DETAILED ACTION
1. This office action is in response to the Application No. 17311730 filed on 05/20/2026. Claims 2-5 and 10-19 has been cancelled. Claims 1, 6-9 and 20 are presented for examination and are currently pending.
Notice of Pre-AIA or AIA Status
2. 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
3. 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has been withdrawn pursuant to 37 CFR 1.114.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
05/20/2026 has been entered.
Claim Objections
4. Claim 20 is objected to because of the following informalities:
Claim 20 recites “The content classification metho”. It should be “The content classification method”.
Appropriate correction is required.
Response to Arguments
5. The claim amendments of 05/20/2026 has overcome the 112(b) rejection of 02/20/2026. As a result, the 112(b) rejection has been withdrawn.
The Applicants argument on page 1 of the remarks that “Applicant requests reconsideration and withdrawal of this rejection because neither Beers, Adibowo, Polatkan, nor any proper combination of the three references describes or suggests "generating a plurality of first classification models by machine learning using a plurality of learning contents each provided with a first feature and a learning label" and "generating a second classification model with the use of a plurality of feature vectors generated by the plurality of first classification models," as recited in claim 1” has been considered and the Examiner is withdrawing the rejections in the previous Office action because Applicant’s amendment necessitated new grounds of rejection presented in this Office Action.
It is noted that Hemani has now been newly applied as primary reference in independent claim 1 to teach the limitation “generating a plurality of first classification models by machine learning using a plurality of learning contents each provided with a first feature and a learning label" and "generating a second classification model with the use of a plurality of feature vectors generated by the plurality of first classification models”.
However, Beers is still relevant as prior art, and is now being applied as a secondary reference because Beers teaches patent number comprising at least one of a state of a family, an application type, and a number of abandoned applications in a family (The system maintains a database of raw patent factors that are derived from the patent publication [0033]; Claim type “A” refers to an apparatus claim, claim type “S” to a system claim, claim type “C” to a claim for a compound, and claim type “M” refers to a method claim [0043]. The Examiner notes the apparatus claim tells us the application type),
and wherein the learning label comprises an information of the patent number is abandoned (In Table 2, legal status code refers to events during the lifetime of the patent. These include office actions, change of ownership, abandonment, maintenance and expiration [0043]), then
It would have been obvious to a person having ordinary skill in the art before the effective filing date of the claimed invention to have modified the method of Hemani and Luo to incorporate the teachings of Beers for the benefit of using of machine learning when identifying input factors and computing the classification model that can be continuously updated in response to changes in the market (Beers [0038])
Furthermore, Luo and Dwane which were applied in the previous Office Action are still relevant to the instant claims. As a result, their teachings have been used in this Office Action.
The dependent claims 6-9 and 20 which depend directly or indirectly from independent claim 1 are still obvious over the prior art of record.
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.
6. Claims 1 and 6-8 are rejected under 35 U.S.C. 103 as being unpatentable over Hemani et al. (US20190005043 filed 06/29/2017) in view of Luo et al. US20180197087) in view of Beers et al. (US20150206069)
Regarding claim 1, Hemani teaches a content classification method (The first-vocabulary set may employ a taxonomy to classify the digital assets based on respective characteristics [0018]; digital assets (e.g., digital images) with multiple visual content classes ... are used to train a model [0156]; These extracted features are then used as a basis to train custom models (e.g., SVM classification models) using training digital assets in compliance with a first-vocabulary set [0019])
comprising the steps of: generating a plurality of first classification models by machine learning (The second vocabulary training data 204 is then employed ... to train models 212, 214 ... In an implementation, each of the models 212, 214 is trained for a respective digital asset characteristic [0039]. The Examiner notes second vocabulary training data 204 including training data digital assets 206 [0038] is used to generate classification models 212, 214 in Fig. 2) using a plurality of learning contents (input is the training digital assets 206 [0058]; training digital assets 206 having associated second-vocabulary tags [0038]) each provided with a first feature and a learning label (1: For each asset “a” in the repository: 2: Compute the set of generic tags “G(a)” ... 3: Compute a feature vector “f(a);”[0059-0061]. The Examiner notes the feature vector is a first feature and a tag is a label);
generating a second classification model (models 222, 224 [0040; Fig. 2) with the use of a plurality of feature vectors generated by the plurality of first classification models (Activations from nodes of the neural network of the models 212, 214 ... are then extracted as features 216 (e.g., as a 1024-dimensional feature vector) ... The features 216 are then employed ... to train models 222, 224 for each of the first-vocabulary tags 228 in the first-vocabulary set 114. In this way, the models 212, 214 trained in accordance with the second-vocabulary set 124 may be leveraged to generate models 222, 224 [0040]);
and performing display on a graphical user interface (user interface of a computing device [0038]; Examples of output devices include a display device (e.g., a monitor or projector) [0186]),
Hemani does not explicitly teach providing judgment data for a plurality of contents each provided with a second feature with the use of the second classification model and updating the first feature and the second feature based on a change over time predicted by a user, wherein the first feature and the second feature comprise metadata of patent number comprising at least one of a state of a family, an application type, and a number of abandoned applications in a family, wherein the learning label comprises an information of the patent number is abandoned, and wherein the plurality of first classification models model and the second classification model are configured to include the change over time by the updated first feature and the updated second feature.
Luo teaches a content classification method (Current model 110 analyzes the document contents including words, phrases, and n-grams within the document to classify and generate accurate security labels for the documents [0041]) comprising the steps of: generating a plurality of first classification models by machine learning (generating initial/current classification models [0024]; In some implementations, server 202 generates a first set of top features that are used to generate an initial/first current classification model, such as current model 110 [0063]; top features can be based on respective feature sets that include top text/word features (i.e., content data) [0064]; generating the first classification model includes using machine learning logic to train the first classification model to determine the classification of the data item [0012]) using a plurality of learning contents each provided with a first feature and a learning label (multiple documents 103, can identify that the word “Whitehouse” is a top content data feature that contributes to a top secret security label of multiple documents [0065]. The Examiner notes content data is a first feature and top secret security label is a learning label);
generating a second classification model with the use of a plurality of feature vectors (generating, by the computing system, ... the second classification model based on the identified text based content data [0010]; top features can be based on respective feature sets that include top text/word features (i.e., content data) [0064]) generated by the plurality of first classification models (a first set of top features that are used to generate an initial/first current classification model, such as current model 110 [0063]);
providing judgment data for a plurality of contents (security classification labels for new documents 224 [0054]. The Examiner notes contents refers to documents contents and instant specification discloses: “the judgment data preferably includes a classification label” (instant specification US20220027799 [0022])) each provided with a second feature with the use of the second classification model (the impact estimate is based on an impact scope that indicates, for example, the extent to which a second set of top features generated by server 202 [0066]; The calculated impact estimate (or metric) [0071]; The second classification model is generated in response to the generated impact metric [0107]) and
performing display on a graphical user interface (when a user creates a new document using device 216, model 260 is used to predict a document security label (e.g., secret, top secret, sensitive, classified, export controlled). User device 216 can cause the recommended security label to be displayed in an example application program such as MS Outlook or a related MS Office application program [0091]), and
updating the first feature and the second feature based on a change over time predicted by a user (features can correspond to document changes or modifications that have occurred over time [0098]; when a user creates new documents 224, security label prediction model 220 will be available locally for the user to execute/launch and classify/generate security labels for new documents 224 [0055]. The Examiner notes document changes or modification is an update that adds something new),
wherein the first feature and the second feature comprise metadata (top features among modified metadata attributes (e.g., document title/owner) [0067])
and wherein the plurality of first classification models and the second classification model are configured to include the change over time (the impact estimate provides a control mechanism that server 202 can use to efficiently determine an appropriate time to trigger retraining and generate a new/updated classification model [0070]) by the updated first feature and the updated second feature (features can correspond to document changes or modifications that have occurred over time [0098]; Training/modeling 308 processes the one or more extracted features that are provided by feature extraction logic 306 and can tune the first model iteration using, for example, the baseline extracted top features [0097]).
It would have been obvious to a person having ordinary skill in the art before the effective filing date of the claimed invention to have modified the method of Hemani to incorporate the teachings of Luo for the benefit of enabling a current classification model to be continually improved based on new information received (Luo [0082])
Hemani and Luo does not explicitly teach patent number comprising at least one of a state of a family, an application type, and a number of abandoned applications in a family, wherein the learning label comprises an information of the patent number is abandoned,
Beers teaches patent number comprising at least one of a state of a family, an application type, and a number of abandoned applications in a family (The system maintains a database of raw patent factors that are derived from the patent publication [0033]; Claim type “A” refers to an apparatus claim, claim type “S” to a system claim, claim type “C” to a claim for a compound, and claim type “M” refers to a method claim [0043]. The Examiner notes the apparatus claim tells us the application type),
and wherein the learning label comprises an information of the patent number is abandoned (In Table 2, legal status code refers to events during the lifetime of the patent. These include office actions, change of ownership, abandonment, maintenance and expiration [0043]),
It would have been obvious to a person having ordinary skill in the art before the effective filing date of the claimed invention to have modified the method of Hemani and Luo to incorporate the teachings of Beers for the benefit of using of machine learning when identifying input factors and computing the classification model that can be continuously updated in response to changes in the market (Beers [0038])
Regarding claim 6, Hemani, Luo and Beers teaches the content classification method according to claim 1, Luo teaches further comprising: providing the learning contents with classification data (a security classification of data items such as electronic files or documents including text and image content [0004]); and
selecting a content having the judgment data which is the same as the classification data from the plurality of contents (select top features or attributes that are identified as important to determining a particular security classification [0062]; the system generates a second classification model for determining a classification of the data item [0005]) which are provided with classification labels with the use of an output of the second classification model (classifications/security labels generated for documents created and stored within the example computer network [0030]) and
displaying the content having the judgment data on the graphical user interface (User device 216 can cause the recommended security label to be displayed in an example application program such as MS Outlook or a related MS Office application program [0091]).
The same motivation to combine independent claim 1 applies here.
Regarding claim 7, Hemani, Luo and Beers teaches the content classification method according to claim 1, Beers teaches wherein features provided for the learning contents and the contents are management parameters assigned to patent numbers (The system then proceeds to compute the input features to the classifier using the raw factors from the patent record … A list of raw factors can be found in Table 1 [0041]; The system computes the model by first computing a set of features from the electronic patent data stored in the database. The features fall into two categories. The first category is the raw factors on a patent basis from Table 1. The second are features that are computed over multiple records of patent data (i.e., over the entire set or over a subset). A list of the features considered when training the model is listed in Table 2 [0042]. The Examiner notes the raw factors are management parameters assigned to patent numbers).
The same motivation to combine independent claim 1 applies here.
Regarding claim 8, Hemani, Luo and Beers teaches the content classification method according to claim 1, Hemani teaches wherein the judgment data includes a classification label or a score (“f(m)” is a 1024 dimensional feature vector of real numbers ρ“Gm” is the set of predicted tags for the asset “m”. Each element of “Gm,” is an ordered pair (tag, confidence score) [0129-0130]).
7. Claim 9 is rejected under 35 U.S.C. 103 as being unpatentable over Hemani et al. (US20190005043 filed 06/29/2017) in view of Luo et al. (US20180197087) in view of Beers et al. (US20150206069) and further in view of Dwane et al. (US20190208056 filed 01/04/2018)
Regarding claim 9, Hemani, Luo and Beers teaches the content classification method according to claim 8, Beers teaches further comprising designating, by the graphical user interface (a user interface receiving the patent information for the target patent; … the user interface providing to a user a signal representing the estimate of patent quality [0016]; The computer system or systems that enable the user to interact with content or features can include a GUI (Graphical User Interface) [0071]),
and displaying a corresponding content in a list form (A list of raw factors can be found in Table 1 [0041]; The system computes the model by first computing a set of features from the electronic patent data stored in the database. The features fall into two categories. The first category is the raw factors on a patent basis from Table 1. The second are features that are computed over multiple records of patent data (i.e., over the entire set or over a subset). A list of the features considered when training the model is listed in Table 2 [0042]).
It would have been obvious to a person having ordinary skill in the art before the effective filing date of the claimed invention to have modified the method of Hemani and Luo to incorporate the teachings of Beers for the benefit of using of machine learning when identifying input factors and computing the classification model that can be continuously updated in response to changes in the market (Beers [0038])
Hemani, Luo and Beers does not explicitly teach designating a particular numerical range of the score
Dwane teaches designates a particular numerical range of the score and displays a corresponding content in a list form (The customer effort variable (SetllaRating) represents a customer effort score. In certain embodiments, the customer service score is based on a numerical rating such as a 1-5 rating [0044]; a user device 204 refers to an information handling system such as a personal computer, … the user device is configured to present an estimation user interface 240 [0030]; A web page is a document which is accessible via a browser which displays the web page via a display device of an information handling system [0034]. The Examiner notes the numerical rating such as a 1-5 rating is displayed in a list form)
It would have been obvious to a person having ordinary skill in the art before the effective filing date of the claimed invention to have modified the method of Hemani, Luo and Beers to incorporate the teachings of Dwane for the benefit of evaluating machine learning operations (Dwane [0062])
8. Claim 20 is rejected under 35 U.S.C. 103 as being unpatentable over Hemani et al. (US20190005043 filed 06/29/2017) in view of Luo et al. (US20180197087) in view of Beers et al. (US20150206069) and further in view of Walsh et al. (US20210011935 filed PCT filed 11/29/2018)
Regarding claim 20, Hemani, Luo and Beers teaches the content classification metho according to claim 1, Beers teaches further comprising: obtaining from each of the plurality of first classification models a judgment label or score (assign a standardized scaled score to each binary classifier of the list of binary classifiers [0009]), and
Hemani, Luo and Beers does not explicitly teach forming a plurality of feature vectors, each feature vector comprising the label or score output by a respective first classification model.
Walsh teaches forming a plurality of feature vectors, each feature vector comprising the label or score output by a respective first classification model (The reduced dimension feature vectors 416 and 418 are concatenated to generate stacked feature vector 420, which may then be evaluated by a machine learning binary classifier 422 to generate a score(ij) for the patent i and standards document j pair [0035]).
It would have been obvious to a person having ordinary skill in the art before the effective filing date of the claimed invention to have modified the method of Hemani, Luo and Beers to incorporate the teachings of Walsh for the benefit of using one or more training sets to improve the performance of the binary classifier in producing more relevant and useful scores (Walsh [0035])
Conclusion
Any inquiry concerning this communication or earlier communications from the examiner should be directed to MORIAM MOSUNMOLA GODO whose telephone number is (571)272-8670. The examiner can normally be reached Monday-Friday 8:00am-5:00pm EST.
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, Michelle T. Bechtold can be reached on (571) 431-0762. 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.
/M.G./Examiner, Art Unit 2148
Prosecution Insights