DETAILED ACTION
The present application, filed on or after March 16, 2013, is being examined under the first inventor to file provisions of the AIA .
Status of Claims
This Final Office Action is in reply to the application filed on 13 November 2025
Claims 1-20 are currently pending and have been examined.
Previous Claim Rejections - 35 USC § 101
The examiner withdraws the 35 U.S.C. 101 of claims 1-20. The independent claims have been amended to include the additional elements, generating a selectable sub-item based on the confidence metric, the sub-item and the label; making the selectable sub-item user-selectable at the user device; receiving at least one second electronic communication from the user device, the at least one second electronic communication comprising user input relating to a selection of the selectable sub-item; and determining a moderation outcome based on the at least one second electronic communication including the selection of the selectable sub-item.
As described in the specification [0049] Similarly, if the machine learning model output indicates that the image sub-item includes nudity or obscenity, sub-item generator 152 sends instructions to reporting system 114 to display a bounding box around the image sub-item and render the image content within the bounding box user-selectable. Paragraph [0067]), “The image segmentation model processes the image to determine if the image includes content that is harmful or malicious. In some embodiments, the image segmentation model identifies objects by classifying objects using training that is performed using a set of training images and object labels. The image segmentation model outputs a set of objects and corresponding confidence values that an object is identified accurately. Each object is identified by bounding box data. Bounding box data includes, for example, two- or three-dimensional coordinates that identify the outer boundaries of the portion of the image that contains the identified object. Paragraph [0070] At operation 320, the sub-item generator 152 requests additional information from the user system. In some embodiments, the sub-item generator 152 provides at least one bounding box (e.g., a sub-item) that includes an identified object. The sub-item generator 152 requests 152 requests the user system to identify the object of the electronic communication that is identified as harmful or malicious. As illustrated by FIG. 3, selectable sub-items include the bounding boxes 320A and 320B and are provided to the user system with bounding box 320A including an overexposed human (e.g., nudity) and bounding box 320B including a person but not a harmful object.
The additional elements when considered in combination integrate the abstract idea into a practical application because they provide meaningful limitations that improve the machine learning-assisted reporting system.
Claim Rejections - 35 USC § 103
In the event the determination of the status of the application as subject to AIA 35 U.S.C. 102 and 103 (or as subject to pre-AIA 35 U.S.C. 102 and 103) is incorrect, any correction of the statutory basis (i.e., changing from AIA to pre-AIA ) for the rejection will not be considered a new ground of rejection if the prior art relied upon, and the rationale supporting the rejection, would be the same under either status.
The following is a quotation of 35 U.S.C. 103 which forms the basis for all obviousness rejections set forth in this Office action:
A patent for a claimed invention may not be obtained, notwithstanding that the claimed invention is not identically disclosed as set forth in section 102, if the differences between the claimed invention and the prior art are such that the claimed invention as a whole would have been obvious before the effective filing date of the claimed invention to a person having ordinary skill in the art to which the claimed invention pertains. Patentability shall not be negated by the manner in which the invention was made.
Claim(s) 1-2, 4, 6-8, 11-12, 14, 16-17 and 20 are rejected under 35 U.S.C. 103 as being unpatentable over Rubinstein et al (US 2014/0365382 A1, hereinafter “Rubinstein”) in view of Lindsay (US 2023/0413019 A1) in view of Stroganov (US 2016/0239493 A1).
Claims 1 and 11: Rubinstein discloses a method and a system comprising: at least one memory device; and a processing device, operatively coupled to the at least one memory device, to (see [0049]: computing system and components to perform the operations):
receiving a first electronic communication from a user device, the first electronic communication comprising a user-generated submission of a content item (See [0022]: A user connected through a user device 100, to the social networking system 101, may also create a content report to complain about content posted by another user on the social networking system 100. The social networking system 101 initiates a content report management process to deal with such content reports, as described in more detail herein);
selecting a trained machine learning model from a plurality of trained machine learning models based on the user-generated submission (see [0037]: The machine-learning module 111 may provide different machine-learned algorithms to calculate CS based on the type of inappropriate content identified by the reporting user. For example, different machine-learned algorithms may be selected to generate CS based on the content being identified as a spam photo versus a racist photo.);
applying the selected trained machine learning model to the content item (see [0037]: using machine-learned algorithms that use social data to determine the probability that content is inappropriate. The machine-learned algorithms may be trained by the machine-learning module 111. The machine-learned algorithms are trained to recognize the characteristics of content that are correlated with various sorts of inappropriate content like pornography, spam, etc.);
receiving at least one second electronic communication from the user device, and determining a moderation outcome based on the at least one second electronic communication (see [0041]: The content review manager 105 then requests 212 that the reporting user reconfirm the content report. If the reporting user does not reconfirm the report, then the content review process is terminated, and the content report is considered resolved. If the reporting user does reconfirm the content report, then RC is queued for manual review by a human agent) .
Rubinstein discloses, [0041] If CS is less than the minimum confidence threshold, then the content review manager 105 provides 212 educational information to the reporting user that reminds the reporting user about what types of content are inappropriate. The content review manager 105 then requests 212 that the reporting user reconfirm the content report. If the reporting user does not reconfirm the report, then the content review process is terminated, and the content report is considered resolved. If the reporting user does reconfirm the content report, then RC is queued for manual review by a human agent. Rubinstein does not expressly disclose receiving, from the selected trained machine learning model, output that (i) identifies at least one sub-item of the content item and (ii) comprises, for a sub-item, a label associated with the sub-item and a confidence metric associated with the label; based on the confidence metric, making the sub-item user-selectable at the user device, the at least one second electronic communication comprising user input relating to the sub-item but Lindsay which also discloses a content moderation system teaches, receiving, from the selected trained machine learning model, output that (i) identifies at least one sub-item of the content item and (ii) comprises, for a sub-item, a label associated with the sub-item and a confidence metric associated with the label; (See [0076]: By way of example, with some embodiments, the video clip may be processed using a pre-trained machine learned model that outputs a score indicating a probability that images within the video clip include objectionable content (e.g., nudity, obscene hand gestures, etc.). . If, for example, a probability score representing a likelihood that an image within the video is objectionable exceeds some threshold, then the relevant video clip may be directed to the review queue for review and approval by the admin end-user.), the at least one second electronic communication comprising user input relating to the sub-item (see [0078]: review and approval by the admin end-user).
It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to include in the content report management of Rubinstein, receiving, from the selected trained machine learning model, output that (i) identifies at least one sub-item of the content item and (ii) comprises, for a sub-item, a label associated with the sub-item and a confidence metric associated with the label as taught by Lindsay “to ensure that outbound messages, created by agent end-users, comply with quality control requirements” (Lindsay, [0019]).
Rubinstein and Lindsay do not expressly disclose generating a selectable sub-item based on the confidence metric, the sub-item and the label; making the selectable sub-item user-selectable at the user device; receiving at least one second electronic communication from the user device, the at least one second electronic communication comprising user input relating to a selection of the selectable sub-item; and determining a moderation outcome based on the at least one second electronic communication including the selection of the selectable sub-item but Stroganov in the same field of endeavor teaches, generating a selectable sub-item based on the confidence metric, the sub-item and the label; making the selectable sub-item user-selectable at the user device; receiving at least one second electronic communication from the user device, the at least one second electronic communication comprising user input relating to a selection of the selectable sub-item; and determining a moderation outcome based on the at least one second electronic communication including the selection of the selectable sub-item (see Fig. 5, [0052]: the video moderation server 108 selects content items from the content statistics database 107, which have largest content suspicion scores calculated at step 304 of the process 300. At step 403, the moderation server 108 creates an HTML report with ordered list of selected content items ranked in accordance with the calculated content suspicion scores. It should be noted that the invention is not limited to only HTML reports and reports of any other encoding, format or type may be used. This report is transmitted to the moderator using the network, such as Internet. The moderator reviews the report and manually classifies the content. Moderator's decision is sent back to the moderation server 108 and analyzed at step 404. If the content is determined by the moderator to have the pornographic material, the content is blocked in the content database 202, by means of sending the content block request 212 to the database 202 illustrated in FIG. 2, see step 405. [0053]: incorporates a ranked list of content items 501, which are to be reviewed by the administrator (moderator). To this end, the administrator (moderator) is provided with controls 503 for playing the listed content items. The administrator (moderator) marks the content as approved or not approved (e.g. containing pornography, blocked) using the control elements 505 and 504, respectively, see FIG. 5. To assist the administrator (moderator) in content review, the graphical user interface 500 additionally incorporates the content rankings 506 as well as calculated content suspicion scores 502.)
Therefore, it would have been obvious to one of ordinary skill in the art before the effective filing date to combine Rubinstein, and Lindsay with the system and method of generating a selectable sub-item based on the confidence metric, the sub-item and the label; making the selectable sub-item user-selectable at the user device; receiving at least one second electronic communication from the user device, the at least one second electronic communication comprising user input relating to a selection of the selectable sub-item; and determining a moderation outcome based on the at least one second electronic communication including the selection of the selectable sub-item as taught by Stroganov because it would help determine how the content is handled in the future (Stroganov, [0050]).
Claims 2 and 12: The combination of Rubinstein, Lindsay and Stroganov discloses the claimed invention as applied to claims 1 and 11 above. Rubinstein further teaches, wherein the content item comprises at least one of text, image, video, or multimodal content that is extracted from at least one of a user profile, a feed, a post, or a message (See [0022], [0028]: content posted include photo).
Claims 4 and 14: The combination of Rubinstein, Lindsay and Stroganov discloses the claimed invention as applied to claims 1 and 11 above. Rubinstein further teaches wherein applying the selected trained machine learning model to the content item comprises generating a classification of the content item and a confidence score associated with the classification (See [0029] The confidence score generator 110 may utilize the services of the machine-learning module 111 to determine the probability that a particular piece of content is inappropriate content. The machine-learning module 111 trains machine-learned models that can generate the confidence score for the reported content).
Claim 6 and 16: The combination of Rubinstein, Lindsay and Stroganov discloses the claimed invention as applied to claims 1 and 11 above. Rubinstein further teaches, further comprising at least one of: removing an access of the user device to the sub-item; and/or sending a notification to the user device (See [0040]: If RC is determined to be inappropriate the human agent may delete RC. In addition, the human agent may enroll the content owner in a "timeout" administered through the access restrictions manager 112. The timeout administered through the access restrictions manager 112 restricts the content owner's ability to post new content for a period of time. [0039]: Requests like this, as well as others, from the social networking system 101 may be communicated to the content owner through the social networking system interface managed by the user interface manager 103, or they may be sent through another form such as email, SMS message, etc.).
Claims 7 and 20: The combination of Rubinstein, Lindsay and Stroganov discloses the claimed invention as applied to claims 1 and 11 above. Rubinstein further teaches, restricting an access of the user device to the sub-item; and/or generating an escalation of review (see [0040]: If RC is determined to be inappropriate the human agent may delete RC. In addition, the human agent may enroll the content owner in a "timeout" administered through the access restrictions manager 112. The timeout administered through the access restrictions manager 112 restricts the content owner's ability to post new content for a period of time. [0041]: If the reporting user does reconfirm the content report, then RC is queued for manual review by a human agent as described above).
Claim 8 and 17: The combination of Rubinstein, Lindsay and Stroganov discloses the claimed invention as applied to claims 1 and 11 above. Rubinstein further discloses, generating a request for additional information, wherein the request for additional information comprises a prompt for the at least one second electronic communication (See [0041]: content review manager 105 then requests 212 that the reporting user reconfirm the content report). Lindsay further teaches, additional information about the sub-item (see [0078]: If, for example, a probability score representing a likelihood that an image within the video is objectionable exceeds some threshold, then the relevant video clip may be directed to the review queue for review and approval by the admin end-user).
Claim(s) 3, 5, 9-10, 13, 15, 18 and 19 are rejected under 35 U.S.C. 103 as being unpatentable over Rubinstein, Lindsay and Stroganov as applied to claims 2 and 12 above, and further in view of Zhdanoz et al (US Patent #11,501,210 B1).
Claims 3 and 13: The combination of Rubinstein, Lindsay and Stroganov discloses the claimed invention as applied to claims 1 and 11 above. Rubinstein discloses [0037]: The machine-learning module 111 may provide different machine-learned algorithms to calculate CS based on the type of inappropriate content identified by the reporting user. For example, different machine-learned algorithms may be selected to generate CS based on the content being identified as a spam photo versus a racist photo. Rubinstein, Lindsay and Stroganov do not expressly disclose wherein selecting the trained machine learning model from the plurality of trained machine learning models comprises at least one of: selecting a text classification model for the text; or selecting an image segmentation model for the image but Zhdanov in the same field of endeavor teaches, herein selecting the trained machine learning model from the plurality of trained machine learning models comprises at least one of: selecting a text classification model for the text; or selecting an image segmentation model for the image (See col. 2 lines 38-47: For example, the predictions may include text classification or labeling (e.g., assigning tags, categorizing text, mining text, etc.), image classification (e.g., categorizing images into classes), object detection (e.g., locating objects in images via bounding boxes), or semantic segmentation (e.g., locating objects in images with pixel-level precision) associated with the content. Col. 11 lines 23-30: To perform the request of the user 102, the content review service 106 may include various components, such as a text analysis component 140, an image analysis component 142, and a threshold component 144. In some instances, based on the request of the user 102 and/or the content 114 being analyzed, the content review service 106 may select a corresponding component. In some instances, the component may be determined based on the task(s) 136 to be performed.).
Therefore, it would have been obvious to one of ordinary skill in the art before the effective filing date to combine Rubinstein, Lindsay and Stroganov with the system and method of selecting a text classification model for the text; and/or selecting an image segmentation model for the image as taught by Zhdanov with the motivation of “accurately inferring or predicting subject matter, fields of interest, or material within various forms of content, such as text, images, videos, or audio” (Zhdanov, col. 2 lines 35-38).
Claims 5 and 15: The combination of Rubinstein, Lindsay and Stroganov discloses the claimed invention as applied to claims 4 and 14 above. Rubinstein, Lindsay and Stroganov do not expressly disclose, wherein generating a classification of the content item and a confidence score associated with the classification comprises: cropping the image to a region of interest within the image; and computing a classification and a confidence score for the region of interest but Zhdanov in the same field of endeavor teaches wherein generating a classification of the content item and a confidence score associated with the classification comprises: cropping the image to a region of interest within the image; and computing a classification and a confidence score for the region of interest (see col. 3 lines 30-45: he ML models may perform image classification, box bounding, semantic segmentation, label verification, and so forth. As a first operation, the ML models may determine whether the image contains foxes and if so, may draw a box (e.g., bounding box) around all the individual foxes (i.e., each fox may be represented with a bounding box). The bounding boxes may be used to identify a position of the objects of interest within the image. The ML models may determine a confidence score associated with bounding boxes being around all the foxes in the image. After drawing a box around all the fox(es), the ML model may determine whether all the foxes have a box. If so, the ML model may determine a confidence score associated with all the fox(es) in the image being represented within a bounding box).
Therefore, it would have been obvious to one of ordinary skill in the art before the effective filing date to combine Rubinstein, Lindsay and Stroganov with the system and method of cropping the image to a region of interest within the image; and computing a classification and a confidence score for the region of interest as taught by Zhdanov in order to “accurately inferring or predicting subject matter, fields of interest, or material within various forms of content, such as text, images, videos, or audio” (Zhdanov, col. 2 lines 35-38).
Claims 9 and 18: The combination of Rubinstein, Lindsay and Stroganov discloses the claimed invention as applied to claims 4 and 14 above. Rubinstein, Lindsay and Stroganov do not expressly disclose, wherein the prompt for the at least one second electronic communication includes a set of selectable sub-items, the prompt further comprising: requesting a user selection of at least one selectable sub-item of the set of selectable sub-items; and comparing the user selection of at least one selectable sub-item with the label associated with the sub-item and a confidence metric associated with the label but Zhdanov in the same field of endeavor teaches, wherein the prompt for the at least one second electronic communication includes a set of selectable sub-items, the prompt further comprising: requesting a user selection of at least one selectable sub-item of the set of selectable sub-items; and comparing the user selection of at least one selectable sub-item with the label associated with the sub-item and a confidence metric associated with the label (see col. 15 lines 25-38: Displaying the reviews and the content may include highlighting or otherwise indicating (e.g., boxes, outlines, etc.), within the content 114, where the reviewer 104 is to review the fields of interest or what the reviewer 104 is to review. Such indications may assist the reviewer 104 in locating his or her reviews within the content 114 for verifying or adjusting the results (e.g., predictions) of the content review service 106. For example, in the example of locating offensive language, the content 114 (e.g., document) or portion of the content 114 that allegedly contains the offensive language may be presented on the display 156. Also on the display 156, the term, object, symbol, text, field of interest etc. that the ML model(s) 126 predicted below the confidence level may be displayed with a box, outline, or highlight. Col. 16 lines 20-28: After reviewing the review(s), the reviewer 104 may transmit the review(s) to the content review service 106. The content review service 106 may utilize the review(s), or the results of the review(s) to further training the ML model(s) 126 via a training component 158. For example, the review(s) received from the reviewer 104 may indicate whether the item(s) predicted by the ML model(s) 126, as corresponding to the request of the user 102, were correct or incorrect. Col. 22 lines 1-15: In some instances, the reviews performed by the reviewer may be used to update the confidence associated with the ML model. For example, if the results provided by the reviewer match the results (or prediction) of the ML model(s), the confidence of the ML model(s) may be increased. ).
Therefore, it would have been obvious to one of ordinary skill in the art before the effective filing date to combine Rubinstein, Lindsay and Stroganov with the system and method of wherein the prompt for the at least one second electronic communication includes a set of selectable sub-items, the prompt further comprising: requesting a user selection of at least one selectable sub-item of the set of selectable sub-items; and comparing the user selection of at least one selectable sub-item with the label associated with the sub-item and a confidence metric associated with the label as taught by Zhdanov with the motivation of improving the accuracy and quality of the machine learning models (Zhdanov, col. 21 lines 56-57).
Claims 10 and 19: The combination of Rubinstein, Lindsay and Zhdanov discloses the claimed invention as applied to claims 9 and 18 above. Lindsay further teaches, wherein the at least one selectable sub-item includes a portion of text or an object depicted in an image (see [0078]: the relevant video clip may be directed to the review queue for review and approval by the admin end-user).
Response to Arguments
Applicant’s arguments with respect to the 35 USC 103 rejections of claims 1-20 have been considered but are moot because the new ground of rejection does not rely on any reference applied in the prior rejection of record for any teaching or matter specifically challenged in the argument.
Conclusion
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 MAAME BALLOU whose telephone number is (571)270-1359. The examiner can normally be reached Monday-Friday 9am-5pm.
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, Lynda Jasmin can be reached at 571-272-6782. 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.
MAAME BALLOU
Examiner
Art Unit 3629
/MAAME BALLOU/ Examiner, Art Unit 3629
/JESSICA LEMIEUX/ Supervisory Patent Examiner, Art Unit 3626