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 .
DETAILED ACTION
2. This office action is in response to the rce filed on 05/08/2026. Claim 1-20 are pending and have been considered below.
Claim Rejections - 35 USC § 103
3. 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.
4. Claim(s) 1-2, 6-9, 13-16 and 20 is/are rejected under 35 U.S.C. 103 as being unpatentable over Dang et al. (US 2018/0293488) in view of Fletcher et al. (US 2008/0244251).
Claim 1. Dang discloses a method for parallel predictive modelling, the method comprising:
receiving a configuration file associated with a specified predictive concept of a plurality of predictive concepts at a production layer of a predictive modelling platform (The input review data 102(2) may be distributed to the various layers for parallel processing, to save time and/or more efficiently use computing resources.") ([0049]), the predictive modelling platform comprising the production layer and a consumption layer ("As shown in the example, the structure may include convolution pooling layers 402 at varying levels of specificity. For example, the layers 402 may include a phrase feature map group, a sentence feature map group, a paragraph feature map group, and/or a context feature map group.") ([0049], fig. 4), wherein the production layer and the consumption layer are communicatively connected by a distributed messaging system ("The input review data 102(2) may be distributed to the various layers for parallel processing, to save time and/or more efficiently use computing resources. A slacked layer may be generated as output from each layer 402.") ([0049], fig. 4);
pulling, by the consumption layer, data from one or more heterogeneous data storage units based on a data source location specified by the configuration file (The prediction engine 110 may retrieve the review data 102(1) and the review data 102(2) from the data storage 108… The data storage 108 may be any suitable type of data storage, such as a cloud service, data warehouse, distributed big data platform, relational and/or non-relational databases, and so forth) ([0036])..(…the data storage 108 may be external to the server computing device(s) 104, and accessible over one or more networks) ([0034])[wherein there must be some sort of directives in the engine regarding location and type of data];
identifying, by the production layer, a job request based on the configuration file (For example, the layers 402 may include a phrase feature map group, a sentence feature map group, a paragraph feature map group, and/or a context feature map group. The input review data 102(2) may be distributed to the various layers for parallel processing, to save time and/or more efficiently use computing resources.") ([0049], fig. 4);
sending, by the distributed messaging system, the job request to the consumption layer, as one of a plurality of job requests to be passed to a predictive model of a plurality of predictive models implemented by a processing container (For example, the layers 402 may include a phrase feature map group, a sentence feature map group, a paragraph feature map group, and/or a context feature map group. The input review data 102(2) may be distributed to the various layers for parallel processing, lo save time and/or more efficiently use computing resources.") ([0049], fig. 4)…( to train the predictive model(s)) ([0041]), wherein the predictive model of the plurality of predictive models is specified by the configuration file (Ratings posted with reviews 102(1) may be used to train one or more models 120. The model(s) 120 may be employed to predict ratings 122 for posted reviews 102(2) that are not initially associated with ratings.") ([0042], fig. 2);
obtaining, from the processing container, an output of the predictive model comprising a value generated based on the data pulled by the consumption layer (A stacked layer may be generated as output from each layer 402.) ([0049], fig. 4);
sending, by the distributed messaging system, the forecast value to the production layer (Accordingly, one or more stacked convolutional and/or pooling layers, with the same or different structures, can be additionally inserted into the layers 402.) ([0049], fig. 4); and
determining, by the production layer, one or more values of the specified predictive concept by applying one or more operators to the forecasted value (user moment), the one or more operators specified by the configuration file (the combined features are reduced through over-fitting reduction 406, and provided to a (e.g., full) connection layer 408 which is then employed to generate the prediction results 122.) ([0050], fig. 4)…( a user moment is a characteristic or attribute of an individual that is related to their behavior online, such as their search habits, products or topics they have expressed an interest in, and so forth) ([0038])…( Incorporating user moment information into the prediction process can further improve the prediction accuracy) ([0069],[0071]).
Dang does not explicitly disclose the configuration file defining a specified plurality of operations to be performed by the predictive modelling platform to determine the specified predictive concept; forecasted value of a future condition.
However, Fletcher discloses the configuration file defining a specified plurality of operations to be performed by the predictive modelling platform to determine the specified predictive concept ([0043], abstract); forecasted value of a future condition ([0058]). Therefore, it would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to have modified the teachings of Dang to run autonomously at an end user's site to construct and maintain one or more predictive models without the intervention of expert personnel.
Claim 2. Dang and Fletcher disclose the method of claim 1, Dang further discloses wherein the predictive model is a deep learning model (Implementations provide for a prediction engine 110 that builds the prediction models 120 through distributed parallel model building, using deep learning that employs CNNs. With respect to distributed parallel model building, implementations provide a framework to build large scale models in parallel using deep learning technology) ([0038]).
Claim 6. Dang and Fletcher disclose the method of claim 1, Dang further discloses wherein the configuration file specifies a feature set for the predictive model used to generate the one or more values of the specified predictive concept.
wherein the configuration file specifies a feature set for the predictive model used to generate the one or more values of the predictive concept (the layers 402 may include a phrase feature map group, a sentence feature map group, a paragraph feature map group, and/or a context feature map group.). ([0049], fig. 4).
Claim 7. Dang and Fletcher disclose the method of claim 1, Dang further discloses comprising:
generating, by the production layer, a control command comprising a parameter based on the determined one or more values of the predictive concept (The prediction engine 110 may build one or more prediction models 120, and use the prediction model(s) 120 to generate prediction results 122. The prediction results 122 may be stored on the server computing device(s) 104, and/or elsewhere, and may be transmitted to one or more prediction output devices 124 for display and use by data consumers, such as marketing professionals.) ([0037]); and
sending the control command, via a network to an external system (The prediction results 122 may be stored on the server computing device(s) 104, and/or elsewhere, and may be transmitted to one or more prediction output devices 124 for display and use by data consumers, such as marketing professionals.) ([0037]).
Claims 8-9, 13-16 and 20 represent the platform and medium of claims 1-2 and 6-7, respectively and are rejected along the same rationale.
6. Claims 3, 5, 10, 12, 17 and 19 are rejected under 35 U.S.C. 103 as being unpatentable over Dang et al. (US 2018/0293488) in view of Fletcher et al. (US 2008/0244251) and further in view of Lin et al. (US 8,364,613).
Claim 3. Dang and Fletcher disclose the method of claim 1, but fail to explicitly disclose further comprising: identifying, by the production layer, a model training request based on the configuration file; sending, by the distributed messaging system, to the consumption layer, a request for updated training data from a data mapper provided by the predictive modelling platform; and training the predictive model based on the updated training data.
However, Lin discloses identifying, by the production layer, a model training request based on the configuration file (In some implementations, models are trained by a training system 416 which receives requests from the prediction API 408 to initiate training and check the status of training ... ) (Col. 10, ll.47-50); sending, by the distributed messaging system, to the consumption layer, a request for updated training data from a data mapper provided by the predictive modelling platform ("In some implementations, models are trained by a training system 416 which receives requests from the prediction API 408 to initiate training and check the status of training ... ") (Col. 10, ll.47-50)…(The prediction API 408 provides the training system 416 with the location of training data 320 to be used in training a particular model. For example, the training data, such as a range of cells in a spreadsheet, can be obtained from the application data 318 through use of the web application API 406 and then provided to the training system ... ") (Col 10, In 50-56); and training the predictive model based on the updated training data (Computer programs embodying such machine learning algorithms can be operable to: input previously trained predictive models and additional training data; implement the machine learning algorithm to generate an updated predictive model that is representative of the original training dataset and the additional training data; and output the updated predictive model in a suitable computer readable and executable format.") (Col. 4, ll. 46-53). Therefore, it would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to have modified the teachings of Dang to request a model training to update the training data, such that the system can identify new parameters associated with the training model, and update the model to reflect the changes.
Claim 5. Dang Fletcher and Lin disclose the method of claim 3, Dang discloses wherein the one or more heterogeneous data storage units comprise a plurality of heterogeneous data storage units (The prediction engine 110 may retrieve the review data 102(1) and the review data 102(2) from the data storage 108… The data storage 108 may be any suitable type of data storage, such as a cloud service, data warehouse, distributed big data platform, relational and/or non-relational databases, and so forth) ([0036])..(…the data storage 108 may be external to the server computing device(s) 104, and accessible over one or more networks) ([0034])) and Lin further discloses wherein the data mapper comprises a data store for the predictive modelling platform (The servers can communicate with each other and with storage systems (e.g., application data storage system 318 and training data storage system 320) at various times using one or more computer networks or other communication means.) (Col 9, ln 64 - Col 10, ln 2), and wherein the data mapper assigns a plurality of connections between a plurality of heterogeneous data storage units (A map-reduce system includes application-independent map modules configured to read input data and to apply at least one applicationspecific map operation to the input data to produce intermediate data values. The map operation is automatically parallelized across multiple servers. Intermediate data structures are used to store the intermediate data values.) (Col 9, ln 39-45). One would have been motivated to do so such that the system can identify new parameters associated with the training model, and update the model to reflect the changes
Claims 10, 12, 17 and 19 represent the platform and medium of claims 3 and 5, respectively and are rejected along the same rationale.
7. Claims 4, 11 and 18 are rejected under 35 U.S.C. 103 as being unpatentable over Dang et al. (US 2018/0293488) in view of Fletcher et al. (US 2008/0244251) in view of Lin et al. (US 8,364,613) and further in view of ELKABETZ et al. (US 2019/0339416).
Claim 4. Dang Fletcher and Lin disclose the method of claim 3, but fail to explicitly disclose wherein the configuration file comprises a history of predictive models used to generate the one or more values of the specified predictive concept. However, ELKABETZ discloses the configuration file comprises a history of predictive models used to generate the one or more values of the predictive concept (The ML training module (678) retrieves an untrained, partially trained, or previously trained ML model from the system database (320), retrieves ML training data from the ML training data store (683), and uses the training data to train or retrain the ML model… ). Therefore, it would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to have modified the teachings of Dang and Lin to include the historical training models in the configuration files, such that the system can use previous training models to train or retrain new models.
Claims 11 and 18 represent the platform and medium of claims 4, respectively and are rejected along the same rationale.
Response to Arguments
8. Applicant’s arguments and amendments filed on 05/08/2026 have been fully considered but are moot in light of new ground of rejection(s).
Applicant argued However, Dang's "user moment information" is not the same as the "one or more operators" recited in claim 1. Unlike the "user moment information," the "one or more operators" are "a set of rules based operators applied to the value of the forecasted parameter" to "computationally efficiently and algorithmically" determine a predictive value. (See Specification, paragraph [0093]).
In response to applicant's argument that the references fail to show certain features of the invention, it is noted that the features upon which applicant relies (i.e., Unlike the "user moment information," the "one or more operators" are "a set of rules based operators applied to the value of the forecasted parameter" to "computationally efficiently and algorithmically) are not recited in the rejected claim(s). Although the claims are interpreted in light of the specification, limitations from the specification are not read into the claims. See In re Van Geuns, 988 F.2d 1181, 26 USPQ2d 1057 (Fed. Cir. 1993).
9. The prior art made of record and not relied upon is considered pertinent to applicant's disclosure (See PTO-892).
Any inquiry concerning this communication or earlier communications from the examiner should be directed to Phenuel S. Salomon whose telephone number is (571) 270-1699. The examiner can normally be reached on Mon-Fri 7:00 A.M. to 4:00 P.M. (Alternate Friday Off) EST.
If attempts to reach the examiner by telephone are unsuccessful, the examiner’s supervisor, Usmaan Saeed can be reached on (571) 272-4046. The fax phone number for the organization where this application or proceeding is assigned is 571-273-3800.
Information regarding the status of an application may be obtained from the Patent Application Information Retrieval (PAIR) system. Status information for published applications may be obtained from either Private PAIR or Public PAIR. Status information for unpublished applications is available through Private PAIR only. For more information about the PAIR system, see http://pair-direct.uspto.gov. Should you have questions on access to the Private PAIR system, contact the Electronic Business Center (EBC) at 866-217-9197 (toll-free). If you would like assistance from a USPTO Customer Service Representative or access to the automated information system, call 800-786-9199 (IN USA OR CANADA) or 571-272-1000.
/PHENUEL S SALOMON/Primary Examiner, Art Unit 2146
Prosecution Insights