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 .
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.
Claims 1-3, 6, 9, 11-13, 16, 19 are rejected under 35 U.S.C. 103 as being unpatentable over Gangadharaiah (US 10860629 B1), in view of Perkins (US 20220253647 A1), and Xie (US 20210157974 A1).
Regarding claim 1, Gangadharaiah discloses “receiving, via a message generation interface, first message attribute data;” (See [Column 2, Lines 12-17], [Column 4, Lines 5-15]; Gangadharaiah discloses receiving a user utterance (first message attribute data) and also discloses that an agent system (interface) is used to receive user utterances)
“inputting the first message attribute data to a first machine learning model, wherein the first machine learning model is configured to generate and output suggested message content based on first correlations between message content and message acceptance data;” (See [Column 2, Lines 12-17]; Gangadharaiah discloses taking an input for the mentioned ML model 116 (first machine learning model) and outputting a generation. Under the broadest reasonable interpretation, message acceptance data covers the reward system for progression and completion of a dialog. As the dialog progresses, there is an implicit “acceptance” of the messages by the user as they reply to them.)
“generating, by the first machine learning model, based on the first message attribute data, a first set of message content suggestions;” (See [Column 4, Lines 20-25], [Column 4, Lines 44-48]; Gangadharaiah discloses generating one or more candidate responses (message content) suggestions based on the given user dialogs (message attribute data) using a chatbot system. This chatbot system utilizes a ML model 116 and can either be part of the chatbot system itself or implemented separately and gets accessed by the chatbot system)
“selecting, by the first machine learning model, based on message evaluation data received by the first machine learning model… at least one message content suggestion from the first set of message content suggestions” (See [Column 4, Lines 5-15]; Gangadharaiah discloses selecting one or more content suggestions from a provided set of suggestions by a chatbot system using a ML model.)
“receiving, via the message generation interface, in response to a presentation at the message generation interface of the selected at least one message content suggestion, feedback data related to the selected at least one message content suggestion;” (See [Column 3, Lines 62-67], [Column 4, Lines 1-15], [Column 7, Lines 17-23]; Gangadharaiah discloses that a selected agent utterance (content suggestion) from the set of candidate utterances is sent to the chatbot system 114 as feedback for analysis as a result of an utterance being selected. The chatbot system 114 can receive this through an agent system 112 (message generation interface))
“generating, by the tuned first machine learning model, a second set of message content suggestions based on the first message attribute data.” (See [Column 7, Lines 1-17]; Gangadharaiah discloses using the retrained (tuned) model to generate another set of responses based on user utterances and previous outputs (first message attribute data))
Gangadharaiah fails to explicitly disclose, “selecting… based on message evaluation data received by the first machine learning model from a second machine learning model”.
Perkins teaches “selecting… based on message evaluation data received by the first machine learning model from a second machine learning model” (See [0039]; Instead of relying on a single ML model, Perkins teaches that a second ML model is used to evaluate one or more predictions (potential content suggestions) from a set of data and evaluates the predictions based on certain criteria.)
Therefore, it would have been obvious to a person having ordinary skill in the art before the effective filing date of the claimed invention having Gangadharaiah and Perkins before them to modify Gangadharaiah to use a second ML model to evaluate predictions. One would be motivated to do so in order to separate prediction evaluation and selecting a message content suggestion to simplify and reduce workload done on the first ML model to produce more reliable results, see e.g., [0001], [0002], where Perkins teaches that some machine learning models may produce unreliable results, and to address this, Perkins describes a solution that uses two machine learning models for more reliable results.
Gangadharaiah fails to explicitly disclose, “configured to generate the message evaluation data by predicting a probability of message acceptance for the at least one message content suggestion”
“tuning the first machine learning model based on the feedback data”.
Xie teaches “configured to generate the message evaluation data by predicting a probability of message acceptance for the at least one message content suggestion” (See [0025]; Xie discloses receiving a generated engagement score that indicates the predicted probability of a user opening (accepting) a received communication message (message content suggestion)).
“tuning the first machine learning model based on the feedback data” (See [0017]; Xie discloses updating (tuning) a model using user feedback data).
Therefore, it would have been obvious to a person having ordinary skill in the art before the effective filing date of the claimed invention having Gangadharaiah and Xie before them to modify Gangadharaiah to generate message evaluation data by predicting the probability of a message’s acceptance as well as using the received feedback to improve a model. One would be motivated to generate evaluation data from predicting the probability of a message’s acceptance in order to determine how successful the message’s content was at attracting a user’s attention, see e.g., [0016], where Xie teaches generating an engagement score based on user interaction to gauge user engagement with a message. One would also be motivated to tune a model using user feedback data to update and improve the ML model based on how users interact with the given message content suggestions to tailor the model’s suggestions to the preferences of the user, see e.g., [0017], where Xie teaches updating a model after gathering user feedback to improve the model’s message content.
Regarding claim 2, Gangadharaiah discloses “the second set of message content suggestions comprises at least one of a reworded version of a message content suggestion of the first set of message content suggestions, a rephrasing of the message content suggestion, or an alternative version of the message content suggestion.” (See [Column 4, Lines 44-50]; Gangadharaiah discloses modifying a response (content suggestion) based on message history to rephrase or reword the response)
Regarding claim 3, Gangadharaiah discloses “receiving, via the message generation interface, second message attribute data;” (See [Column 7, Lines 17-23]; Gangadharaiah discloses that the second message attribute data is provided from the model)
“based on the second message attribute data, generating, by the first machine learning model, at least one of a reworded version of a message content suggestion of the first set of message content suggestions, a rephrasing of the message content suggestion, or an alternative version of the message content suggestion.” (See [Column 4, Lines 44-50]; Gangadharaiah discloses the agent (model) modifying one or more selected generated responses (content suggestion) after the responses are generated by the chatbot)
Regarding claim 6, Gangadharaiah fails to explicitly disclose, “tuning the second machine learning model based on the feedback data”.
Xie teaches “tuning the second machine learning model based on the feedback data” (See [0017]; Xie discloses updating (tuning) a model using user feedback data).
Therefore, it would have been obvious to a person having ordinary skill in the art before the effective filing date of the claimed invention having Gangadharaiah and Xie before them to modify Gangadharaiah to tune the second ML model based on the given user feedback data. One would be motivated to tune a model using user feedback data to update and improve the ML model based on how users interact with the given message content suggestions to tailor the model’s suggestions to the preferences of the user, see e.g., [0017], where Xie teaches updating a model after gathering user feedback to improve the model’s message content.
Regarding claim 9, Gangadharaiah discloses “the first machine learning model comprises a first encoder-decoder model architecture.” (See [Column 2, Lines 12-17], [Column 3, Lines 25-28]; Gangadharaiah discloses that the ML model generates outputs using a Seq2Seq ML model, which comprises an encoder-decoder model architecture).
Regarding claim 11, this claim is similar in scope to claim 1.
Regarding claim 12, this claim is similar in scope to claim 2.
Regarding claim 13, this claim is similar in scope to claim 3.
Regarding claim 16, this claim is similar in scope to claim 6.
Regarding claim 19, this claim is similar in scope to claim 9.
Claim Rejections - 35 USC § 103
Claim(s) 4, 7, 8, 14, 17, 18 are rejected under 35 U.S.C. 103 as being unpatentable over Gangadharaiah (US 10860629 B1), in view of Perkins (US 20220253647 A1), and Xie (US 20210157974 A1), and further in view of Shah (US 20230306049 A1) and Johnsen (US 20190043063 A1).
Regarding claim 4, Gangadharaiah fails to explicitly disclose, “outputting… estimated recipient acceptance data associated with the at least one message content suggestion”.
Shah teaches “outputting… estimated recipient acceptance data associated with the at least one message content suggestion” (See [0145], [0153]; Shah discloses a set of automated response options (content suggestions) being received from a sender, and also discloses that a recipient's response evaluation (acceptance data) can be recorded).
Shah fails to explicitly disclose, “outputting” is done by “the second machine learning model”.
Perkins teaches “outputting, by the second machine learning model” (See [0040]; Perkins discloses that one or more machine learning models can output information based on inputs, so the recipient's acceptance data could be outputted by a machine learning model).
Therefore, it would have been obvious to a person having ordinary skill in the art before the effective filing date of the claimed invention having Gangadharaiah, Shah, and Perkins before them to modify Gangadharaiah to output estimated message acceptance data using a second ML model. One would be motivated to do so in order to obtain the estimated message acceptance data after collecting and processing the message statistics from sending suggested messages to recipients. One would also be motivated to use a second ML model to allocate the work required to process the message statistics to lower the workload on the first ML model.
Gangadharaiah fails to explicitly disclose, “presenting the estimated recipient acceptance data to a prospective message sender via the message generation interface”.
Johnsen teaches “presenting the estimated recipient acceptance data to a prospective message sender via the message generation interface” (See [0043]; Johnsen discloses displaying information about engagement metrics (including recipient acceptance data) through a graphical user interface (message generation interface)).
Therefore, it would have been obvious to a person having ordinary skill in the art before the effective filing date of the claimed invention having Gangadharaiah and Johnsen before them to modify Gangadharaiah to display the information about the engagement metrics through a visual interface. One would be motivated to do so in order to evaluate the metrics through a convenient visual interface.
Regarding claim 7, Gangadharaiah fails to explicitly disclose, “the feedback data is based on at least one interaction of a prospective message sender with the message generation interface in response to a presentation by the message generation interface of the at least one message content suggestion prior to a sending of a message comprising the at least one message content suggestion by the prospective message sender to at least one prospective message recipient”.
Shah teaches “the feedback data is based on at least one interaction of a prospective message sender with the message generation interface in response to a presentation by the message generation interface of the at least one message content suggestion prior to a sending of a message comprising the at least one message content suggestion by the prospective message sender to at least one prospective message recipient” (See [0145], [0146]; Shah discloses a an interaction with a user (prospective message sender) and the message generation interface where the user is presented with generated message suggestions before sending a message to a prospective recipient. The user can choose to select, combine, edit, or ignore the responses before sending a message. The interaction that the user takes is used to train the model as feedback data).
Therefore, it would have been obvious to a person having ordinary skill in the art before the effective filing date of the claimed invention having Gangadharaiah and Shah before them to modify Gangadharaiah to collect feedback data after presenting users with editable message suggestions before they send a message. One would be motivated to do so in order to evaluate what kinds of choices a user makes when a user is offered a selection of editable messages to send to recipients before they send an initial message.
Regarding claim 8, Gangadharaiah fails to explicitly disclose, “the feedback data is based on at least one interaction of a prospective message recipient with a message receiving interface in response to a presentation by the message receiving interface of a message comprising the at least one message content suggestion to the prospective message recipient”.
Shah teaches “the feedback data is based on at least one interaction of a prospective message recipient with a message receiving interface in response to a presentation by the message receiving interface of a message comprising the at least one message content suggestion to the prospective message recipient” (See [0153]; Shah discloses recording feedback data for the model whenever the recipient leaves a positive response after receiving a message and/or a conversation).
Therefore, it would have been obvious to a person having ordinary skill in the art before the effective filing date of the claimed invention having Gangadharaiah and Shah before them to modify Gangadharaiah to record user interaction in response to a sent message. One would be motivated to do so in order to analyze how users interact with generated messages.
Regarding claim 14, this claim is similar in scope to claim 4.
Regarding claim 17, this claim is similar in scope to claim 7.
Regarding claim 18, this claim is similar in scope to claim 8.
Claim Rejections - 35 USC § 103
Claims 5, 15 are rejected under 35 U.S.C. 103 as being unpatentable over Gangadharaiah (US 10860629 B1), in view of Perkins (US 20220253647 A1), and Xie (US 20210157974 A1), and further in view of Lu (US 20210011919 A1) and Hirate (US 20230316307 A1).
Regarding claim 5, Gangadharaiah fails to explicitly disclose, “determining, based on a social graph, a link between a first entity and a second entity, wherein at least one of the first entity or the second entity represents, in the social graph, a prospective message recipient”.
Lu teaches “determining, based on a social graph, a link between a first entity and a second entity, wherein at least one of the first entity or the second entity represents, in the social graph, a prospective message recipient” (See [0113], [0130]; Lu discloses that a social network can be represented with a network graph, and also that a user's link between other prospective message recipients can be determined by a user defining a friend list that the user wants to potentially exchange messages with).
Lu fails to explicitly disclose, “based on the link, determining the first message attribute data”.
Hirate teaches “based on the link, determining the first message attribute data” (See [0085]; Hirate discloses determining attributes based on the link between users in a relationship graph (social graph)).
Therefore, it would have been obvious to a person having ordinary skill in the art before the effective filing date of the claimed invention having Gangadharaiah, Lu, and Hirate before them to modify Gangadharaiah to model a social network of friends and colleagues with a network graph and determining user attributes based on the links between users. One would be motivated to do so in order to easily determine who is related to other people in a social network with links and also to determine what kind of attributes connect one person to another based on the link between certain people.
Regarding claim 15, this claim is similar in scope to claim 5.
Claim Rejections - 35 USC § 103
Claims 10, 20 are rejected under 35 U.S.C. 103 as being unpatentable over Gangadharaiah (US 10860629 B1), in view of Perkins (US 20220253647 A1), and Xie (US 20210157974 A1), and further in view of Paiement (US 20230206096 A1).
Regarding claim 10, Gangadharaiah fails to disclose “the second machine learning model comprises a second encoder-decoder model architecture.”
Paiement teaches “the second machine learning model comprises a second encoder-decoder model architecture.” (See [0068]; Paiement discloses that a second MLM (Machine Learning Model) may comprise an encoder-decoder neural network)
Therefore, it would have been obvious to a person having ordinary skill in the art before the effective filing date of the claimed invention having Gangadharaiah and Paiement before them to modify Gangadharaiah to use a second machine learning model with the same encoder-decoder model architecture as the first model. One would be motivated to do so in order to improve performance and reduce workload on a single model by using a second model in conjunction with the first model.
Regarding claim 20, this claim is similar in scope to claim 10.
Response to Arguments
Applicant's arguments regarding the 35 USC 103 rejection are moot in view of the new grounds of rejection necessitated by applicant's amendments.
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 DAVID KIM whose telephone number is (571)272-4331. The examiner can normally be reached 7:30 AM - 4:30 PM.
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, Matthew Ell can be reached at (571) 270-3264. 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.
/D.K./Examiner, Art Unit 2141
/MATTHEW ELL/Supervisory Patent Examiner, Art Unit 2141
Prosecution Insights