DETAILED ACTION
Notice of Pre-AIA or AIA Status
The present application, filed on or after March 16, 2013, is being examined under the first inventor to file provisions of the AIA .
Specification
The disclosure is objected to because of the following informalities:
Paragraph [0020]: This allows various models to up updated using a data. The use of up updated is incorrect.
Appropriate correction is required.
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.
Claim 1-20 are rejected under 35 U.S.C. 101 because the claimed invention are directed to abstract ides without significantly more.
Regarding Claim 1:
Step 1 - Is the claim directed to a process, a machine, manufacture or composition of matter? -Yes, the claim is directed to a process.
Step 2A - Prong 1 - Does the claim recite an abstract idea, law of nature, or natural phenomenon? - Yes, the claim recites the abstract ideas:
detecting a new domain of data at a first node included in the nodes, the first node associated with a first model: - This limitation is directed to the abstract idea of a mental process, as the process of detecting a new domain of data is a thought process that can be performed in a human mind by observing and judging (concepts performed in the human mind (including an observation, evaluation, judgment, opinion) (see MPEP § 2106.04(a)(2), subsection III)).
Step 2A - Prong 2 - Does the claim recite additional elements that integrate the judicial exception into a practical application? - No, there are no additional elements that integrate the judicial exception into a practical application. The additional elements:
serving a machine learning model to nodes in a computing system: - This limitation invokes a computer merely as a tool for performing an existing process (see MPEP 2106.05(f)(2)) and therefore fails to integrate the exception into a practical application.
and adapting the first model to learn the new domain without forgetting previously learned domains, wherein adapting the first model includes retrieving sample data from other nodes that is similar to data of the new domain and training the first model with the sample data: - This limitation is directed to mere data gathering and outputting. The courts (as per Ultramercial, 772 F.3d at 715, 112 USPQ2d at 1754) have recognized mere data gathering and outputting as insignificant extra-solution activity (see MPEP 2106.05(g)(3)) and therefore fails to integrate the exception into a practical application.
Step 2B - Does the claim recite additional elements that amount to significantly more than the judicial exception? - No, there are no additional elements that amount to significantly more than the judicial exception. The additional elements:
serving a machine learning model to nodes in a computing system: - This limitation invokes a computer merely as a tool for performing an existing process (see MPEP 2106.05(f)(2)) and therefore fails to amount to significantly more than the judicial exception.
and adapting the first model to learn the new domain without forgetting previously learned domains, wherein adapting the first model includes retrieving sample data from other nodes that is similar to data of the new domain and training the first model with the sample data: - This limitation is directed to receiving or transmitting data over a network. The courts (as per Symantec, 838 F.3d at 1321, 120 USPQ2d at 1362) have recognized receiving or transmitting data over a network as well-understood, routine, and conventional functions when they are claimed in a merely generic manner (e.g., at a high level of generality) (see MPEP 2106.05(d) II).
Regarding Claim 2:
Step 1 - Is the claim directed to a process, a machine, manufacture or composition of matter? - Yes, the claim is directed to a process.
Step 2A - Prong 1 - Does the claim recite an abstract idea, law of nature, or natural phenomenon? - Yes, the claim is dependent on claim 1 which included an abstract (see rejection for claim 1).
Step 2A - Prong 2 - Does the claim recite additional elements that integrate the judicial exception into a practical application? - No, there are no additional elements that integrate the judicial exception into a practical application. The additional element:
training the machine learning model at a central server prior to serving the machine learning model to the nodes, wherein the model served to the nodes is associated with learned domains, domain summarizations, and a rehearsal data: - This limitation recites the training of machine learning model in high level of generality, as this limitation merely indicates a field of use or technological environment (see MPEP 2106.05(h)), therefore it fails to integrate the judicial exception into a practical application.
Step 2B - Does the claim recite additional elements that amount to significantly more than the judicial exception? - No, there are no additional elements that amount to significantly more than the judicial exception. The additional element:
training the machine learning model at a central server prior to serving the machine learning model to the nodes, wherein the model served to the nodes is associated with learned domains, domain summarizations, and a rehearsal data: - This limitation recites the training of machine learning model in high level of generality as this limitation merely indicates a field of use or technological environment (see MPEP 2106.05(h)), therefore it fails to amount to significantly more than the judicial exception.
Regarding Claim 3:
Step 1 - Is the claim directed to a process, a machine, manufacture or composition of matter? - Yes, the claim is directed to a process.
Step 2A - Prong 1 - Does the claim recite an abstract idea, law of nature, or natural phenomenon? - Yes, the claim is dependent on claim 2, claim 2 is dependent on claim 1 which included an abstract (see rejection for claim 1).
Step 2A - Prong 2 - Does the claim recite additional elements that integrate the judicial exception into a practical application? - No, there are no additional elements that integrate the judicial exception into a practical application. The additional element:
storing information for each of the nodes in a table at the central server, the information including a summarization of unlabeled data, a score of the rehearsal data, and an insertion order for domains subsequently learned: - This limitation is directed to an insignificant extra solution activity, as it is merely storing data which is a conventional computer function, therefore, it does not impose meaningful limits on the claim such that it is not nominally or tangentially related to the invention per 2106.05(g) (2).
Step 2B - Does the claim recite additional elements that amount to significantly more than the judicial exception? - No, there are no additional elements that amount to significantly more than the judicial exception. The additional element:
storing information for each of the nodes in a table at the central server, the information including a summarization of unlabeled data, a score of the rehearsal data, and an insertion order for domains subsequently learned: - This limitation amounts to storing and retrieving information in memory, as it is considered well-understood, routine and conventional under MPEP 2106.05(d) II (iv).
Regarding Claim 4:
Step 1 - Is the claim directed to a process, a machine, manufacture or composition of matter? - Yes, the claim is directed to a process.
Step 2A - Prong 1 - Does the claim recite an abstract idea, law of nature, or natural phenomenon? - Yes, the claim is dependent on claim 1 which included an abstract (see rejection for claim 1).
Step 2A - Prong 2 - Does the claim recite additional elements that integrate the judicial exception into a practical application? - No, there are no additional elements that integrate the judicial exception into a practical application. The additional element:
receiving a request for a new domain at the central server, the request including a summarization of data of the new domain: - This limitation is directed to mere data gathering and outputting. The courts (as per Ultramercial, 772 F.3d at 715, 112 USPQ2d at 1754) have recognized mere data gathering and outputting as insignificant extra-solution activity (see MPEP 2106.05(g)(3)) and therefore fails to integrate the exception into a practical application.
Step 2B - Does the claim recite additional elements that amount to significantly more than the judicial exception? - No, there are no additional elements that amount to significantly more than the judicial exception. The additional element:
receiving a request for a new domain at the central server, the request including a summarization of data of the new domain: - This limitation is directed to receiving or transmitting data over a network. The courts (as per Symantec, 838 F.3d at 1321, 120 USPQ2d at 1362) have recognized receiving or transmitting data over a network as well-understood, routine, and conventional functions when they are claimed in a merely generic manner (e.g., at a high level of generality) (see MPEP 2106.05(d) II).
Regarding Claim 5:
Step 1 - Is the claim directed to a process, a machine, manufacture or composition of matter? - Yes, the claim is directed to a process.
Step 2A - Prong 1 - Does the claim recite an abstract idea, law of nature, or natural phenomenon? - Yes, the claim is dependent on claim 4, claim 4 is dependent on claim 1 which included an abstract (see rejection for claim 1). Additionally, claim5 recites the abstract idea:
determining whether other nodes have learned the new domain by comparing the summarization data with summarization data of the other nodes stored in the table: - This limitation is directed to the abstract idea of a mental process, as the process of comparing the summarization data is a thought process that can be performed in a human mind by observing, evaluating and judging (concepts performed in the human mind (including an observation, evaluation, judgment, opinion) (see MPEP § 2106.04(a)(2), subsection III)).
Step 2A - Prong 2 - Does the claim recite additional elements that integrate the judicial exception into a practical application? - No, there are no additional elements that integrate the judicial exception into a practical application.
Step 2B - Does the claim recite additional elements that amount to significantly more than the
judicial exception? - No, there are no additional elements that amount to significantly more than the judicial exception.
Regarding Claim 6:
Step 1 - Is the claim directed to a process, a machine, manufacture or composition of matter? - Yes, the claim is directed to a process.
Step 2A - Prong 1 - Does the claim recite an abstract idea, law of nature, or natural phenomenon? - Yes, the claim is dependent on claim 5 which included an abstract (see rejection for claim 5). Additionally, claim 6 recites the abstract idea:
generating a labeled dataset including labeled data from the new domain from some of the other nodes: - This limitation is directed to the abstract idea of a mental process, as the process of labeling unlabeled data is a thought process that can be performed in a human mind by observing and judging (concepts performed in the human mind (including an observation, evaluation, judgment, opinion) (see MPEP § 2106.04(a)(2), subsection III)).
Step 2A - Prong 2 - Does the claim recite additional elements that integrate the judicial exception into a practical application? - No, there are no additional elements that integrate the judicial exception into a practical application.
Step 2B - Does the claim recite additional elements that amount to significantly more than the
judicial exception? - No, there are no additional elements that amount to significantly more than the
judicial exception.
Regarding Claim 7:
Step 1 - Is the claim directed to a process, a machine, manufacture or composition of matter? - Yes, the claim is directed to a process.
Step 2A - Prong 1 - Does the claim recite an abstract idea, law of nature, or natural phenomenon? - Yes, the claim is dependent on claim 6 which included an abstract (see rejection for claim 6).
Step 2A - Prong 2 - Does the claim recite additional elements that integrate the judicial exception into a practical application? - No, there are no additional elements that integrate the judicial exception into a practical application. The additional element:
training the first model using the labeled dataset and the data from the new domain at the first node: - This limitation recites the training of machine learning model in high level of generality, as this limitation merely indicates a field of use or technological environment (see MPEP 2106.05(h)), therefore it fails to integrate the judicial exception into a practical application.
Step 2B - Does the claim recite additional elements that amount to significantly more than the judicial exception? - No, there are no additional elements that amount to significantly more than the judicial exception. The additional element:
training the first model using the labeled dataset and the data from the new domain at the first node: - This limitation recites the training of machine learning model in high level of generality as this limitation merely indicates a field of use or technological environment (see MPEP 2106.05(h)), therefore it fails to amount to significantly more than the judicial exception.
Regarding Claim 8:
Step 1 - Is the claim directed to a process, a machine, manufacture or composition of matter? - Yes, the claim is directed to a process.
Step 2A - Prong 1 - Does the claim recite an abstract idea, law of nature, or natural phenomenon? - Yes, the claim is dependent on claim 7, claim 7 is dependent on claim 6 which included an abstract (see rejection for claim 6). Additionally, claim 8 recites the abstract idea:
selecting nodes from the other nodes to acquire the labeled dataset by identifying a set of nodes containing the new domain that achieve a performance greater than a threshold and selecting the nodes whose intersection of its learned domains and learned domains of the first node are highest: - This claim is directed to a mathematical concept, as the process of identifying a set of nodes that achieve a performance greater than a threshold is a step of “determining” a variable or number using mathematical method (see MPEP 2106.04(a)(2) subsection C).
Step 2A - Prong 2 - Does the claim recite additional elements that integrate the judicial exception into a practical application? - No, there are no additional elements that integrate the judicial exception into a practical application.
Step 2B - Does the claim recite additional elements that amount to significantly more than the judicial exception? - No, there are no additional elements that amount to significantly more than the judicial exception.
Regarding Claim 9:
Step 1 - Is the claim directed to a process, a machine, manufacture or composition of matter? - Yes, the claim is directed to a process.
Step 2A - Prong 1 - Does the claim recite an abstract idea, law of nature, or natural phenomenon? - Yes, the claim is dependent on claim 1 which included an abstract (see rejection for claim 1). Additionally, claim 9 recites the abstract idea:
labeling data at the first node using the updated first model: - This limitation is directed to the abstract idea of a mental process, as the process of labeling unlabeled data is a thought process that can be performed in a human mind by observing and judging (concepts performed in the human mind (including an observation, evaluation, judgment, opinion) (see MPEP § 2106.04(a)(2), subsection III)).
Step 2A - Prong 2 - Does the claim recite additional elements that integrate the judicial exception into a practical application? - No, there are no additional elements that integrate the judicial exception into a practical application.
Step 2B - Does the claim recite additional elements that amount to significantly more than the judicial exception? - No, there are no additional elements that amount to significantly more than the judicial exception.
Regarding Claim 10:
Step 1 - Is the claim directed to a process, a machine, manufacture or composition of matter? - Yes, the claim is directed to a process.
Step 2A - Prong 1 - Does the claim recite an abstract idea, law of nature, or natural phenomenon? - Yes, the claim is dependent on claim 9 which included an abstract (see rejection for claim 9).
Step 2A - Prong 2 - Does the claim recite additional elements that integrate the judicial exception into a practical application? - No, there are no additional elements that integrate the judicial exception into a practical application. The additional element:
sending scores of the updated first model to the central server: - This limitation is directed to mere data gathering and outputting. The courts (as per Ultramercial, 772 F.3d at 715, 112 USPQ2d at 1754) have recognized mere data gathering and outputting as insignificant extra-solution activity (see MPEP 2106.05(g)(3)) and therefore fails to integrate the exception into a practical application.
Step 2B - Does the claim recite additional elements that amount to significantly more than the judicial exception? - No, there are no additional elements that amount to significantly more than the judicial exception. The additional element:
sending scores of the updated first model to the central server: - This limitation is directed to receiving or transmitting data over a network. The courts (as per Symantec, 838 F.3d at 1321, 120 USPQ2d at 1362) have recognized receiving or transmitting data over a network as well-understood, routine, and conventional functions when they are claimed in a merely generic manner (e.g., at a high level of generality) (see MPEP 2106.05(d) II).
Regarding claim 11, this claim is directed to a non-transitory medium and is rejected on the same basis as claim 1 since they are analogous.
Regarding claim 12, this claim is directed to a non-transitory medium and is rejected on the same basis as claim 2 since they are analogous.
Regarding claim 13, this claim is directed to a non-transitory medium and is rejected on the same basis as claim 3 since they are analogous.
Regarding claim 14, this claim is directed to a non-transitory medium and is rejected on the same basis as claim 4 since they are analogous.
Regarding claim 15, this claim is directed to a non-transitory medium and is rejected on the same basis as claim 5 since they are analogous.
Regarding claim 16, this claim is directed to a non-transitory medium and is rejected on the same basis as claim 6 since they are analogous.
Regarding claim 17, this claim is directed to a non-transitory medium and is rejected on the same basis as claim 7 since they are analogous.
Regarding claim 18, this claim is directed to a non-transitory medium and is rejected on the same basis as claim 8 since they are analogous.
Regarding claim 19, this claim is directed to a non-transitory medium and is rejected on the same basis as claim 9 since they are analogous.
Regarding claim 20, this claim is directed to a non-transitory medium and is rejected on the same basis as claim 10 since they are analogous.
Claim Rejections - 35 USC § 103
In the event the determination of the status of the application as subject to AIA 35 U.S.C. 102 and 103 (or as subject to pre-AIA 35 U.S.C. 102 and 103) is incorrect, any correction of the statutory basis (i.e., changing from AIA to pre-AIA ) for the rejection will not be considered a new ground of rejection if the prior art relied upon, and the rationale supporting the rejection, would be the same under either status.
The following is a quotation of 35 U.S.C. 103 which forms the basis for all obviousness rejections set forth in this Office action:
A patent for a claimed invention may not be obtained, notwithstanding that the claimed invention is not identically disclosed as set forth in section 102, if the differences between the claimed invention and the prior art are such that the claimed invention as a whole would have been obvious before the effective filing date of the claimed invention to a person having ordinary skill in the art to which the claimed invention pertains. Patentability shall not be negated by the manner in which the invention was made.
The factual inquiries for establishing a background for determining obviousness under 35 U.S.C. 103 are summarized as follows:
1. Determining the scope and contents of the prior art.
2. Ascertaining the differences between the prior art and the claims at issue.
3. Resolving the level of ordinary skill in the pertinent art.
4. Considering objective evidence present in the application indicating obviousness or nonobviousness.
This application currently names joint inventors. In considering patentability of the claims the examiner presumes that the subject matter of the various claims was commonly owned as of the effective filing date of the claimed invention(s) absent any evidence to the contrary. Applicant is advised of the obligation under 37 CFR 1.56 to point out the inventor and effective filing dates of each claim that was not commonly owned as of the effective filing date of the later invention in order for the examiner to consider the applicability of 35 U.S.C. 102(b)(2)(C) for any potential 35 U.S.C. 102(a)(2) prior art against the later invention.
Claims 1-3, 4-5, 7, 10-11, 14-15, 17 are rejected under 35 U.S.C. 103 as being unpatentable over Zhang et al (CN 115081532A - hereinafter Zhang ) in view of Yu et al (US Patent 12147879 - hereinafter Yu).
Referring to Claim 1, Zhang teaches a method comprising:
serving a machine learning model to nodes in a computing system (see Zhang Paragraph 23: “Step 6.1: Select up to 5 clients and distribute the constructed federated learning global network and pseudo-sample set to each client participating in this round of training”. Examiner interprets the distribution of the constructed federated learning global network to each client to be equivalent as the claimed “serving a machine learning model to nodes”);
detecting a new domain of data at a first node included in the nodes, the first node associated with a first model (see Zhang Paragraph 11: “Step 1.3: Using the same method as in Step 1.2, obtain local sample sets for at least 10 clients. Set the task sample sets for each client to arrive in sequence, and discard the sample sets of old tasks immediately when a new task sample set arrives”. Examiner interprets the arrival of new task sample set in sequence to be equivalent as the claimed “detecting a new domain of data at a first node included in the nodes, the first node associated with a first model”).
However, Zhang fails to teach:
and adapting the first model to learn the new domain without forgetting previously learned domains, wherein adapting the first model includes retrieving sample data from other nodes that is similar to data of the new domain and training the first model with the sample data.
Yu teaches, in analogous system,
adapting the first model to learn the new domain without forgetting previously learned domains, wherein adapting the first model includes retrieving sample data from other nodes that is similar to data of the new domain and training the first model with the sample data (see Yu Column 7 Lines 23 to 46: “As stated previously, in some illustrative embodiments, the grouping is based on data statistical measures/characteristics for the local datasets as indicated in the dataset sketch commitment data structures. For example, a clustering, e.g., K-means clustering, of dataset sketch commitment data structures (which again are vector data structures), may be performed on the vectors to identify groupings of participants. The result is that participants having similar local datasets are identified through the clustering or grouping performed based on the statistical characteristics of their individual local datasets used to train their own individual local computer models that are the basis for the updates provided as part of the federated machine learning. Having identified groupings of participants, within each group, aggregation over the updates from those participants in that group (or cluster) is performed. The updates from the participants within the group are more alike other participants within the group than with participants that are not within the group, i.e., part of another group. The aggregation, e.g., robust aggregation such as trimmed-mean aggregation, median-based aggregation, or the like or scoring-based aggregation, is performed within each group from approximately identical independent distribution (iid) datasets”. Examiner interprets the fact that participants having similar local datasets are identified through the clustering to be equivalent as the claimed “retrieving sample data from other nodes that is similar to data of the new domain” and performing the aggregation over the updates from those participants in the group is interpreted to be equivalent as the claimed “adapting the first model to learn the new domain without forgetting previously learned domains”).
It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to modify the teachings of Zhang with the above teachings of Yu by distributing the learning global network to each client and clients use memory replay method to train the distributed global network when new tasks arrive and aggregate updates to the global network, as taught by Zhang, and providing statistical characteristics of dataset to a participant to train the federated machine learning model locally and the modified set of updates generated is used to update the federated machine learning model, as taught by Yu. The modification would have been obvious because one of ordinary skill in the art would be motivated to identify participants having similar local datasets through grouping (as suggested by Yu at Column 7 Lines 30 to 35: “The result is that participants having similar local datasets are identified through the clustering or grouping performed based on the statistical characteristics of their individual local datasets used to train their own individual local computer models that are the basis for the updates provided as part of the federated machine learning”).
Referring to Claim 2, Zhang – Yu teaches the method of claim 1, Zhang further teaches:
further comprising training the machine learning model at a central server prior to serving the machine learning model to the nodes, wherein the model served to the nodes is associated with learned domains, domain summarizations, and a rehearsal data (see Zhang at Paragraph 43: “Based on the digital image dataset in step 1, this invention constructs a seven-layer convolutional neural network as a federated learning global network on the server. Its structure is as follows: first convolutional layer, first pooling layer, second convolutional layer, second pooling layer, third convolutional layer, first fully connected layer, and second fully connected layer”. Examiner interprets the construction on the server of a seven-layer convolutional neural network as a federated learning global network to be equivalent as the claimed “training the machine learning model at a central server prior to serving the machine learning model to the nodes”) and (see Zhang at Paragraph 45: “Step 4.1: In this invention, a Generative Adversarial Network (GAN) is constructed on the server to accumulate client knowledge and generate old task data”, where Examiner interprets the old task and the client knowledge to be equivalent as the claimed “learned domains” and “domain summarizations” respectively) and further (see Zhang at Paragraph 85: “As can be seen from Table 1, the classification accuracy of the federated learning global network for Task 1 in this invention remained at 88.6% after training for 5 tasks, without a significant decrease. Moreover, the accuracy was higher than that of the other 3 existing technologies, proving that this invention prevents the federated learning global network from forgetting knowledge of old tasks in the task flow by using the method of old task memory replay and local differential privacy”, where Examiner interprets old task memory replay to be equivalent as the claimed “rehearsal”).
Referring to Claim 3, Zhang – Yu teaches the method of claim 2, Zhang further teaches:
storing information for each of the nodes in a table at the central server, the information including a summarization of unlabeled data, a score of the rehearsal data, and an insertion order for domains subsequently learned (see Zhang at Paragraph 6: “The technical approach to achieving the objective of this invention is as follows: This invention maintains a memory generator model composed of a generative adversarial network (GAN) in a central server, continuously accumulates client task knowledge using data uploaded by the client, and sends the generated task pseudo data to the client. The client mixes the received pseudo data with the current task data according to the importance ratio for training, thereby enabling the effective recovery of old task knowledge without increasing the client's computational burden”. Examiner interprets the accumulation of client task knowledge in a memory generator at central server and the data uploaded by client as well as the importance ratio to be equivalent as the claimed “storing information for each of the nodes in a table at the central server”, “summarization of unlabeled data” and “score of the rehearsal data” respectively) and further (see Zhang at Paragraph 39 “Step 1.3: Using the same method as in Step 1.2, obtain local sample sets for at least 10 clients. Set the task sample sets for each client to arrive in sequence, and discard the sample sets of old tasks immediately when a new task sample set arrives” where Examiner interprets the setting of the task sample sets for each client to arrive in sequence to be equivalent as the claimed “insertion order for domains subsequently learned”).
Referring to Claim 4, Zhang – Yu teaches the method of claim 1, Zhang further teaches:
receiving a request for a new domain at the central server, the request including a summarization of data of the new domain (see Zhang at Paragraph 39: “Step 1.3: Using the same method as in Step 1.2, obtain local sample sets for 10 clients. Set the task sample sets for each client to arrive in sequence, and discard the sample sets of old tasks immediately when a new task sample set arrives”. Examiner interprets obtaining local sample sets to be equivalent as the claimed “receiving a request for a new domain at the central server”, the sample sets is interpreted as including the claimed “data summarization”).
Referring to Claim 5, Zhang teaches the method of claim 4.
However, Zhang fails to teach:
further comprising determining whether other nodes have learned the new domain by comparing the summarization data with summarization data of the other nodes stored in the table.
Yu teaches, in analogous system,
further comprising determining whether other nodes have learned the new domain by comparing the summarization data with summarization data of the other nodes stored in the table (see Yu Column 8 Line 67 and Column 9 Lines 1 to 6: “That is, after grouping participants in the manner described previously, data samples may be selected from prior public datasets which match the dataset sketch commitment data structure for the group. These samples may be used to produce a reference update against which individual participant updates may be compared to determine deviations from the reference update”. Examiner interprets the reference update to be equivalent as the claimed “summarization data”).
It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to modify the teachings of Zhang with the above teachings of Yu by distributing the learning global network to each client and clients use memory replay method to train the distributed global network when new tasks arrive and aggregate updates to the global network, as taught by Zhang, and providing statistical characteristics of dataset to a participant to train the federated machine learning model locally and the modified set of updates generated is used to update the federated machine learning model, as taught by Yu. The modification would have been obvious because one of ordinary skill in the art would be motivated to produce reference updates and compare participant updates to reference updates in order to determine deviations (as suggested by Yu at Column 9 Lines 3 to 6: “These samples may be used to produce a reference update against which individual participant updates may be compared to determine deviations from the reference update”).
Referring to Claim 7, Zhang - Yu teaches the method of claim 6, Zhang further teaches:
training the first model using the labeled dataset and the data from the new domain at the first node (see Zhang at Paragraph 39: “Step 6.2: Combine the partial pseudo-sample set Dp with the partial local sample set according to the importance ratio of each task to obtain the mixed local sample set. Step 6.3: All clients participating in this round of training use the mixed local sample set and local stochastic gradient descent to iteratively update the parameters of each layer of the federated learning global network until the total loss function of the federated learning global network training converges, thus obtaining the trained federated learning global network”. Examiner interprets the pseudo-sample set and the local sample set as well as the trained federated learning global network to be equivalent as the claimed “labeled dataset”, “data from new domain” and “first model” respectively).
Referring to Claim 10, Zhang - Yu teaches the method of claim 9, Zhang further teaches:
further comprising sending scores of the updated first model to the central server (see Zhang Paragraph 67: Therefore, this weighted aggregation method, by integrating the local model gradients trained by various local users, eliminates to some extent the impact of the imbalanced sample size problem of local users on the accuracy of the global network, which is beneficial to the optimization of the global network in federated learning”. Examiner interprets the integration of the local model gradients to be equivalent as the claimed “sending scores”).
Referring to independent Claim 11, this claim is rejected on the same basis as independent claim 1 since they are analogous claims.
Referring to dependent Claim 12, this claim is rejected on the same basis as dependent claim 2 since they are analogous claims.
Referring to dependent Claim 13, this claim is rejected on the same basis as dependent claim 3 since they are analogous claims.
Referring to dependent Claim 14, this claim is rejected on the same basis as dependent claim 4 since they are analogous claims.
Referring to dependent Claim 15, this claim is rejected on the same basis as dependent claim 5 since they are analogous claims.
Referring to dependent Claim 17, this claim is rejected on the same basis as dependent claim 7 since they are analogous claims.
Referring to dependent Claim 20, this claim is rejected on the same basis as dependent claim 10 since they are analogous claims.
Claims 6, 8-9, 16, 18-19 are rejected under 35 U.S.C. 103 as being unpatentable over Zhang et al (CN 115081532A - hereinafter Zhang) in view of Yu et al (US Patent 12147879 - hereinafter Yu) and in further view of Watanabe et al(US Patent 12008075 - hereinafter Watanabe).
Referring to Claim 6, the combination of Zhang and Yu teaches the method of claim 5.
However, the combination of Zhang and Yu fails to teach,
further comprising generating a labeled dataset including labeled data from the new domain from some of the other nodes.
Watanabe teaches, in analogous system,
further comprising generating a labeled dataset including labeled data from the new domain from some of the other nodes (see Watanabe Column 6 Lines 24 to 31: “In some embodiments, the additional training data includes augmented data that is automatically generated based on a small number of labeled data samples. Data augmentation can be performed by central training server 105 and/or by distributed computing nodes 130A-130N, and can produce a large volume of labeled training data using only a small set of manually labeled data”. Examiner interprets the production of a large volume of labeled training data by computing nodes 130A-130N to be equivalent as the claimed “generating a labeled dataset including labeled data from the new domain from some of the other nodes”).
It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to modify the combination of Zhang and Yu with the above teachings of Watanabe by distributing the learning global model to each client and clients use memory replay method with statistical characteristics of dataset to train the distributed global model when new tasks arrive and aggregate updates to the global model, as taught by Zhang and Yu, and including labeled data from the new domain from some other nodes, as taught by Watanabe. The modification would have been obvious because one of ordinary skill in the art would be motivated to produce a large volume of labeled training data using a small number of labeled data samples (as suggested by Watanabe at Column 6 Lines 27 to 32: “Data augmentation can be performed by central training server 105 and/or by distributed computing nodes 130A-130N, and can produce a large volume of labeled training data using only a small set of manually labeled data. Data augmentation is depicted and described in further detail with respect to FIG. 4”).
Referring to Claim 8, the combination of Zhang and Yu teaches the method of claim 7.
However, the combination of Zhang and Yu fails to teach,
further comprising selecting nodes from the other nodes to acquire the labeled dataset by identifying a set of nodes containing the new domain that achieve a performance greater than a threshold and selecting the nodes whose intersection of its learned domains and learned domains of the first node are highest.
Watanabe teaches, in analogous system,
further comprising selecting nodes from the other nodes to acquire the labeled dataset by identifying a set of nodes containing the new domain that achieve a performance greater than a threshold and selecting the nodes whose intersection of its learned domains and learned domains of the first node are highest (see Watanabe Column 6 Lines 33 to 42 : “Threshold selection module 125 may determine a threshold value for node analysis module 120 to employ in the analysis of the statistical data that is obtained from distributed computing nodes 130A-130N. In particular, threshold selection module 125 may experimentally determine a threshold value by iteratively generating a biased set of training data, training a model on that data, testing the model, and gradually adding samples to the training data to reduce bias until a model is obtained that achieves a desired level of performance”. Examiner interprets the generation of biased set of training data at some nodes based on the determined threshold value to be equivalent as the claimed “selecting nodes from the other nodes to acquire the labeled dataset by identifying a set of nodes containing the new domain that achieve a performance greater than a threshold and selecting the nodes whose intersection of its learned domains and learned domains of the first node are highest” ).
It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to modify the combination of Zhang and Yu with the above teachings of Watanabe by distributing the learning global model to each client and clients use memory replay method with statistical characteristics of dataset to train the distributed global model when new tasks arrive and aggregate updates to the global model, as taught by Zhang and Yu, and selecting nodes from the nodes to acquire the labeled dataset, as taught by Watanabe. The modification would have been obvious because one of ordinary skill in the art would be motivated to achieve a desired level of performance (as suggested by Watanabe at Column 6 Lines 36 to 42: “In particular, threshold selection module 125 may experimentally determine a threshold value by iteratively generating a biased set of training data, training a model on that data, testing the model, and gradually adding samples to the training data to reduce bias until a model is obtained that achieves a desired level of performance”).
Referring to Claim 9, the combination of Zhang and Yu teaches the method of claim 1.
However, the combination of Zhang and Yu fails to teach,
further comprising labeling data at the first node using the updated first model.
Watanabe teaches, in analogous system,
further comprising labeling data at the first node using the updated first model (see Watanabe Column 6 Lines 24 to 31: “In some embodiments, the additional training data includes augmented data that is automatically generated based on a small number of labeled data samples. Data augmentation can be performed by central training server 105 and/or by distributed computing nodes 130A-130N, and can produce a large volume of labeled training data using only a small set of manually labeled data”. Examiner interprets the production of a large volume of labeled training data by the computing nodes to be equivalent as the claimed “labeling data at the first node using the updated first model”).
It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to modify the combination of Zhang and Yu with the above teachings of Watanabe by distributing the learning global model to each client and clients use memory replay method with statistical characteristics of dataset to train the distributed global model when new tasks arrive and aggregate updates to the global model, as taught by Zhang and Yu, and labeling data at the first node using the updated first model, as taught by Watanabe. The modification would have been obvious because one of ordinary skill in the art would be motivated to produce a large volume of labeled training data using a small number of labeled data samples (as suggested by Watanabe at Column 6 Lines 27 to 31: “Data augmentation can be performed by central training server 105 and/or by distributed computing nodes 130A-130N, and can produce a large volume of labeled training data using only a small set of manually labeled data”).
Referring to dependent Claim 16, this claim is rejected on the same basis as dependent claim 6 since they are analogous claims.
Referring to dependent Claim 18, this claim is rejected on the same basis as dependent claim 8 since they are analogous claims.
Referring to dependent Claim 19, this claim is rejected on the same basis as dependent claim 9 since they are analogous claims.
Conclusion
Any inquiry concerning this communication or earlier communications from the examiner should be directed to AWADAGBE G HOUNTON whose telephone number is (571)270-0670. The examiner can normally be reached Monday-Friday 8am-5pm.
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, David Yi can be reached at (571) 270-7519. 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.
/AWADAGBE G HOUNTON/Examiner, Art Unit 2126
/DAVID YI/Supervisory Patent Examiner, Art Unit 2126
Prosecution Insights