DETAILED ACTION
This Non-Final communication is in response to Application No. 18/524,493 filed 11/30/2025. The present application, filed on or after March 16, 2013, is being examined under the first inventor to file provisions of the AIA . The Request for Continued Examination and Amendment presented on 5/14/2026 which provides amendments to claims 1, 8, and 14, is hereby acknowledged. Claims 1-20 are currently pending.
Claim Rejections – Withdrawn
The previous 35 U.S.C § 112(b) rejection of claim 12 has been withdrawn.
Response to Arguments
While the scope changing amendments to the independent claims necessitate and new consideration and search, the Examiner maintains that Cox is still applicable to the claims.
Applicant asserts that Cox does not teach or suggest “monitoring, by a software module executed locally on a user device, a plurality of user interactions with one or more items on one or more interfaces, the plurality of user interactions being exclusive of item level data of the one or more items”.
However, the Examiner maintains that Cox performs a monitoring of user interactions exclusive of item level data. Specifically, Figure 8 and description describe the monitoring of user gaze activity and logging points/positions of interest within the user interface. The specific content associated with the points/positions of interest is not determined until the process of Figure 9, after several points/positions of interest are determined by the process of Figure 8. Therefore, the process of Figure 8 of Cox is monitoring, by a software module executed locally on a user device, a plurality of user interactions with one or more items on one or more interfaces, the plurality of user interactions being exclusive of item level data of the one or more items. Additionally, as shown in the updated rejections below, it would be obvious in view of Cox to capture item level data only after a user interest score exceeds a threshold because Cox captures data with several interest levels, but then a user preference can be to utilize only the captured data with a certain interest level, therefore, having the captured data be a certain interest level (threshold score) would save on storage space.
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) 1-6 and 14-19 is/are rejected under 35 U.S.C. 103 as being unpatentable over Cox et al. (US 2017/0153797 A1, hereinafter “Cox”), and further in view of Krupat et al. (US 10,401,860 B2, hereinafter “Krupat”).
Regarding claim 1, Cox teaches a computer-implemented method for determining user interest, comprising:
monitoring, by a software module executed locally on a user device, a plurality of user interactions with one or more items on one or more interfaces, the monitoring excluding item level data of the one or more items; More specifically, activity monitor 110 may monitor for user interaction behaviors that indicate the user has an interest in particular selections of content within a display interface (Cox, [0024]). User interactions with particular selections of content can include gazing and scrolling (Cox, [0023], [0038], gazing and scrolling are user interactions “excluding item level data”).
…identify user interest behavior patterns and output a user interest score; More specifically, model 116 may indicate a level of user interest for each of the content elements based on additional activity characteristics identified in association with the content elements in points of interest log 112, such as an amount of time that a user spends gazing at each content element (Cox, [0025]).
receiving, by the one or more processors, …, the user interest score based on the plurality of user interactions; determining, by the one or more processors, that the user interest score exceeds a user interest score threshold; More specifically, model 116 may categorize the amount of time that a user spends gazing at each content element into threshold categories matching the amount of time, such as a high threshold for content elements gazed at for an amount of time greater than a high threshold time, a medium threshold for content elements gazed at for an amount of time greater than a low threshold time and less than a high threshold time, and a low threshold for content elements gazed at for an amount of time less than the low threshold time (Cox, [0025]).
triggering, by the one or more processors, a capturing of the item level data by the user device of the one or more interfaces based on the user interest score …, wherein the item level data of the one or more interfaces is stored locally on the user device. More specifically, Figure 9 depicts the process of capturing the content associated with the locations of interest and according to an interest threshold that occurs after the process of Figure 8 where interest levels are identified (Cox, [0076]-[0079]).
However, Cox may not explicitly teach every aspect of
[the capturing is] based on the user interest score exceeding the user interest score threshold;
With Cox additionally disclosing that content of interest stored in the model are categorized into high, medium, and low levels of interest (Cox, [0025]), and a user specification that defines which level of interest of content to access/utilize/apply from the model for additional processes, such as flow analyzer 120 (Cox, [0030]), it would have been obvious to one of ordinary skill in the art before the effective filling date of the claimed invention given the teachings of Cox that a method for determining user interest scores with content in a user interface and capturing the content based on the user interest score would include [the capturing is] based on the user interest score exceeding the user interest score threshold. With Cox disclosing models determining interest levels of a user by monitoring user interactions with content, storing the content with the interest level, and utilizing content based on the content’s interest score meeting a threshold, one of ordinary skill in the art of implementing a method for determining user interest scores with content in a user interface and capturing the content based on the user interest score would include [the capturing is] based on the user interest score exceeding the user interest score threshold because if Cox is only going to utilize content meeting a threshold interest score, then it would be obvious to modify Cox to only store content meeting the threshold interest score resulting in saving storage space. One would therefore be motivated to combine these teachings as in doing so would create this method for determining user interest levels with content in a user interface.
However, Cox may not explicitly teach or suggest every aspect of
transmitting, by one or more processors of the user device, to a remote server device, the plurality of user interactions as inputs to a machine-learning model, wherein the machine-learning model has been trained, using one or more gathered and/or simulated sets of user interactions, to identify user interest behavior patterns and output a user interest score.
Krupat discloses a system for collecting cognitive state data (construed as user interactions) of an individual viewing website content on one device, sending the cognitive state data to another device for analysis and returning the result of the analysis to the device of the individual which is then evaluated and sent to content provider (Krupat, abstract). The cognitive state data, including facial data, collected of the individual by one device includes data collected by a camera (Krupat, col 8, line 35 – col 9, line 5). Voice data is also collected by microphone during the interaction with website content. The cognitive state data, augmented with voice data, is analyzed by a remote computer to determine emotional mood using data gathered of other individuals and a machine/deep learning neural network (Krupat, col 9, lines 6-28). Figure 3 depicts distinct devices of the individual 310, content provider 312, and aggregation machine 320. The aggregation machine analyzes the cognitive state data received from the individual device to determine emotional engagement, the emotional engagement is returned to the individual device and evaluated to be sent to the content provider (Krupat, col 12, line 23 – col 13, line 23). The returned emotional mood/engagement is construable to be a score (Krupat, col 10, line 33 – col 11, line 6; col 25, line 11-13). The machine learning that identifies the emotional mood/engagement score is trained with data of other individuals to identify trends and patterns (Krupat, col 26, lines 5-21; col 29, line 38 - col 30, line 9; col 38, lines 23-47). Therefore, Krupat discloses a system for determining user engagement scores with currently presented content, where a remote device receives user interaction data from an individual’s device without any content-identifying information and, using machine learning, determines and returns the user engagement scores to the individual’s device which is further evaluated and provided to a content provider.
It would have been obvious to one of ordinary skill in the art before the effective filling date of the claimed invention given the teachings of Cox and Krupat that a method for determining user interest levels with content in a user interface would include transmitting user interaction data to a remote server with a machine learning model to determine interest levels that are returned. With Cox and Krupat disclosing models determining interest/engagement levels of a user by monitoring user interactions with content, and with Krupat additionally disclosing that a remote device receives user cognitive state/reaction/interaction data without any content-identifying information, a machine learning model on the remote device performs the determination of a user engagement/interest score, and returning the engagement/interest score, one of ordinary skill in the art of implementing a method for determining user interest levels with content in a user interface would include transmitting user interaction data to a remote server with a machine learning model to determine interest levels that are returned because machine learning models have been shown to be a faster, more intelligent method for processing information. One would therefore be motivated to combine these teachings as in doing so would create this method for determining user interest levels with content in a user interface.
Regarding claim 2, Cox and Krupat teach the computer-implemented method of claim 1, wherein the software module executed locally on the user device is one of an extension or a plugin. More specifically, the capturing of cognitive state/interaction data is executed with a plug-in or browser extension (Krupat, col 7, lines 33-62; col 8, lines 19-34).
Regarding claim 3, Cox and Krupat teach the computer-implemented method of claim 1, wherein the plurality of user interactions comprise one or more of interactions with one or more links embedded in the one or more interfaces or scrolling the one or more interfaces. More specifically, scrolling the user interface is criteria of determining user interests (Cox, [0023]).
Regarding claim 4, Cox and Krupat teach the computer-implemented method of claim 1, wherein the item level data comprises a plurality of attributes of the one or more interfaces, text of the one or more interfaces, and/or images of the one or more interfaces. More specifically, the content of interest could be images and/or text (Cox, [0034]; Krupat, col 32, line 44 – col 33, line 11).
Regarding claim 5, Cox and Krupat teach the computer-implemented method of claim 1, further comprising associating the item level data of the one or more interfaces with a user point of interest. More specifically, each of the points of interest identified within interface is mapped to a separate content element displayed within the interface to form a model correlating each separate content element with a user interest of the particular user (Cox, abstract). Additionally, the engagement/interest can be correlated to a point in time of video content (Krupat, col 10, line 33 - col 11, line 6; col 11, lines 49-52).
Regarding claim 6, Cox and Krupat teach the computer-implemented method of claim 5, further comprising outputting, by the one or more processors, to an output device of the user, an element associated with the user point of interest. More specifically, based on the model, within a stream accessed for review by the particular user, a flow of a selection of entries of interest that meet the user interest is identified from among multiple entries in the stream. A separate selectable navigation breakpoint is selectively displayed with each of the selection of entries of interest within the stream, wherein selection of each separate selectable navigation breakpoint steps through the flow of the selection of entries of interest only (Cox, abstract). See also Krupat, Figures 4 and 6.
Regarding claims 14-19, these claims recite the system that performs the steps of the method of claims 1-6, therefore, the same rationale of rejection is applicable.
Claim(s) 7 and 20 is/are rejected under 35 U.S.C. 103 as being unpatentable over Cox and Krupat, and further in view of Jung et al. (US 2021/0144445 A1, hereinafter “Jung”).
Regarding claim 7, Cox and Krupat teach the computer-implemented method of claim 1, however, may not explicitly teach every aspect of wherein the item level data of the one or more interfaces is captured for a predetermined amount of time.
Jung discloses providing content of interest within broadcast media (Jung, abstract). The display apparatus may record a content of interest for a pre-set time and store the recorded content of interest (Jung, [0052]).
It would have been obvious to one of ordinary skill in the art before the effective filling date of the claimed invention given the teachings of Cox and Dorai with Jung that a method for determining user interest levels with content in a user interface would include capturing the content for a predetermined period of time. With Cox, Dorai, and Jung disclosing determining content of interest for a user, and with Jung additionally disclosing capturing the content for a predetermined period of time, one of ordinary skill in the art of implementing a method for determining user interest levels with content in a user interface would include capturing the content for a predetermined period of time in order to provide efficient storage management where capturing of content of interest could be continuous. One would therefore be motivated to combine these teachings as in doing so would create this method for determining user interest levels with content in a user interface.
Regarding claim 20, this claim recites the system that performs the steps of the method of claim 7, therefore, the same rationale of rejection is applicable.
Claim(s) 8-13 is/are rejected under 35 U.S.C. 103 as being unpatentable over Dorai-Raj et al. (US 2020/0226418 A1, hereinafter “Dorai”), in view of Cox, and further in view of Krupat.
Regarding claim 8, Dorai teaches a computer-implemented method for training a machine-learning model for determining user interest, comprising:
providing, by one or more processors, one or more gathered and/or simulated sets of user interactions to one or more machine-learning algorithms … as one or more sets of training data, …; More specifically, receiving a set of user interaction data indicating a group of multiple different users' interactions with one or more UI elements, generating, from the received set of user interaction data, a set of user group training data, inputting the set of user group training data to the machine learning model (Dorai, [0006]).
determining, by the one or more machine-learning algorithms, associations between the one or more gathered and/or simulated sets of user interactions and one or more user interest behavior patterns; More specifically, entities that can have a user interest score include metrics (Dorai, [0053]). A metric measures performance, behavior, and/or activity (Dorai, [0054]). UIM 110 receives user interaction data from user device 104 over network 102, and predicts user interest in digital components (Dorai, [0059]).
modifying one or more of a layer, a weight, a synapse, or a node of a machine-learning model based on the associations between the one or more gathered and/or simulated sets of user interactions and one or more user interest behavior patterns; More specifically, training data generator 108 determines training example weights in one of several ways. In some implementations, training data generator 108 can weight each training example equally, and sample only data points within a predetermined period of time. In some implementations, training data generator 108 can using decay functions to weight training examples differently depending on how recent the example is (Dorai, [0074]).
outputting, by the one or more processors, the machine-learning model, wherein the machine-learning model is trained to identify user interest behavior patterns based on a plurality of user interactions and output a user interest score based on the plurality of user interactions and the modified one or more of the layer, the weight, the synapse, or the node of the machine-learning model. More specifically, the UIM can then weight metrics more heavily in its user interest scoring model (Dorai, [0090]). The system generates, using the trained machine learning model, a set of user interest scores for the particular user, wherein each of the user interest score is indicative of the user's interest in accessing information corresponding to a UI element of the application (410). In this particular example, UIM 110 uses neural network 111a to generate user interest scores for each entity referenced by the UI element to provide an overall interest score for the UI element (Dorai, [0110]).
However, Dorai may not explicitly teach every aspect of
user interactions excluding item level data of one or more items;
triggering, by the one or more processors, a user device to capture item level data of one or more items based on a plurality of user interactions used to determine a user interest score exceeds a user interest score threshold, wherein the item level data is stored locally on the user device.
Cox discloses an activity monitor 110 may monitor for user interaction behaviors that indicate the user has an interest in particular selections of content within a display interface (Cox, [0024]). User interactions with particular selections of content can include gazing and scrolling (Cox, [0023], [0038], gazing and scrolling are user interactions “excluding item level data”). Model 116 may indicate a level of user interest for each of the content elements based on additional activity characteristics identified in association with the content elements in points of interest log 112, such as an amount of time that a user spends gazing at each content element (Cox, [0025]). Model 116 may categorize the amount of time that a user spends gazing at each content element into threshold categories matching the amount of time, such as a high threshold for content elements gazed at for an amount of time greater than a high threshold time, a medium threshold for content elements gazed at for an amount of time greater than a low threshold time and less than a high threshold time, and a low threshold for content elements gazed at for an amount of time less than the low threshold time (Cox, [0025]). A user specification that defines which level of interest of content to access/utilize/apply from the model for additional processes, such as flow analyzer 120 (Cox, [0030]). Figure 9 depicts the process of capturing the specific content on the local device associated with the position/points of interest within the user interface and logged according to an interest threshold that occurs after the process of Figure 8 where only interest levels in locations/points are identified (Cox, [0076]-[0079]).
It would have been obvious to one of ordinary skill in the art before the effective filling date of the claimed invention given the teachings of Dorai and Cox that a method for training a machine learning model with user interactions for determining user interest levels with content in a user interface would include wherein the plurality of user interactions is exclusive of item level data of the one or more interfaces, and the triggering of capturing of the item level data to the user devce when the user interest score exceeds a user interest threshold. With Dorai and Cox disclosing models determining interest levels of a user by monitoring user interactions with content, and with Cox additionally disclosing the interest level in a position/point on a user interface is determined first, then the content/item detail for the interest point is determined, and user preferences that specify a certain level of interest to utilize for further processing, one of ordinary skill in the art of implementing a method for training a machine learning model with user interactions for determining user interest levels with content in a user interface would include wherein the plurality of user interactions is exclusive of item level data of the one or more interfaces, and the triggering of capturing of the item level data to the user device when the user interest score exceeds a user interest threshold because the system would benefit from not having to capture the content until interest is definitively determined, saving storage space and processing time. One would therefore be motivated to combine these teachings as in doing so would create this method for training a machine learning model with user interactions for determining user interest levels with content in a user interface.
However, Dorai and Cox may not explicitly teach every aspect of
[the] user interactions excluding item level data of one or more items [are provided to] one or more machine learning algorithms of a remote server device;
Krupat discloses a system for collecting cognitive state data (construed as user interactions) of an individual viewing website content on one device, sending the cognitive state data to another device for analysis and returning the result of the analysis to the device of the individual and then evaluated and sent to content provider (Krupat, abstract). The cognitive state data, including facial data, collected of the individual by one device includes data collected by a camera (Krupat, col 8, line 35 – col 9, line 5). Voice data is also collected by microphone during the interaction with website content. The cognitive state data, augmented with voice data, is analyzed by a remote computer to determine emotional mood using data gathered of other individuals and a machine/deep learning neural network (Krupat, col 9, lines 6-28). Figure 3 depicts distinct devices of the individual 310, content provider 312, and aggregation machine 320. The aggregation machine analyzes the cognitive state data received from the individual device to determine emotional engagement, the emotional engagement is returned to the individual device and evaluated to be sent to the content provider (Krupat, col 12, line 23 – col 13, line 23). The returned emotional mood/engagement is construable to be a score (Krupat, col 10, line 33 – col 11, line 6; col 25, line 11-13). The machine learning that identifies the emotional mood/engagement score is trained with data of other individuals to identify trends and patterns (Krupat, col 26, lines 5-21; col 29, line 38 - col 30, line 9; col 38, lines 23-47). Therefore, Krupat discloses a system for determining user engagement scores with currently presented content, where a remote device receives user interaction data from an individual’s device without any content-identifying information and, using machine learning, determines and returns the user engagement scores to the individual’s device which is further evaluated and provided to a content provider.
It would have been obvious to one of ordinary skill in the art before the effective filling date of the claimed invention given the teachings of Dorai and Cox with Krupat that a method for training a machine-learning model for determining user interest would include [the] user interactions excluding item level data of one or more items [are provided to] one or more machine learning algorithms of a remote server device. With Dorai, Cox, and Krupat disclosing models determining interest/engagement levels of a user by monitoring user reactions/interactions with content, and with Krupat additionally disclosing that a remote device receives user cognitive state/reaction/interaction data without any content-identifying information, a machine learning model on the remote device performs the determination of a user engagement/interest score and returns the engagement/interest score, one of ordinary skill in the art of implementing a method for training a machine-learning model for determining user interest would include [the] user interactions excluding item level data of one or more items [are provided to] one or more machine learning algorithms of a remote server device because a remote machine learning model can be utilized by other systems for more purposes and continuously be updated, which have been shown to be a faster, more intelligent method for processing information. One would therefore be motivated to combine these teachings as in doing so would create this method for determining user interest levels with content in a user interface.
Regarding claim 9, Dorai and Cox with Krupat teach the computer-implemented method of claim 8, wherein the one or more gathered and/or simulated sets of user interactions comprise one or more of interactions with one or more links embedded in one or more interfaces or scrolling the one or more interfaces. More specifically, user interaction data can come in other forms, such as event data (e.g., application listeners track events such as mouse clicks, scrolling) (Dorai, [0040]). The action of scrolling such that content item 204c is within viewport 202 for at least a predetermined period of time can be considered a positive indication of interest in content item 204c (Dorai, [0102]).
Regarding claim 10, Dorai and Cox with Krupat teach the computer-implemented method of claim 8, further comprising monitoring, by a software module stored locally on a user device, the plurality of user interactions with one or more interfaces, wherein the plurality of user interactions is exclusive of item level data of the one or more interfaces. More specifically, Figure 9 depicts the process of capturing the specific content on the local device associated with the position/points of interest within the user interface and logged according to an interest threshold that occurs after the process of Figure 8 where only interest levels in locations/points are identified (Cox, [0076]-[0079]).
Regarding claim 11, Dorai and Cox with Krupat teach the computer-implemented method of claim 10, further comprising: providing, by one or more processors, the plurality of user interactions to the machine-learning model; and outputting, by the machine-learning model, the user interest score based on the plurality of user interactions. More specifically, the UIM can then weight metrics more heavily in its user interest scoring model (Dorai, [0090]). The system generates, using the trained machine learning model, a set of user interest scores for the particular user, wherein each of the user interest score is indicative of the user's interest in accessing information corresponding to a UI element of the application (410). In this particular example, UIM 110 uses neural network 111a to generate user interest scores for each entity referenced by the UI element to provide an overall interest score for the UI element (Dorai, [0110]). Additionally, model 116 may indicate a level of user interest for each of the content elements based on additional activity characteristics identified in association with the content elements in points of interest log 112, such as an amount of time that a user spends gazing at each content element (Cox, [0025]). Further, Figure 3 depicts distinct devices of the individual 310, content provider 312, and aggregation machine 320. The aggregation machine analyzes the cognitive state data received from the individual device to determine emotional engagement, the emotional engagement is returned to the individual device and evaluated to be sent to the content provider (Krupat, col 12, line 23 – col 13, line 23). The returned emotional mood/engagement is construable to be a score (Krupat, col 10, line 33 – col 11, line 6; col 25, line 11-13).
Regarding claim 12, Dorai and Cox with Krupat teach the computer-implemented method of claim 11, further comprising triggering, by the one or more processors, a capturing of the item level data of the one or more interfaces based on the user interest score exceeding a user interest score threshold, wherein the item level data of the one or more interfaces is stored locally on the user device. More specifically, with Cox additionally disclosing that content of interest stored in the model are categorized into high, medium, and low levels of interest (Cox, [0025]), and a user specification that defines which level of interest of content to access/utilize/apply from the model for additional processes, such as flow analyzer 120 (Cox, [0030]), it would have been obvious to one of ordinary skill in the art before the effective filling date of the claimed invention given the teachings of Cox that a method for determining user interest scores with content in a user interface and capturing the content based on the user interest score would include [the capturing is] based on the user interest score exceeding the user interest score threshold as shown above with respect to claim 1.
Regarding claim 13, Dorai and Cox with Krupat teach the computer-implemented method of claim 8, further comprising gathering, by a second machine-learning model, the one or more gathered and/or simulated sets of user interactions. More specifically it would be obvious that more than one model can be involved as suggested by Cox at [0010], [0076], [0108] is order to better structure the allocation of functions. Further, a plurality of neural networks are suggested for use in determining the user engagement/emotion score (Krupat, col 36, line 1 – col 40, line 30).
Pertinent Prior Art
The prior art made of record on form PTO-892 and not relied upon is considered pertinent to applicant's disclosure. Applicant is required under 37 C.F.R. § 1.111(c) to consider these references fully when responding to this action.
Lee (US 2015/0089520 A1) – determining a user’s level of interest in content by monitoring user interactions using machine learning.
Rivera (US 8,943,526 B2) – determining a user’s level of engagement with presented content by analyzing non-content data sensed about the user.
Hammit (US 2020/0327398 A1) – determining a user’s level of engagement with presented content by analyzing non-content data sensed about the user with a neural network.
Sloo (US 2020/0327398 A1) – determining a user’s level of interest and whether it satisfies a criterion, and marking content of interest accordingly.
Conclusion
Any inquiry concerning this communication or earlier communications from the examiner should be directed to PATRICK F RIEGLER whose telephone number is (571)270-3625. The examiner can normally be reached M-F 9:30am-6:00pm, ET.
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, Kieu Vu can be reached at (571) 272-4057. 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.
/PATRICK F RIEGLER/ Primary Examiner, Art Unit 2171
Prosecution Insights