Notice of Pre-AIA or AIA Status
The present application, filed on or after March 16, 2013, is being examined under the first inventor to file provisions of the AIA .
Claim Rejections - 35 USC § 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 1-10, and 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 federated learning method, applied to a first node, wherein the method comprises” and is thus a process, one of the four statutory categories of patentable subject matter.
Subject Matter Eligibility Analysis Step 2A Prong 1:
“selecting, ” (a mental process that can be performed in the human mind, i.e. judgement; Selecting client machines with datasets containing data information and class type that is similar to the required training data type.)
“indicating, ” (a mental process that can be performed in the human mind, i.e. judgement; Identifying the client machines that will perform a training process.)
Claim 1 therefore recites an abstract idea.
Subject Matter Eligibility Analysis Step 2A Prong 2:
"obtaining, by the first node, data distribution information of the plurality of second nodes based on a target data feature required by a training task, wherein data distribution information of any second node indicates a data class of service data that is both locally stored in the second node and that satisfies the target data feature” (This step is directed to data gathering, which is understood to be insignificant extra solution activity - see MPEP 2106.05(g))
“... by the first node ...” (mere instructions to apply the exception using a generic computer component - see MPEP 2106.05(f))
"” (This step is directed to data gathering, which is understood to be insignificant extra solution activity - see MPEP 2106.05(g))
The additional elements as disclosed above alone or in combination do not integrate the judicial exception into practical application as they are mere insignificant extra solution activity in combination of generic computer functions being implemented with generic computer elements in a high level of generality to perform the disclosed abstract idea above. Therefore, Claim 1 is directed to the abstract idea.
Subject Matter Eligibility Analysis Step 2B:
"obtaining, by the first node, data distribution information of the plurality of second nodes based on a target data feature required by a training task, wherein data distribution information of any second node indicates a data class of service data that is both locally stored in the second node and that satisfies the target data feature” (This step is directed to transmitting or receiving information, which is understood to be insignificant extra solution activity and well understood, routine and conventional activity of transmitting and receiving data as identified by the court - see MPEP 2106.05(d))
“... by the first node ...” (mere instructions to apply the exception using a generic computer component - see MPEP 2106.05(f))
"” (This step is directed to transmitting or receiving information, which is understood to be insignificant extra solution activity and well understood, routine and conventional activity of transmitting and receiving data as identified by the court - see MPEP 2106.05(d))
The additional elements as disclosed above alone or in combination do not recite significantly more than the abstract idea itself as they are mere insignificant extra solution activity in combination of generic computer functions being implemented with generic computer elements in a high level of generality to perform the disclosed abstract idea above. Therefore, Claim 1 is subject-matter ineligible.
Regarding Claim 2:
Subject Matter Eligibility Analysis Step 2A Prong 1:
“” (a mental process that can be performed in the human mind, i.e. judgement)
Subject Matter Eligibility Analysis Step 2A Prong 2 & 2B:
“wherein at least one data analysis model is deployed in each second node, and each data analysis model corresponds to one data feature group and ” (mere instructions to apply the exception using a generic computer component - see MPEP 2106.05(f))
“the obtaining, by the first node, data distribution information of the plurality of second nodes based on the target data feature required by the training task comprises” (This step is directed to data gathering, which is understood to be insignificant extra solution activity (2106.05(g) in step 2A prong 2) and well understood, routine and conventional activity of transmitting and receiving data as identified by the court (2106.05(d) in step 2B))
“sending, by the first node, a first query message to each of the plurality of second nodes based on the target data feature, wherein the first query message sent to each second node comprises an identifier of the target data feature and an identifier of a target data analysis model, and the target data analysis model corresponds to the target data feature” (This step is directed to data gathering, which is understood to be insignificant extra solution activity (2106.05(g) in step 2A prong 2) and well understood, routine and conventional activity of transmitting and receiving data as identified by the court (2106.05(d) in step 2B))
“separately receiving, by the first node, the corresponding data distribution information from the plurality of second nodes, wherein data distribution information of each second node indicates an identifier of at least one data class and data information of service data that is stored in the second node and that separately belongs to the at least one data class” (This step is directed to data gathering, which is understood to be insignificant extra solution activity (2106.05(g) in step 2A prong 2) and well understood, routine and conventional activity of transmitting and receiving data as identified by the court (2106.05(d) in step 2B))
Regarding Claim 3:
Subject Matter Eligibility Analysis Step 2A Prong 1: None
Subject Matter Eligibility Analysis Step 2A Prong 2 & 2B:
“wherein the first query message sent by the first node to each second node further comprises an identifier of the target data class, and the data distribution information fed back by the second node comprises the identifier of the target data class and data information of the target service data that is stored in the second node and that belongs to the target data class” (merely specifies a particular technological environment in which the abstract idea is to take place, ie. a field of use, and thus does not integrate the abstract idea into a practical application nor cannot provide significantly more than the abstract idea itself - see MPEP 2106.05(h))
Regarding Claim 4:
Subject Matter Eligibility Analysis Step 2A Prong 1: None
Subject Matter Eligibility Analysis Step 2A Prong 2 & 2B:
“sending, by the first node, a data analysis model deployment message to each of the plurality of second nodes, wherein the data analysis model deployment message sent to each second node comprises an identifier of the at least one data analysis model and a model file of the at least one data analysis model” (This step is directed to data gathering, which is understood to be insignificant extra solution activity (2106.05(g) in step 2A prong 2) and well understood, routine and conventional activity of transmitting and receiving data as identified by the court (2106.05(d) in step 2B))
Regarding Claim 5:
Subject Matter Eligibility Analysis Step 2A Prong 1: None
Subject Matter Eligibility Analysis Step 2A Prong 2 & 2B:
“sending, by the first node, a model training message to each of the at least two target second nodes, wherein the model training message sent to each target second node comprises an identifier of a target artificial intelligence AI model, and the target AI model corresponds to the target data class” (This step is directed to data gathering, which is understood to be insignificant extra solution activity (2106.05(g) in step 2A prong 2) and well understood, routine and conventional activity of transmitting and receiving data as identified by the court (2106.05(d) in step 2B))
“obtaining, by the first node based on updated AI models respectively received from the at least two target second nodes, the federated learning model that is in the training task and that corresponds to the target data class” (This step is directed to data gathering, which is understood to be insignificant extra solution activity (2106.05(g) in step 2A prong 2) and well understood, routine and conventional activity of transmitting and receiving data as identified by the court (2106.05(d) in step 2B))
Regarding Claim 6:
Subject Matter Eligibility Analysis Step 2A Prong 1: None
Subject Matter Eligibility Analysis Step 2A Prong 2 & 2B:
“wherein the model training message sent to any target second node further comprises the identifier of the target data class and the identifier of the target data analysis model” (merely specifies a particular technological environment in which the abstract idea is to take place, ie. a field of use, and thus does not integrate the abstract idea into a practical application nor cannot provide significantly more than the abstract idea itself - see MPEP 2106.05(h))
Regarding Claim 7:
Subject Matter Eligibility Analysis Step 2A Prong 1: None
Subject Matter Eligibility Analysis Step 2A Prong 2 & 2B:
“sending, by the first node, a model evaluation message to each of the at least two target second nodes, wherein the model evaluation message sent to each target second node comprises an identifier and an evaluation indicator of a target evaluation model, and the target evaluation model corresponds to the target data class” (This step is directed to data gathering, which is understood to be insignificant extra solution activity (2106.05(g) in step 2A prong 2) and well understood, routine and conventional activity of transmitting and receiving data as identified by the court (2106.05(d) in step 2B))
“separately receiving, by the first node, corresponding model evaluation results from the at least two target second nodes” (This step is directed to data gathering, which is understood to be insignificant extra solution activity (2106.05(g) in step 2A prong 2) and well understood, routine and conventional activity of transmitting and receiving data as identified by the court (2106.05(d) in step 2B))
Regarding Claim 8:
Subject Matter Eligibility Analysis Step 2A Prong 1: None
Subject Matter Eligibility Analysis Step 2A Prong 2 & 2B:
“wherein the model evaluation message sent to each target second node further comprises the identifier of the target data class and the identifier of the target data analysis model” (merely specifies a particular technological environment in which the abstract idea is to take place, ie. a field of use, and thus does not integrate the abstract idea into a practical application nor cannot provide significantly more than the abstract idea itself - see MPEP 2106.05(h))
Regarding Claim 9:
Subject Matter Eligibility Analysis Step 2A Prong 1: None
Subject Matter Eligibility Analysis Step 2A Prong 2 & 2B:
“wherein the first node is part of a federated learning system, and the federated learning system is a wireless AI model–driven network system; the first node comprises a model management function (MMF) module; and any second node comprises a model training function (MTF) module, a data management function (DMF) module, and a model evaluation function (MEF) module” (merely specifies a particular technological environment in which the abstract idea is to take place, ie. a field of use, and thus does not integrate the abstract idea into a practical application nor cannot provide significantly more than the abstract idea itself - see MPEP 2106.05(h))
“wherein the at least one data analysis model is deployed in the DMF module or the MTF module” (mere instructions to apply the exception using a generic computer component - see MPEP 2106.05(f))
“the sending, by the first node, a first query message to each second node comprises: sending, by the MMF module, the first query message to the DMF module or the MTF module of each second node” (This step is directed to data gathering, which is understood to be insignificant extra solution activity (2106.05(g) in step 2A prong 2) and well understood, routine and conventional activity of transmitting and receiving data as identified by the court (2106.05(d) in step 2B))
Regarding Claim 10:
Subject Matter Eligibility Analysis Step 2A Prong 1: None
Subject Matter Eligibility Analysis Step 2A Prong 2 & 2B:
“sending, by the first node, a mapping relationship table to each of the plurality of second nodes, wherein the mapping relationship table sent to each second node is used for recording a mapping relationship between an identifier of a data feature, an identifier of an AI model, an identifier of a data analysis model, and an identifier of a data class” (This step is directed to data gathering, which is understood to be insignificant extra solution activity (2106.05(g) in step 2A prong 2) and well understood, routine and conventional activity of transmitting and receiving data as identified by the court (2106.05(d) in step 2B))
Regarding Claim 20:
The claim recites a system that performs the method as described in claim 1. Therefore, claim 20 is rejected for the same reasons as disclosed for claim 1. The limitations for additional elements of claim 20 are analyzed below.
Subject Matter Eligibility Analysis Step 2A Prong 1:
Please see Step 2A Prong 1 analysis of claim 1
Subject Matter Eligibility Analysis Step 2A Prong 2 & 2B:
“A federated learning apparatus, comprising a processor, wherein the processor is coupled to a memory, the memory is configured to store a program or instructions, and when the program or the instructions are executed by the processor, the apparatus is enabled to perform” (mere instructions to apply the exception using a generic computer component - see MPEP 2106.05(f))
Claim Rejections - 35 USC § 103
The following is a quotation of 35 U.S.C. 103 which forms the basis for all obviousness rejections set forth in this Office action:
A patent for a claimed invention may not be obtained, notwithstanding that the claimed invention is not identically disclosed as set forth in section 102, if the differences between the claimed invention and the prior art are such that the claimed invention as a whole would have been obvious before the effective filing date of the claimed invention to a person having ordinary skill in the art to which the claimed invention pertains. Patentability shall not be negated by the manner in which the invention was made.
Claims 1-8, 11-18 and 20 are rejected under 35 U.S.C. 103 as being unpatentable over Akdeniz (US20230068386A1) in view of Kumar (US20230297844A1).
Regarding claim 1, Akdeniz teaches:
“A federated learning method, applied to a first node, wherein the method comprises” (abstract, Federated machine learning training is performed using an apparatus and a plurality of clients of the edge computing network.)
“obtaining, by the first node, data distribution information of the plurality of second nodes based on a target data feature required by a training task, wherein data distribution information of any second node indicates a data class of service data that is both locally stored in the second node and that satisfies the target data feature” ([0205-0207, 0318, 0365-0370], The server (first node) may collect information regarding data distribution at each client (second nodes) based on statistics and the data distribution information is sent to the server by the client. The server identifies clients with that have independent and identically distributed (IID) data by determining the divergence parameters between the data distribution of the client and the target data at the server. The KL divergence may be computed to determine the distance between the client data distribution and the target distribution (satisfies the target data feature). The distribution of datasets is non-IID across edge computing nodes and the training data samples contains a label corresponding to the class. The training data samples is weighed based on expected computing delay, signal quality of raw data, etc. (service data).)
“selecting, by the first node, at least two target second nodes from the plurality of second nodes based on a target data ” ([0365-0372, 0384], The server may provide the target data distribution to the clients for client selection in each training round. The KL divergence is computed between the two distributions to determine similarities and differences in the data distribution. Each client stores a local dataset. The server may select K clients for each training iteration using a round robin technique. In the experiment, K may be 10 clients. Akdeniz does not explicitly disclose target data class, but the computation of the KL divergence implies that the data distributions include data points of one or more categories.)
“indicating, by the first node, the at least two target second nodes to perform federated learning, to obtain a federated learning model that is in the training task and that corresponds to the target data ” ([0122, 0127-0128, 0365-0372, 0384], The server may select clients based on data-distribution criteria and the selected clients receive the global model from the server to update the global model using the local data. The server receives the global model updates from the clients and aggregate the updates to generate a final global model.)
Akdeniz does not explicitly disclose an implementation of “the target data class”. However, Kumar discloses in the same field of endeavor:
“selecting, by the first node, at least two target second nodes from the plurality of second nodes based on a target data class ..., wherein the at least two target second nodes locally store target service data that satisfies the target data feature and that belongs to the target data class” ([0040, 0043-0046], The plurality of users has their own local training data for federated learning and each dataset may have different data distribution and different labels. The server may have a global user label set (target data class) that contains all the users’ labels and selects users having a subset of the labels for federated learning. Each user trains the local model using their local data that contains training samples having a subset of the global user label. The users send their computed probabilities to the global user for updating the global model.)
“indicating, by the first node, the at least two target second nodes to perform federated learning, to obtain a federated learning model that is in the training task and that corresponds to the target data class” ([0058-0060], In the example, 3 operators are selected to perform the federated learning. Each of the 3 user train their local model using their local dataset and provide the updates to the global user. Each user have data consisting of a subset of the public dataset.)
It would be obvious to one of the ordinary skill in the art before the effective filing date of the claimed invention to incorporate the teaching of “the target data class” from Kumar into the teaching of Akdeniz. Doing so can improve the federated learning training process by implementing a distributed learning process at a local computing device to train the model using datasets consisting of heterogeneous labels (Kumar, abstract).
Regarding claim 2, Akdeniz teaches:
“wherein at least one data analysis model is deployed in each second node, and each data analysis model corresponds to one data feature group ” ([0120, 0123, 0152, 0167], In federated learning, training may be performed using gradient descent algorithms and linear regression, which may be used for classification tasks. The central server sends the current model to the selected client nodes and the local nodes independently perform linear regression on its own data. Each local node contains a local dataset and data may be different for each local node. The deployed model performs the analysis corresponding to the features in the local dataset.)
“the obtaining, by the first node, data distribution information of the plurality of second nodes based on the target data feature required by the training task comprises” ([0318, 0365-0370], The server (first node) may collect information regarding data distribution at each client (second nodes) based on statistics and the data distribution information is sent to the server by the client. The server identifies clients with that have independent and identically distributed (IID) data by determining the divergence parameters between the data distribution of the client and the target data at the server.)
“sending, by the first node, a first query message to each of the plurality of second nodes based on the target data feature, wherein the first query message sent to each second node comprises an identifier of the target data feature and an identifier of a target data analysis model, and the target data analysis model corresponds to the target data feature” ([0264, 0269, 0280-0282, 0286, 0299, Fig. 17], The server sends a client capability request notification to all the clients in the federated learning process. The notification is requesting the compute rate, communication time, and number of training examples for each client nodes. The compute rate determines the time each local client completes the training of the local model. The number of training samples corresponds to the identifier of the target data features because each node may have a different data distribution and the compute rate is an identifier of a target data analysis model because the server may only select clients with the shortest completion times.)
“separately receiving, by the first node, the corresponding data distribution information from the plurality of second nodes, wherein data distribution information of each second node indicates an identifier of at least one data class and data information of service data that is stored in the second node and that separately belongs to the at least one data class” ([0287-0291], The server may receive the compute rate, communication time, and number of training examples from each client. The collection of the compute rate shows a distribution of the data across a plurality of client devices. The compute rate, communication time, and number of training examples are all data belonging its own separate category of information.)
Akdeniz does not explicitly disclose an implementation of “data analysis model corresponds to one data feature group and identifies a data class of service data that satisfies the corresponding data feature group”. However, Kumar discloses in the same field of endeavor:
“wherein at least one data analysis model is deployed in each second node, and each data analysis model corresponds to one data feature group and identifies a data class of service data that satisfies the corresponding data feature group” ([0040, 0043-0046, 0071], The plurality of users has their own local training data for federated learning and each dataset may have different data distribution and different labels. The local model may be a classifier-type model and may be used to identify a sub-set of the global labels from the local dataset.)
It would be obvious to one of the ordinary skill in the art before the effective filing date of the claimed invention to incorporate the teaching of “data analysis model corresponds to one data feature group and identifies a data class of service data that satisfies the corresponding data feature group” from Kumar into the teaching of Akdeniz. Doing so can improve the federated learning training process by implementing a distributed learning process at a local computing device to train the model using datasets consisting of heterogeneous labels (Kumar, abstract).
Regarding claim 3, Akdeniz in view of Kumar teaches:
“wherein the first query message sent by the first node to each second node further comprises an identifier of the target data class, and the data distribution information fed back by the second node comprises the identifier of the target data class and data information of the target service data that is stored in the second node and that belongs to the target data class” ([0318, 0365-0370], In some embodiment, the server may provide the target data distribution to the clients and request a computed distance from each client. The local clients may compute a KL divergence between the target distribution and local data distribution. The local data is stored locally on the clients and the computed distance determines the similarity of data distributions having different data classes. Kumar [0040, 0043-0046, 0071] may be combined to further teach the local model is a classifier-type model and may be used to identify a sub-set of the global labels from the local dataset.)
Regarding claim 4, Akdeniz teaches:
“sending, by the first node, a data analysis model deployment message to each of the plurality of second nodes, wherein the data analysis model deployment message sent to each second node comprises an identifier of the at least one data analysis model and a model file of the at least one data analysis model” ([0127-0130, 0312], The client compute nodes may receive commands from a central server to perform machine learning. The server may send the global model (model file) to each of the client computing nodes and the global model may also consists of model parameters (identifier of the at least one data analysis model). The clients train the model based on their local datasets and obtain updated model parameters.)
Regarding claim 5, Akdeniz in view of Kumar teaches:
“sending, by the first node, a model training message to each of the at least two target second nodes, wherein the model training message sent to each target second node comprises an identifier of a target artificial intelligence AI model, and the target AI model corresponds to the target data ” ([0127-0130, 0334, 0341, 0353-0360, Fig. 17], In federated learning systems, a server has a global model and each client nodes may train a local model using their local dataset. The global model is updated based on the aggregated values from all the clients. The target artificial intelligence AI model is interpreted to be the final global model and the data analysis model is the local model that is trained on local data. The server may select clients based on training loss, which needs to be initiated by training process in each client nodes. The global model is provided to the client nodes and in each training epoch, the global model parameters are provided to the client nodes. Kumar [0071] may be combined to further teach the local model is a classifier-type model and may be used to train a global model in federated learning process.)
“obtaining, by the first node based on updated AI models respectively received from the at least two target second nodes, the federated learning model that is in the training task and that corresponds to the target data ” ([0122, 0127-0128, 0365-0372, 0384], The server may select clients based on data-distribution criteria and the selected clients receive the global model from the server to update the global model using the local data. The server receives the global model updates from the clients and aggregate the updates to generate a final global model. Kumar [0071] may be combined to further teach the local model is a classifier-type model and may be used to train a global model in federated learning process.)
Regarding claim 6, Akdeniz in view of Kumar teaches:
“wherein the model training message sent to any target second node further comprises the identifier of the target data class and the identifier of the target data analysis model” ([Kumar, 0043-0045, 0071], The local users train their local models on their local dataset that contains a subset of the global labels set. The local model may be a classifier model and it is trained to identify the classes of data that is present in the local dataset.)
Regarding claim 7, Akdeniz in view of Kumar teaches:
“sending, by the first node, a model evaluation message to each of the at least two target second nodes, wherein the model evaluation message sent to each target second node comprises an identifier and an evaluation indicator of a target evaluation model, and the target evaluation model corresponds to the target data ” ([0341, 0352-0355, 0382-0385], The server may request information from the clients during the capability exchange phase such as training loss (evaluation indicator). The Specification (par. 350) of the claimed invention discloses that the evaluation model may be an updated AI model after the model training. It is interpreted that the evaluation model is a trained global model. Table TBR1 shows the test accuracy, which is the performance evaluation of the model. Kumar [0046] may be combined to further teach the local model is a classifier-type model and may be used to test their local models using a public dataset.
“separately receiving, by the first node, corresponding model evaluation results from the at least two target second nodes” ([0336], The server may receive the training loss from clients.)
Regarding claim 8, Akdeniz in view of Kumar teaches:
“wherein the model evaluation message sent to each target second node further comprises the identifier of the target data class and the identifier of the target data analysis model” ([Kumar, 0043-0046, 0071], The local users train their local models on their local dataset that contains a subset of the global labels set. The local model may be a classifier model and it is trained to identify the classes of data that is present in the local dataset. The trained local models are tested using the public dataset, which contains all labels.)
Regarding claim 11, Akdeniz teaches:
“A federated learning method, applied to a second node in a federated learning system, wherein the method comprises” (abstract, Federated machine learning training is performed using an apparatus and a plurality of clients of the edge computing network.)
“receiving, by the second node, a first query message from a first node, wherein the first query message indicates a target data feature required by a training task” ([0286, 0366-0367], The server send a target data distribution to the client nodes.)
“sending, by the second node, data distribution information to the first node based on the target data feature, wherein the data distribution information indicates a data class of service data that is locally stored in the second node and that satisfies the target data feature” ([0318, 0365-0370], The server (first node) may collect information regarding data distribution at each client (second nodes) based on statistics and the data distribution information is sent to the server by the client. The server identifies clients with that have independent and identically distributed (IID) data by determining the divergence parameters between the data distribution of the client and the target data at the server. The computation of the KL divergence implies that the data distributions include data points of one or more categories)
“receiving, by the second node and from the first node, an indication to train a target artificial intelligence (AI) model” ([0122, 0127-0128, 0365-0372, 0384], The server may select clients based on data-distribution criteria and the selected clients receive the global model from the server to update the global model using the local data during training iterations.)
“training, by the second node as indicated by the first node and by using stored target service data that belongs to a target data class, the target AI model corresponding to the target data class, to obtain an updated AI model” ([0122, 0127-0128, 0365-0372, 0384], The server may select clients based on data-distribution criteria and the selected clients receive the global model from the server to update the global model using the local data during training iterations. The local data may be different across different client nodes and the training of the local model provides updates to the server.)
“sending, by the second node, the updated AI model to the first node” ([0152], The client computing nodes send the update model to the central server.)
Regarding claim 12, the claim limitations are similar to claim 2 without providing additional features in the claims and thus are rejected on the same basis as claim 2. The claims recite sending and receiving of information between a first node and a second node and it would be obvious to one having ordinary skills in the arts to interpret claims 12 and 2 to be similar. The slight difference in language indicates which of the node is receiving or performing the sending operations.
Regarding claims 13, and 14, the claim limitations are similar to claims 3 and 4 without providing additional features in the claims and thus are rejected on the same basis as claims 3 and 4. The claims recite sending and receiving of information between a first node and a second node and it would be obvious to one having ordinary skills in the arts to interpret these claims to be similar.
Regarding claim 15, some claim limitations are similar to claim 5 and thus are rejected on the same basis as claim 5. The additional limitations are addressed by Akdeniz in view of Kumar:
“obtaining, by the second node based on the identifier of the AI model, stored target service data that satisfies the target data feature and that belongs to the target data class” ([Kumar, 0045], The local data in the local user environment may vary in each iteration and the local data may correspond with one of the subset of labels in a collection of global labels. The local user trains the local model using the local data consisting of a subset of labels.)
“training, by the second node, the AI model based on the target service data, to obtain the updated AI model” ([Akdeniz, 0122, 0127-0128, 0365-0372, 0384], The server may select clients based on data-distribution criteria and the selected clients receive the global model from the server to update the global model using the local data during training iterations. The local data may be different across different client nodes and the training of the local model provides updates to the server.)
Regarding claims 16, 17, and 18, the claim limitations are similar to claims 6, 7, and 8 without providing additional features in the claims and thus are rejected on the same basis as claims 6, 7, and 8. The claims recite sending and receiving of information between a first node and a second node and it would be obvious to one having ordinary skills in the arts to interpret these claims to be similar.
Regarding claim 20:
Claim 20 recites a system that performs the same process as described in Claim 1. Therefore claim 20 is rejected under the same reasons mention for claim 1. The additional elements of claim 20 is addressed below:
“A federated learning apparatus, comprising a processor, wherein the processor is coupled to a memory, the memory is configured to store a program or instructions, and when the program or the instructions are executed by the processor, the apparatus is enabled to perform” ([abstract, 0102-0103], The federated learning system may comprise of a processor and memory.)
Claims 9 and 19 are rejected under 35 U.S.C. 103 as being unpatentable over Akdeniz (US20230068386A1) in view of Kumar (US20230297844A1) and Sheller (US20190042937A1).
Regarding claim 9, Akdeniz in view of Kumar teaches:
“wherein the first node is part of a federated learning system, and the federated learning system is a wireless AI model–driven network system; the first node comprises a model management function (MMF) ” ([Akdeniz, 0053, 0063-0065, 0657-0660, 0751-0753, Figure 5], The federated learning system may consist of client endpoints in communication with the server over a wireless network. The server consists of the global model in a federated learning system and the server may aggregate the updates from all clients to update the global ML model. Akdeniz does not explicitly disclose a module to perform the operations of managing the global model on the server, but it is implied the central server consists of a processor to perform the operations of updating the global model. The client nodes have their own local dataset and it is implied the local data is stored in a database on the client device. The client nodes also include processor to perform the training of the local model on the local dataset. The training losses is computed and provided to the server, which implies that a model evaluation function is performed on the performance of the local model.)
“the sending, by the first node, a first query message to each second node comprises: sending, by the MMF ” ([Akdeniz, 0264, 0286], The server sends a client capability request notification to all the clients in the federated learning process.)
Akdeniz in view of Kumar does not explicitly disclose an implementation of “the first node comprises a model management function (MMF) module; and any second node comprises a model training function (MTF) module, a data management function (DMF) module, and a model evaluation function (MEF) module”. However, Sheller discloses in the same field of endeavor:
“...the first node comprises a model management function (MMF) module; and any second node comprises a model training function (MTF) module, a data management function (DMF) module, and a model evaluation function (MEF) module ...” ([0027-0028, 0038, 0040, Figure 2 & 3], In federated learning systems, the server consists of an aggregator, which receives the model updates from edge devices and update the global model. In each edge device, there is a local data store that manages local model information and updates the information. The edge device also consists of a neural network trainer component. The neural network processor may implement a model and determine the model updates which are provided to the aggregator.)
“the sending, by the first node, a first query message to each second node comprises: sending, by the MMF ” ([0037-0038], The model receiver obtains the model from the central server and stores it into the local model data store.)
It would be obvious to one of the ordinary skill in the art before the effective filing date of the claimed invention to incorporate the teaching of “the first node comprises a model management function (MMF) module; and any second node comprises a model training function (MTF) module, a data management function (DMF) module, and a model evaluation function (MEF) module” from Sheller into the teaching of Akdeniz in view of Kumar. Doing so can improve the federated learning training process by implementing communications between different components between the server and a trusted edge device (Sheller, abstract).
Regarding claim 19, the claim limitations are similar to claim 9 without providing additional features in the claims and thus are rejected on the same basis as claim 9. The claims recite sending and receiving of information between a first node and a second node and it would be obvious to one having ordinary skills in the arts to interpret claims 19 and 2 to be similar. The slight difference in language indicates which of the node is receiving or performing the sending operations.
Claim 10 is rejected under 35 U.S.C. 103 as being unpatentable over Akdeniz (US20230068386A1) in view of Kumar (US20230297844A1), and Dirac (US20150379424A1).
Akdeniz in view of Kumar does not explicitly disclose an implementation of “sending, by the first node, a mapping relationship table to each of the plurality of second nodes, wherein the mapping relationship table sent to each second node is used for recording a mapping relationship between an identifier of a data feature, an identifier of an AI model, an identifier of a data analysis model, and an identifier of a data class”. However, Dirac discloses in the same field of endeavor:
“sending, by the first node, a mapping relationship table to each of the plurality of second nodes, wherein the mapping relationship table sent to each second node is used for recording a mapping relationship between an identifier of a data feature, an identifier of an AI model, an identifier of a data analysis model, and an identifier of a data class” ([0032, 0035, 0056-0059, Figure 1], A MLS manager may interact with clients to process requests for a variety of ML tasks. Data source artifacts may be generated to identify certain data features and data type. Artifacts may also represent models and it may comprise of an identifier.)
It would be obvious to one of the ordinary skill in the art before the effective filing date of the claimed invention to incorporate the teaching of “sending, by the first node, a mapping relationship table to each of the plurality of second nodes, wherein the mapping relationship table sent to each second node is used for recording a mapping relationship between an identifier of a data feature, an identifier of an AI model, an identifier of a data analysis model, and an identifier of a data class” from Dirac into the teaching of Akdeniz in view of Kumar. Doing so can improve a machine learning service by implementing a correlation between data sources, models, and aliases for a variety of ML operations (Dirac, abstract).
Conclusion
Any inquiry concerning this communication or earlier communications from the examiner should be directed to GARY MAC whose telephone number is (703)756-1517. The examiner can normally be reached Monday - Friday 8:00 AM - 5:00 PM.
Examiner interviews are available via telephone, in-person, and video conferencing using a USPTO supplied web-based collaboration tool. To schedule an interview, applicant is encouraged to use the USPTO Automated Interview Request (AIR) at http://www.uspto.gov/interviewpractice.
If attempts to reach the examiner by telephone are unsuccessful, the examiner’s supervisor, Abdullah Kawsar can be reached at (571) 270-3169. 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.
/GARY MAC/Examiner, Art Unit 2127
/JEREMY L STANLEY/Examiner, Art Unit 2127