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 .
Response to Arguments
Applicant’s arguments with respect to claims 1 and 16 have been considered but are
moot because the new ground of rejection.
Applicant’s Argument A:
“ Classifying the original content from the plurality of platforms into one or more topics, and, for each topic, creating recommended content. Concerns clustering historical conversational utterances for training a chatbot model, rather than curating recent original content from a plurality of different platforms into topic-specific recommended content.” [Reply; page 10].
In response to applicants argument:
The claim is silent as to why the classification is performed or what downstream use the classification serves. It only requires the classification of original content from a plurality of platforms into topics and creation of recommended content per topic. Moya discloses both steps utterances (original content) sourced from multi-tenant and multi-channel platform (Moya, [0024, 0105,0024,0081 and 0105]) are classified into topic grouping via a clustering algorithm ( Moya, [Abstract, 0067, 0069] and for each such topic the system generates output content. Moya’s ultimate application of this classify and generate process is to train or improve a chatbot model does not remove Moya’s disclosure from the scope of claim, because claim 1 does not require that the “recommended content” be put to any particular end use, nor does it exclude recommended content that is subsequently used to train or configure a model. Moya’s clustering is not limited to model training. It also output topics specific content. A suggested topic label generated from the clustered utterances (Moya, [0069]) and “suggested response” (Moya, [0074]).
Applicant’s Argument B:
“ transforming an initial chatbot response into a domain-specific response that mimics a style unique to a particular domain. Summarizing original content classified by topic to create recommended content. Generate chatbot responses having domain-specific styles which is different from the claimed topic-based summarization and content-curation process”. [Reply; page 11].
In response to applicants argument:
Moya discloses classifying recent original content received from a plurality of different platforms (Moya, [0024]) by topic via clustering algorithm (Moya, [Abstract, 0067-0080]).
Banerjee is relied upon not for topic classification, but for its techniques of aggregating content associated with a given category from multiple sources/ platform (Banerjee, [0098]) and generally synthesizing representative output content from that aggregated content via scored pattern selection and combination (Banerjee, [0119, 0144]), which reads on summarizing classified content to create recommended content.
Applicant’s Argument C:
"Silverstein's analysis and management of responses in real-time text-based discourse does not disclose or suggest creating topic-specific recommended content based on recent original content from a plurality of different content-providing platforms and therefore does not cure the deficiencies of Moya. Silverstein also does not disclose collecting feedback after a user has read recommended content or providing an effect report or a statistical report concerning the recommended content to a creator of the original content. [Reply; page 12].
In response to applicants argument:
Silverstein reference was not cited in the 103 rejection of claim 1 therefore the arguments are moot. Also, the arguments towards Silverstein reference not disclosing collecting feedback or providing a report are conclusionary.
Applicant's arguments fail to comply with 37 CFR 1.111(b) because they amount to a general allegation that the claims define a patentable invention without specifically pointing out how the language of the claims patentably distinguishes them from the references.
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, 5-7, 9-10 and 15-20 is/are rejected under 35 U.S.C. 103 as being
unpatentable over Moya et al. (Pub. No.: US20230298568A1), hereinafter “Moya” in view of Banerjee et al. (Pub. No.: US20180365212 A1), hereinafter "Ban".
As to claim 1, Moya discloses A content curation method implemented on a
computer device including at least one processor (Moya, [0002], A chatbot is a software application that executes on the site and that is used to interact with the user, often in lieu of a direct human interaction. [0024], a multi-turn conversation is carried out between an end user 100, and a conversational bot software application 102 that executes in a network-accessible computing platform 104. [0097], A machine implementing the techniques herein comprises a hardware processor, and non-transitory computer memory holding computer program instructions that are executed by the processor to perform the above-described methods.).
providing, by the at least one processor, the recommended content to a user
using a chatbot for content curation (Moya, [0057], the system receives input data, e.g., from a human designer, and in response configures a directed acyclic graph (DAG) that represents a conversation flow. [0056], the bot is controlled to be more deliberate in how it drives the conversation forward proactively.;
collecting, by the at least one processor, a user response to the recommended
content provided by the user to the chatbot (Moya, [0024], a multi-turn conversation is carried out between an end user 100, and a conversational bot software application 102. [0020], a non-linguistic action taken by an actor, e.g., clicking a button or a link on a Graphical User Interface (GUI), entering data in a form, or the like. [0022], Utterance: a sequence of words that is grammatically complete; usually one sentence.).;
and reflecting, by the at least one processor, the user response in at least one of a user's personalization recommendation and a report related to the recommended content (Moya, [0031], the data model keeps track of any number of events, all of which can be actively “extended” at any time. [0067], the service provider typically analyzes a customer's historical data (e.g., a set of historical human or bot conversational transcripts over some time period) and uses that data for model training. By manually clustering groups of utterances, the provider trains a classifier to identify topics.).
Moya however is silent to disclose explicitly, creating, by the at least one processor, recommended content for one or more topics using original content produced over a recent period of time on each of plurality of platforms, wherein the plurality of platforms are different content-providing platforms. Wherein the creating comprises classifying the original content from the plurality of platforms into the one or more topics; and for each topic of the one or more topics, creating the recommended content for the respective topic by summarizing, using generative artificial intelligence (AI), the original content classified into the respective topic.
Banerjee discloses the similar concept in same field of endeavor including creating, by the at least one processor, recommended content for one or more topics using original content produced over a recent period of time on each of a plurality of platforms, wherein the plurality of platforms are different content-providing platforms (Moya , [0002], A chatbot is a software application that executes on the site and that is used to interact with the user, often in lieu of a direct human interaction. [0031], the data model keeps track of any number of events, all of which can be actively “extended” at any time. [0030], the data model comprises the observation history, namely, a hierarchical set of events that have been determined to represent the conversation up to at least one conversation turn (and typically many turns). (Ban, [0006], computerized techniques to construct domain-specific word-graphs using tweets posted from Twitter® accounts (and/or any other type of network accessible platform/resource that enables learning/training of a system to understand language styles) that belong to users from specific domains and use the graph to generate word-patterns. [0098], any type of network accessible platform/resource from which language processing can occur can be utilized as a basis for formulating the word-graphs, for example, but not limited to, other social networking sites (e.g., Facebook®), email (e.g., Yahoo! Mail®), blogs, articles, instant messaging platforms (e.g., Yahoo! Messenger®, WhatsApp®, and the like), web portals, and the like, or some combination thereof.).;
Wherein the creating comprises classifying the original content from the plurality of platforms into the one or more topics (Moya, [0003], The method begins by configuring a conversational bot using a machine learning model trained to classify utterances into topics. [0067], the service provider typically analyzes a customer's historical data (e.g., a set of historical human or bot conversational transcripts over some time period) and uses that data for model training. By manually clustering groups of utterances, the provider trains a classifier to identify topics (or, more generally, user intents). (Ben, [0098], other social networking sites (e.g., Facebook®), email (e.g., Yahoo! Mail®), blogs, articles, instant messaging platforms (e.g., Yahoo! Messenger®, WhatsApp®, and the like), web portals, and the like, or some combination thereof.);
and for each topic of the one or more topics, creating the recommended content for the respective topic by summarizing, using generative artificial intelligence (AI), the original content classified into the respective topic (Ban, [0006], a method is disclosed for a novel, computerized framework for automatically generating and/or transforming chatbot responses to produce domain-specific responses that mimic native styles unique to particular domains. [0042], instead of simply outputting this response, as in conventional systems, the disclosed systems and methods automatically transform the initial chatbot response to produce domain-specific response that mimics a native style unique to the particular domain from which the query was entered.
Therefore, before the effective filling date of this instant application it would have been obvious to one of the ordinary skilled in the art to incorporate the teachings of “Banerjee” in those of Moya to source original content from plurality of different content providing platforms and automatically generate topic-specific recommended content using generative AI.
As to claim 5, Mayo-Ban discloses the providing of the recommended content
comprises providing reference information indicating the original content, and the original content is a source of the recommended content (Moya, [0027], the data model comprises an observation history, together with a set of events that have been determined to represent the conversation. [0037], candidate interpretation is a pointer identifying specific lines of historical data in the set of inter-related tables that comprise that relational database 402.).
As to claim 6, Mayo-Ban discloses the reference information includes information on
at least one of a link and a creator of the original content (Moya, [0037], the candidate interpretation is a pointer identifying specific lines of historical data in the set of inter-related tables. [0030], the data model is persisted (and in the depicted tree grows right-ward), the conversation history between the user and the bot is represented. [0026], organized as clusters of utterances 212.).
As to claim 7, Mayo-Ban discloses the recommended content is randomly selected
for the user from among one or more recommended contents classified by topic (Moya, [0048], ActionSelectorContinueLog-this is the simplest action selector. Every time critics approve a candidate interpretation, the system records in the data model which line of which transcript that candidate pointed at. This action selector blindly proposes that the next thing to say is whatever was said next in that particular transcript.).
As to claim 9, Mayo-Ban discloses the reflecting comprises extracting a user's
personalization information for content recommendation based on the user response (Moya, [0067], the service provider typically analyzes a customer's historical data (e.g., a set of historical human or bot conversational transcripts over some time period) and uses that data for model training. By manually clustering groups of utterances, the provider trains a classifier to identify topics (or, more generally, user intents). [0080], The approach herein enables unsupervised Al-based self-serve model improvement wherein topics are exposed and added to a base AI model (or some other model) to improve its operation with respect to future human/bot conversations.).
As to claim 10, Mayo-Ban discloses the user's personalization information includes
at least one of preference by topic and a subscription status (Moya, [0067], the provider trains a classifier to identify topics (or, more generally, user intents). [0003], The method begins by configuring a conversational bot using a machine learning model trained to classify utterances into topics.).
As to claim 20, Mayo-Ban discloses wherein the at least one processor is caused to
provide at least one of an effect report by a user and a statistical report that compiles user responses to the recommended content by topic to a creator of the original content (Moya, [0067], the service provider typically analyzes a customer's historical data (e.g., a set of historical human or bot conversational transcripts over some time period) and uses that data for model training. [0068], These utterances may be exposed to interested and authorized persons. e.g., through a customer dashboard “discovery” tab. [0076], cross-validation is used to compute internal metrics, e.g., about the precision, recall, and F-score of one or more topics. From these results, a confusion matrix can be computed that reports how often each topic is misclassified with respect to one or more other topics. This information may be reported to a user. [0078 -0079], If the system sees that many site visitors drop off from chat after getting a response to a particular topic, this may indicate that the input is not getting correctly recognized, or that the currently configured response is unsatisfying. In such circumstance, and via email or in the dashboard, the system may be configured to inform users of topics that have unusually high drop off. The above-described examples, which provide several diagnostic tools that further improve the AI model, are exemplary.).
As to claim 15, (Mayo-Ban: [0093-0095], CRM) is rejected for same rationale as
applied to claim 1 above.
As to claim 16 is rejected for same rationale as applied to claim 1 above.
As to claim 18 is rejected for same rationale as applied to claims 5 and 6 above.
As to claim 19 is rejected for same rationale as applied to claim 10 above.
Claim(s) 8, 11-14 is/are rejected under 35 U.S.C. 103 as being unpatentable over
Mayo-Ban as applied to claim 1,5,9,12 and 16 above, and further in view of
Silverstein et al. (Pub. No.: US 20210334471 A1), hereinafter “Sil”.
As to claim 8, The combined system of Mayo and Ban disclose the invention as
applied above. Mayo and Ban, however, are silent to disclose explicitly, the collecting comprises, when it is determined that the user read the recommended content, collecting the user response to the recommended content through a conversation between the chatbot and the user after a certain period of time elapses from a corresponding point in time.
Sil however discloses a similar concept in the same field of endeavor including, the collecting comprises, when it is determined that the user read the recommended content, collecting the user response to the recommended content through a conversation between the chatbot and the user after a certain period of time elapses from a corresponding point in time (Silverstein, Fig 5A [0081], the communications server 402 determines a frequency of responses (author post frequency) posted by a respondent over time. In implementations, author post frequency comprises an inter-arrival time between message postings of a participant. In embodiments, the communications server 402 takes a time of a post "n" and subtracts the time of the previous n-1 post.).
Therefore, before the effective filing date of the instant application it would have
been obvious to one of the ordinary skilled in the art to incorporate the teachings of
“Sil” in those of Mayo and Ban to determine a frequency of the responses of the at least one respondent over time and determine an evasiveness score for each of the responses based on natural language processing of the responses.
As to claim 11, Mayo-Ban-Sil discloses the reflecting comprises providing an effect
report by a user to a creator of the original content that is a source of the recommended content (Silverstein, [0088], the communications server 402 optionally scores participants based on aggregate evasiveness of their responses over time. [0069], the ranking module 412 is configured to: utilize the bridged discourse model to rank evasiveness of individual responses, determine a display order of responses based on the ranking and change the display order as needed, insert an indicator of evasiveness in a virtual window of the text-based discourse session, score participants based on aggregate evasiveness, and manage participation based on participant scores. [0097], the communication server 402 retrieves metrics from those participants, such as a base evasiveness score based on semantic rating of the chat, the participants' expertise in the topic of the thread, and the velocity of conversation of the topic. The communication server 402 utilizes the bridged discourse model to rank each of the given responses to the question.).
As to claim 12, Mayo-Ban-Sil discloses the reflecting comprises providing a
statistical report that compiles user responses to the recommended content by topic to a creator of the original content (Silverstein, [0098], linguistics analytics of step 501 of FIG. 5A along with the evasiveness score for the topic/question and the frequency of responses to output evasiveness rankings for participants.).
As to claim 13, Mayo-Ban-Sil discloses the statistical report includes a summary of
positive content and negative content among conversations exchanged with users about the recommended content (Silverstein, [0095], the communications server 402 optionally determines if evasiveness of a response is malicious or non-malicious based on sentiment analysis of the text-based discourse session.).
As to claim 14, Mayo-Ban-Sil discloses sharing, by the at least one processor, the
user response through an open chat related to the recommended content (Silverstein, [0065], the communications server 402 enables chat sessions to take place via one or more virtual rooms or channels using on- screen text, typed in real-time. [0086], responses to a preceding question are original displayed based on a time the responses were posted to the text-based discourse (e.g., chat) session.),
wherein the sharing includes, when the recommended content is created based on recent conversations of the open chat, summarizing conversations exchanged with users about the content, and sharing the summarized conversations on the opencast (Silverstein, [0080], the communications server 402 analyzes the text- based discourse and determines that the term "system" collocates with the term "and" and determines that the term "and" collocates with the term "methods".) [0079], At step 501, the communications server 402 analyzes the text-based discourse accessed at step 500 using a corpus of linguistics analysis (corpus linguistics). [0086], Responses to a preceding question are original displayed based on a time the responses were posted to the text-based discourse (e.g., chat) session.
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 MUSADDIQUE WAHID whose telephone number is (571)270-0865. The examiner can normally be reached Monday Friday, 8 a.m. 5 p.m. 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, Umar Cheema can be reached at 5712703037. The fax phone number for the organization where this application or proceeding is assigned is 571-273-8300.
Information regarding the status of published or unpublished applications may be obtained from Patent Center. Unpublished application information in Patent Center is available to registered users. To file and manage patent submissions in Patent Center, visit: https://patentcenter.uspto.gov. Visit https://www.uspto.gov/patents/apply/patent-center for more information about Patent Center and https://www.uspto.gov/patents/docx for information about filing in DOCX format. For additional questions, contact the Electronic Business Center (EBC) at 866-217-9197 (toll-free). If you would like assistance from a USPTO Customer Service Representative, call 800-786-9199 (IN USA OR CANADA) or 571-272-1000.
/M.W./Examiner, Art Unit 2458
/UMAR CHEEMA/Supervisory Patent Examiner, Art Unit 2458
Prosecution Insights