DETAILED ACTION
This action is in response to the amendments filed 12/30/2025. Claims 1-20 are pending and have been examined.
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 .
Information Disclosure Statement
The information disclosure statements (IDS) submitted on 10/13/2025, 11/21/2025, and 04/06/2026 are in compliance with the provisions of 37 CFR 1.97. Accordingly, the information disclosure statements are being considered by the examiner.
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.
Regarding claim 1:
Subject Matter Eligibility Analysis Step 1:
Claim 1 recites “A method of managing a distribution of inference models hosted by data processing systems”, thus it is a process, one of the four statutory categories of patentable subject matter.
Subject Matter Eligibility Analysis Step 2A Prong 1:
Claim 1 recites the step:
“prior to the future period of time, updating the inference models based on the deployment plan to obtain an updated distribution of inference models”: This involves a human updating inference models based on the deployment plan in order to obtain updated models prior to a future point in time. Hence this is a mental process.
Thus, claim 1 recites abstract ideas.
Subject Matter Eligibility Analysis Step 2A Prong 2:
Claim 1 recites the additional elements:
“obtaining condition data for the data processing systems, the condition data indicating how the data processing systems are likely to operate during a future period of time”: This element does not integrate the abstract idea above into a practical application because the element recites an insignificant extra solution activity of data transmission (MPEP 2106.05(g)).
“obtaining a deployment plan for the inference models based on the condition data, the distribution, and inference model types available for deployment, the deployment plan specify an operating mode, how each of the inference models is divided into portions, and where the portions are distributed in the data processing systems”: This element does not integrate the abstract idea into a practical application because the element recites an insignificant extra solution activity of data transmission (MPEP 2106.05(g)).
“obtaining, during the period of time, an inference using the updated distribution of the inference models and the operating mode, the operating mode indicating at least one automatically implemented change to the updated distribution of the inference models responsive to an occurrence of a condition”: This element does not integrate the abstract idea above into a practical application because the element recites generic computing components (inference models) to implement the abstract ideas (MPEP 2106.05(f)).
Therefore, claim 1 is directed to the abstract ideas.
Subject Matter Eligibility Analysis Step 2B:
The additional elements in claim 1 do not provide significantly more than the abstract ideas themselves, taken alone and in combination because:
“obtaining condition data for the data processing systems, the condition data indicating how the data processing systems are likely to operate during a future period of time”: This element mentions a well-understood, routine, and conventional activity of “receiving or transmitting data over a network” (MPEP 2106.05(d)(I), Intellectual Ventures v. Symantec, 838 F.3d 1307, 1321; 120 USPQ2d 1353, 1362 (Fed. Cir. 2016) [utilizing an intermediary computer to forward information])
“obtaining a deployment plan for the inference models based on the condition data, the distribution, and inference model types available for deployment, the deployment plan specify an operating mode, how each of the inference models is divided into portions, and where the portions are distributed in the data processing systems: This element mentions a well-understood, routine, and conventional activity of “receiving or transmitting data over a network” (MPEP 2106.05(d)(I), Intellectual Ventures v. Symantec, 838 F.3d 1307, 1321; 120 USPQ2d 1353, 1362 (Fed. Cir. 2016) [utilizing an intermediary computer to forward information])
“obtaining, during the period of time, an inference using the updated distribution of the inference models and the operating mode, the operating mode indicating at least one automatically implemented change to the updated distribution of the inference models responsive to an occurrence of a condition”: This element recites generic computing components (inference models) to implement the abstract ideas (MPEP 2106.05(f)).
Since there is no nexus between the additional elements that could cause the combination to provide an inventive concept, claim 1 is subject-matter ineligible.
Regarding claim 2:
Subject Matter Eligibility Analysis Step 1:
Claim 2 is a process as in claim 1
Subject Matter Eligibility Analysis Step 2A Prong 1:
In addition to the mental concepts in claim 1, claim 2 recites:
“identifying a time of day associated with the future period of time”: This involves a human identifying a time in the future; hence this is a mental process.
“identifying an occurrence of an event”: This could involve a human determining the occurrence of an event. Thus, this is also a mental process.
“identifying a geographic area in which at least a portion of the data processing systems are likely to reside during the future period of time”: This could involve a human identifying the location that a portion of the data processing systems are located during a future point in time. Hence, this is a mental process.
Claim 2 therefore recites abstract ideas.
Subject Matter Eligibility Analysis Step 2A Prong 2:
Claim 2 recites the same additional elements as claim 1, thus claim 2 is directed to the abstract ideas.
Subject Matter Eligibility Analysis Step 2B:
Since claim 2 recites the same additional elements as claim 1 and there is no nexus between the additional elements that could cause the combination to provide an inventive concept, claim 2 is subject-matter ineligible.
Regarding claim 3:
Subject Matter Eligibility Analysis Step 1:
Claim 3 is a process as in claim 2
Subject Matter Eligibility Analysis Step 2A Prong 1:
In addition to the mental concepts in claim 2, claim 3 recites:
identifying, based on the inference model types, a first inference model type associated with the time of the day”: This involves a human identifying a first inference model that is associated with a certain time in the future; hence this is a mental process.
identifying a first entry of a previous deployment plan for a second inference model type that is not associated with the time of the day, the distribution of the inference models being based on the previous deployment plan”: This could involve a human identifying a first entry of a previous deployment plan for another inference model that is not associated with a certain time in the day. Thus, this is also a mental process.
replacing the first entry with a second entry to obtain the deployment plan, the second entry indicating that an instance of the second inference model type should be present in the distribution of the inference models”: This could involve a human replacing the first entry from step (II) with a second entry from the second model in step (II). Hence, this is a mental process.
Claim 3 therefore recites abstract ideas.
Subject Matter Eligibility Analysis Step 2A Prong 2:
Claim 3 recites the same additional elements as claim 2, thus claim 3 is directed to the abstract ideas.
Subject Matter Eligibility Analysis Step 2B:
Since claim 3 recites the same additional elements as claim 2 and there is no nexus between the additional elements that could cause the combination to provide an inventive concept, claim 3 is subject-matter ineligible.
Regarding claim 4:
Subject Matter Eligibility Analysis Step 1:
Claim 4 is a process as in claim 2
Subject Matter Eligibility Analysis Step 2A Prong 1:
In addition to the mental concepts in claim 2, claim 4 recites:
“identifying, based on the inference model types, a first inference model type associated with the event”: This involves a human identifying a first inference model that is associated with a certain event; hence this is a mental process.
“identifying a first entry of a previous deployment plan for a second inference model type that is not associated with the event, the distribution of the inference models being based on the previous deployment plan”: This could involve a human identifying a first entry of a previous deployment plan for another inference model that is not associated with a certain event. Thus, this is also a mental process.
“replacing the first entry with a second entry to obtain the deployment plan, the second entry indicating that an instance of the second inference model type should be present in the distribution of the inference models”: This could involve a human replacing the first entry from step (II) with a second entry from the second model in step (II). Hence, this is a mental process.
Claim 4 therefore recites abstract ideas.
Subject Matter Eligibility Analysis Step 2A Prong 2:
Claim 4 recites the same additional elements as claim 2, thus claim 4 is directed to the abstract ideas.
Subject Matter Eligibility Analysis Step 2B:
Since claim 4 recites the same additional elements as claim 2 and there is no nexus between the additional elements that could cause the combination to provide an inventive concept, claim 4 is subject-matter ineligible.
Regarding claim 5:
Subject Matter Eligibility Analysis Step 1:
Claim 5 is a process as in claim 4.
Subject Matter Eligibility Analysis Step 2A Prong 1:
In addition to the mental concepts in claim 4, claim 5 recites:
“adding a new entry to the previous deployment plan to obtain the deployment plan, the new entry indicating that another instance of the second inference model type should be present in the distribution of the inference models”: This involves a human adding a new entry to the previous deployment plan in claim 4 in order to obtain a new deployment plan, therefore this is a mental process.
Claim 5 therefore recites abstract ideas.
Subject Matter Eligibility Analysis Step 2A Prong 2:
Claim 5 recites the same additional elements as claim 4, thus claim 5 is directed to the abstract ideas.
Subject Matter Eligibility Analysis Step 2B:
Since claim 5 recites the same additional elements as claim 4 and there is no nexus between the additional elements that could cause the combination to provide an inventive concept, claim 5 is subject-matter ineligible.
Regarding claim 6:
Subject Matter Eligibility Analysis Step 1:
Claim 6 is a process as in claim 2
Subject Matter Eligibility Analysis Step 2A Prong 1:
In addition to the mental concepts in claim 2, claim 6 recites:
“identifying, based on the inference model types, a first inference model type associated with the geographic location”: This involves a human identifying a first inference model that is associated with a certain geographic location; hence this is a mental process.
“identifying a first entry of a previous deployment plan for a second inference model type that is not associated with the geographic location, the distribution of the inference models being based on the previous deployment plan”: This could involve a human identifying a first entry of a previous deployment plan for another inference model that is not associated with a certain geographic location. Thus, this is also a mental process.
“replacing the first entry with a second entry to obtain the deployment plan, the second entry indicating that an instance of the second inference model type should be present in the distribution of the inference models”: This could involve a human replacing the first entry from step (II) with a second entry from the second model in step (II). Hence, this is a mental process.
Claim 6 therefore recites abstract ideas.
Subject Matter Eligibility Analysis Step 2A Prong 2:
Claim 6 recites the same additional elements as claim 2, thus claim 6 is directed to the abstract ideas.
Subject Matter Eligibility Analysis Step 2B:
Since claim 6 recites the same additional elements as claim 2 and there is no nexus between the additional elements that could cause the combination to provide an inventive concept, claim 6 is subject-matter ineligible.
Regarding claim 7:
Subject Matter Eligibility Analysis Step 1:
Claim 7 is a process as in claim 2.
Subject Matter Eligibility Analysis Step 2A Prong 1:
Claim 7 recites the same mental concepts as claim 2, thus claim 7 recites abstract ideas.
Subject Matter Eligibility Analysis Step 2A Prong 2:
In addition to the additional elements in claim 2, claim 7 recites:
“the distribution of the inference models comprises an inference model over fitted for a second time of the day, the updated distribution of the inference model comprises a second inference model overfitted for the time of the days, and the inference model is not a member of the updated distribution of the inference models”: This element does not integrate the abstract ideas into a practical application because the element recites a technological environment (machine learning model overfitting) in which to apply a judicial exception (MPEP 2106.05(h)).
Subject Matter Eligibility Analysis Step 2B:
The additional elements in claim 7 do not provide significantly more than the abstract ideas themselves taken alone and in combination because:
“the distribution of the inference models comprises an inference model over fitted for a second time of the day, the updated distribution of the inference model comprises a second inference model overfitted for the time of the days, and the inference model is not a member of the updated distribution of the inference models”: This element recites a technological environment (machine learning model overfitting) in which to apply a judicial exception (MPEP 2106.05(h)).
Since there is no nexus between the additional elements that could cause the combination to provide an inventive concept, claim 7 is subject-matter ineligible.
Regarding claim 8:
Subject Matter Eligibility Analysis Step 1:
Claim 8 recites “A non-transitory machine-readable medium having instructions stored therein, which when executed by a processor, cause the processor to perform operations…”, thus it is an article of manufacture, one of the four statutory categories of patentable subject matter.
Subject Matter Eligibility Analysis Step 2A Prong 1:
Claim 8 recites the step:
“prior to the future period of time, updating the inference models based on the deployment plan to obtain an updated distribution of inference models”: This involves a human updating inference models based on the deployment plan in order to obtain updated models prior to a future point in time. Hence this is a mental process.
Thus, claim 8 recites abstract ideas.
Subject Matter Eligibility Analysis Step 2A Prong 2:
Claim 8 recites the additional elements:
“A non-transitory machine-readable medium having instructions stored therein, which when executed by a processor, cause the processor to perform operations for managing a distribution of inference models hosted by the data processing systems”: This element does not integrate the abstract idea in Step 2A Prong 1 into a practical application because the element recites a generic computing component (MPEP 2106.05(f)).
“obtaining condition data for the data processing systems, the condition data indicating how the data processing systems are likely to operate during a future period of time”: This element does not integrate the abstract idea above into a practical application because the element recites an insignificant extra solution activity of data transmission (MPEP 2106.05(g)).
“obtaining a deployment plan for the inference models based on the condition data, the distribution, and inference model types available for deployment, the deployment plan specify an operating mode, how each of the inference models is divided into portions, and where the portions are distributed in the data processing systems”: This element does not integrate the abstract idea into a practical application because the element recites an insignificant extra solution activity of data transmission (MPEP 2106.05(g)).
“obtaining, during the period of time, an inference using the updated distribution of the inference models and the operating mode, the operating mode indicating at least one automatically implemented change to the updated distribution of the inference models responsive to an occurrence of a condition”: This element does not integrate the abstract idea above into a practical application because the element recites generic computing components (inference models) to implement the abstract ideas (MPEP 2106.05(f)).
Therefore, claim 8 is directed to the abstract ideas.
Subject Matter Eligibility Analysis Step 2B:
The additional elements in claim 8 do not provide significantly more than the abstract ideas themselves, taken alone and in combination because:
“A non-transitory machine-readable medium having instructions stored therein, which when executed by a processor, cause the processor to perform operations for managing a distribution of inference models hosted by the data processing systems”: This element recites a generic computing component (MPEP 2106.05(f)).
“obtaining condition data for the data processing systems, the condition data indicating how the data processing systems are likely to operate during a future period of time”: This element mentions a well-understood, routine, and conventional activity of “receiving or transmitting data over a network” (MPEP 2106.05(d)(I), Intellectual Ventures v. Symantec, 838 F.3d 1307, 1321; 120 USPQ2d 1353, 1362 (Fed. Cir. 2016) [utilizing an intermediary computer to forward information])
“obtaining a deployment plan for the inference models based on the condition data, the distribution, and inference model types available for deployment, the deployment plan specify an operating mode, how each of the inference models is divided into portions, and where the portions are distributed in the data processing systems”: This element mentions a well-understood, routine, and conventional activity of “receiving or transmitting data over a network” (MPEP 2106.05(d)(I), Intellectual Ventures v. Symantec, 838 F.3d 1307, 1321; 120 USPQ2d 1353, 1362 (Fed. Cir. 2016) [utilizing an intermediary computer to forward information])
“obtaining, during the period of time, an inference using the updated distribution of the inference models and the operating mode, the operating mode indicating at least one automatically implemented change to the updated distribution of the inference models responsive to an occurrence of a condition”: This element recites generic computing components (inference models) to implement the abstract ideas (MPEP 2106.05(f)).
Since there is no nexus between the additional elements that could cause the combination to provide an inventive concept, claim 8 is subject-matter ineligible.
Regarding claim 9:
Claim 9 is an article of manufacture as claim 8 and is rejected for the same reasons as claim 2.
Regarding claim 10:
Claim 10 is an article of manufacture as claim 9 and is rejected for the same rationale as claim 3.
Regarding claim 11:
Claim 11 is an article of manufacture as claim 9. Additionally claim 11 is rejected for the same reasons as claim 4.
Regarding claim 12:
Claim 12 is an article of manufacture as claim 11 and is rejected for the same reasons as claim 5.
Regarding claim 13:
Claim 13 is an article of manufacture as claim 9 and is rejected for the same rationale as claim 6.
Regarding claim 14:
Claim 14 is an article of manufacture as claim 9. Additionally claim 14 is rejected for the same reasons as claim 7.
Regarding claim 15:
Subject Matter Eligibility Analysis Step 1:
Claim 15 recites “A data processing system comprising…a processor; and a memory coupled to the processor….”, thus it is an article of manufacture, one of the four statutory categories of patentable subject matter.
Subject Matter Eligibility Analysis Step 2A Prong 1:
Claim 15 recites the step:
“prior to the future period of time, updating the inference models based on the deployment plan to obtain an updated distribution of inference models”: This involves a human updating inference models based on the deployment plan in order to obtain updated models prior to a future point in time. Hence this is a mental process.
Thus, claim 15 recites abstract ideas.
Subject Matter Eligibility Analysis Step 2A Prong 2:
Claim 15 recites the additional elements:
“a processor”: This element does not integrate the abstract idea in Step 2A Prong 1 into a practical application because the element recites a generic computing component (MPEP 2106.05(f)).
“a memory coupled to the processor to store instructions, which when executed by the processor, cause the processor to perform operations for managing a distribution of inference models hosted by data processing systems”: This element does not integrate the abstract idea in Step 2A Prong 1 into a practical application because the element recites a generic computing component (MPEP 2106.05(f)).
“obtaining condition data for the data processing systems, the condition data indicating how the data processing systems are likely to operate during a future period of time”: This element does not integrate the abstract idea above into a practical application because the element recites an insignificant extra solution activity of data transmission (MPEP 2106.05(g)).
“obtaining a deployment plan for the inference models based on the condition data, the distribution, and inference model types available for deployment, the deployment plan specify an operating mode, how each of the inference models is divided into portions, and where the portions are distributed in the data processing systems”: This element does not integrate the abstract idea into a practical application because the element recites an insignificant extra solution activity of data transmission (MPEP 2106.05(g)).
“obtaining, during the period of time, an inference using the updated distribution of the inference models and the operating mode, the operating mode indicating at least one automatically implemented change to the updated distribution of the inference models responsive to an occurrence of a condition”: This element does not integrate the abstract idea above into a practical application because the element recites generic computing components (inference models) to apply the abstract ideas (MPEP 2106.05(f)).
Therefore, claim 15 is directed to the abstract ideas.
Subject Matter Eligibility Analysis Step 2B:
The additional elements in claim 15 do not provide significantly more than the abstract ideas themselves, taken alone and in combination because:
“a processor”: This element recites a generic computing component (MPEP 2106.05(f)).
“a memory coupled to the processor to store instructions, which when executed by the processor, cause the processor to perform operations for managing a distribution of inference models hosted by data processing systems”: This element recites a generic computing component (MPEP 2106.05(f)).
“obtaining condition data for the data processing systems, the condition data indicating how the data processing systems are likely to operate during a future period of time”: This element mentions a well-understood, routine, and conventional activity of “receiving or transmitting data over a network” (MPEP 2106.05(d)(I), Intellectual Ventures v. Symantec, 838 F.3d 1307, 1321; 120 USPQ2d 1353, 1362 (Fed. Cir. 2016) [utilizing an intermediary computer to forward information])
“obtaining a deployment plan for the inference models based on the condition data, the distribution, and inference model types available for deployment, the deployment plan specify an operating mode, how each of the inference models is divided into portions, and where the portions are distributed in the data processing systems”: This element mentions a well-understood, routine, and conventional activity of “receiving or transmitting data over a network” (MPEP 2106.05(d)(I), Intellectual Ventures v. Symantec, 838 F.3d 1307, 1321; 120 USPQ2d 1353, 1362 (Fed. Cir. 2016) [utilizing an intermediary computer to forward information])
“obtaining, during the period of time, an inference using the updated distribution of the inference models and the operating mode, the operating mode indicating at least one automatically implemented change to the updated distribution of the inference models responsive to an occurrence of a condition”: This element recites generic computing components (inference models) to apply the abstract ideas (MPEP 2106.05(f)).
Since there is no nexus between the additional elements that could cause the combination to provide an inventive concept, claim 15 is subject-matter ineligible.
Regarding claim 16:
Claim 16 is an article of manufacture as claim 15 and is rejected for the same reasons as claims 2
and 9.
Regarding claim 17:
Claim 17 is an article of manufacture as claim 15 and is rejected for the same rationale as claims 3 and 10.
Regarding claim 18:
Claim 18 is an article of manufacture as claim 15. Additionally claim 18 is rejected for the same reasons as claims 4 and 11.
Regarding claim 19:
Claim 19 is an article of manufacture as claim 18 and is rejected for the same reasons as claims 5 and 12.
Regarding claim 20:
Claim 20 is an article of manufacture as claim 16 and is rejected for the same rationale as claims 6 and 13.
Claim Rejections - 35 USC § 102
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 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.
Claims 1-20 are rejected under 35 U.S.C. 102 (a) (1) as being anticipated over Eck et al. (Scalable Deployment of AI Time-series Models for IoT, published 2020, arXiv:2003.12141v1)
Regarding claim 1, Eck et al. teaches “A method of managing a distribution of inference models hosted by data processing systems” (Eck et al. Section 1 “The remainder of the paper is organised as follows. Section 2 overviews the system architecture and workflow. The focus is then put on the AI model preparation and deployment, in Section 3.”; “workflow” corresponds to “A method”) “the method comprising:”
“obtaining condition data for the data processing systems, the condition data indicating how the data processing systems are likely to operate during a future period of time” (Eck et al., Section 2 “Invocations of certain micro-services are triggered based on real-world events, such as the availability of time-series data (1). The IoT data is ingested often at irregular frequencies, with many devices submitting data in parallel. Ingested time-series are stored in a knowledge-based data store. As new time-series become available, data scientists can provide additional semantics (2) to contextualise the time-series [Chen et al., 2018]. IoT data ingestion and semantic definition occur out-of-band, and do not directly affect the main modelling work flow”; “provide…semantics” corresponds to “obtaining condition data”; “IoT data” corresponds to “data processing systems”; Eck et al. Listing 2 defines a “scoring_deployment” field with the parameters “time” and “repeatEvery” which signals the time and repeated time intervals to perform model scoring; hence it corresponds to “during a future period of time”);
“obtaining a deployment plan for the inference models based on the condition data, the distribution, and inference model types available for deployment, the deployment plan specifying: an operating mode” (Eck et al., Section 2 “Based on semantics (2) and desired model training/scoring schedules, model deployments…are written (5) and registered with the system (6)”; Listing 2 defines a deployment plan containing a “context”; Section 3.1 “…. context provides semantic information associated with the time-series targeted by a specific model instance, in terms of concepts of signals (what physical quantity, what unit) and entity (what location, what name, type, GIS coordinates,…) …modelID and modelVersion point to the model and model version data”; Section 2 “The model deployments…are stored in a database and specify the execution of the model implementations against specific instances of the application semantic concepts. A model scheduling micro-service (7) periodically….determines if each model in question is due for training or scoring…”’;“model deployments…are written” correspond to “obtaining a deployment plan for the inference models”; “desired model training/scoring schedules” correspond to “the distribution”; “modelID and modelVersion” correspond to “inference model types available for deployment”; “specify the execution of the model implementations” correspond to “the deployment plan specify an operating mode”; “semantics” correspond to “condition data”);
“the deployment plan specifying … , how each of the inference models is divided into portions, and where the portions are distributed in the data processing systems” (Eck et al., Figure 3 “Semantic representation of the IoT data”
PNG
media_image1.png
610
711
media_image1.png
Greyscale
; Figure 5 “Deployed models and trained versions for a context”
PNG
media_image2.png
479
593
media_image2.png
Greyscale
; Section 4.2 “Specific model instances were deployed by preparing model deployment configurations, as outlined in Section 3.2. The semantic context was set to the target entity (e.g. SUBSTATION S1) and signal (e.g. ENERGY LOAD), the training and scoring schedules were defined (hourly and weekly, respectively), custom user parameters were used to specify a training window, the time and the desired forecasting window … The models deployed for a given context can be explored in the semantic graph, as shown in Fig. 5”; the target entity corresponds to “where the portions are distributed in the data processing systems”; Section 4.3 “The deployment of hundreds of models was semi-automated by programmatically setting the deployment configurations for the AI model implementations defined in Section 4.2 to all required semantic targets. For example, the 174 models of site 3 are based on only 6 model implementations”; six inference models are split into 174 portions across a network, thus the context comprises “how each of the inference models is divided into portions” and “where the portions are distributed in the data processing systems”.
“prior to the future period of time, updating the inference models based on the deployment plan to obtain an updated distribution of the inference models” (Eck et al., Section 2 “A model scheduling micro-service (7) periodically loads the registered model deployments and determines if each model in question is due for training or scoring, based on the user-specified schedules. Model execution involves automatically installing the model implementation (green) from the PyPi repository (8) and running the relevant model implementation routines (3) with specific model deployment configuration (5) and semantic data (2) loaded from the system data store. At model training, historical data (9) are retrieved from the time-series micro-service suite and a new model version object (pink) is produced, including a model object with fitted parameters (for example the weights of a neural network) and some training metadata, such as train time. The model version is saved into the data store and mapped to the semantic context of the model. At model scoring, the model version object is loaded from the system data store and made available to the scoring routine, which produces a time-series forecast (blue). The forecast micro-service (10) handles the persistence of prediction time-series in a database.”; “each model” and “a new model version object…with fitted parameters” correspond to “updating the inference models”; “running the relevant model implementation routines (3) with specific model deployment configuration (5)” corresponds to “based on the deployment plan”; “a new model version object…is produced” corresponds to “an updated distribution of the inference models”; “At model scoring” corresponds to a “future period of time”; since “running the relevant model implementation routines” occurs prior to model scoring, this corresponds to “prior to the future period of time” ); and
“obtaining, during the period of time, an inference using the updated distribution of the inference models and the operating mode, the operating mode indicating at least one automatically implemented change to the updated distribution of the inference models responsive to an occurrence of a condition” (Eck et al., Section 2 “At model training, historical data (9) are retrieved from the time-series micro-service suite and a new model version object (pink) is produced, including a model object with fitted parameters (for example the weights of a neural network) and some training metadata, such as train time. The model version is saved into the data store and mapped to the semantic context of the model. At model scoring, the model version object is loaded from the system data store and made available to the scoring routine, which produces a time-series forecast (blue). The forecast micro-service (10) handles the persistence of prediction time-series in a database.”; “At model scoring” corresponds to “during the period of time”; “produces a time-series forecast” and “the system data store” which store multiple models for model scoring corresponds to “an inference using the updated distribution of the inference models”; “a new model version object” that was used at model scoring corresponds to “the operating mode”; “a model object with fitted parameters…and some training metadata, such as train time” corresponds to “the operating mode indicating the at least one automatically implemented change to the updated distribution of the inference models”; “historical data (9) are retrieved from the time-series micro-service suite and a new model version object…is produced” corresponds to “responsive to an occurrence of a condition”)
Regarding claim 2, the rejections of claim 1 are incorporated. Eck et al. further recites:
“wherein obtaining the condition data for the data processing systems comprises at least one selected from a group consisting of…identifying a time of day associated with the future period of time” (Eck et al. Listing 2 defines a “scoring_deployment” field with the parameters “time” and “repeatEvery” which signals the time and repeated time intervals to perform model scoring; hence it corresponds to “identifying a time of day associated with the future period of time”);
“wherein obtaining the condition data for the data processing systems comprises at least one selected from a group consisting of…identifying an occurrence of an event” (Eck et al. Listing 2 describes a “user_parameters” field with the parameter “train_time” which signals the times when a model was trained, hence “train_time” corresponds to “an occurrence of an event”); and
“wherein obtaining the condition data for the data processing systems comprises at least one selected from a group consisting of…identifying a geographic area in which at least a portion of the data processing systems are likely to reside during the future period of time.” (Eck et al. Section 2 “Based on semantics (2) and desired model training/scoring schedules, model deployments…are written (5) and registered with the system (6)”; Listing 2 defines a deployment plan containing a “context”; Section 3.1 “…. context provides semantic information associated with the time-series targeted by a specific model instance, in terms of concepts of signals (what physical quantity, what unit) and entity (what location, what name, type, GIS coordinates,…) …modelID and modelVersion point to the model and model version data”; Section 2 “The model version is saved into the data store and mapped to the semantic context of the model. At model scoring, the model version object is loaded from the system data store and made available to the scoring routine, which produces a time-series forecast (blue).”; “entity (what location…GIS coordinates,…)” corresponds to “identifying a geographic area in which at least a portion of the data processing systems are likely to reside”; “model scoring” corresponds to “future period of time”)
Regarding claim 3, the rejections of claim 2 are incorporated. Eck et al. further discloses “obtaining the deployment plan comprises:”
“identifying, based on the inference model types, a first inference model type associated with the time of the day” (Eck et al. Section 2 “The model deployments (yellow) are stored in a database and specify the execution of the model implementations against specific instances of the application semantic concepts. A model scheduling micro-service (7) periodically loads the registered model deployments and determines if each model in question is due for training or scoring, based on the user-specified schedules. Model execution involves automatically installing the model implementation (green) from the PyPi repository (8) and running the relevant model implementation routines (3) with specific model deployment configuration (5) and semantic data (2) loaded from the system data store.”; Eck et al. Listing 2 defines an example of a model deployment for each model with fields “scoring_deployment” and “user_parameters” with the parameters “frequency”, “time” and “repeatEvery” which signals the time and repeated time intervals to perform model scoring; “A model scheduling micro-service” that “determines if each model…is due for…scoring based on the user-specified schedules” and “user_parameters” in each model deployment corresponds to “identifying, based on the inference model types, a first model type”; “time” corresponds to “associated with the time of the day”);
“identifying a first entry of a previous deployment plan for a second inference model type that is not associated with the time of the day, the distribution of the inference models being based on the previous deployment plan” (Eck et al. Section 2 “The model deployments (yellow) are stored in a database and specify the execution of the model implementations against specific instances of the application semantic concepts. A model scheduling micro-service (7) periodically loads the registered model deployments and determines if each model in question is due for training or scoring, based on the user-specified schedules. Model execution involves automatically installing the model implementation (green) from the PyPi repository (8) and running the relevant model implementation routines (3) with specific model deployment configuration (5) and semantic data (2) loaded from the system data store. At model training, historical data (9) are retrieved from the time-series micro-service suite and a new model version object (pink) is produced, including a model object with fitted parameters (for example the weights of a neural network) and some training metadata, such as train time. The model version is saved into the data store and mapped to the semantic context of the model. At model scoring, the model version object is loaded from the system data store and made available to the scoring routine, which produces a time-series forecast (blue).”; “running the relevant model implementation routines…with specific model deployment configuration…a new model version object…is produced” corresponds to “identifying a first entry of a previous deployment plan for a second inference model type that is not associated with the time of the day” in which “a new model version object” corresponds to “a second inference model type that is not associated with the time of the day” since the model version object is not part of the “specific model deployment configuration” which includes “time” as a parameter which corresponds to “the time of the day” as stated above; “specific model deployment configuration” corresponds to “a first entry of a previous deployment plan”; “the data store” corresponds to “the distribution of the inference models being based on the previous deployment plan”) ; and
“replacing the first entry with a second entry to obtain the deployment plan, the second entry indicating that an instance of the second inference model type should be present in the distribution of the inference models.” (Eck et al. Section 2 “The model deployments (yellow) are stored in a database and specify the execution of the model implementations against specific instances of the application semantic concepts. A model scheduling micro-service (7) periodically loads the registered model deployments and determines if each model in question is due for training or scoring, based on the user-specified schedules. Model execution involves automatically installing the model implementation (green) from the PyPi repository (8) and running the relevant model implementation routines (3) with specific model deployment configuration (5) and semantic data (2) loaded from the system data store. At model training, historical data (9) are retrieved from the time-series micro-service suite and a new model version object (pink) is produced, including a model object with fitted parameters (for example the weights of a neural network) and some training metadata, such as train time. The model version is saved into the data store and mapped to the semantic context of the model.”; “a new model version object…with fitted parameters…and some training metadata” corresponds to “a second entry”; since the semantic context includes model deployments, “mapped to the semantic context” corresponds to “the deployment plan”; “The model version is saved into the data store” corresponds to “an instance of the second inference model type should be present in the distribution of the inference models”)
Regarding claim 4, the rejections of claim 2 are incorporated. Eck et al. further teaches:
“obtaining the deployment plan comprises:”
“identifying, based on the inference model types, a first inference model type associated with the event” (Eck et al. Section 2 “The model deployments (yellow) are stored in a database and specify the execution of the model implementations against specific instances of the application semantic concepts. A model scheduling micro-service (7) periodically loads the registered model deployments and determines if each model in question is due for training or scoring, based on the user-specified schedules. Model execution involves automatically installing the model implementation (green) from the PyPi repository (8) and running the relevant model implementation routines (3) with specific model deployment configuration (5) and semantic data (2) loaded from the system data store.”; Eck et al. Listing 2 describes a “user_parameters” field with the parameter “train_time” which signals the times when a model was trained; “A model scheduling micro-service” that “determines if each model…is due for…scoring based on the user-specified schedules” and “user_parameters” in each model deployment corresponds to “identifying, based on the inference model types, a first model type”; “train_time” corresponds to “associated with the event”);
“identifying a first entry of a previous deployment plan for a second inference model type that is not associated with the event, the distribution of the inference models being based on the previous deployment plan” (Eck et al. Section 2 “The model deployments (yellow) are stored in a database and specify the execution of the model implementations against specific instances of the application semantic concepts. A model scheduling micro-service (7) periodically loads the registered model deployments and determines if each model in question is due for training or scoring, based on the user-specified schedules. Model execution involves automatically installing the model implementation (green) from the PyPi repository (8) and running the relevant model implementation routines (3) with specific model deployment configuration (5) and semantic data (2) loaded from the system data store. At model training, historical data (9) are retrieved from the time-series micro-service suite and a new model version object (pink) is produced, including a model object with fitted parameters (for example the weights of a neural network) and some training metadata, such as train time. The model version is saved into the data store and mapped to the semantic context of the model. At model scoring, the model version object is loaded from the system data store and made available to the scoring routine, which produces a time-series forecast (blue).”; “running the relevant model implementation routines…with specific model deployment configuration…a new model version object…is produced” corresponds to “identifying a first entry of a previous deployment plan for a second inference model type that is not associated with the event” in which “a new model version object” corresponds to “a second inference model type that is not associated with the event” since the model version object is not part of the “specific model deployment configuration” which includes “train_time” as a parameter which corresponds to “the event” as stated above; “specific model deployment configuration” corresponds to “a first entry of a previous deployment plan”; “the data store” corresponds to “the distribution of the inference models being based on the previous deployment plan”); and
“replacing the first entry with a second entry to obtain the deployment plan, the second entry indicating that an instance of the second inference model type should be present in the distribution of the inference models.” (Eck et al. Section 2 “The model deployments (yellow) are stored in a database and specify the execution of the model implementations against specific instances of the application semantic concepts. A model scheduling micro-service (7) periodically loads the registered model deployments and determines if each model in question is due for training or scoring, based on the user-specified schedules. Model execution involves automatically installing the model implementation (green) from the PyPi repository (8) and running the relevant model implementation routines (3) with specific model deployment configuration (5) and semantic data (2) loaded from the system data store. At model training, historical data (9) are retrieved from the time-series micro-service suite and a new model version object (pink) is produced, including a model object with fitted parameters (for example, the weights of a neural network) and some training metadata, such as train time. The model version is saved into the data store and mapped to the semantic context of the model.”; “a new model version object…with fitted parameters…and some training metadata” corresponds to “a second entry”; since the semantic context includes model deployments, “mapped to the semantic context” corresponds to “the deployment plan”; “The model version is saved into the data store” corresponds to “an instance of the second inference model type should be present in the distribution of the inference models”)
Regarding claim 5, the rejections of claim 4 are incorporated. Eck et al. further mentions “obtaining the deployment plan comprises:”
“adding a new entry to the previous deployment plan to obtain the deployment plan, the new entry indicating that another instance of the second inference model type should be present in the distribution of the inference models.” (Eck et al. Section 2 “The model deployments (yellow) are stored in a database and specify the execution of the model implementations against specific instances of the application semantic concepts. A model scheduling micro-service (7) periodically loads the registered model deployments and determines if each model in question is due for training or scoring, based on the user-specified schedules. Model execution involves automatically installing the model implementation (green) from the PyPi repository (8) and running the relevant model implementation routines (3) with specific model deployment configuration (5) and semantic data (2) loaded from the system data store. At model training, historical data (9) are retrieved from the time-series micro-service suite and a new model version object (pink) is produced, including a model object with fitted parameters (for example, the weights of a neural network) and some training metadata, such as train time. The model version is saved into the data store and mapped to the semantic context of the model.”; “training metadata” corresponds to “a new entry”; since the semantic context includes model deployments, “model version is…mapped to the semantic context” and corresponds to “the previous deployment plan to obtain the deployment plan”; “The model version is saved into the data store” which includes saving “training metadata” corresponds to “another instance of the second inference model type should be present in the distribution of the inference models”)
Regarding claim 6, the rejections of claim 2 are incorporated. Eck et al. further discloses
“obtaining the deployment plan comprises:”
“identifying, based on the inference model types, a first inference model type associated with the geographic location” (Eck et al. Section 2 “The model deployments (yellow) are stored in a database and specify the execution of the model implementations against specific instances of the application semantic concepts. A model scheduling micro-service (7) periodically loads the registered model deployments and determines if each model in question is due for training or scoring, based on the user-specified schedules. Model execution involves automatically installing the model implementation (green) from the PyPi repository (8) and running the relevant model implementation routines (3) with specific model deployment configuration (5) and semantic data (2) loaded from the system data store.”; Listing 2 defines a deployment plan containing a “context”; Section 3.1 “…. context provides semantic information associated with the time-series targeted by a specific model instance, in terms of concepts of signals (what physical quantity, what unit) and entity (what location, what name, type, GIS coordinates,…) …modelID and modelVersion point to the model and model version data”;“A model scheduling micro-service” that “determines if each model…is due for…scoring based on the user-specified schedules” and “user_parameters” in each model deployment corresponds to “identifying, based on the inference model types, a first model type”; “entity” corresponds to “associated with the geographic location”);
“identifying a first entry of a previous deployment plan for a second inference model type that is not associated with the geographic location, the distribution of the inference models being based on the previous deployment plan” (Eck et al. Section 2 “The model deployments (yellow) are stored in a database and specify the execution of the model implementations against specific instances of the application semantic concepts. A model scheduling micro-service (7) periodically loads the registered model deployments and determines if each model in question is due for training or scoring, based on the user-specified schedules. Model execution involves automatically installing the model implementation (green) from the PyPi repository (8) and running the relevant model implementation routines (3) with specific model deployment configuration (5) and semantic data (2) loaded from the system data store. At model training, historical data (9) are retrieved from the time-series micro-service suite and a new model version object (pink) is produced, including a model object with fitted parameters (for example the weights of a neural network) and some training metadata, such as train time. The model version is saved into the data store and mapped to the semantic context of the model. At model scoring, the model version object is loaded from the system data store and made available to the scoring routine, which produces a time-series forecast (blue).”; “running the relevant model implementation routines…with specific model deployment configuration…a new model version object…is produced” corresponds to “identifying a first entry of a previous deployment plan for a second inference model type that is not associated with the event” in which “a new model version object” corresponds to “a second inference model type that is not associated with the event” since the model version object is not part of the “specific model deployment configuration” which includes “entity” as a field which corresponds to “the geographic location” as stated above; “specific model deployment configuration” corresponds to “a first entry of a previous deployment plan”; “the data store” corresponds to “the distribution of the inference models being based on the previous deployment plan”); and
“replacing the first entry with a second entry for the first inference model type to obtain the deployment plan.” (Eck et al. Section 2 “The model deployments (yellow) are stored in a database and specify the execution of the model implementations against specific instances of the application semantic concepts. A model scheduling micro-service (7) periodically loads the registered model deployments and determines if each model in question is due for training or scoring, based on the user-specified schedules. Model execution involves automatically installing the model implementation (green) from the PyPi repository (8) and running the relevant model implementation routines (3) with specific model deployment configuration (5) and semantic data (2) loaded from the system data store. At model training, historical data (9) are retrieved from the time-series micro-service suite and a new model version object (pink) is produced, including a model object with fitted parameters (for example, the weights of a neural network) and some training metadata, such as train time. The model version is saved into the data store and mapped to the semantic context of the model.”; “a new model version object…with fitted parameters…and some training metadata” corresponds to “a second entry”; since the semantic context includes model deployments, “mapped to the semantic context” corresponds to “the deployment plan”; since a new model version for the original model is saved into the data store this correspond to “replacing the first entry with a second entry for the first inference model type”)
Regarding claim 7, the rejections of claim 2 are incorporated. Eck et al. further recites:
“the distribution of the inference models comprises an inference model over fitted for a second time of the day, the updated distribution of the inference model comprises a second inference model overfitted for the time of the days, and the inference model is not a member of the updated distribution of the inference models” (Eck et al. Section 2 “A model implementation (green) can then be used for execution in many model deployments. Based on semantics (2) and desired model training/scoring schedules, model deployments (see Section 3.2) are written (5) and registered with the system (6). The model deployments (yellow) are stored in a database and specify the execution of the model implementations against specific instances of the application semantic concepts. Model execution involves automatically installing the model implementation (green) from the PyPi repository (8) and running the relevant model implementation routines (3) with specific model deployment configuration (5) and semantic data (2) loaded from the system data store. At model training, historical data (9) are retrieved from the time-series micro-service suite and a new model version object (pink) is produced, including a model object with fitted parameters (for example the weights of a neural network) and some training metadata, such as train time. The
model version is saved into the data store and mapped to the semantic context of the model.”; Listing 2 defines a “train_time” parameter which includes at least two dates; “PyPi repository” corresponds to “the distribution of the inference models”; “running the relevant model implementation…a new model version object…is produced, including a model object with fitted parameters” corresponds to “an inference model overfitted”; since the “model deployment” contains the “train_time” parameter which includes at least two dates, “train_time” corresponds to “a second time of the day”; “The
model version is saved into the data store and mapped to the semantic context of the model” corresponds to “the updated distribution of the inference model comprises a second inference model overfitted for the time of the days, and the inference model is not a member of the updated distribution of the inference models”)
Regarding claim 8, Eck et al. teaches “A non-transitory machine-readable medium having instructions stored therein, which when executed by a processor, cause the processor to perform operations for managing a distribution of inference models hosted by data processing systems” (Eck et al. Section 4.3 “As detailed in Section 2, model execution leverages the serverless cloud computing framework. The resources allocated to the individual serverless jobs was 2 CPUs and 2GB of RAM memory. The resource specification of the knowledge-based time-series services were 1 CPU Core, 4GB RAM for the relational database and 1 CPU, 0.5GB RAM for the graph database.”; “RAM” corresponds to “A non-transitory machine-readable medium”; “CPU” corresponds to “processor”) “the operations comprising:”
“obtaining condition data for the data processing systems, the condition data indicating how the data processing systems are likely to operate during a future period of time” (Eck et al., Section 2 “Invocations of certain micro-services are triggered based on real-world events, such as the availability of time-series data (1). The IoT data is ingested often at irregular frequencies, with many devices submitting data in parallel. Ingested time-series are stored in a knowledge-based data store. As new time-series become available, data scientists can provide additional semantics (2) to contextualise the time-series [Chen et al., 2018]. IoT data ingestion and semantic definition occur out-of-band, and do not directly affect the main modelling work flow”; “provide…semantics” corresponds to “obtaining condition data”; “IoT data” corresponds to “data processing systems”; Eck et al. Listing 2 defines a “scoring_deployment” field with the parameters “time” and “repeatEvery” which signals the time and repeated time intervals to perform model scoring; hence it corresponds to “during a future period of time”);
“obtaining a deployment plan for the inference models based on the condition data, the distribution, and inference model types available for deployment, the deployment plan specifying: an operating mode” (Eck et al., Section 2 “Based on semantics (2) and desired model training/scoring schedules, model deployments…are written (5) and registered with the system (6)”; Listing 2 defines a deployment plan containing a “context”; Section 3.1 “…. context provides semantic information associated with the time-series targeted by a specific model instance, in terms of concepts of signals (what physical quantity, what unit) and entity (what location, what name, type, GIS coordinates,…) …modelID and modelVersion point to the model and model version data”; Section 2 “The model deployments…are stored in a database and specify the execution of the model implementations against specific instances of the application semantic concepts. A model scheduling micro-service (7) periodically….determines if each model in question is due for training or scoring…”’;“model deployments…are written” correspond to “obtaining a deployment plan for the inference models”; “desired model training/scoring schedules” correspond to “the distribution”; “modelID and modelVersion” correspond to “inference model types available for deployment”; “specify the execution of the model implementations” correspond to “the deployment plan specify an operating mode”; “determines if each model in question is due for training or scoring”; “semantics” correspond to “condition data”);
“the deployment plan specifying: … , how each of the inference models is divided into portions, and where the portions are distributed in the data processing systems” (Eck et al., Figure 3 “Semantic representation of the IoT data”
PNG
media_image1.png
610
711
media_image1.png
Greyscale
; Figure 5 “Deployed models and trained versions for a context”
PNG
media_image2.png
479
593
media_image2.png
Greyscale
; Section 4.2 “Specific model instances were deployed by preparing model deployment configurations, as outlined in Section 3.2. The semantic context was set to the target entity (e.g. SUBSTATION S1) and signal (e.g. ENERGY LOAD), the training and scoring schedules were defined (hourly and weekly, respectively), custom user parameters were used to specify a training window, the time and the desired forecasting window … The models deployed for a given context can be explored in the semantic graph, as shown in Fig. 5”; the target entity corresponds to “where the portions are distributed in the data processing systems”; Section 4.3 “The deployment of hundreds of models was semi-automated by programmatically setting the deployment configurations for the AI model implementations defined in Section 4.2 to all required semantic targets. For example, the 174 models of site 3 are based on only 6 model implementations”; six inference models are split into 174 portions across a network, thus the context comprises “how each of the inference models is divided into portions” and “where the portions are distributed in the data processing systems”.
“prior to the future period of time, updating the inference models based on the deployment plan to obtain an updated distribution of the inference models” (Eck et al., Section 2 “A model scheduling micro-service (7) periodically loads the registered model deployments and determines if each model in question is due for training or scoring, based on the user-specified schedules. Model execution involves automatically installing the model implementation (green) from the PyPi repository (8) and running the relevant model implementation routines (3) with specific model deployment configuration (5) and semantic data (2) loaded from the system data store. At model training, historical data (9) are retrieved from the time-series micro-service suite and a new model version object (pink) is produced, including a model object with fitted parameters (for example the weights of a neural network) and some training metadata, such as train time. The model version is saved into the data store and mapped to the semantic context of the model. At model scoring, the model version object is loaded from the system data store and made available to the scoring routine, which produces a time-series forecast (blue). The forecast micro-service (10) handles the persistence of prediction time-series in a database.”; “each model” and “a new model version object…with fitted parameters” correspond to “updating the inference models”; “running the relevant model implementation routines (3) with specific model deployment configuration (5)” corresponds to “based on the deployment plan”; “a new model version object…is produced” corresponds to “an updated distribution of the inference models”; “At model scoring” corresponds to a “future period of time”; since “running the relevant model implementation routines” occurs prior to model scoring, this corresponds to “prior to the future period of time” ); and
“obtaining, during the period of time, an inference using the updated distribution of the inference models and the operating mode, the operating mode indicating at least one automatically implemented change to the updated distribution of the inference models responsive to an occurrence of a condition” (Eck et al., Section 2 “At model training, historical data (9) are retrieved from the time-series micro-service suite and a new model version object (pink) is produced, including a model object with fitted parameters (for example the weights of a neural network) and some training metadata, such as train time. The model version is saved into the data store and mapped to the semantic context of the model. At model scoring, the model version object is loaded from the system data store and made available to the scoring routine, which produces a time-series forecast (blue). The forecast micro-service (10) handles the persistence of prediction time-series in a database.”; “At model scoring” corresponds to “during the period of time”; “produces a time-series forecast” and “the system data store” which store multiple models for model scoring corresponds to “an inference using the updated distribution of the inference models”; “a new model version object” that was used at model scoring corresponds to “the operating mode”; “a model object with fitted parameters…and some training metadata, such as train time” corresponds to “the operating mode indicating the at least one automatically implemented change to the updated distribution of the inference models”; “historical data (9) are retrieved from the time-series micro-service suite and a new model version object…is produced” corresponds to “responsive to an occurrence of a condition”)
Regarding claim 9, the rejections of claim 8 are incorporated and is rejected for the same rationale as claim 2.
Regarding claim 10, the rejections of claim 9 are incorporated. Additionally, claim 10 is rejected for the same reasons as claim 3.
Regarding claim 11, the rejections of claim 9 are incorporated and is rejected for the same reasons as claim 4.
Regarding claim 12, the rejections of claim 11 are incorporated. Additionally, claim 12 is rejected for the same rationale as claim 5.
Regarding claim 13, the rejections of claim 9 are incorporated and is rejected for the same reasons as claim 6.
Regarding claim 14, the rejections of claim 9 are incorporated and is rejected for the same rationale as claim 7.
Regarding claim 15, Eck et al. recites “A data processing system” (Eck et al. Section 1 “Leveraging the key technologies of serverless computing and knowledge-based data representation, a system for the management and deployment of large quantities of AI timeseries forecasting models in IoT applications was designed.”) “, comprising:”
“a processor” (Eck et al. Section 4.3 “As detailed in Section 2, model execution leverages the serverless cloud computing framework. The resources allocated to the individual serverless jobs was 2 CPUs and 2GB of RAM memory. The resource specification of the knowledge-based time-series services were 1 CPU Core, 4GB RAM for the relational database and 1 CPU, 0.5GB RAM for the graph database.”; “CPU corresponds to “a processor”) ; and
“a memory coupled to the processor to store instructions, which when executed by the processor, cause the processor to perform operations for managing a distribution of inference models hosted by data processing systems,” (Eck et al. Section 4.3 “As detailed in Section 2, model execution leverages the serverless cloud computing framework. The resources allocated to the individual serverless jobs was 2 CPUs and 2GB of RAM memory. The resource specification of the knowledge-based time-series services were 1 CPU Core, 4GB RAM for the relational database and 1 CPU, 0.5GB RAM for the graph database.”; “RAM” corresponds to “a memory”) “the operations comprising”:
“obtaining condition data for the data processing systems, the condition data indicating how the data processing systems are likely to operate during a future period of time” (Eck et al., Section 2 “Invocations of certain micro-services are triggered based on real-world events, such as the availability of time-series data (1). The IoT data is ingested often at irregular frequencies, with many devices submitting data in parallel. Ingested time-series are stored in a knowledge-based data store. As new time-series become available, data scientists can provide additional semantics (2) to contextualise the time-series [Chen et al., 2018]. IoT data ingestion and semantic definition occur out-of-band, and do not directly affect the main modelling work flow”; “provide…semantics” corresponds to “obtaining condition data”; “IoT data” corresponds to “data processing systems”; Eck et al. Listing 2 defines a “scoring_deployment” field with the parameters “time” and “repeatEvery” which signals the time and repeated time intervals to perform model scoring; hence it corresponds to “during a future period of time”);
“obtaining a deployment plan for the inference models based on the condition data, the distribution, and inference model types available for deployment, the deployment plan specifying: an operating mode” (Eck et al., Section 2 “Based on semantics (2) and desired model training/scoring schedules, model deployments…are written (5) and registered with the system (6)”; Listing 2 defines a deployment plan containing a “context”; Section 3.1 “…. context provides semantic information associated with the time-series targeted by a specific model instance, in terms of concepts of signals (what physical quantity, what unit) and entity (what location, what name, type, GIS coordinates,…) …modelID and modelVersion point to the model and model version data”; Section 2 “The model deployments…are stored in a database and specify the execution of the model implementations against specific instances of the application semantic concepts. A model scheduling micro-service (7) periodically….determines if each model in question is due for training or scoring…”’;“model deployments…are written” correspond to “obtaining a deployment plan for the inference models”; “desired model training/scoring schedules” correspond to “the distribution”; “modelID and modelVersion” correspond to “inference model types available for deployment”; “specify the execution of the model implementations” correspond to “the deployment plan specify an operating mode”; “semantics” correspond to “condition data”);
“the deployment plan specifying: … , how each of the inference models is divided into portions, and where the portions are distributed in the data processing systems” (Eck et al., Figure 3 “Semantic representation of the IoT data”
PNG
media_image1.png
610
711
media_image1.png
Greyscale
; Figure 5 “Deployed models and trained versions for a context”
PNG
media_image2.png
479
593
media_image2.png
Greyscale
; Section 4.2 “Specific model instances were deployed by preparing model deployment configurations, as outlined in Section 3.2. The semantic context was set to the target entity (e.g. SUBSTATION S1) and signal (e.g. ENERGY LOAD), the training and scoring schedules were defined (hourly and weekly, respectively), custom user parameters were used to specify a training window, the time and the desired forecasting window … The models deployed for a given context can be explored in the semantic graph, as shown in Fig. 5”; the target entity corresponds to “where the portions are distributed in the data processing systems”; Section 4.3 “The deployment of hundreds of models was semi-automated by programmatically setting the deployment configurations for the AI model implementations defined in Section 4.2 to all required semantic targets. For example, the 174 models of site 3 are based on only 6 model implementations”; six inference models are split into 174 portions across a network, thus the context comprises “how each of the inference models is divided into portions” and “where the portions are distributed in the data processing systems”.
“prior to the future period of time, updating the inference models based on the deployment plan to obtain an updated distribution of the inference models” (Eck et al., Section 2 “A model scheduling micro-service (7) periodically loads the registered model deployments and determines if each model in question is due for training or scoring, based on the user-specified schedules. Model execution involves automatically installing the model implementation (green) from the PyPi repository (8) and running the relevant model implementation routines (3) with specific model deployment configuration (5) and semantic data (2) loaded from the system data store. At model training, historical data (9) are retrieved from the time-series micro-service suite and a new model version object (pink) is produced, including a model object with fitted parameters (for example the weights of a neural network) and some training metadata, such as train time. The model version is saved into the data store and mapped to the semantic context of the model. At model scoring, the model version object is loaded from the system data store and made available to the scoring routine, which produces a time-series forecast (blue). The forecast micro-service (10) handles the persistence of prediction time-series in a database.”; “each model” and “a new model version object…with fitted parameters” correspond to “updating the inference models”; “running the relevant model implementation routines (3) with specific model deployment configuration (5)” corresponds to “based on the deployment plan”; “a new model version object…is produced” corresponds to “an updated distribution of the inference models”; “At model scoring” corresponds to a “future period of time”; since “running the relevant model implementation routines” occurs prior to model scoring, this corresponds to “prior to the future period of time” ); and
“obtaining, during the period of time, an inference using the updated distribution of the inference models and the operating mode, the operating mode indicating at least one automatically implemented change to the updated distribution of the inference models responsive to an occurrence of a condition” (Eck et al., Section 2 “At model training, historical data (9) are retrieved from the time-series micro-service suite and a new model version object (pink) is produced, including a model object with fitted parameters (for example the weights of a neural network) and some training metadata, such as train time. The model version is saved into the data store and mapped to the semantic context of the model. At model scoring, the model version object is loaded from the system data store and made available to the scoring routine, which produces a time-series forecast (blue). The forecast micro-service (10) handles the persistence of prediction time-series in a database.”; “At model scoring” corresponds to “during the period of time”; “produces a time-series forecast” and “the system data store” which store multiple models for model scoring corresponds to “an inference using the updated distribution of the inference models”; “a new model version object” that was used at model scoring corresponds to “the operating mode”; “a model object with fitted parameters…and some training metadata, such as train time” corresponds to “the operating mode indicating the at least one automatically implemented change to the updated distribution of the inference models”; “historical data (9) are retrieved from the time-series micro-service suite and a new model version object…is produced” corresponds to “responsive to an occurrence of a condition”)
Regarding claim 16 , the rejections of claim 15 are incorporated and is rejected for the same rationale as claims 2 and 9.
Regarding claim 17, the rejections of claim 16 are incorporated. Additionally, claim 17 is rejected for the same reasons as claim 3 and 10.
Regarding claim 18, the rejections of claim 16 are incorporated and is rejected for the same reasons as claims 4 and 11.
Regarding claim 19, the rejections of claim 18 are incorporated. Additionally, claim 19 is rejected for the same rationale as claim 5 and 12.
Regarding claim 20, the rejections of claim 16 are incorporated and is rejected for the same reasons as claims 6 and 13.
Response to Arguments
The following responses address arguments and remarks made in the instant remarks dated 12/30/2025.
Objections
Previous objections are withdrawn in light of the instant amendments.
101 Rejections
On pages 9-10 of the instant remarks, the Applicant argues that the claimed invention reflects an improvement to technology:
“As described above, the claimed "deployment plan," which is obtained based on "condition
data," specifies "how each of the inference models is divided into portions, and where the portions
are distributed in the data processing systems." Based on the "deployment plan," the distribution
of the inference models is updated, taking into account the "condition data" that indicates "how
the data processing systems are likely to operate during a future period of time."
With these limitations, the claimed invention makes it possible to timely ("prior to the
future period of time") update the model distribution based on an expected performance of the data
processing systems. Accordingly, the data processing systems are "better able to dynamically
respond to changes conditions impacting data processing systems hosting inference models," and
hence are "more likely to provide inferences of desired accuracy and at desirable levels of
reliability while limiting the computing resources expended for inference generation." Original
specification, ¶¶[15]-[16]. In short, the claim limitations reflect an improvement to the technology
of distributed computing, and thus integrate the allegedly abstract idea into a practical application”
Regarding the Applicant’s arguments above, the Examiner respectfully disagrees. The purported improvement of a claimed invention must be sufficiently detailed, as noted in MPEP 2106.05(a): the disclosure must provide sufficient details such that one of ordinary skill in the art would recognize the claimed invention as providing an improvement. The specification need not explicitly set forth the improvement, but it must describe the invention such that the improvement would be apparent to one of ordinary skill in the art. Conversely, if the specification explicitly sets forth an improvement but in a conclusory manner (i.e., a bare assertion of an improvement without the detail necessary to be apparent to a person of ordinary skill in the art), the examiner should not determine the claim improves technology.
The improvement argued by the Applicant is represented only at a very high level of generality. It’s not clear what the “condition data” comprises, how deployment plans are obtained based on the condition data, or how specifically this deployment plan will positively impact the accuracy and reliability of the system. One of ordinary skill in the art would not recognize definitively how to implement the claimed invention in such a way that the argued improvements are realized.
Thus, no rejections are withdrawn on this basis. See the 101 rejections section for more detail.
102 Rejections
On page 11 of the instant remarks, the Applicant argues that the relied upon references do not fully disclose limitations of the amended claims:
“Without acquiescing to the analysis in the Office action, Applicant notes that Eck fails to disclose
the deployment plan specifying "how each of the inference models is divided into portions, and
where the portions are distributed in the data processing systems." In fact, Eck is silent with
respect to dividing inference models into portions and distribute the portions in the data processing
systems. Therefore, Eck fails to anticipate amended independent claim 1 and likewise fails to
anticipate amended independent claims 8 and 15 and all dependent claims”
Regarding the Applicant’s arguments above, the Examiner respectfully disagrees. Eck discloses partitioning instantiations (portions) of a small pool of models across different parts of a distributed network, each model’s network location being defined by its context (part of the deployment plan) (Eck, Section 4.2 & Section 4.3). See the 102 rejections section for more detail. No rejections are withdrawn on this basis.
Conclusion
The prior art made of record and not relied upon is considered pertinent to applicant's disclosure:
Chen et al (Castor: Contextual IoT Time Series Data and Model Management at Scale, published 2019, arXiv:1811.08566v3) also discusses on training and saving models in a database and scheduling model execution.
Vu et al. (Predicting condition of a host for cybersecurity applications, published 5/18/2021, US20200145448A1) teaches obtaining condition data for data processing systems.
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 Aaron P Gormley whose telephone number is (571)272-1372. The examiner can normally be reached Monday - Friday 12:00 PM - 8:00 PM EST.
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, Michelle T Bechtold can be reached at (571) 431-0762. 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.
/AG/Examiner, Art Unit 2148 /MICHELLE T BECHTOLD/Supervisory Patent Examiner, Art Unit 2148
Prosecution Insights