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 Objections
Claim 20 is objected to under 37 CFR 1.75 as being a substantial duplicate of claim 17. When two claims in an application are duplicates or else are so close in content that they both cover the same thing, despite a slight difference in wording, it is proper after allowing one claim to object to the other as being a substantial duplicate of the allowed claim. See MPEP § 608.01(m).
Claim Rejections - 35 USC § 101
35 U.S.C. 101 reads as follows:
Whoever invents or discovers any new and useful process, machine, manufacture, or composition of matter, or any new and useful improvement thereof, may obtain a patent therefor, subject to the conditions and requirements of this title.
Claims 1-20 are rejected under 35 U.S.C. 101 because the claimed invention is
directed to an abstract idea without significantly more.
Step 1: The claims recite a method, system, and non-transitory machine-readable storage medium, each of which are one of the four categories of eligible subject matter.
Claims 1, 12, and 16
Step 2A Prong 1: The claims recite the following limitations:
using the at least one input to determine an entity identity (Mental Process); … causing the automated agent to machine-learn a supervision level via the context data, wherein the machine-learned supervision level indicates a level of supervision of the automated agent by an entity associated with the entity identity (Mental Process).
Under the broadest reasonable interpretation of the claim language, determining an entity identity and determining a supervision level are mental processes because a human
mind can practically perform the processes with the aid of a pencil, paper, and data.
Accordingly, the claims recite an abstract idea.
Step 2A Prong 2: The judicial exceptions are not integrated into a practical
application. The claims recite the following additional elements:
receiving at least one input via at least one device; …using the entity identity to create an automated agent and load context data associated with the entity identity into at least one layer of a multi-layer memory of the automated agent; …and configuring the automated agent to execute a task on behalf of the entity and in accordance with the machine-learned supervision level.
The processors and memory are generic computing components recited at a high
level as a means to apply the judicial exception, as discussed in MPEP 2106.05(f). Creating an automated agent and configuring the agent to execute tasks is generally linking the abstract ideas to the technological environment of machine learning,
as discussed in MPEP 2106.05(h). Receiving inputs is mere data gathering, which is an insignificant extra-solution activity as discussed in MPEP 2106.05(g). The claims are directed towards an abstract idea.
Step 2B: The claims do not include additional elements that are sufficient to amount to significantly more than the judicial exception. The processors and memory are generic computing components recited at a high level as a means to apply the judicial exception, as discussed in MPEP 2106.05(f). Creating an automated agent and configuring the agent to execute tasks is generally linking the abstract ideas to the technological environment of machine learning, as discussed in MPEP 2106.05(h). Receiving inputs is mere data gathering, which is an insignificant extra-solution activity as discussed in MPEP 2106.05(g). The claims are not patent eligible.
Dependent Claims
Claims 2-3, 6, 9-10, 13-14, 17-18, 20: These claims recite further abstract ideas (mental processes) and thus are ineligible.
Claims 4-5, 7-8, 11, 15, 19: These claims recite further mere data gathering and
generally linking the abstract ideas to the technological environment of machine learning
and as explained above these do not provide a practical application or inventive concept
and thus are ineligible.
Claim Rejections - 35 USC § 103
In the event the determination of the status of the application as subject to AIA 35 U.S.C. 102 and 103 (or as subject to pre-AIA 35 U.S.C. 102 and 103) is incorrect, any correction of the statutory basis (i.e., changing from AIA to pre-AIA ) for the rejection will not be considered a new ground of rejection if the prior art relied upon, and the rationale supporting the rejection, would be the same under either status.
The following is a quotation of 35 U.S.C. 103 which forms the basis for all obviousness rejections set forth in this Office action:
A patent for a claimed invention may not be obtained, notwithstanding that the claimed invention is not identically disclosed as set forth in section 102, if the differences between the claimed invention and the prior art are such that the claimed invention as a whole would have been obvious before the effective filing date of the claimed invention to a person having ordinary skill in the art to which the claimed invention pertains. Patentability shall not be negated by the manner in which the invention was made.
Claims 1-3, 7-13, 15-17, and 19-20 are rejected under 35 U.S.C. 103 as being unpatentable over Sharifi et al (Pub. No.: US 20220130379 A1), hereafter Sharifi in view of Vasylyev (Pub. No.: US 20240412720 A1). Examiner notes that Vasylyev has a priority date of 05/13/2024 due to provisional application 63/646605 reciting the teachings of Vasylyev below.
Regarding claims 1, 12, and 16, Sharifi teaches receiving at least one input via at least one device (user interface input devices of device 110 may collect inputs, P0015, P0034); using the at least one input to determine an entity identity (“In implementations that identify instance(s) of data for a particular user, the instance(s) of data and the associated past interaction(s) can be associated with the particular user utilizing one or more user verification techniques (e.g., speaker verification, facial verification, and/or other verification technique(s)).”, P0027); using the entity identity to create an automated agent and load context data associated with the entity identity into at least one layer of a … memory of the automated agent (“the system identifies one or more instances of data from one or more past assistant interactions that are each of a particular class and that are each initiated by a particular user and/or a particular assistant device”, P0027. “the system can determine to adapt automated assistant functionality or functionalities for any class(es) that have an associated class proficiency measure that satisfies a threshold. For instance, the system can determine to adapt automated assistant functionality or functionalities, for a particular class and for a particular user and/or particular user device”, P0030); causing the automated agent to machine-learn a supervision level via the context data, wherein the machine-learned supervision level indicates a level of supervision of the automated agent by an entity associated with the entity identity (“the system processes the individual metrics, using a trained machine learning model, to generate output, and determines the class proficiency measure based on the output. The output can, for example, be a measure from 0 to 1 whose magnitude indicates a degree of proficiency (e.g., with 1 indicating the most proficiency and 0 indicating the least). Such a machine learning model can be trained utilizing training instances that each include individual metrics for a corresponding particular user and/or corresponding particular user device”, P0057); and configuring the automated agent to execute a task on behalf of the entity and in accordance with the machine-learned supervision level (“After generating class proficiency measure(s), the system can proceed to block 258. At block 258, the system determines, for one or more of the class(es) for which a class proficiency measure has been generated and based on the corresponding class proficiency measure(s), whether to adapt automated assistant functionality/functionalities for the particular user device and/or the particular user”, P0030. “the system activates one or more capabilities that are specific to the class(es). The activated capabilities for a class can include previously inactive intent(s) and/or parameter(s) that are specific to the class. For example, an intent that enables renaming of smart devices via spoken input can be activated for a smart device control class. As another example, an intent that enables changing the color temperature of lights can be activated for a smart light control class”, P0034).
Sharifi does not appear to explicitly teach a “multi-layer memory”.
Vasylyev teaches a multi-layer memory (“the memory structure of assistant system 2 may be composed of multiple layers each designed to store and process information at different time frames, contextual background or levels of abstraction. This multilayer memory structure may be configured as a hierarchical memory structure designed to efficiently manage a vast amount of data, segregating it based on relevance and complexity”, P0564).
Accordingly, it would have been obvious to a person having ordinary skill in the
art before the effective filing date of the claimed invention, having the teachings of
Sharifi and Vasylyev before them, to include Vasylyev’s specific teaching of a hierarchical memory structure for an assistant system in Sharifi’s system of Adapting Automated Assistant Functionality Based On Generated Proficiency Measures. One would have been motivated to make such a combination of a hierarchical memory structure for an assistant system (see Vasylyev P0564) and selectively loading data associated with particular users onto memory (see Sharifi P0043) for reduced latency and improved accuracy and user experience within an AI assistant system (see Vasylyev P0009).
Regarding claims 2, and 13, Sharifi in view of Vasylyev teaches the limitations of claims 1 and 12 as outlined above. Sharifi further teaches encoding the automated agent with a first objective of executing the task in accordance with the machine-learned supervision level (trained machine learning model may be used to aid the automated assistant, P0057); and configuring the automated agent to, during execution of the task, test whether the first objective is met (machine learning model may verify that prompts/intents are completed. If unresolved, the NLU engine 124 can optionally work in concert with a dialog manager engine (not illustrated) that determines unresolved intent(s) and/or parameter(s) and/or generates corresponding prompt(s). The NLU engine 124 can utilize one or more NLU machine learning models in determining intent(s) and/or parameter(s). P0019).
Regarding claim 3, Sharifi in view of Vasylyev teaches the limitations of claim 1 as outlined above. Sharifi further teaches identifying a workflow associated with the task (“generating a rarity metric, of the individual metrics, based on comparison of: a particular measure that is based on usage of certain intents or certain parameters in the past interactions indicated by the data and a population-based measure that is based on usage of the certain intents or certain parameters in the other past interactions, of the class of interactions, from the population of users”, P0075); and training the automated agent to machine-learn at least one modification to the workflow via the context data (class proficiency measurements and metrics corresponding to intents and interactions with users are used to train the automated assistant, P0057, P0075).
Regarding claim 7, Sharifi in view of Vasylyev teaches the limitations of claim 1 as outlined above. Sharifi further teaches storing data obtained via an interaction between the entity and the automated agent during execution of the task in a first layer of the multi-layer memory (interaction data between a particular user and particular assistant are stored in the system, P0027); creating a compressed version of the data obtained via the interaction (user interaction between a particular user and particular assistant may be shortened, P0037).
Vasylyev further teaches storing the compressed version of the data obtained via the interaction in a second layer of the multi-layer memory (hierarchical memory may store compressed data, P0573, P0125).
Regarding claim 8, Sharifi in view of Vasylyev teaches the limitations of claim 7 as outlined above. Sharifi further teaches using the compressed version of the data obtained via the interaction stored in the second layer of the multi-layer memory to machine-learn at least one of the supervision level or a modification to a workflow associated with the task or a modification of the supervision level (automated assistants may implement adaptations of one or more blocks 264A-F, with block 264C being shortening interactions between a particular user and particular assistant, and block 264D being storing the shortened interaction, P0037, P0041-P0042, P0047. Block 264D includes machine learning based on user-assistant interactions, P0041).
Regarding claim 9, Sharifi in view of Vasylyev teaches the limitations of claim 1 as outlined above. Sharifi further teaches wherein configuring the automated agent to execute a task comprises: mapping the context data to at least one argument of a prompt (user input may be mapped to intents, P0044); and using a language model and the prompt including the at least one argument to generate and output a plan for executing the task to the automated agent (natural language understanding models carry out intents when mapped user input is processed by the automated assistant, P0019, P0044-P0046).
Regarding claim 10, Sharifi in view of Vasylyev teaches the limitations of claim 1 as outlined above. Sharifi further teaches querying at least one data resource, wherein the at least one data resource comprises at least one of entity profile data associated with the entity via an online system or entity interaction data associated with interactions of the entity with at least one of the automated agent or a different automated agent or the online system (particular users may interact with the automated assistant using computing systems communicating with an assistant device via one or more networks such as the internet, P0010); and using a language model to select the context data from among the entity profile data and the entity interaction data (natural language models may be used to process inputs of a particular user when interacting with an automated assistant, P0041-P0042).
Regarding claim 11, Sharifi in view of Vasylyev teaches the limitations of claim 1 as outlined above. Sharifi further teaches causing the automated agent to execute the task in accordance with the machine-learned supervision level (machine learning models may be used to determine intents, resolve intents, and determine parameters of intents, P0019); and storing data obtained during execution of the task in the multi-layer memory of the automated agent (in step 264D, data corresponding to intents and parameters of intents, P0042).
Regarding claims 15 and 19, Sharifi in view of Vasylyev teaches the limitations of claims 12 and 16 as outlined above. Sharifi further teaches storing data obtained via an interaction between the entity and the automated agent during execution of the task in a first layer of the multi-layer memory (interaction data between a particular user and particular assistant are stored in the system, P0027); creating a compressed version of the data obtained via the interaction (user interaction between a particular user and particular assistant may be shortened, P0037); storing the compressed version of the data obtained via the interaction in a second layer of the multi-layer memory (the system stores the shortened interaction, P0041-P0042); and using the compressed version of the data obtained via the interaction stored in the second layer of the multi-layer memory to machine-learn at least one of the supervision level or a modification to a workflow associated with the task or a modification of the supervision level (automated assistants may implement adaptations of one or more blocks 264A-F, with block 264C being shortening interactions between a particular user and particular assistant, and block 264D being storing the shortened interaction, P0037, P0041-P0042, P0047. Block 264D includes machine learning based on user-assistant interactions, P0041).
Regarding claims 17 and 20, Sharifi in view of Vasylyev teaches the limitations of claim 16 as outlined above. Sharifi further teaches encode the automated agent with a first objective of executing the task in accordance with the machine-learned supervision level (natural language understanding models determine actions of the automated assistant as intents, P0019); and configure the automated agent to, during execution of the task, test whether the first objective is met (the natural language understanding models can resolve intents based on user inputs processed by the automated assistant. If intents are unresolved, natural language understanding models can work with the dialog manager engine to generate corresponding prompts to ensure intents are executed based on user inputs, P0019).
Claims 4-6, 14, and 18 are rejected under 35 U.S.C. 103 as being unpatentable over Sharifi in view of Vasylyev and further in view of Anders et al (Pub. No.: US 20190325864 A1), hereafter Anders.
Regarding claim 4, Sharifi in view of Vasylyev teaches the limitations of claim 1 as outlined above. Vasylyev further teaches storing data obtained via an interaction between the entity and the automated agent in a first layer of the multi-layer memory (user-assistant interactions may be stored in memory units within the multilayer memory structure, P0090, P0564)… storing the machine-learned difference between the data obtained via the interaction and a machine-generated probable interaction between the entity and the automated agent in a second layer of the multi-layer memory (training data such as interaction logs may be stored in memory units within the multilayer memory structure, P0175, P0564).
Sharifi does not appear to explicitly teach “training the automated agent to machine-learn a difference between the data obtained via the interaction and a machine-generated probable interaction between the entity and the automated agent”.
Anders teaches training the automated agent to machine-learn a difference between the data obtained via the interaction and a machine-generated probable interaction between the entity and the automated agent (neural network may be trained based on the difference between the generated output corresponding to an associated probability and a label of a training example based on a user’s input provided to an automated assistant, P0044, P0008).
Accordingly, it would have been obvious to a person having ordinary skill in the
art before the effective filing date of the claimed invention, having the teachings of
Sharifi, Vasylyev, and Anders before them, to include Anders’s specific teaching of training a neural network based on a difference in an output corresponding to an associated probability and a label of a training example based on a user’s input provided to an automated assistant in Sharifi’s system of Adapting Automated Assistant Functionality Based On Generated Proficiency Measures. One would have been motivated to make such a combination of training a neural network based on a difference in an output corresponding to an associated probability and a label of a training example based on a user’s input provided to an automated assistant (see Anders P0044) and training a machine learning model based on an automated assistant (see Sharifi P0057) for more efficient consumption of computer and/or network resources (see Anders P0003).
Regarding claim 5, Sharifi in view of Vasylyev and further in view of Anders teaches the limitations of claim 4 as outlined above. Sharifi further teaches using the machine-learned difference between the interaction and the machine-generated probable interaction stored in the second layer of the multi-layer memory, training the automated agent to machine-learn at least one of the supervision level or a modification to a workflow associated with the task or a modification to the supervision level (“the system processes the individual metrics, using a trained machine learning model, to generate output, and determines the class proficiency measure based on the output. The output can, for example, be a measure from 0 to 1 whose magnitude indicates a degree of proficiency (e.g., with 1 indicating the most proficiency and 0 indicating the least). Such a machine learning model can be trained utilizing training instances that each include individual metrics for a corresponding particular user and/or corresponding particular user device”, P0057).
Regarding claim 6, Sharifi in view of Vasylyev and further in view of Anders teaches the limitations of claim 5 as outlined above. Anders further teaches generating a prior probability distribution using historical interactions between the entity and the automated agent (feed forward neural network may be used to generate probability distributions based on user interactions with the automated assistant, P0008); and using the prior probability distribution to produce the machine-generated probable interaction (neural network may generate outputs corresponding to probability distributions based on interactions between users and the automated assistant, P0044, P0008).
Regarding claims 14 and 18, Sharifi in view of Vasylyev teaches the limitations of claims 12 and 16 as outlined above. Sharifi further teaches storing data obtained via an interaction between the entity and the automated agent in a first layer of the multi-layer memory (interaction data between a particular user and particular assistant are stored in the system, P0027).
Sharifi does not appear to explicitly teach “storing a machine-learned difference between the data obtained via the interaction and a machine-generated probable interaction between the entity and the automated agent in a second layer of the multi-layer memory; using the machine-learned difference between the interaction and the machine-generated probable interaction stored in the second layer of the multi-layer memory to machine-learn at least one of the supervision level or a modification to a workflow associated with the task or a modification to the supervision level; generating a prior probability distribution using historical interactions between the entity and the automated agent; and using the prior probability distribution to produce the machine-generated probable interaction”.
Anders teaches storing a machine-learned difference between the data obtained via the interaction and a machine-generated probable interaction between the entity and the automated agent in a second layer of the multi-layer memory (database 113 stores the trained model including the difference between the generated output and the training label based on a user’s input provided to an automated assistant, P0044); using the machine-learned difference between the interaction and the machine-generated probable interaction stored in the second layer of the multi-layer memory to machine-learn at least one of the supervision level or a modification to a workflow associated with the task or a modification to the supervision level (neural network may be trained based on the difference between the generated output corresponding to an associated probability and a label of a training example based on a user’s input provided to an automated assistant, P0044, P0008); generating a prior probability distribution using historical interactions between the entity and the automated agent (feed forward neural network may be used to generate probability distributions based on user interactions with the automated assistant, P0008); and using the prior probability distribution to produce the machine-generated probable interaction (neural network may generate outputs corresponding to probability distributions based on interactions between users and the automated assistant, P0044, P0008).
Accordingly, it would have been obvious to a person having ordinary skill in the
art before the effective filing date of the claimed invention, having the teachings of
Sharifi, Vasylyev, and Anders before them, to include Anders’s specific teaching of training a neural network based on a difference in an output corresponding to an associated probability and a label of a training example based on a user’s input provided to an automated assistant in Sharifi’s system of Adapting Automated Assistant Functionality Based On Generated Proficiency Measures. One would have been motivated to make such a combination of training a neural network based on a difference in an output corresponding to an associated probability and a label of a training example based on a user’s input provided to an automated assistant (see Anders P0044) training a machine learning model based on an automated assistant (see Sharifi P0057) for more efficient consumption of computer and/or network resources (see Anders P0003).
Conclusion
The prior art made of record and not relied upon is considered pertinent to applicant's disclosure.
US 20200104357 A1 (BELLEGARDA et al) teaches a system of automated assistants including training neural networks using differences in a probability distribution of machine generated content and a probability distribution of user errors.
US 20120173464 A1 (Tur et al) teaches a method for an automated assistant.
Any inquiry concerning this communication or earlier communications from the examiner should be directed to ISHAN MOUNDI whose telephone number is (703)756-1547. The examiner can normally be reached 8:30 A.M. - 5 P.M..
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.
/I.M./Examiner, Art Unit 2141
/MATTHEW ELL/Supervisory Patent Examiner, Art Unit 2141
Prosecution Insights