Notice of Pre-AIA or AIA Status
The present application, filed on or after March 16, 2013, is being examined under the first inventor to file provisions of the AIA .
Information Disclosure Statement
The information disclosure statements (IDS) submitted on 12/23/2023 was filed before the mailing date of the first office action. The submission is in compliance with the provisions of 37 CFR 1.97. Accordingly, the information disclosure statement is being considered by the examiner.
Specification
The title of the invention is not descriptive. A new title is required that is clearly indicative of the invention to which the claims are directed.
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 1-12 are rejected under 35 U.S.C. 101. Claims 1-10 are directed to a system, claim 11 is directed to a method, and claim 12 is directed to a non-transitory computer-readable medium; therefore, claims 1-12 fall within one of the four statutory categories (i.e., process, machine, manufacture, or composition of matter). However, claims 1-12 fall within the judicial exception of an abstract idea, specifically the abstract ideas of “Mental Processes” (including observation, evaluation, and opinion) and “Mathematical Concepts (including mathematical calculations and relationships)”.
Claim 1:
Claim 1 is directed to a system; therefore, the claim does fall within one of the four statutory categories (i.e., process, machine, manufacture, or composition of matter).
Claim 1 recites the following abstract ideas:
Step 2A Prong 1:
divide parameter data related to a network model of a neural network into a plurality of blocks (mental step directed to observation, evaluation – a person could divide observed parameter data related to a neural network model into a plurality of blocks in their mind, potentially assisted by pen and paper (see MPEP 2106.04(a)(2)(III)),
and generate embedded information based on connection information for each block obtained by converting, for the respective blocks, data in each block including the parameter data and connection information for another block (mental step directed to observation, evaluation – a person could convert observed or mentally determined blocks comprising observed parameter data and observed connection information for other blocks in their mind, potentially assisted by pen and paper, to generate embedded information);
Claim 1 recites the following additional elements:
an information processing apparatus comprising: one or more processors; and
one or more memories that store a computer-readable instruction configured to be executed by the one or more processors; and perform output setting for using data with the generated embedded information embedded in an output from the network model, as an output of the network model.
Step 2A Prong 2:
The information processing apparatus comprising one or more processors and memories are interpreted as aspects of the technological environment used to merely apply the aforementioned abstract ideas. Performing output setting for using data with the generated information embedded in an output from the neural network is interpreted as insignificant extra-solution activity directed to mere data gathering. These additional elements, when considered as a whole with the aforementioned abstract ideas, do not integrate those abstract ideas into a practical application (see MPEP 2106.05(f) and MPEP 2106.05(g)).
Step 2B:
The information processing apparatus comprising one or more processors and memories are interpreted as generic computer components used to merely apply the aforementioned abstract ideas. Performing output setting for using data with the generated information embedded in an output from the neural network is interpreted as well-understood, routine conventional activity directed to transmitting the output data from the neural network model over a network. These additional elements, when considered as a whole with the aforementioned abstract ideas, do not amount to significantly more than those abstract idea (see MPEP 2106.05(d)(II) and MPEP 2106.05(f)).
Claim 2 recites wherein the parameter is a weight coefficient in the network model and configuration information indicating a configuration of the network model, or the weight coefficient in the network model.
This limitation is interpreted as an additional element directed to a further description of the kind of parameter being divided in the field of use or technological environment of claim 1. This additional element, when considered as a whole with the aforementioned abstract ideas, does not integrate those abstract ideas into a practical application or amount to significantly more than those abstract idea (see MPEP 2106.05(h)).
Claim 3 recites wherein the connection information is a hash value obtained by converting data in each block.
This limitation is interpreted as an additional element directed to a further description of the kind of connection information being generated in the field of use or technological environment of claim 1. This additional element, when considered as a whole with the aforementioned abstract ideas, does not integrate those abstract ideas into a practical application or amount to significantly more than those abstract idea (see MPEP 2106.05(h))
Claim 4 recites wherein the one or more processors convert, for the respective blocks, data in each block including the parameter data and connection information for another block into the connection information.
Converting data in each block into the connection information for another block including parameter data and connection information for the respective block is interpreted as an abstract idea directed to a mental step of observation, evaluation – a person could convert observed parameter data and connection information from a given block into connection information for another block in their mind, potentially assisted by pen and paper.
Wherein the one or more processors perform this conversion step is interpreted as an additional element directed to mere instructions to apply the aforementioned abstract idea with generic computer components. This additional element, when considered as a whole with the aforementioned abstract ideas, does not integrate those abstract ideas into a practical application or amount to significantly more than those abstract idea (see MPEP 2106.05(f)).
Claim 5 recites wherein the one or more processors convert, for the respective blocks, data in each block including the parameter data and connection information for a preceding block into the connection information to blockchain the plurality of blocks.
Converting data in each block into the connection information for a preceding block including parameter data and connection information for the respective block is interpreted as an abstract idea directed to a mental step of observation, evaluation – a person could convert observed parameter data and connection information from a given block into connection information for a preceding block in their mind, potentially assisted by pen and paper.
Wherein the one or more processors perform this conversion step is interpreted as an additional element directed to mere instructions to apply the aforementioned abstract idea with generic computer components. This additional element, when considered as a whole with the aforementioned abstract ideas, does not integrate those abstract ideas into a practical application or amount to significantly more than those abstract idea (see MPEP 2106.05(f)).
Claim 6 recites wherein the one or more processors blockchain blocks corresponding to a plurality of network models divided into a plurality of blocks, across the plurality of the network models.
The one or more processors are interpreted as an additional element directed to an aspect of the technological environment in which the abstract ideas identified in claim 5 are performed and as well-understood, routine, conventional activity directed to generic computer components. Blockchain-ing blocks corresponding to a plurality of network models divided into a plurality of blocks is interpreted as an additional element directed to an aspect of the technological environment in which the abstract ideas identified in claim 5 are performed, and as well-understood, routine, conventional activity directed to saving information in memory. This additional element, when considered as a whole with the aforementioned abstract ideas, does not integrate those abstract ideas into a practical application or amount to significantly more than those abstract idea (see MPEP 2106.05(h) and MPEP 2106.05(d)(II)).
Claim 7 recites wherein the network model outputs a classification result, and wherein the output setting is for use of data with the embedded information embedded in an output indicating the classification result from the network model, as an output of the network model.
Wherein the network model outputs a classification result is interpreted as an additional element directed to mere data gathering and transmitting data over a network. This additional element, when considered as a whole with the aforementioned abstract ideas, does not integrate those abstract ideas into a practical application or amount to significantly more than those abstract idea (see MPEP 2106.05(d)(II) and MPEP 2106.05(g)).
Wherein the output setting is “for use” of data with the information embedded as an output of the neural network is interpreted as the intended use of this neural network output indicating a classification result and does not provide additional patentable weight to this claim (see MPEP 2103).
Claim 8 recites wherein the output setting is for embedding of the embedded information in a softmax output or a logits output from the network model.
Examiner notes that the broadest reasonable interpretation of this limitation includes interpreting that wherein the output setting “is for” embedding information in a softmax output or logits output is merely the intended use or necessary outcome of outputting information from the network model (see MPEP 2103). In this interpretation, no additional patentable weight would be given to this limitation. However; for purposes of compact prosecution, Examiner notes that the broadest reasonable interpretation of performing output setting for information embedded in an softmax or logits output from the neural network could also be interpreted as insignificant extra-solution activity directed to mere data gathering and as well-understood, routine conventional activity directed to transmitting the output data from the neural network model over a network. This additional element, when considered as a whole with the aforementioned abstract ideas, does not integrate those abstract ideas into a practical application or amount to significantly more than those abstract idea (see MPEP 2106.05(d)(II) and MPEP 2106.05(g)).
Claim 9 recites wherein the network model outputs a feature quantity, and wherein the output setting is for use of data with the embedded information embedded in a feature quantity output from the network model, as an output of the network model.
Wherein the network model outputs a feature quantity is interpreted as an additional element directed to insignificant extra-solution activity directed to mere data gathering and as well-understood, routine, conventional activity directed to transmitting data over a network. This additional element, when considered as a whole with the aforementioned abstract ideas, does not integrate those abstract ideas into a practical application or amount to significantly more than those abstract idea (see MPEP 2106.05(d)(II) and MPEP 2106.05(g)).
Wherein the output setting is “for use” of data with the information embedded in a feature quantity output of the neural network is interpreted as the intended use of this neural network output and does not provide additional patentable weight to this claim (see MPEP 2103).
Claim 10 recites wherein the embedded information is generated by repetitively arranging the connection information for each block in a predetermined order.
Generating embedded information by repetitively arranging the connection information in each block in a predetermined order is interpreted as an abstract idea directed to a mental step of observation, evaluation – a person could repetitively arrange observed or mentally determined connection information in each block in a predetermined order in their mind to generate embedded information.
Claim 11 is a method claim and its limitation is included in claim 1. The only difference is that claim 11 requires a method. Therefore, claim 11 is rejected for the same reasons as claim 1.
Claim 12 is a non-transitory computer-readable medium claim and its limitation is included in claim 1. The only difference is that claim 12 requires a non-transitory computer-readable medium, which is interpreted as a generic computer component used to merely implement the abstract ideas as identified in the analysis of claim 1 (see MPEP 2106.05(f)). Therefore, claim 12 is rejected for the same reasons as claim 1.
Viewed as a whole, these additional claim elements do not provide meaningful limitations to transform the abstract idea into a patent eligible application of the abstract idea such that the claims amount to significantly more than the abstract idea itself. Therefore, the claims are rejected under 35 U.S.C. 101.
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 1-12 are rejected under 35 U.S.C. 103 as being unpatentable over Goel et al* (“DeepRing: Protecting Deep Neural Network with Blockchain”, herein Goel) in view of Zhang et al (US 20220138550 A1, herein Zhang).
*a copy of this document was provided with the IDS dated 12/12/2023
Regarding claim 1, Goel teaches dividing parameter data related to a network model of a neural network into a plurality of blocks (fig. 3-4 and section 3.2 para. 2 recite “Figure 3 represents a block of the DeepRing. It comprises of the hashes of the current and previous block, public and private keys of the current layer, public keys of the layers appearing immediately before and after the current layer and AES key and model parameters of the current and the next layer” (i.e., model parameters can be stored in a block like the one shown in fig. 3, and divided up into a plurality of blocks as shown in fig. 4)),
and generating embedded information based on connection information for each block obtained by converting, for the respective blocks, data in each block including the parameter data and connection information for another block (section 3.2 para. 2-3 recite “Just like blocks in a blockchain, blocks in DeepRing have a shared common ledger which stores the state of the model. The hash associated with a block is a function of the hash of the previous block and the parameters of the current and next layer parameters. Hash of a block j which corresponds to layer i in the DNN architecture is given by: (EQ1). Here, ϕ is any suitable hash function such as SHA256” (i.e., generating hash values, or embedded connection information, by converting hash values between blocks in the blockchain));
and performing output setting for using data with the generated embedded information embedded in an output from the network model, as an output of the network model (section 3.2 para. 1 recites “blocks of DeepRing serve the following purposes: store the parameters of the layer; compute the output of the layer; update the ledger after output computation; and validate the output of the next layer”. Section 3.4.2 para. 1 recites “After processing the input to a block, each block updates the ledger with the following four items: layer output encrypted by its AES key; AES key encrypted by the public key of the next layer; signature of the layer; and hash of the output of the next layer” (i.e., outputting the hash values, or generated embedded information, as an output of the neural network model)).
However, while one of ordinary skill in the art would recognize that the methods from Goel would be implemented on a computing apparatus, Goel does not explicitly teach an information processing apparatus comprising: one or more processors; and one or more memories that store a computer-readable instruction configured to be executed by the one or more processors.
Zhang teaches an information processing apparatus comprising: one or more processors; and one or more memories that store a computer-readable instruction configured to be executed by the one or more processors, thereby the computer-readable instruction causing the information processing apparatus to (fig. 6A and at least para. [0089] teach an information processing apparatus comprising one or more processors and memories. At least para. [0153] describes the computer system, or information processing apparatus, in the context of executable instructions).
Therefore, it would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to combine these teachings by implementing the method of training a neural network model on a blockchain from Goel using the information processing apparatus from Zhang. Goel and Zhang are both directed to methods of training a neural network model on a blockchain, and at least fig. 2 of Goel and para. [0038] of Zhang describe the benefits of training an AI model on a blockchain. One of ordinary skill in the art would recognize that the known method from Goel could be implemented using the known apparatus from Zhang to perform a similar neural network training.
Regarding claim 2, the combination of Goel and Zhang teaches the information processing apparatus according to claim 1 as mentioned above, wherein the parameter is a weight coefficient in the network model and configuration information indicating a configuration of the network model, or the weight coefficient in the network model (Goel section 3.1 describes the weight parameters for an ith and i+1th layer in the network model. Goel section 4.1 para. 2 recites “Keeping the trained parameters of the network fixed, we associate each layer with a logistic importance parameter p, and train the network again to compute the importance of each layer weights” (i.e., a parameter is a neural network weight used to determine the configuration of the neural network model)).
Regarding claim 3, the combination of Goel and Zhang teaches the information processing apparatus according to claim 1 as mentioned above, wherein the connection information is a hash value obtained by converting data in each block (Goel section 3.2 para. 2 recites “Figure 3 represents a block of the DeepRing. It comprises of the hashes of the current and previous block, public and private keys of the current layer, public keys of the layers appearing immediately before and after the current layer and AES key and model parameters of the current and the next layer. Just like blocks in a blockchain, blocks in DeepRing have a shared common ledger which stores the state of the model. The hash associated with a block is a function of the hash of the previous block and the parameters of the current and next layer parameters. Hash of a block j which corresponds to layer i in the DNN architecture is given by: (EQ1)” (i.e., storing a hash value, or connection information, in each block of the blockchain)).
Regarding claim 4, the combination of Goel and Zhang teaches the information processing apparatus according to claim 1 as mentioned above, wherein the one or more processors convert, for the respective blocks, data in each block including the parameter data and connection information for another block into the connection information (Goel fig. 4 and section 3.2 para. 1 recite “Figure 4 shows the transformation of a model from DNN architecture to a DeepRing architecture. The architecture of DeepRing is inspired by that of a blockchain”. Goel fig. 3 and section 3.2 para. 2 recite “Figure 3 represents a block of the DeepRing. It comprises of the hashes of the current and previous block, public and private keys of the current layer, public keys of the layers appearing immediately before and after the current layer and AES key and model parameters of the current and the next layer. Just like blocks in a blockchain, blocks in DeepRing have a shared common ledger which stores the state of the model. The hash associated with a block is a function of the hash of the previous block and the parameters of the current and next layer parameters” (i.e., parameter and hash values, or connection information, for each block comprises parameter and hash values, or connection information, for at least another block in the blockchain)).
Regarding claim 5, the combination of Goel and Zhang teaches the information processing apparatus according to claim 4 as mentioned above, wherein the one or more processors convert, for the respective blocks, data in each block including the parameter data and connection information for a preceding block into the connection information to blockchain the plurality of blocks (Goel fig. 4 and section 3.2 para. 1 recite “Figure 4 shows the transformation of a model from DNN architecture to a DeepRing architecture. The architecture of DeepRing is inspired by that of a blockchain”. Goel fig. 3 and section 3.2 para. 2 recite “Figure 3 represents a block of the DeepRing. It comprises of the hashes of the current and previous block, public and private keys of the current layer, public keys of the layers appearing immediately before and after the current layer and AES key and model parameters of the current and the next layer. Just like blocks in a blockchain, blocks in DeepRing have a shared common ledger which stores the state of the model. The hash associated with a block is a function of the hash of the previous block and the parameters of the current and next layer parameters” (i.e., parameter and hash values, or connection information, for each block comprises parameter and hash values, or connection information, for at least a preceding block in the blockchain)).
Regarding claim 6, the combination of Goel and Zhang teaches the information processing apparatus according to claim 5 as mentioned above, wherein the one or more processors blockchain blocks corresponding to a plurality of network models divided into a plurality of blocks, across the plurality of the network models (Goel section 3.2 para. 1 recites “Figure 4 shows the transformation of a model from DNN architecture to a DeepRing architecture. The architecture of DeepRing is inspired by that of a blockchain. A blockchain is a linked list of ever-growing blocks with transaction records. DeepRing, on the other hand, is a closed chain of a finite number of blocks. Each block (except the ouroboros block) represents a layer of the deep neural network” (i.e., blocks corresponding to the model layers, or sub-models, of the network model stored in the blockchain. Examiner notes that at least para. [0049] of Zhang discusses how layers of a neural network can comprise sub-models)).
Regarding claim 7, the combination of Goel and Zhang teaches the information processing apparatus according to claim 1 as mentioned above, wherein the network model outputs a classification result (at least tables 2-3 in Goel section 4.2 show the network model utilized for an object detection, or classification task), and wherein the output setting is for use of data with the embedded information embedded in an output indicating the classification result from the network model, as an output of the network model (Examiner notes that wherein the output setting is “for use” of data with the information embedded in a output of the neural network indicating a classification result is interpreted as the intended use of this neural network output and does not provide additional patentable weight to this claim (see MPEP 2103)).
Regarding claim 8, the combination of Goel and Zhang teaches the information processing apparatus according to claim 7 as mentioned above, wherein the output setting is for embedding of the embedded information in a softmax output or a logits output from the network model (Examiner notes that the broadest reasonable interpretation of this limitation includes interpreting that wherein the output setting “is for” embedding information in a softmax output or logits output is merely the intended use or necessary outcome of outputting information from the network model (see MPEP 2103). In this interpretation, no additional patentable weight would be given to this limitation. However; for purposes of compact prosecution, Examiner notes that at least section 5 para. 1 of Goel teaches using a softmax activation at the output layer).
Regarding claim 9, the combination of Goel and Zhang teaches the information processing apparatus according to claim 1 as mentioned above, wherein the network model outputs a feature quantity (Zhang para. [0105] recites “the training information 752 may include the results of training a sub-model that are stored by a blockchain peer. For example, the results may include the change to the parameters of the AI model's algorithm as a result of the training data and the sub-model being executed”. Zhang para. [0110] recites “FIG. 7C, the blockchain 770 includes a number of blocks 7781, 7782, . . ., 778N cryptographically linked in an ordered sequence, where N ≥ 1. The blocks 7781, 7782, . . ., 778N are subject to a hash function which produces n-bit alphanumeric outputs (where n is 256 or another number) from inputs that are based on information in the blocks. Examples of such a hash function include, but are not limited to, a SHA-type (SHA stands for Secured Hash Algorithm) algorithm” (i.e., the results of the model training, or output, can be stored used a feature hashing algorithm such as SHA)), and wherein the output setting is for use of data with the embedded information embedded in a feature quantity output from the network model, as an output of the network model (Examiner notes that wherein the output setting is “for use” of data with the information embedded in a feature quantity output of the neural network is interpreted as the intended use of this neural network output and does not provide additional patentable weight to this claim (see MPEP 2103)).
Regarding claim 10, the combination of Goel and Zhang teaches the information processing apparatus according to claim 9 as mentioned above, wherein the embedded information is generated by repetitively arranging the connection information for each block in a predetermined order (Zhang para. [0005] recites “FIG. 1 is a diagram illustrating a blockchain network for iterative training of an artificial intelligence model according to example embodiments”. Zhang para. [0032] recites “the peers may execute a consensus protocol to validate block chain storage transactions, group the storage transactions into blocks, and build a hash chain over the blocks. This process forms the ledger by ordering the storage transactions, as is necessary, for consistency”. Zhang para. [0033] recites “After validation, the transactions enter an ordering phase in which a consensus protocol is used to produce an ordered sequence of endorsed transactions grouped into blocks” (i.e., ordering, or arranging the transactions, or connection information for each block, in a predetermined order in the repetitive, or iterative training method as shown in at least fig. 1 of Zhang)).
Claim 11 is a method claim and its limitation is included in claim 1. The only difference is that claim 11 requires a method (Zhang para. [0003] teaches a method). Therefore, claim 11 is rejected for the same reasons as claim 1.
Claim 12 is a non-transitory computer-readable medium claim and its limitation is included in claim 1. The only difference is that claim 12 requires a non-transitory computer-readable medium (Zhang para. [0004] teaches a non-transitory computer-readable medium). Therefore, claim 12 is rejected for the same reasons as claim 1.
Conclusion
The prior art made of record and not relied upon is considered pertinent to applicant's disclosure.
US 20010020837 A1 (Yamashita et al) teaches a neural network based face detection system based on extracted feature quantities from a recognized image.
US 20180232639 A1 (Lin et al) teaches a method for training a neural network and storing parameters of the neural network in a blockchain.
US 20240048703 A1 (Beye et al) teaches a neural network encoder-decoder model for processing images such that a feature quantity can be the output of the model.
Any inquiry concerning this communication or earlier communications from the examiner should be directed to LEAH M FEITL whose telephone number is (571) 272-8350. The examiner can normally be reached on M-F 0900-1700 EST.
Examiner interviews are available via telephone, in-person, and video conferencing using a USPTO supplied web-based collaboration tool. To schedule an interview, applicant is encouraged to use the USPTO Automated Interview Request (AIR) at http://www.uspto.gov/interviewpractice.
If attempts to reach the examiner by telephone are unsuccessful, the examiner’s supervisor, Viker Lamardo can be reached on (571) 270-5871. 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.
/L.M.F./ Examiner, Art Unit 2147
/VIKER A LAMARDO/Supervisory Patent Examiner, Art Unit 2147
Prosecution Insights