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 . Claims 1-20 are presented in this case.
Priority
Applicant's claim for the benefit of a prior-filed Provisional application 63/313,657 and 63/313,658 filed on 12/24/2022 is acknowledged.
Claim Objections
Claims 3, 8-9 and 13 are objected to because of the following informalities:
Claim 3, 13 line 2 recites the phrase “after deploying” which should be “after the deploying”
Claim 8-9, 17, 20 line 2 recites the phrase “user interest or user preferences” which should be “user interest and user preferences”
For the informalities above and wherever else they may occur appropriate correction is required.
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 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 of this title, 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, 5-6, 11 and 15 are rejected under 35 U.S.C. 103 as being unpatentable over Johnsson et al. (US 20240135247 A1 hereinafter Johnsson) in view of Lupesko et al. (US 11423283 B1 hereinafter Lupesko)
As to independent claim 1, Johnsson teaches a computer-implemented method of deploying a machine learning model, comprising: [deploying models ¶32-34]
receiving a user request for deploying a machine learning model, for an application, to an edge device; [receives a request for a model (to be deployed) ¶9 "receiving, by an apparatus, a request for a machine learning model solving a task T" ]
determining a device constraint type associated with the edge device, wherein the device constraint type is one of a number of device constraint types associated with a plurality of edge devices capable of running the application; [determines constraints for the execution environment (edge computer¶24) ¶9 "resource constraints of the execution environment."; four types (number) ¶53 "hardware constraints, software constraints, sampling requirements and resource usage of the execution environment."]
identifying a machine learning model corresponding to the device constraint type of the edge device, [selects suitable model based on constraints ¶9 "determining, from the first set of machine learning models, at least one suitable machine learning model to be deployed, wherein the determining is based on the calculated complexity and the resource constraints of the execution environment"]
deploying the machine learning model to the edge device. [deploys accordingly ¶61 "highest ranked ML model is deployed in the execution environment 102"]
Johnsson does not specifically teach wherein the machine learning model is one of a number of tiers of machine learning models developed for the application according to the device constraint types.
However, Lupesko teaches wherein the machine learning model is one of a number of tiers of machine learning models developed for the application according to the device constraint types; [model tiers (variants) that match performance characteristics Col. 5 ln. 50-66 "If the model does not meet desired performance characteristics, in some embodiments a model variant is dynamically generated at 809. For example, some embodiments, a profile of a model variant that will meet the characteristics is selected (one that meets the desired performance characteristics) and used to generate a model variant from the evaluated (or original) model."]
Accordingly, it would have been obvious to a person of ordinary skill in the art before the effective filling date of the claimed invention to modify the model development disclosed by Johnsson by incorporating the wherein the machine learning model is one of a number of tiers of machine learning models developed for the application according to the device constraint types disclosed by Lupesko because both techniques address the same field of machine learning and by incorporating Lupesko into Johnsson provide more optimized models that are more adaptable to conditions in a more cost-effective system [Lupesko Col. 1-2 ln. 53-4]
As to dependent claim 5, the rejection of claim 1 is incorporated, Johnsson and Lupesko further teach wherein the device constraint types and the tiers of machine learning models have a one-to-one correspondence. [Lumpesko variants set with change to models and characteristics Col. 5 ln. 19-25, Col. 5-6 ln . 65-9 " a profile of a model variant that will meet the characteristics is selected (one that meets the desired performance characteristics) and used to generate a model variant "]
As to dependent claim 6, the rejection of claim 1 is incorporated, Johnsson and Lupesko further teach wherein the edge device is an enterprise server. [Johnsson server apparatus as edge node ¶28]
As to independent claim 11, Johnsson teaches a system for deploying a machine learning model, comprising: [apparatus with a system for deploying models ¶32-34]
a processor; and [processing unit ¶33]
a memory, coupled to the processor, configured to store executable instructions that, when executed by the processor, cause the processor to perform operations including: [memory with instructions ¶33]
receiving a user request for deploying a machine learning model, for an application, to an edge device; [receives a request for a model (to be deployed) ¶9 "receiving, by an apparatus, a request for a machine learning model solving a task T"]
determining a device constraint type associated with the edge device, wherein the device constraint type is one of a number of device constraint types associated with a plurality of edge devices capable of running the application; [determines constraints for the execution environment (edge computer¶24) ¶9 "resource constraints of the execution environment."; four types (number) ¶53 "hardware constraints, software constraints, sampling requirements and resource usage of the execution environment."]
identifying a machine learning model corresponding to the device constraint type of the edge device, [selects suitable model based on constraints ¶9 "determining, from the first set of machine learning models, at least one suitable machine learning model to be deployed, wherein the determining is based on the calculated complexity and the resource constraints of the execution environment"]
deploying the machine learning model to the edge device. [deploys accordingly ¶61 "
highest ranked ML model is deployed in the execution environment 102"]
Johnsson does not specifically teach wherein the machine learning model is one of a number of tiers of machine learning models developed for the application according to the device constraint types.
However, Lupesko teaches wherein the machine learning model is one of a number of tiers of machine learning models developed for the application according to the device constraint types; [model tiers (variants) that match performance characteristics Col. 5 ln. 50-66 "If the model does not meet desired performance characteristics, in some embodiments a model variant is dynamically generated at 809. For example, some embodiments, a profile of a model variant that will meet the characteristics is selected (one that meets the desired performance characteristics) and used to generate a model variant from the evaluated (or original) model."]
Accordingly, it would have been obvious to a person of ordinary skill in the art before the effective filling date of the claimed invention to modify the model development disclosed by Johnsson by incorporating the wherein the machine learning model is one of a number of tiers of machine learning models developed for the application according to the device constraint types disclosed by Lupesko because both techniques address the same field of machine learning and by incorporating Lupesko into Johnsson for more optimized models that are more adaptable to conditions in a more cost-effective system [Lupesko Col. 1-2 ln. 53-4]
As to dependent claim 15, the rejection of claim 11 is incorporated, Johnsson and Lupesko further teach wherein the device constraint types and the tiers of machine learning models have a one-to-one correspondence. [Lumpesko variants set with change to models and characteristics Col. 5 ln. 19-25, Col. 5-6 ln . 65-9 " a profile of a model variant that will meet the characteristics is selected (one that meets the desired performance characteristics) and used to generate a model variant "]
Claims 2 and 12 are rejected under 35 U.S.C. 103 as being unpatentable over Johnsson in view of Lupesko, as applied in claim 1 and 11 above, and further in view of Kuo et al. (US 20190156246 A1 hereinafter Kuo)
As to dependent claim 2, Johnsson and Lupesko teach the method of claim 1 above that is incorporated,
Johnsson and Lupesko do not specifically teach wherein the machine learning models are developed and trained on a cloud device.
However, Kuo teaches wherein the machine learning models are developed and trained on a cloud device. [trained/developed in the cloud ¶15 "Machine learning models may be trained in the cloud (e.g., by a provider network) "]
Accordingly, it would have been obvious to a person of ordinary skill in the art before the effective filling date of the claimed invention to modify the learning process disclosed by Johnsson and Lupesko by incorporating the wherein the machine learning models are developed and trained on a cloud device disclosed by Kuo because all techniques address the same field of machine learning and by incorporating Kuo into Johnsson and Lupesko alleviate concerns with excessive use of computing resources and time [Kuo ¶16-17]
As to dependent claim 12, Johnsson and Lupesko teach the method of claim 11 above that is incorporated,
Johnsson and Lupesko do not specifically teach wherein the machine learning models are developed and trained on a cloud device.
However, Kuo teaches wherein the machine learning models are developed and trained on a cloud device. [trained/developed in the cloud ¶15 "Machine learning models may be trained in the cloud (e.g., by a provider network) "]
Accordingly, it would have been obvious to a person of ordinary skill in the art before the effective filling date of the claimed invention to modify the learning process disclosed by Johnsson and Lupesko by incorporating the wherein the machine learning models are developed and trained on a cloud device disclosed by Kuo because all techniques address the same field of machine learning and by incorporating Kuo into Johnsson and Lupesko alleviate concerns with excessive use of computing resources and time [Kuo ¶16-17]
Claims 3-4 and 13-14 are rejected under 35 U.S.C. 103 as being unpatentable over Johnsson in view of Lupesko, as applied in claim 1 and 11 above, and further in view of Choudhary et al. (US 20190385043 A1 hereinafter Choudhary)
As to dependent claim 3, Johnsson and Lupesko teach the method of claim 1 above that is incorporated,
Johnsson and Lupesko do not specifically teach wherein the machine learning model is trained on the edge device after deploying to the edge device.
However, Choudhary teaches wherein the machine learning model is trained on the edge device after deploying to the edge device. [trains local models on client devices (edge) ¶21-22 "trains a machine learning model across client devices that implement local versions of the model"]
Accordingly, it would have been obvious to a person of ordinary skill in the art before the effective filling date of the claimed invention to modify the learning process disclosed by Johnsson and Lupesko by incorporating the wherein the machine learning model is trained on the edge device after deploying to the edge device by Choudhary because all techniques address the same field of machine learning and by incorporating Choudhary into Johnsson and Lupesko provide more efficient consumption of computing resources [Choudhary ¶4]
As to dependent claim 4, the rejection of claim 3 is incorporated, Johnsson, Lupesko and Choudhary further teach wherein, prior to determining the device constraint type associated with the edge device, the method further comprises: receiving, from the edge device, device information for the edge device; and [Lumpesko receives edge characteristics (device information) Col. 5 ln. 19-25 "receives edge device characteristics 721 including, for example, FLOPS, GPU RAM, CPU RAM, CPU speed, power, network capabilities (wired, wireless, and their types), memory, etc"]
determining the device constraint type associated with the edge device based on the received device information for the edge device. [Lumpesko determines a change to a model (constraint) according to information Col. 5 ln. 19-25]
As to dependent claim 13, Johnsson and Lupesko teach the method of claim 1 above that is incorporated,
Johnsson and Lupesko do not specifically teach wherein the machine learning model is trained on the edge device after deploying to the edge device.
However, Choudhary teaches wherein the machine learning model is trained on the edge device after deploying to the edge device. [trains local models on client devices (edge) ¶21-22 "trains a machine learning model across client devices that implement local versions of the model"]
Accordingly, it would have been obvious to a person of ordinary skill in the art before the effective filling date of the claimed invention to modify the learning process disclosed by Johnsson and Lupesko by incorporating the wherein the machine learning model is trained on the edge device after deploying to the edge device by Choudhary because all techniques address the same field of machine learning and by incorporating Choudhary into Johnsson and Lupesko provide more efficient consumption of computing resources [Choudhary ¶4]
As to dependent claim 14, the rejection of claim 13 is incorporated, Johnsson, Lupesko and Choudhary further teach wherein, prior to determining the device constraint type associated with the edge device, the method further comprises: receiving, from the edge device, device information for the edge device; and [Lumpesko receives edge characteristics (device information) Col. 5 ln. 19-25 "receives edge device characteristics 721 including, for example, FLOPS, GPU RAM, CPU RAM, CPU speed, power, network capabilities (wired, wireless, and their types), memory, etc"]
determining the device constraint type associated with the edge device based on the received device information for the edge device. [Lumpesko determines a change to a model (constraint) according to information Col. 5 ln. 19-25]
Claims 7 and 16 are rejected under 35 U.S.C. 103 as being unpatentable over Johnsson in view of Lupesko, as applied in claim 1 and 11 above, and further in view of O'Neill (US 6832373 B2)
As to dependent claim 7, Johnsson and Lupesko teach the method of claim 1 above that is incorporated,
Johnsson and Lupesko do not specifically teach wherein a quantity of the device constraint types is determined based on device information of the plurality of edge devices capable of running the application.
However, O'Neill teaches wherein a quantity of the device constraint types is determined based on device information of the plurality of edge devices capable of running the application. [package or size of update (quantity) determined from client devices (edge) Col. 7 ln. 36-50 "receives the identity information 113 from the client device 104 and subsequently generates the desired update package 110"]
Accordingly, it would have been obvious to a person of ordinary skill in the art before the effective filling date of the claimed invention to modify the learning process disclosed by Johnsson and Lupesko by incorporating the wherein a quantity of the device constraint types is determined based on device information of the plurality of edge devices capable of running the application by O'Neill because all techniques address the same field of machine learning and by incorporating O'Neill into Johnsson and Lupesko provides a more convenient and reliable update process [O'Neill Col. 3 ln. 42-60]
As to dependent claim 16, Johnsson and Lupesko teach the method of claim 15 above that is incorporated,
Johnsson and Lupesko do not specifically teach wherein a quantity of the device constraint types is determined based on device information of the plurality of edge devices capable of running the application.
However, O'Neill teaches wherein a quantity of the device constraint types is determined based on device information of the plurality of edge devices capable of running the application. [package or size of update (quantity) determined from client devices (edge) Col. 7 ln. 36-50 "receives the identity information 113 from the client device 104 and subsequently generates the desired update package 110"]
Accordingly, it would have been obvious to a person of ordinary skill in the art before the effective filling date of the claimed invention to modify the learning process disclosed by Johnsson and Lupesko by incorporating the wherein a quantity of the device constraint types is determined based on device information of the plurality of edge devices capable of running the application by O'Neill because all techniques address the same field of machine learning and by incorporating O'Neill into Johnsson and Lupesko provides a more convenient and reliable update process [O'Neill Col. 3 ln. 42-60]
Claims 8-9 and 17 are rejected under 35 U.S.C. 103 as being unpatentable over Johnsson in view of Lupesko, as applied in claim 1 and 11 above, and further in view of Feuz et al. (US 20190163667 A1 hereinafter Feuz)
As to dependent claim 8, Johnsson and Lupesko teach the method of claim 1 above that is incorporated,
Johnsson and Lupesko do not specifically teach wherein the machine learning models are trained based on user data reflecting one or more of user interests or user preferences of a user.
However, Feuz teaches wherein the machine learning models are trained based on user data reflecting one or more of user interests or user preferences of a user. [personalized training using user data ¶29 "models that capture the user preferences", ¶55 "update the machine-learned model over time as additional data (e.g., user-specific data) is received"]
Accordingly, it would have been obvious to a person of ordinary skill in the art before the effective filling date of the claimed invention to modify the learning process disclosed by Johnsson and Lupesko by incorporating the wherein the machine learning models are trained based on user data reflecting one or more of user interests or user preferences of a user by Feuz because all techniques address the same field of machine learning and by incorporating Feuz into Johnsson and Lupesko enhances the trust in models while maintaining high accuracy [Feuz ¶26-27]
As to dependent claim 9, the rejection of claim 8 is incorporated, Johnsson, Lupesko and Feuz further teach wherein an output of the trained machine learning models is tuned towards one or more of the user interests or user preferences of the user. [Feuz output based on input ¶13, user preferences ¶29 "models that capture the user preferences"]
As to dependent claim 17, Johnsson and Lupesko teach the method of claim 11 above that is incorporated,
Johnsson and Lupesko do not specifically teach wherein the machine learning models are trained based on user data reflecting one or more of user interests or user preferences of a user.
However, Feuz teaches wherein the machine learning models are trained based on user data reflecting one or more of user interests or user preferences of a user. [personalized training using user data ¶29 "models that capture the user preferences", ¶55 "update the machine-learned model over time as additional data (e.g., user-specific data) is received"]
Accordingly, it would have been obvious to a person of ordinary skill in the art before the effective filling date of the claimed invention to modify the learning process disclosed by Johnsson and Lupesko by incorporating the wherein the machine learning models are trained based on user data reflecting one or more of user interests or user preferences of a user by Feuz because all techniques address the same field of machine learning and by incorporating Feuz into Johnsson and Lupesko enhances the trust in models while maintaining high accuracy [Feuz ¶26-27]
Claim 10 are rejected under 35 U.S.C. 103 as being unpatentable over Johnsson in view of Lupesko and Choudhary, as applied in claim 3 above, and further in view of Feuz.
As to dependent claim 10, Johnsson, Lupesko and Choudhary teach the method of claim 3 above that is incorporated,
Johnsson, Lupesko and Choudhary do not specifically teach wherein the machine learning models are trained based on user data reflecting one or more of user interests or user preferences of a user.
However, Feuz teaches wherein the machine learning models are trained based on user data reflecting one or more of user interests or user preferences of a user. [personalized training using user data ¶29 "models that capture the user preferences", ¶55 "update the machine-learned model over time as additional data (e.g., user-specific data) is received"]
Accordingly, it would have been obvious to a person of ordinary skill in the art before the effective filling date of the claimed invention to modify the learning process disclosed by Johnsson, Lupesko and Choudhary by incorporating the wherein the machine learning models are trained based on user data reflecting one or more of user interests or user preferences of a user by Feuz because all techniques address the same field of machine learning and by incorporating Feuz into Johnsson, Lupesko and Choudhary enhances the trust in models while maintaining high accuracy [Feuz ¶26-27]
Claim 18 is rejected under 35 U.S.C. 103 as being unpatentable over Kuo in view of Beaudoin (US 11455572 B2)
As to independent claim 18, Kuo teaches a machine learning system, comprising: [machine learning system ¶18]
a cloud training pipeline, [trained/developed in the cloud ¶15 "Machine learning models may be trained in the cloud (e.g., by a provider network) "]
a deployment engine; and [deployment service ¶12]
an edge inference pipeline, wherein [inference application ¶12]
the deployment engine is configured to deploy one of the number of tiers of machine learning models to an edge device based on a device constraint type of the edge device; and [different sized model deployed according to edge resources ¶53]
the edge inference pipeline is configured to access a machine learning model deployed to the edge device to process received input to generate a prediction. [deploys models ¶25, that generates inferences and predications based on data input (collected data) ¶28-29 "process the collected data to generate inference data (e.g., one or more inferences and/or one or more predictions)"
Kuo does not specifically teach the cloud training pipeline is configured to train a number of tiers of machine learning models for an application, a quantity of the number of tiers of machine learning models corresponding to a quantity of device constraint types for a plurality of edge devices capable of running the application;
However, Beaudoin teaches the cloud training pipeline is configured to train a number of tiers of machine learning models for an application, a quantity of the number of tiers of machine learning models corresponding to a quantity of device constraint types for a plurality of edge devices capable of running the application; [Trains tiers of models for particular types of systems and capabilities Col. 4 ln. 49-67 "the version installed on the various data processing systems may be stripped down versions of the trained model. As an example, the stripped down version of a fully trained neural network may have fewer hidden layers than the full version"]
Accordingly, it would have been obvious to a person of ordinary skill in the art before the effective filling date of the claimed invention to modify the model development disclosed by Kuo by incorporating the cloud training pipeline is configured to train a number of tiers of machine learning models for an application, a quantity of the number of tiers of machine learning models corresponding to a quantity of device constraint types for a plurality of edge devices capable of running the application disclosed by Beaudoin because both techniques address the same field of machine learning and by incorporating Beaudoin into Kuo reduces latency in learning systems for delivering model results [Beaudoin Col. 2 ln. 23-47]
Claim 19 is rejected under 35 U.S.C. 103 as being unpatentable over Kuo in view of Beaudoin, as applied in claim 18 above, and further in view of O'Neill.
As to dependent claim 19, Kuo and Beaudoin teach the method of claim 18 above that is incorporated,
Kuo and Beaudoin do not specifically teach wherein the quantity of device constraint types is determined based on device information of the plurality of edge devices capable of running the application.
However, O'Neill teaches wherein the quantity of device constraint types is determined based on device information of the plurality of edge devices capable of running the application. [package or size of update (quantity) determined from client devices (edge) Col. 7 ln. 36-50 "receives the identity information 113 from the client device 104 and subsequently generates the desired update package 110"]
Accordingly, it would have been obvious to a person of ordinary skill in the art before the effective filling date of the claimed invention to modify the learning process disclosed by Kuo and Beaudoin by incorporating the wherein the quantity of device constraint types is determined based on device information of the plurality of edge devices capable of running the application by O'Neill because all techniques address the same field of machine learning and by incorporating O'Neill into Kuo and Beaudoin provides a more convenient and reliable update process [O'Neill Col. 3 ln. 42-60]
Claim 20 is rejected under 35 U.S.C. 103 as being unpatentable over Kuo in view of Beaudoin, as applied in claim 18 above, and further in view of Feuz.
As to dependent claim 20, Kuo and Beaudoin teach the method of claim 18 above that is incorporated,
Kuo and Beaudoin do not specifically teach wherein the machine learning models are trained based on user data associated with a user, the user data reflecting one or more of user interests or user preferences of the user.
However, Feuz teaches wherein the machine learning models are trained based on user data associated with a user, the user data reflecting one or more of user interests or user preferences of the user. [personalized training using user data ¶29 "models that capture the user preferences", ¶55 "update the machine-learned model over time as additional data (e.g., user-specific data) is received"]
Accordingly, it would have been obvious to a person of ordinary skill in the art before the effective filling date of the claimed invention to modify the learning process disclosed by Kuo and Beaudoin by incorporating the wherein the machine learning models are trained based on user data associated with a user, the user data reflecting one or more of user interests or user preferences of the user by Feuz because all techniques address the same field of machine learning and by incorporating Feuz into Kuo and Beaudoin enhances the trust in models while maintaining high accuracy [Feuz ¶26-27]
Conclusion
The prior art made of record and not relied upon is considered pertinent to applicant's disclosure. Applicant is required under 37 C.F.R. § 1.111(c) to consider these references fully when responding to this action.
SCHMIDT et al. (US 20220156642 A1) teaches deployed models according to input features (see ¶61)
It is noted that any citation to specific pages, columns, lines, or figures in the prior art references and any interpretation of the references should not be considered to be limiting in any way. A reference is relevant for all it contains and may be relied upon for all that it would have reasonably suggested to one having ordinary skill in the art. In re Heck, 699 F.2d 1331, 1332-33, 216 U.S.P.Q. 1038, 1039 (Fed. Cir. 1983) (quoting In re Lemelson, 397 F.2d 1006, 1009, 158 U.S.P.Q. 275, 277 (C.C.P.A. 1968)).
Any inquiry concerning this communication or earlier communications from the examiner should be directed to Beau Spratt whose telephone number is 571 272 9919. The examiner can normally be reached 8:30am to 5:00pm (PST).
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, Jennifer Welch can be reached at 571 272 7212. The fax phone number for the organization where this application or proceeding is assigned is 571 483 7388.
Information regarding the status of an application may be obtained from the Patent Application Information Retrieval (PAIR) system. Status information for published applications may be obtained from either Private PAIR or Public PAIR. Status information for unpublished applications is available through Private PAIR only. For more information about the PAIR system, see http://pair-direct.uspto.gov. Should you have questions on access to the Private PAIR system, contact the Electronic Business Center (EBC) at 866 217 9197 (toll-free). If you would like assistance from a USPTO Customer Service Representative or access to the automated information system, call 800 786 9199 (IN USA OR CANADA) or 571 272 1000.
/BEAU D SPRATT/Primary Examiner, Art Unit 2143