architecture comprises at least in part the identified multimodal model combination (abstract, para [0010], [0084], [0104], and displaying, by one or more computer processors, output results of the new model architecture (fig. 1, item 106).
It would have been obvious to one of ordinary skill in the art before the time of the effective filing date of the claimed invention to have included the concept of executing, by one or more computer processors, the two or more multimodal model combinations with the received data and the associated metadata to identify a new model architecture for the machine learning task, wherein the two or more multimodal model combinations are evaluated to identify a multimodal model combination that is more accurate at performing the machine learning task as compared to other multimodal model combinations in the two or more multimodal model combinations, and the new model architecture comprises at least in part the identified multimodal model combination, and displaying, by one or more computer processors, output results of the new model architecture as suggested in Howard into the combination of Le and Polleri since all of these systems are addressing the need of generating new machine learning model. By incorporating the teaching of Howard into the combination of Le and Polleri’s system would improve the integrity of Le’s system by providing an intelligent adaptive system that combines input data types, processing history and objective, research knowledge and situation context to determine what is the most appropriate mathematical model, choose the most appropriate computing infrastructure on which to perform learning and propose the best solution for a given problem (Howard para [0007].
Regarding claim 5, the combination of Le, Polleri and Howard discloses the computer implemented method of claim 1. Howard additionally discloses wherein the received data is multimodal data, and selecting the pre-compiled model further comprises transforming features of the multimodal data match input requirements of the pre-compiles models using a conversion model trained to transform the features of the multimodal data (para [0010], [0079], [0084], [0084]).
Regarding claim 6, the combination of Le, Polleri and Howard discloses the computer implemented method of claim 1. Le additionally discloses processing, by one or more computer processors, the received data and the associated metadata using a chatbot to read textual information (Le, para [0063-64], conversational interaction with a user interface represents a chatbot. Data and feature information received); and
applying, by one or more computer processors, one or more natural language processing techniques to the received data and the associated metadata to extract information corresponding to the task context and the problem domain (Le, para [0038, 40-41], Line 1-4 System 300 may also receive information, which may use a Bidirectional Encoder Representations from Transformers (BERT) language model for performing natural language processing),
Regarding claim 7, the combination of Le, Polleri and Howard discloses the computer-implemented method of claim 1. Howard additional discloses wherein generating the two or more multimodal model combinations further comprises: generating, for each metric in a set of accuracy metrics, two or more multimodal model combinations; and evaluating, for each metric in the set of accuracy metrics, the respective two or more multimodal model combinations to identify a multimodal model combination that is more accurate at performing the machine learning task based on the metric as compared to other multimodal model combinations in the two or more multimodal model combinations (para [0010], [0095], [0104].
Regarding claims 8 and 12-14, they are product claims that correspond to method claims 1 and 5-7. Therefore, they are rejected for the same reason as method claims 1 and 5-7 above.
Regarding claims 15 and 19-20, they are system claims that correspond to method claims 1 and 5-6. Therefore, they are rejected for the same reason as method claims 1 and 5-6 above.
Claim(s) 2-3, 9-10 and 16-17 are rejected under 35 U.S.C. 103 as being unpatentable over Le United States Patent Application Publication US 2022/0094647 in view of Polleri United States Patent Application Publication US 2021/0081819, in view of Howard United States Patent Application Publication 2020/0057964, as applied in claims 1, 8 and 15 above, and further in view of Orhan United States Patent US 11,748,568.
Le, Polleri and Orhan were cited in the previous office action.
Regarding claim 2, the combination of Le, Polleri and Howard discloses the computer implemented method of claim 1. Le further discloses a level of error and a backpropagation process (Le, para [0058]). However, the combination of Le, Polleri and Howard does not disclose responsive to determining the level of error associated with the results is not acceptable to the user, iteratively repeating, by one or more computer processors, a process of generating and executing the one or more multimodal models until the level of error associated with the results is acceptable to the user.
Ohran discloses responsive to determining the level of error associated with the results is not acceptable to the user, iteratively repeating, by one or more computer processors, a process of generating and executing the one or more multimodal models until the level of error associated with the results is acceptable to the user (Col 17 Line 15-24 Clients on whose behalf the automated anomaly detection is being performed may provide client feedback 695, indicating whether the clients found the particular combinations of metrics and statistics that were selected using the model 610 useful or not (determining if the level of error is acceptable to the user or not). Such client feedback 695 may also be used to improve the model 610 over time, e.g., by labeling combinations of statistics and metrics used for re-training the model (iteratively repeating process of generating and executing the models) based on whether the client found them useful or not (re-training until results is acceptable to the user)).
It would have been obvious to one of ordinary skill in the art before the time of the effective filing date of the claimed invention to incorporate the concept of responsive to determining the level of error associated with the results is not acceptable to the user, iteratively repeating, by one or more computer processors, a process of generating and executing the one or more multimodal models until the level of error associated with the results is acceptable to the user as suggested in Orhan into the combination of Le, Polleri and Howard since all of these system are addressing the need of generating machine learning model. By incorporating Orhan into the combination of Le, Polleri and Howard would improve the user experience of application administrators and improve the user experience of application end users. (Orhan, Col 4 Line 20-35).
Regarding claim 3, the combination of Le, Polleri, howard and Orhan discloses the computer implemented method of claim 2. Orhan additionally discloses further comprising: receiving, by one or more computer processors, additional data from the user to improve the results (Orhan, col 5, rows 13-41, user defined anomaly specification represents data from a user that is additional that is deemed an anomaly and therefore used to improve results).
Regarding claims 9-10, they are product claims that correspond to method claims 2-3. Therefore, they are rejected for the same reason as method claims 2-3 above.
Regarding claims 16-17, they are system claims that correspond to method claims 2-3. Therefore, they are rejected for the same reason as method claims 2-3 above.
Claim(s) 4, 11 and 18 is/are rejected under 35 U.S.C. 103 as being unpatentable over Le United States Patent Application Publication US 2022/0094647 in view of Polleri United States Patent Application Publication US 2021/0081819, in view of Howard United States Patent Application Publication 2020/0057964, as applied in claims 1, 8 and 15 above, and further in view of Arici United States Patent 11,928,182.
Le, Polleri, Arici were cited in the previous office action.
Regarding claim 4, the combination of Le, Polleri and Howard discloses the computer implemented method of claim 1. The combination of Le, Polleri and Howard does not disclose decomposing, by one or more computer processors, the received data into two or more data sub-types and based on the data sub-types, selecting, by one or more computer processors, a subset of the selected at least two of the one or more pre-compiled models.
However, Arici discloses further comprising decomposing, by one or more computer processors, the received data into two or more data sub-types (Arici, col 8, rows 42-46, If a sufficient amount of labeled training data 201 is available, the labeled training data may be split up into two subsets 207A and 207B (interpreted as decomposing received data into data sub-types)); and
based on the data sub-types, selecting, by one or more computer processors, a subset of the selected at least two of the one or more pre-compiled models (Arici, col 8, rows 45-57, The system first selects the base models and trains them on subset 207A. It then selects the stacking model 213 to be trained on predictions from 207B where base models and stacking meta-model are a subset of pre-compiled models).
It would have been obvious to one of ordinary skill in the art before the time of the effective filing date of the claimed invention to include the concept of decomposing, by one or more computer processors, the received data into two or more data sub-types and based on the data sub-types, selecting, by one or more computer processors, a subset of the selected at least two of the one or more pre-compiled models as suggested by Arici into the combination of Le, Polleri and Howard since al;l of these system are addressing the need of generating machine learning models . By incorporating the teaching of Arici into the combination of Le, Polleri and Howard ‘system would avoid overfitting, which is a known problem for some traditional types of stacking-based ensemble preparation approaches (Arici, Col 2 Line 46-50).
Regarding claim 11, it is product claim that corresponds to method claim 4. Therefore, it is rejected for the same reason as method claim 4 above.
Regarding claim 18, it is system claim that corresponds to method claim 4. Therefore, it is rejected for the same reason as method claim 4 above.
Response to Arguments
Applicant’s arguments, see pages 9-11, filed 01/02/2026, with respect to rejection of claims 1-20 under 35 USC 101 have been fully considered and are persuasive. Therefore, the rejection of claims 1-20 under 35 USC 101 has been withdrawn.
Applicant’s arguments with respect to claim(s) 1-20 under 35 USC 103 have been considered but are moot because the new ground of rejection (see rejection above).
Conclusion
The prior art made of record and not relied upon is considered pertinent to applicant's disclosure.
Polleri et al. (WO 2021/050382 A1) disclose a system and method for enabling the user to generate a machine learning system using AI chatbot.
Tran et al. (US Publication No. 2021/0073376 A1) disclose a system for learning input preprocessing to harden machine learning models.
Polleri et al. (US Patent No. 11526267 B2) a system and method for enabling the user to generate a machine learning system using AI chatbot.
Le et al. (US Patent No. 11405337 B2) disclose a system for generating dynamic interface options using machine learning models.
Brown et al., ("Uncertainty Quantification in Multimodal Ensembles of Deep Learners", FLAIRS Conference, 2020) disclose a system that analyze a novel technique to measure epistemic uncertainty from deep ensembles of modalities.
Applicant's amendment necessitated the new ground(s) of rejection presented in this Office action. Accordingly, THIS ACTION IS MADE FINAL. See MPEP § 706.07(a). Applicant is reminded of the extension of time policy as set forth in 37 CFR 1.136(a).
A shortened statutory period for reply to this final action is set to expire THREE MONTHS from the mailing date of this action. In the event a first reply is filed within TWO MONTHS of the mailing date of this final action and the advisory action is not mailed until after the end of the THREE-MONTH shortened statutory period, then the shortened statutory period will expire on the date the advisory action is mailed, and any nonprovisional extension fee (37 CFR 1.17(a)) pursuant to 37 CFR 1.136(a) will be calculated from the mailing date of the advisory action. In no event, however, will the statutory period for reply expire later than SIX MONTHS from the mailing date of this final action.
Any inquiry concerning this communication or earlier communications from the examiner should be directed to JENNIFER N WELCH whose telephone number is (571)272-7212. The examiner can normally be reached M-T 5:30AM - 4:00PM.
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, David Wiley, can be reached at (571) 272-4150. 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.
JENNIFER N. WELCH
Supervisory Patent Examiner
Art Unit 2143
/JENNIFER N WELCH/Supervisory Patent Examiner, Art Unit 2143
Prosecution Insights