DETAILED ACTION 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. Drawings The drawings are objected to as failing to comply with 37 CFR 1.84(p)(4) because reference character “ 308 ” has been used to designate “Natural Language Processing Model” , “Embedding Theme Classifying Machine-Learning Model” , “Theme Classification”, “Performing An Action Based On The Theme Classification”, “Determining An Action Based On The Theme Classification”, “Predict An Increase In Experience Score”, and “Determining An Action Based On The Subtheme Classification” , and reference character “ 3 14” has been used to designate “Unstructured Text”, “Classification Embeddings”, “Loss Function”, “Model Fitting”, and “Training Theme Classification” . The drawings are objected to as failing to comply with 37 CFR 1.84(p)(5) because they do not include the following reference signs mentioned in the description : “ environment 100” in paragraph 0041, line 5, paragraph 0042, line 1, paragraph 0045, lines 1 and 5, and paragraph 0048, lines 1 and 5 “ experience score 512” in paragraph 0005, lines 2-3 . The drawings are objected to as failing to comply with 37 CFR 1.84(p)(5) because they include the following reference characters not mentioned in the description: “Classification Embeddings 314” in Figure 3 “Suggest A Subtheme 408” in Figure 4 . The drawings are objected to because : In Figure 1, “Client Application 110n” should read “Client Application 112a”. In Figure 3, “Natural Language Processing Model 308” should read “Natural Language Processing Model 304”. In Figure 3, “Unstructured Text 314” should read “Unstructured Text 302”. In Figure 3, “Loss Function 314” should read “Loss Function 316”. In Figure 3, “Model Fitting 314” should read “Model Fitting 318”. In Figure 5, “Theme Classification 308” should read “Theme Classification 502”. In Figure 5, “Performing An Action Based On The Theme Classification 308” should read “Performing An Action Based On The Theme Classification 504”. In Figure 5, “Determining An Action Based On The Theme Classification 308” should read “Determining An Action Based On The Theme Classification 506”. In Figure 5, “Predict An Increase In Experience Score 308” should read “Predict An Increase In Experience Score 508”. Corrected drawing sheets in compliance with 37 CFR 1.121(d) are required in reply to the Office action to avoid abandonment of the application. Any amended replacement drawing sheet should include all of the figures appearing on the immediate prior version of the sheet, even if only one figure is being amended. The figure or figure number of an amended drawing should not be labeled as “amended.” If a drawing figure is to be canceled, the appropriate figure must be removed from the replacement sheet, and where necessary, the remaining figures must be renumbered and appropriate changes made to the brief description of the several views of the drawings for consistency. Additional replacement sheets may be necessary to show the renumbering of the remaining figures. Each drawing sheet submitted after the filing date of an application must be labeled in the top margin as either “Replacement Sheet” or “New Sheet” pursuant to 37 CFR 1.121(d). If the changes are not accepted by the examiner, the applicant will be notified and informed of any required corrective action in the next Office action. The objection to the drawings will not be held in abeyance. Specification The disclosure is objected to because of the following informalities: In paragraph 0031, lines 6-7, the meaning of “the natural language processing machine-learning model is trained on large” is not clear. In paragraph 0050, lines 7-8, “an option to select a number reflecting likely they are to recommend” should read “an option to select a number reflecting how likely they are to recommend”. In paragraph 0070, line 6, “feature” should read “feature.”. In paragraph 0071, line 6, “regularization” should read “regularization.”. In paragraph 0099, line 4, “client application 110n” should read “client application 112a”. Appropriate correction is required. Claim Rejections - 35 USC § 102 The following is a quotation of the appropriate paragraphs of 35 U.S.C. 102 that form the basis for the rejections under this section made in this Office action: A person shall be entitled to a patent unless – (a)(1) the claimed invention was patented, described in a printed publication, or in public use, on sale , or otherwise available to the public before the effective filing date of the claimed invention. (a)(2) the claimed invention was described in a patent issued under section 151, or in an application for patent published or deemed published under section 122(b), in which the patent or application, as the case may be, names another inventor and was effectively filed before the effective filing date of the claimed invention. Claims 1 , 10 and 16 are rejected under 35 U.S.C. 102 (a)(2) as being anticipated by Shi et al. (US Patent No. 12 , 346 , 940 ), hereinafter Shi . Regarding claim 1, Shi discloses a computer-implemented method comprising: receiving experience data comprising an experience score associated with unstructured text ( Column 1, lines 23-29, "The system can determine a group of input data, comprising a group of score data indicative of respective ratings of respective experiences of respective users with a website, a corresponding group of text data indicative of respective accounts of the respective experiences of the respective users with the website" ; A group of score data indicative of respective ratings of respective experiences of respective users reads on an experience score , and a corresponding group of text data indicative of respective accounts of the respective experiences of the respective users reads on unstructured text .); providing the unstructured text to a theme-classifying machine-learning model, the theme-classifying machine-learning model trained to classify unstructured text into at least one theme from a defined taxonomy of themes ( Column 11, lines 6-13, " A machine learning model (e.g., as implemented by quantifying user experience component 402 of FIG. 4) can predict from a user's verbatim (e.g., verbatim 212 of FIG. 2) to a question of, “What would you like to see improved on the website?” and classify the verbatim into intent/intuition categories based on word choice, syntax, and/or tone. In some examples, such as in system architecture 500, there can be four categories"; Column 13, lines 28-46, "Operation 606 depicts categorizing respective text data of the corresponding group of text data based on an intent of respective intents of the respective accounts of the respective experiences, to produce a group of categorized text data. That is, verbatims can be processed to be categorized based on intent, similar to as discussed with respect to system architecture 500 of FIG. 5. In some examples, the categorizing of the respective text data comprises categorizing the respective text data as one of a complaint, a suggestion, a query, or a compliment. These are example categories, and can be distinguished from a sentiment categorization (e.g., positive, neutral, or negative). In some examples, operation 606 comprises training a machine learning model with supervised learning, wherein input data to the machine learning model comprising respective learning text and corresponding respective category labels. That is, a machine learning model can be trained with a labelled data set to categorize verbatims." ; A machine learning model classify ing a user’s verbatim int o four categories reads on providing the unstructured text to a theme-classifying machine-learning model , the theme-classifying machine-learning model trained to classify unstructured text into at least one theme from a defined taxonomy of themes , where the four intent categories of a complaint, a suggestion, a query, and a compliment read on a defined taxonomy of themes , and training a machine learning model with learning text and corresponding respective category labels reads on the theme-classifying machine-learning model being trained to classify unstructured text .) ; generating, using the theme-classifying machine-learning model, a theme classification for the unstructured text ( Column 13, lines 28-46, "Operation 606 depicts categorizing respective text data of the corresponding group of text data based on an intent of respective intents of the respective accounts of the respective experiences, to produce a group of categorized text data. That is, verbatims can be processed to be categorized based on intent, similar to as discussed with respect to system architecture 500 of FIG. 5. In some examples, the categorizing of the respective text data comprises categorizing the respective text data as one of a complaint, a suggestion, a query, or a compliment. These are example categories, and can be distinguished from a sentiment categorization (e.g., positive, neutral, or negative). In some examples, operation 606 comprises training a machine learning model with supervised learning, wherein input data to the machine learning model comprising respective learning text and corresponding respective category labels. That is, a machine learning model can be trained with a labelled data set to categorize verbatims." ; C ategorizing t ext data as one of a complaint, a suggestion, a query, or a compliment reads on generating a theme classification for unstructured text .); and associating the theme classification for the unstructured text with the experience score ( Column 1, lines 38-45, "The system can further identify a binary sparse matrix of the binary sparse matrices that corresponds to score data of the group of score data that satisfies a defined score criterion, to produce an identified binary sparse matrix. The system can further store an indication of the identified binary sparse matrix, with categorized text data of the group of categorized text data corresponding to the identified binary sparse matrix." ; I dentify ing a binary sparse matrix corresponding to the score data and stor ing the binary sparse matrix corresponding to the categorized text data reads on associating the theme classification for the unstructured text with the experience score .). Regarding claim 10, arguments analogous to claim 1 are applicable. In addition, Shi discloses a non-transitory computer-readable medium storing instructions (Column 1, lines 62-65, “ An example non-transitory computer-readable medium can comprise instructions that, in response to execution, cause a system comprising a processor to perform operations. ”) that, when executed by at least one processor (Column 18, lines 59-61, “ The processing unit 1004 can be any of various commercially available processors. ”) , cause a computer system to perform the steps of claim 1. Regarding claim 16, arguments analogous to claim 1 are applicable. In addition, Shi discloses a system comprising: at least one processor (Column 18, lines 59-61, “ The processing unit 1004 can be any of various commercially available processors. ”) ; and at least one non-transitory computer-readable storage medium storing instructions (Column 1, lines 62-65, “ An example non-transitory computer-readable medium can comprise instructions that, in response to execution, cause a system comprising a processor to perform operations. ”) that, when executed by the at least one processor, cause the system to perform the steps of claim 1. 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 2 , 9, 11 – 13 and 17 are rejected under 35 U.S.C. 103 as being unpatentable over Shi in view of Mathew et al. (US Patent No. 10 , 372 , 741 ), hereinafter Mathew . Regarding claim 2 , Shi discloses the computer-implemented method as claimed in claim 1, but does not specifically disclose further comprising: determining, utilizing the theme-classifying machine-learning model, a subtheme classification associated with the at least one theme classification; and associating the subtheme classification and the theme classification with the experience score. Mathew teaches: determining, utilizing the theme-classifying machine-learning model, a subtheme classification associated with the at least one theme classification ( Column 5, lines 18 31, "Classification engine 113 may identify whether a particular classification category applies to a portion of unstructured text. A classification category may refer to a concept that either summarizes a block of unstructured text or may refer to a concept that is being described in a block of unstructured text. Classification categories may include physical objects (e.g. bathroom sink), people (e.g. waitstaff), locations (e.g. lobby), characteristics of objects (e.g. dirty, torn), characteristics of people (e.g. appearance, attitude), perceptions (e.g. unsafe), emotions (e.g. anger) etc. In some embodiments each classification category is represented by one or many rules. In some embodiments, the rules may be expressed in Boolean logic. In some embodiments, the rules may be represented by a trained machine learning model."; Column 5, lines 39-49, "Theme detection engine 115 may include subsystems to determine categories/themes within a collection of documents. Theme detection engine 115 may determine categories/themes using unsupervised techniques. In some embodiments, theme detection engine 115 may organize themes in a hierarchical structure in which a child theme may belong to a parent theme. In some embodiments, theme detection engine 115 may suggest one or several categorization rules that represent the concept of the theme such that classification engine 113 may identify whether the theme applies to a portion of unstructured text." ; A classification engine reads on a theme-classifying machine-learning model, and organiz ing themes in a hierarchical structure in which a child theme may belong to a parent theme reads on determining a subtheme classification associated with the at least one theme classification , where a child theme reads on a subtheme.); and associating the subtheme classification and the theme classification with the experience score ( Column 5, lines 1-10, "Sentiment scoring engine 112 may identify the general feeling, attitude or opinion that the author of a section of unstructured text is expressing towards a situation or event. In some embodiments, the sentiment scoring engine may classify sentiment as either positive, negative or neutral. In some embodiments, the sentiment scoring engine may assign a numeric sentiment score on a numeric scale ranging from a minimum value representing the lowest possible sentiment to a maximum value representing the highest possible sentiment."; Column 5, lines 32-35, "Reporting engine 114 may report against categories and sentiment expressed in a collection of documents. In some embodiments, the categories used in reporting may include theme detected topics. ; A numeric sentiment score reads on an experience score , and reporting against categories and sentiment , where the categories include theme detected topics , reads on associating the subtheme classification and the theme classification with the experience score .). Mathew is considered to be analogous to the claimed invention because it is in the same field of text classification. Therefore, it would have been obvious to someone of ordinary skill in the art before the effective filing date of the claimed invention to have modified Shi to incorporate the teachings of Mathew to determine them es in unstructured text , organiz e the themes in a hierarchical structure in which a child theme may belong to a parent theme , determine numeric sentiment score s for section s of unstructured text , and report against categories and sentiment , where the categories include theme detected topics . Doing so would allow for automatic unsupervised detection of discussion topics from unstructured feedback text where the results of topic groupings are tagged with meaningful labels ( Mathew ; Column 1, lines 58 61 ). Regarding claim 9 , Shi discloses the computer-implemented method as claimed in claim 1, but does not specifically disclose further comprising: receiving, from the theme-classifying machine-learning model and based on the unstructured text, a suggested theme to add to the defined taxonomy of themes; and adding the suggested theme to the defined taxonomy of themes. Mathew teaches: receiving, from the theme-classifying machine-learning model and based on the unstructured text, a suggested theme to add to the defined taxonomy of themes ( Column 5, lines 18 31, "Classification engine 113 may identify whether a particular classification category applies to a portion of unstructured text. A classification category may refer to a concept that either summarizes a block of unstructured text or may refer to a concept that is being described in a block of unstructured text. Classification categories may include physical objects (e.g. bathroom sink), people (e.g. waitstaff), locations (e.g. lobby), characteristics of objects (e.g. dirty, torn), characteristics of people (e.g. appearance, attitude), perceptions (e.g. unsafe), emotions (e.g. anger) etc. In some embodiments each classification category is represented by one or many rules. In some embodiments, the rules may be expressed in Boolean logic. In some embodiments, the rules may be represented by a trained machine learning model."; Column 5, lines 49-51, "In some embodiments, theme detection engine 115 may suggest a name to identify each determined theme." ; S uggest ing a name to identify each determined theme reads on receiving a suggested theme to add to the defined taxonomy of themes . ); and adding the suggested theme to the defined taxonomy of themes ( Column 5, lines 49-51, "In some embodiments, theme detection engine 115 may suggest a name to identify each determined theme."; Column 21, lines 58-59, "Theme naming 865 may include steps in which identified top-level and second-level themes are assigned labels." ; S uggest ing a name to identify each determined theme and performing theme naming by assigning labels to top-level and second-level themes reads on adding the suggested theme to the defined taxonomy of themes . ). Mathew is considered to be analogous to the claimed invention because it is in the same field of text classification. Therefore, it would have been obvious to someone of ordinary skill in the art before the effective filing date of the claimed invention to have modified Shi to incorporate the teachings of Mathew to s uggest a name to identify each determined theme and performing theme naming by assigning labels to top-level and second-level themes . Doing so would allow for automatic unsupervised detection of discussion topics from unstructured feedback text where the results of topic groupings are tagged with meaningful labels ( Mathew ; Column 1, lines 58 61 ). Regarding claim 11, arguments analogous to claim 2 are applicable. Regarding claim 12 , Shi discloses the computer-readable medium as claimed in claim 1 0 , but does not specifically disclose further comprising: receiving, using the theme-classifying machine-learning model, an additional theme classification and an additional subtheme classification for the unstructured text; and associating, in combination with the theme classification, the additional theme classification and the additional subtheme classification with the experience score. Mathew teaches: receiving, using the theme-classifying machine-learning model, an additional theme classification and an additional subtheme classification for the unstructured text ( Column 5, lines 18 31, "Classification engine 113 may identify whether a particular classification category applies to a portion of unstructured text. A classification category may refer to a concept that either summarizes a block of unstructured text or may refer to a concept that is being described in a block of unstructured text. Classification categories may include physical objects (e.g. bathroom sink), people (e.g. waitstaff), locations (e.g. lobby), characteristics of objects (e.g. dirty, torn), characteristics of people (e.g. appearance, attitude), perceptions (e.g. unsafe), emotions (e.g. anger) etc. In some embodiments each classification category is represented by one or many rules. In some embodiments, the rules may be expressed in Boolean logic. In some embodiments, the rules may be represented by a trained machine learning model."; Column 5, lines 39-49, "Theme detection engine 115 may include subsystems to determine categories/themes within a collection of documents. Theme detection engine 115 may determine categories/themes using unsupervised techniques. In some embodiments, theme detection engine 115 may organize themes in a hierarchical structure in which a child theme may belong to a parent theme. In some embodiments, theme detection engine 115 may suggest one or several categorization rules that represent the concept of the theme such that classification engine 113 may identify whether the theme applies to a portion of unstructured text." ; A classification engine reads on a theme-classifying machine-learning model, and organiz ing themes in a hierarchical structure in which a child theme may belong to a parent theme reads on receiving additional theme classification s and additional subtheme classification s for the unstructured text .); and associating the subtheme classification and the theme classification with the experience score ( Column 5, lines 1-10, "Sentiment scoring engine 112 may identify the general feeling, attitude or opinion that the author of a section of unstructured text is expressing towards a situation or event. In some embodiments, the sentiment scoring engine may classify sentiment as either positive, negative or neutral. In some embodiments, the sentiment scoring engine may assign a numeric sentiment score on a numeric scale ranging from a minimum value representing the lowest possible sentiment to a maximum value representing the highest possible sentiment."; Column 5, lines 32-35, "Reporting engine 114 may report against categories and sentiment expressed in a collection of documents. In some embodiments, the categories used in reporting may include theme detected topics. ; A numeric sentiment score reads on an experience score , and reporting against categories and sentiment , where the categories include theme detected topics , reads on associating the additional theme classification and the additional subtheme classification with the experience score. ). Mathew is considered to be analogous to the claimed invention because it is in the same field of text classification. Therefore, it would have been obvious to someone of ordinary skill in the art before the effective filing date of the claimed invention to have modified Shi to incorporate the teachings of Mathew to determine them es in unstructured text , organiz e the themes in a hierarchical structure in which a child theme may belong to a parent theme , determine numeric sentiment score s for section s of unstructured text , and report against categories and sentiment , where the categories include theme detected topics . Doing so would allow for automatic unsupervised detection of discussion topics from unstructured feedback text where the results of topic groupings are tagged with meaningful labels ( Mathew ; Column 1, lines 58 61 ). Regarding claim 13 , Shi discloses the computer-readable medium as claimed in claim 10, but does not specifically disclose further comprising instructions that, when executed by the at least one processor, cause the computer system to: receiving, from the theme-classifying machine-learning model, one or more suggested subthemes associated with a given theme in the defined taxonomy of themes; and associating the one or more subthemes with the given theme in the defined taxonomy of themes. Mathew teaches: receiving, from the theme-classifying machine-learning model, one or more suggested subthemes associated with a given theme in the defined taxonomy of themes ( Column 5, lines 18 31, "Classification engine 113 may identify whether a particular classification category applies to a portion of unstructured text. A classification category may refer to a concept that either summarizes a block of unstructured text or may refer to a concept that is being described in a block of unstructured text. Classification categories may include physical objects (e.g. bathroom sink), people (e.g. waitstaff), locations (e.g. lobby), characteristics of objects (e.g. dirty, torn), characteristics of people (e.g. appearance, attitude), perceptions (e.g. unsafe), emotions (e.g. anger) etc. In some embodiments each classification category is represented by one or many rules. In some embodiments, the rules may be expressed in Boolean logic. In some embodiments, the rules may be represented by a trained machine learning model."; Column 5, lines 39-49, "Theme detection engine 115 may include subsystems to determine categories/themes within a collection of documents. Theme detection engine 115 may determine categories/themes using unsupervised techniques. In some embodiments, theme detection engine 115 may organize themes in a hierarchical structure in which a child theme may belong to a parent theme. In some embodiments, theme detection engine 115 may suggest one or several categorization rules that represent the concept of the theme such that classification engine 113 may identify whether the theme applies to a portion of unstructured text." ; Column 5, lines 49-51, "In some embodiments, theme detection engine 115 may suggest a name to identify each determined theme." ; S uggest ing a name to identify each determined theme , where the themes are organized in a hierarchical structure in which a child theme may belong to a parent theme , reads on receiving one or more suggested subthemes associated with a given theme in the defined taxonomy of themes .); and associating the one or more subthemes with the given theme in the defined taxonomy of themes ( Column 5, lines 39-49, "Theme detection engine 115 may include subsystems to determine categories/themes within a collection of documents. Theme detection engine 115 may determine categories/themes using unsupervised techniques. In some embodiments, theme detection engine 115 may organize themes in a hierarchical structure in which a child theme may belong to a parent theme. In some embodiments, theme detection engine 115 may suggest one or several categorization rules that represent the concept of the theme such that classification engine 113 may identify whether the theme applies to a portion of unstructured text." ; O rganiz ing themes in a hierarchical structure in which a child theme may belong to a parent theme reads on associating the one or more subthemes with the given theme in the defined taxonomy of themes .). Mathew is considered to be analogous to the claimed invention because it is in the same field of text classification. Therefore, it would have been obvious to someone of ordinary skill in the art before the effective filing date of the claimed invention to have modified Shi to incorporate the teachings of Mathew to s uggest a name to identify each determined theme and o rganiz e themes in a hierarchical structure in which a child theme may belong to a parent theme . Doing so would allow for automatic unsupervised detection of discussion topics from unstructured feedback text where the results of topic groupings are tagged with meaningful labels ( Mathew ; Column 1, lines 58 61 ). Regarding claim 17, arguments analogous to claim 2 are applicable. Claim s 3 – 4 , 14 – 15 , 18 and 20 are rejected under 35 U.S.C. 103 as being unpatentable over Shi in view of Rawat et al. (US Patent No. 11 , 657 , 415 ), hereinafter Rawat . Regarding claim 3 , Shi discloses the computer-implemented method as claimed in claim 1, but does not specifically disclose: wherein receiving experience data comprising an experience score associated with unstructured text comprises: receiving a response to a net promoter score survey comprising a net promoter score; and receiving the unstructured text associated with the net promoter score survey. Rawat teaches: wherein receiving experience data comprising an experience score associated with unstructured text comprises: receiving a response to a net promoter score survey comprising a net promoter score ( Column 5, lines 60-66, "Responsive to a user-initiated request from the user 125a to provide a feedback , an interactive window may be presented to the user 125a to allow the user 125a to input the feedback (e.g., respond to a received survey or an embedded survey). The interactive window may include a rating tool and/or a comment area for the user 125a to provide comments."; Column 10, lines 31-39, "In one implementation, the high priority topic identification module 225 may determine the priority topic by calculating a net promoter score (NPS) uplift for each topic and then selecting a topic with the largest value as the priority topic. The NPS uplift for each topic may quantify the maximum possible improvement in overall product NPS for each topic, where the NPS is a general customer loyalty metric that measures how likely the customers are to recommend a product, service, or website to a friend." ; A net promoter score survey presenting an interactive window to a user to allow the user to input feedback , where the interactive window include s a rating tool , reads on receiving a response to a net promoter score survey comprising a net promoter score .); and receiving the unstructured text associated with the net promoter score survey ( Column 5, lines 60-66, "Responsive to a user-initiated request from the user 125a to provide a feedback , an interactive window may be presented to the user 125a to allow the user 125a to input the feedback (e.g., respond to a received survey or an embedded survey). The interactive window may include a rating tool and/or a comment area for the user 125a to provide comments."; Column 10, lines 31-39, "In one implementation, the high priority topic identification module 225 may determine the priority topic by calculating a net promoter score (NPS) uplift for each topic and then selecting a topic with the largest value as the priority topic. The NPS uplift for each topic may quantify the maximum possible improvement in overall product NPS for each topic, where the NPS is a general customer loyalty metric that measures how likely the customers are to recommend a product, service, or website to a friend." ; A net promoter score survey presenting an interactive window to a user to allow the user to input feedback , where the interactive window include s a comment area for the user to provide comments , reads on receiving the unstructured text associated with the net promoter score survey .). Rawat is considered to be analogous to the claimed invention because it is in the same field of text classification. Therefore, it would have been obvious to someone of ordinary skill in the art before the effective filing date of the claimed invention to have modified Shi to incorporate the teachings of Rawat to implement a net promoter score survey presenting an interactive window to a user to allow the user to input feedback , where the interactive window include s a rating tool and a comment area for the user to provide comments . Doing so would allow for monitoring the content of online user feedbacks in real-time to automatically identify topics from user feedbacks , prioritiz ing a priority topic from the identified topics , and initiating immediate action ( Rawat ; Column 4, lines 7-14 ). Regarding claim 4 , Shi in view of Rawat discloses the computer-implemented method as claimed in claim 3. Rawat further teaches: wherein associating the theme classification with the experience score comprises associating the theme classification with the net promoter score ( Column 15, line 52 - Column 16, line 12, "At block 403, the online user feedback management server 101 identifies a plurality of topics from the online user feedbacks. To identify the topics from the online user feedbacks, the online user feedback management server 101 may first determine whether a user feedback contains verbatim or comments. If a user feedback contains verbatim or comments, the verbatim or comments may be first converted into machine-encoded text. Different techniques may be applied to convert the user feedbacks into the machine-encoded text. These techniques may include, but are not limited to, optical character recognition, machine learning models, etc. The machine-encoded text may be then compared to keywords related to each topic, to identify one or more topics from each user feedback. In some implementations, more than one topic may be identified from single user feedback. By identifying the topics related to all received user feedbacks, the topics for the user feedbacks related to the product may be then identified. At block 405, the online user feedback management server 101 determines a NPS uplift for each identified topic. Here, a NPS uplift is calculated based on a NPS for the product that is calculated based on the rating, and an adjusted NPS (i.e., NPS adj ) that is calculated based on the rating and/or verbatim, as further described in detail in FIG. 5. The calculated NPS uplift for each topic may represent the maximum potential improvement to the overall NPS for the product that can be gained by fully addressing issues within a given topic." ; I dentify ing topics from online user feedback verbatims and determining a net promoter score uplift for each identified topi c based on the rating and the verbatim reads on associating the theme classification with the net promoter score .). Rawat is considered to be analogous to the claimed invention because it is in the same field of text classification. Therefore, it would have been obvious to someone of ordinary skill in the art before the effective filing date of the claimed invention to have modified Shi in view of Rawat to further incorporate the teachings of Rawat to i dentif y topics from online user feedback verbatims and determine a net promoter score uplift for each identified topi c based on the rating and the verbatim . Doing so would allow for monitoring the content of online user feedbacks in real-time to automatically identify topics from user feedbacks , prioritiz ing a priority topic from the identified topics , and initiating immediate action ( Rawat ; Column 4, lines 7-14 ). Regarding claim 14 , Shi discloses the computer-readable medium as claimed in claim 10, but does not specifically disclose further comprising instructions that, when executed by the at least one processor, cause the computer system to perform an action based on the associated theme classification and experience score. Rawat teaches: perform an action based on the associated theme classification and experience score ( Column 6, lines 38-56, "In one implementation, the online user feedback management server 101 may further identify a topic(s) with a priority, where the priority topic means that the issues related to the topic are more important and thus require a more instant action compared to other identified topics. After identification of the priority topic, the online user feedback management server 101 may further implement certain immediate actions to address issues related to the topics. For example, the online user feedback management server 101 may modify the survey, used to collect the feedback, to include more sub-topics related to the topic in the survey, may generate an alert to alarm relevant personnel in charge and/or automatically forward the relevant feedbacks and/or the priority topic to the personnel in charge. In some implementations, the online user feedback management server 101 may provide a reply in response to an issue raised in a feedback . Certain other actions that may be implemented by the online user feedback management server 101 are also possible and contemplated." ; I mplement ing certain immediate actions to address issues reads on perform ing an action based on the associated theme classification and experience score .). Rawat is considered to be analogous to the claimed invention because it is in the same field of text classification. Therefore, it would have been obvious to someone of ordinary skill in the art before the effective filing date of the claimed invention to have modified Shi to incorporate the teachings of Rawat to i mplement certain immediate actions to address issues . Doing so would allow for monitoring the content of online user feedbacks in real-time to automatically identify topics from user feedbacks , prioritiz ing a priority topic from the identified topics , and initiating immediate action ( Rawat ; Column 4, lines 7-14 ). Regarding claim 15 , Shi discloses the computer-readable medium as claimed in claim 10, but does not specifically disclose further comprising instructions that, when executed by the at least one processor, cause the computer system to: determine an action to perform based on the theme classification; and predict that performing the determined action will correlate to an increase in the experience score. Rawat teaches: determine an action to perform based on the theme classification ( Column 6, lines 38 56, "In one implementation, the online user feedback management server 101 may further identify a topic(s) with a priority, where the priority topic means that the issues related to the topic are more important and thus require a more instant action compared to other identified topics. After identification of the priority topic, the online user feedback management server 101 may further implement certain immediate actions to address issues related to the topics. For example, the online user feedback management server 101 may modify the survey, used to collect the feedback, to include more sub-topics related to the topic in the survey, may generate an alert to alarm relevant personnel in charge and/or automatically forward the relevant feedbacks and/or the priority topic to the personnel in charge. In some implementations, the online user feedback management server 101 may provide a reply in response to an issue raised in a feedback . Certain other actions that may be implemented by the online user feedback management server 101 are also possible and contemplated." ; I mplement ing certain immediate actions to address issues reads on determin ing an action to perform based on the theme classification .); and predict that performing the determined action will correlate to an increase in the experience score ( Column 10, lines 31-51, "In one implementation, the high priority topic identification module 225 may determine the priority topic by calculating a net promoter score (NPS) uplift for each topic and then selecting a topic with the largest value as the priority topic. The NPS uplift for each topic may quantify the maximum possible improvement in overall product NPS for each topic, where the NPS is a general customer loyalty metric that measures how likely the customers are to recommend a product, service, or website to a friend. The high priority topic identification module 225 may make two assumptions in calculating the NPS uplift for each topic. First, for a given topic, assume that all feedbacks in the passive and detractor groups that contain the topic in the text feedback are movable by resolving the issues represented in the topic. Second, assume that similar proportions among feedbacks that do not provide text feedback are similarly movable. For the computation of the NPS uplift for a given topic, the first step is to construct a frequency table based on rating scores (promoters, passives, detractors) and the text feedback (no text, feedback related to the topic of interest, feedback unrelated to topic), as can be seen in FIG. 3A." ; C alculating a net promoter score (NPS) uplift for each topic by quantify ing the maximum possible improvement in overall product NPS for each topic based on resolving the issues represented in the topi c reads on predict ing that performing the determined action will correlate to an increase in the experience score .). Rawat is considered to be analogous to the claimed invention because it is in the same field of text classification. Therefore, it would have been obvious to someone of ordinary skill in the art before the effective filing date of the claimed invention to have modified Shi to incorporate the teachings of Rawat to i mplement certain immediate actions to address issues and c alculat e a net promoter score (NPS) uplift for each topic by quantify ing the maximum possible improvement in overall product NPS for each topic based on resolving the issues represented in the topi c. Doing so would allow for monitoring the content of online user feedbacks in real-time to automatically identify topics from user feedbacks , prioritiz ing a priority topic from the identified topics , and initiating immediate action ( Rawat ; Column 4, lines 7-14 ). Regarding claim 18, arguments analogous to claim 14 are applicable. Regarding claim 20, arguments analogous to claim 3 are applicable. Claim 5 is rejected under 35 U.S.C. 103 as being unpatentable over Shi in view of L'Huillier et al. (US Patent No. 9 , 741 , 058 ), hereinafter L'Huillier . Regarding claim 5 , Shi discloses the computer-implemented method as claimed in claim 1, but does not specifically disclose: wherein receiving experience data comprising an experience score associated with unstructured text comprises: receiving, from a third-party media information service, a media experience score; and receiving, from the third-party media information service, unstructured text associated with the media experience score. L'Huillier teaches: wherein receiving experience data comprising an experience score associated with unstructured text comprises: receiving, from a third-party media information service, a media experience score ( Column 5, lines 9-14, "In certain cases, the consumer review information (e.g., reviews, ratings and survey responses) may be accessed from Internet websites using, for example, a web crawler. In certain cases, the consumer reviews, ratings and survey responses may be accessed from a database associated with a promotion and marketing service." ; A ccess ing consumer review information from a database associated with a promotion and marketing service , where the consumer review information includes ratings , reads on receiving a media experience score from a third-party media information service .); and receiving, from the third-party media information service, unstructured text associated with the media experience score ( Column 5, lines 9-14, "In certain cases, the consumer review information (e.g., reviews, ratings and survey responses) may be accessed from Internet websites using, for example, a web crawler. In certain cases, the consumer reviews, ratings and survey responses may be accessed from a database associated with a promotion and marketing service." ; A ccess ing consumer review information from a database associated with a promotion and marketing service , where the consumer review information includes ratings and reviews, reads on receiving unstructured text associated with the media experience score from the third-party media information service .). L'Huillier is considered to be analogous to the claimed invention because it is in the same field of text classification. Therefore, it would have been obvious to someone of ordinary skill in the art before the effective filing date of the claimed invention to have modified Shi to incorporate the teachings of L'Huillier to a ccess consumer review information from a database associated with a promotion and marketing service , where the consumer review information includes ratings and reviews. Doing so would allow for programmatically extracting one or more attribute descriptors regarding a commercial entity or object from consumer reviews with consistency and accuracy ( L'Huillier ; Column 3, lines 33-47 ). Claim s 6 – 7 are rejected under 35 U.S.C. 103 as being unpatentable over Shi in view of Jain et al. (US Patent No. US 11 , 914 , 963 ), hereinafter Jain . Regarding claim 6 , Shi discloses the computer-implemented method as claimed in claim 1. Shi further discloses : providing the unstructured text to a natural language processing model ( Column 7, lines 11-17, "System architecture 200 comprises input 202, artificial intelligence (AI) ensemble modeling 204, machine learning (ML) output 206, output consolidation 208, and outputs 210. In turn, input 202 comprises verbatim 212 and web journey data 214. AI ensemble modeling 204 comprises journey mapping 216, natural language processing (NLP) 218, and deep learning 220."; Column 10, lines 56-57, "Natural language processing can be performed on a verbatim to determine the verbatim's intent." ; A verbatim reads on unstructured text , and an AI ensemble model compris ing natural language processing reads on a natural language processing model .). Shi does not specifically disclose further comprising: receiving, from the natural language processing model, classification embeddings; and providing the classification embeddings to the theme-classifying machine-learning model to receive the theme classification for the unstructured text. Jain teaches: receiving, from the natural language processing model, classification embeddings ( Column 6, lines 48-61, "According to one embodiment, detecting semantic relatedness comprises analyzing text, embedding the text, and determining the semantic relatedness to the concept of a category, where each concept includes a set of words/phrases embedded in a similar fashion. The text embedding can be projected onto each concept embedding and reduced to a number score representing semantic relatedness, which can then be compared to a threshold or otherwise analyzed to determine meaningful semantic relatedness. More particularly, the semantic relatedness of a piece of text to a collection of words or phrases, such as a concept, can be determined using embeddings to transform the set of text and the collection of words and phrases for semantic concepts into numeric representations"; Column 9, lines 26-31, "Various embedding techniques may be used including, but not limited to, neural networks or other natural language processing techniques. By way of example, but not limitation, full word embeddings or sub-word embeddings may be pre-trained using deep neural networks with large amounts of data." ; E mbedding text and using the embeddings to determin e the semantic relatedness of text to the concept of a category reads on classification embedding s.); and providing the classification embeddings to the theme-classifying machine-learning model to receive the theme classification for the unstructured text ( Column 8, lines 51-62, "Semantic classification system 100 includes classifiers to determine the semantic relatedness of text to concepts or categories. For example, semantic classification system 100 includes classifier 102 to classify text according to category 110. While only one category and classifier are illustrated, semantic classification system 100 may include any number of categories and classifiers. In some embodiments, a single classifier may classify according to multiple categories or multiple classifiers may classify according to a single category. In some embodiments, the structure of a category may be considered to be part of the structure of a classifier that classifies text based on that category." ; D etermin ing the semantic relatedness of text to categories reads on receiv ing the theme classification for the unstructured text .). Jain is considered to be analogous to the claimed invention because it is in the same field of text classification. Therefore, it would have been obvious to someone of ordinary skill in the art before the effective filing date of the claimed invention to have modified Shi to incorporate the teachings of Jain to e mbed text and use the embeddings to determine the semantic relatedness of text to categories . Doing so would allow for determin ing semantic relatedness of relatively small segments of text and map relatedness to abstract categories that include one or more concepts where each concept may require more than one word or phrase for a semantic definition ( Jain ; Column 2, lines 24-31 ). Regarding claim 7 , Shi in view of Jain et discloses the computer-implemented method as claimed in claim 6. Jain further teaches: wherein the natural language processing model comprises a bidirectional encoder representations from transformers (BERT) model ( Column 9, lines 26-42, "Various embedding techniques may be used including, but not limited to, neural networks or other natural language processing techniques. By way of example, but not limitation, full word e