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 .
Response to Arguments
Claim Objections
Applicant's arguments filed June 11, 2026 regarding claim 8 objection have been fully considered, Examiner withdraws the objection based on the applicant correction.
35 USC § 101
Applicant's arguments filed June 11, 2026 with respect to the rejection of claims 4-6, 8-11, and 13 under 35 U.S.C. 101 have been fully considered but they are not persuasive.
Applicant argues that features of dependent claims 2 and 3, which were not rejected under 35 U.S.C. 101, have been incorporated into the independent claims from which rejected claims 4-6, 8-11, and 13 depend, and that the rejections should accordingly be withdrawn.
Examiner respectfully disagrees. As an initial matter, the absence of a rejection of previously presented claims 2 and 3 under 35 U.S.C. 101 does not constitute a determination that the limitations of those claims integrate a judicial exception into a practical application under Step 2A Prong Two or amount to significantly more than the judicial exception under Step 2B. Each claim is evaluated as a whole under the framework of MPEP 2106, and no finding of eligibility-conferring subject matter was made with respect to the limitations of claims 2 and 3.
Further, the amendment does not remove or amend the abstract ideas for which claims 4-6, 8-11, and 13 were rejected. Claims 4-6, 8-11, and 13 continue to recite the mental processes identified in the previous Office action, namely collecting and combining data from multiple sources (claim 4), selecting or identifying an item based on received information and performing an estimation (claims 5 and 6), determining a method of adjusting a model (claim 8), determining or evaluating validity by comparing output data with input data (claims 9-11), and determining an abnormality based on model output (claim 13). The eligibility of those claims therefore turns on whether the limitations newly incorporated into independent claim 1 constitute additional elements that integrate the recited mental processes into a practical application or amount to significantly more. They do not.
The limitation "acquire the training data based on a simulation result of the target using the simulator" (this limitation describes data collection/receiving, which is a well-understood, routine, conventional activity, see MPEP 2106.05(d)(II)(i))
The limitation "wherein the time-series data of the item that influences the target correspond to an input of a simulator, the time-series data of the measurement item correspond to an output of the simulator, and the simulator is configured to simulate an operation of the target" (the limitation describes the content and source of the data being collected, which merely specifies the type of data acquired; the recited simulator is invoked as a tool at a high level of generality, and the limitation is directed to mere instruction to apply the abstract idea on a generic computer to process, which is a well-understood, routine, conventional activity, see MPEP 2106.05(d)(II)(i))
The limitation "acquire the training data including data obtained by adding noise to data of the simulation result" (this limitation describes data collection/receiving and specifies the content of the data collected, which is a well-understood, routine, conventional activity, see MPEP 2106.05(d)(II)(i); the addition of noise to training data is a conventional data augmentation technique, as evidenced by Tremblay et al. of record, section 4.1, describing random Gaussian noise among classic augmentations applied during training)
The limitation "train, by using the training data, an inverse model of the simulator that receives an input of time-series data of the measurement item and outputs time-series data of the item that influences the target" (This limitation is directed to mere instruction to apply the abstract idea on a generic computer to process, which is a well-understood, routine, conventional activity, see MPEP 2106.05(d)(II)(i))
The recited simulator and inverse model are claimed at a high level of generality as generic computing tools for performing the data acquisition and model training. Applicant's specification describes the model as configurable using known machine learning techniques such as a recurrent neural network or a long short-term memory (see published paragraphs [0071]-[0072]), thereby describing the element at a high level of generality as conventional machine learning machinery, see MPEP 2106.05(d)(I). The claims do not reflect an improvement to the functioning of a computer or to any other technology or technical field, see MPEP 2106.05(a). The additional elements, considered individually and as an ordered combination with the additional elements previously identified, do not integrate the judicial exception into practical application and do not amount to significantly more than the Judicial exception.
Accordingly, the rejection of claims 4-6, 8-11, and 13 under 35 U.S.C. 101 is maintained.
35 USC § 103
Applicant's arguments filed June 11, 2026 with respect to the rejections of claims 1, 4-15 under 35 U.S.C. 103 have been fully considered but they are not persuasive. Applicant's arguments are directed to the limitations newly added by amendment, which are addressed by the updated grounds of rejection set forth in this Office action as necessitated by amendment.
Regarding claim 1, Applicant argues that Keeler's inputs to the inverse plant model, errors between desired and predicted outputs, and predicted control variables output from the inverse plant model are not indicative of the claimed training data wherein the time-series data of the item that influences the target correspond to an input of a simulator and the time-series data of the measurement item correspond to an output of the simulator.
Examiner respectfully disagrees, for three reasons.
First, Applicant argues the wrong disclosure. The features Applicant identifies the error input to inverse model 76 and the control variables it outputs are not what the rejection relies upon for the claimed training data. For the training data, the rejection relies on Keeler's training set of time-series pairs: "the pattern y(t) is provided as a time series output of a plant for a time series input x(t)" (Keeler, Col 7 L 55). The input x(t) includes the control variables that influence the plant, and y(t) is the measured plant output (Keeler, Col 5 L 6). These are the claimed time-series data of an item that influences the target and time-series data of a measurement item regarding the target.
Second, Keeler teaches the claimed simulator and the claimed input/output correspondence. Keeler's plant predictive model 74 "is deemed to be a relatively accurate model of the operation of the plant 72" (Keeler, FIG. 7a description); it receives the control variables as input and produces the plant's predicted output. A model of the operation of the target that receives the influencing items as input and produces the measurement items as output is, under the broadest reasonable interpretation, a simulator configured to simulate an operation of the target, with the influencing items corresponding to its input and the measurement items corresponding to its output, exactly the claimed correspondence. Keeler likewise teaches an inverse model of that simulator: "an inverse plant model 76 which is identical to the neural network representing the plant predictive model 74," operated by back propagation, which "is similar to an inversion of the network" (Keeler, Col 10 L 5), receiving output-side information and providing the control variables that influence the target.
Third, to the extent Applicant argues that Keeler does not teach training the inverse model on time-series data generated using the simulator, the rejection relies on Green for training a time-series model on simulation-generated training data, and one cannot show nonobviousness by attacking references individually where the rejection is based on a combination of references. Green teaches a trained LSTM performing sequence-to-sequence time-series prediction (Green [0057]) and teaches generating its training data with a simulator: "a 3D model of the tool and physics-based simulation may be used to generate synthetic training data" (Green [0067]), where the simulated conditions are the simulator's input and the resulting synthetic sensor data are its output. In the combination, Keeler's forward/inverse model arrangement trained in the manner taught by Green renders the claimed features obvious. Applicant's assertion that Green "does not remedy" the deficiencies of Keeler is a conclusory statement that does not address the combination as applied.
Regarding dependent claims 4-13, Applicant argues only that these claims are allowable by virtue of their dependencies and that Tremblay and Nakaya do not remedy the alleged deficiencies of Keeler and Green. Because the rejection of claim 1 is maintained for the reasons above, and because Applicant presents no separate arguments directed to the additional features of the dependent claims or to the teachings of Tremblay and Nakaya as applied, these arguments are not persuasive for the same reasons.
Regarding independent claims 14 and 15, Applicant relies on reasons similar to those presented for claim 1. These arguments are not persuasive for the reasons set forth above with respect to claim 1.
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 4-6, 8-11, and 13 rejected under 35 U.S.C. 101 because the claimed invention is directed to an abstract idea without significantly more.
MPEP 2106 (III) sets out steps for evaluating whether a claim is drawn to patent-eligible subject matter. The analysis of claims 4-6, 8-11, and 13, in accordance with these steps, follows.
Step 1 Analysis:
Claims 4-6, 8-11, and 13 are directed to an information processing device (machine). Therefore, claims 4-6, 8-11, and 13 fall into one of four statutory categories (i.e., machine).
As to claim 4 (which incorporates the limitations from claim 1),
Step 2A Prong 1: this claim recites the following abstract ideas:
"acquire the training data including actual measurement data of the measurement item in addition to data based on the simulation result." (the limitation describes collecting and combining data from multiple sources, which is a mental process implemented in the human mind)
Step 2A Prong 2 and 2B: the claim recited the following additional elements:
"a memory configured to store instructions; and" (This limitation is directed to mere instruction to store the abstract idea on a generic memory and apply the abstract idea on a generic computer to process, which is a well-understood, routine, conventional activity, see MPEP 2106.05(d)(II)(i))
"a processor configured to execute the instructions to:" (This limitation is directed to mere instruction to apply the abstract idea on a generic computer to process, which is a well-understood, routine, conventional activity, see MPEP 2106.05(d)(II)(i))
"acquire training data, the training data including time-series data of a measurement item regarding a target and time-series data of an item that influences the target" (this limitation describes data collection/receiving, which is a well-understood, routine, conventional activity, see MPEP 2106.05(d)(II)(i))
"wherein the time-series data of the item that influences the target correspond to an input of a simulator, the time-series data of the measurement item correspond to an output of the simulator, and the simulator is configured to simulate an operation of the target" (the limitation describes the content and source of the data being collected, which merely specifies the type of data acquired; the recited simulator is invoked as a generic computer tool, and the limitation is directed to mere instruction to apply the abstract idea on a generic computer to process, which is a well-understood, routine, conventional activity, see MPEP 2106.05(d)(II)(i))
"train, by using the training data, an inverse model of the simulator that receives an input of time-series data of the measurement item and outputs time-series data of the item that influences the target" (This limitation is directed to mere instruction to apply the abstract idea on a generic computer to process, which is a well-understood, routine, conventional activity, see MPEP 2106.05(d)(II)(i))
"acquire the training data based on a simulation result of the target using the simulator" (this limitation describes data collection/receiving, which is a well-understood, routine, conventional activity, see MPEP 2106.05(d)(II)(i))
"acquire the training data including data obtained by adding noise to data of the simulation result" (this limitation describes data collection/receiving and specifies the content of the data collected, which is a well-understood, routine, conventional activity, see MPEP 2106.05(d)(II)(i))
"wherein the processor is further configured to execute the instructions to" (This limitation is directed to mere instruction to apply the abstract idea on a generic computer to process, which is a well-understood, routine, conventional activity, see MPEP 2106.05(d)(II)(i))
The additional element does not integrate the judicial exception into practical application and does not amount to significantly more than the Judicial exception.
As to claim 5,
Step 2A Prong 1: this claim recites the following abstract ideas:
"receive designation of a non-estimation target item among items influencing the target" (the limitation describes selecting or identifying an item based on received information, which is a mental process implemented in the human mind)
"and estimates a value of an item that is not the non-estimation target." (the limitation describes performing an estimation, which is a mental process implemented in the human mind)
Step 2A Prong 2 and 2B: the claim recited the following additional elements:
"wherein the processor is further configured to execute the instructions to" (This limitation is directed to mere instruction to apply the abstract idea on a generic computer to process, which is a well-understood, routine, conventional activity, see MPEP 2106.05(d)(II)(i))
The additional elements incorporated from claim 1, from which this claim depends, are addressed above with respect to claim 4 and likewise are directed to generic computer implementation and data collection/receiving, which are well-understood, routine, conventional activities, see MPEP 2106.05(d)(II)(i).
The additional element does not integrate the judicial exception into practical application and does not amount to significantly more than the Judicial exception.
As to claim 6,
Step 2A Prong 1: this claim recites the following abstract ideas:
"estimate the value of the item that is not the non-estimation target" (the limitation describes performing an estimation, which is a mental process implemented in the human mind)
Step 2A Prong 2 and 2B: the claim recited the following additional elements:
"wherein the processor is further configured to execute the instructions to" (This limitation is directed to mere instruction to apply the abstract idea on a generic computer to process, which is a well-understood, routine, conventional activity, see MPEP 2106.05(d)(II)(i))
"using the inverse model that is adjusted so that the value of the non-estimation target item is a predetermined value for the item." (This limitation is directed to mere instruction to apply the abstract idea on a generic computer to process, which is a well-understood, routine, conventional activity, see MPEP 2106.05(d)(II)(i))
The additional element does not integrate the judicial exception into practical application and does not amount to significantly more than the Judicial exception.
As to claim 8,
Step 2A Prong 1: this claim recites the following abstract ideas:
"learn a method of adjusting the inverse model for each item that has a possibility of being designated as the non-estimation target item." (the limitation describes learning or determining a method for adjusting a model, which is a mental process implemented in the human mind)
Step 2A Prong 2 and 2B: the claim recited the following additional elements:
"wherein the processor is further configured to execute the instructions to" (This limitation is directed to mere instruction to apply the abstract idea on a generic computer to process, which is a well-understood, routine, conventional activity, see MPEP 2106.05(d)(II)(i))
The additional element does not integrate the judicial exception into practical application and does not amount to significantly more than the Judicial exception.
As to claim 9,
Step 2A Prong 1: this claim recites the following abstract ideas:
"determine validity of output data of the inverse model, based on input data to the inverse model." (the limitation describes determining or evaluating the results by comparing the output data with input data, which is a mental process implemented in the human mind)
Step 2A Prong 2 and 2B: the claim recited the following additional elements:
"wherein the processor is further configured to execute the instructions to" (This limitation is directed to mere instruction to apply the abstract idea on a generic computer to process, which is a well-understood, routine, conventional activity, see MPEP 2106.05(d)(II)(i))
The additional elements incorporated from claim 1, from which this claim depends, are addressed above with respect to claim 4 and likewise are directed to generic computer implementation and data collection/receiving, which are well-understood, routine, conventional activities, see MPEP 2106.05(d)(II)(i).
The additional element does not integrate the judicial exception into practical application and does not amount to significantly more than the Judicial exception.
As to claim 10,
Step 2A Prong 1: this claim recites the following abstract ideas:
"determine the validity of the output data of the inverse model, based on a consistency between an inference result obtained by qualitative inference using a qualitative expression of the output data of the inverse model, and the input data to the inverse model." (the limitation describes evaluating or determining inconsistency between inference results and input data, which is a mental process implemented in the human mind)
Step 2A Prong 2 and 2B: the claim recited the following additional elements:
"wherein the processor is further configured to execute the instructions to" (This limitation is directed to mere instruction to apply the abstract idea on a generic computer to process, which is a well-understood, routine, conventional activity, see MPEP 2106.05(d)(II)(i))
The additional element does not integrate the judicial exception into practical application and does not amount to significantly more than the Judicial exception.
As to claim 11,
Step 2A Prong 1: this claim recites the following abstract ideas:
"determine the validity of the output data of the inverse model by comparing the simulation result with the input data to the inverse model." (the limitation describes determining evaluation by comparison of results, which is a mental process implemented in the human mind)
Step 2A Prong 2 and 2B: the claim recited the following additional elements:
"wherein the processor is further configured to execute the instructions to" (This limitation is directed to mere instruction to apply the abstract idea on a generic computer to process, which is a well-understood, routine, conventional activity, see MPEP 2106.05(d)(II)(i))
The additional element does not integrate the judicial exception into practical application and does not amount to significantly more than the Judicial exception.
As to claim 13,
Step 2A Prong 1: this claim recites the following abstract ideas:
"estimate an abnormality related to the target, based on output data calculated by inputting actual measurement data of the measurement item into the inverse model." (the limitation describes determining an abnormality based on model output, which is a mental process implemented in the human mind)
Step 2A Prong 2 and 2B: the claim recited the following additional elements:
"wherein the processor is further configured to execute the instructions to" (This limitation is directed to mere instruction to apply the abstract idea on a generic computer to process, which is a well-understood, routine, conventional activity, see MPEP 2106.05(d)(II)(i))
The additional elements incorporated from claim 1, from which this claim depends, are addressed above with respect to claim 4 and likewise are directed to generic computer implementation and data collection/receiving, which are well-understood, routine, conventional activities, see MPEP 2106.05(d)(II)(i).
The additional element does not integrate the judicial exception into practical application and does not amount to significantly more than the Judicial exception.
Claim Rejections - 35 USC § 103
Claim(s) 1, 4-10, and 12-15 is/are rejected under 35 U.S.C. 103 as being unpatentable over Keeler et al. (US 6985781 B2) in view of Green et al. (US 20200276680 A1) and Tremblay et al. (Training Deep Networks with Synthetic Data: Bridging the Reality Gap by Domain Randomization, 23 April 2018).
As to claim 1 Keeler teaches an information processing device comprising:
acquire training data, the training data including time-series data of a measurement item regarding a target and time-series data of an item that influences the target, (see Keeler [Col 5 L 6] "there is illustrated a diagrammatic view of a predicted model 10 of a plant 12. The plant 12 is any type of physical, chemical, biological, electronic or economic process with inputs and outputs. ... The input of the model 10 is comprised of an input vector 20 of known plant inputs, which inputs comprise in part manipulated variables referred to as "control" variables, and in part measured or non-manipulated variables referred to as "state" variables. The control variables are the input to the plant 12. When the inputs are applied to the plant 12, an actual output results.", and see Keeler [Col 7 L 55] "Initially, the pattern y(t) is provided as a time series output of a plant for a time series input x(t). The first network, labelled "NET 1" is trained on the pattern y(t) as target values")
wherein the time-series data of the item that influences the target correspond to an input of a simulator, the time-series data of the measurement item correspond to an output of the simulator, and the simulator is configured to simulate an operation of the target; (see Keeler [Col 5 L 27] "By comparison, the output of the model 10 is a predicted output. To the extent that the model 10 is an accurate model, the actual output and the predicted output will be essentially identical.", and see Keeler [Col 9 L 61] "a plant predictive model 74 is developed with a neural network to accurately model the plant in accordance with the function f(c(t),s(t)) to provide an output o.sup.p(t), which represents the predicted output of plant predictive model 74. The inputs to the plant model 74 are the control inputs c(t) and the state variables s(t). For purposes of optimization/control, the plant model 74 is deemed to be a relatively accurate model of the operation of the plant 72.")
train, by using the training data, an inverse model of the simulator that receives an input of time-series data of the measurement item and (see Keeler [Col 10 L 5] "An error is generated between the desired and the predicted outputs and input to an inverse plant model 76 which is identical to the neural network representing the plant predictive model 74, with the exception that it is operated by back propagating the error through the original plant model with the weights of the predictive model frozen. This back propagation of the error through the network is similar to an inversion of the network with the output of the plant model 76 representing a .DELTA.c(t+1)", and see Keeler [Col 10 L67] "a conventional control network 83 is utilized that is trained on a given desired input for receiving the state variables and control variables and generating the control variables that are necessary to provide the desired outputs. ... the weights in the control network 83 of FIG. 7b are frozen and were learned by training the control network 83 on a given desired output.")
Keeler does not explicitly teach "a memory configured to store instructions; and", "a processor configured to execute the instructions to", "acquire the training data based on a simulation result of the target using the simulator; and", and "acquire the training data including data obtained by adding noise to data of the simulation result"
However, Green teaches
a memory configured to store instructions; and (see Green paragraph [0026] "The system may also include wireless connectivity, a memory storage device")
a processor configured to execute the instructions to: (see Green paragraph [0057] "a processor (Processing Device [114])")
outputs time-series data of the item that influences the target; (see Green paragraph [0041] "determines whether detected power tool motion is normal given specific sensor data, e.g., data from the power tool itself, and optionally, data from the user and/or from external sensors. The system measures multiple sensing modalities in three-dimensions to calculate the total kinematic motion of a power tool while in use. ... The system then runs inference on a pre-trained neural network to determine if the sensed motion from the power tool is desirable or undesirable for the particular power tool.", and see Green paragraph [0057] "a pre-trained LSTM running a sequence-to-sequence time-series prediction and classification")
acquire the training data based on a simulation result of the target using the simulator; and (see Green paragraph [0067] "In addition, a 3D model of the tool and physics-based simulation may be used to generate synthetic training data for normal use.", and see Green paragraph [0069] "Additionally, a 3D physics-based simulator may also be programmed to generate synthetic data of a more representative real-world abnormal motion event.")
It would have been obvious to a person of ordinary skill in the art before the effective filing date of the invention to modify the invention of Keeler to implement the models on a processor and memory and to train the time-series model using training data generated based on a simulation result of the target, as taught by Green, because a machine learning method trained on time series data "offers more flexibility in detecting kickback in a larger variety of operational scenarios" and "allows for continuous improvement of the detection algorithms as more and more data is collected for future training" (Green [0024]), and because simulation-generated training data allows data to be obtained for events for which real data cannot safely be collected, "[s]ince it is unsafe to request users to intentionally misuse a power tool" (Green [0068]).
Keeler as modified by Green does not explicitly teach "acquire the training data including data obtained by adding noise to data of the simulation result"
However, Tremblay teaches
acquire the training data including data obtained by adding noise to data of the simulation result. (see Tremblay section [4.1] "During training, we applied the following data augmentations: random brightness, random contrast, and random Gaussian noise. We also included more classic augmentations to our training process, such as random flips, random resizing, box jitter, and random crop.", and see Tremblay section [5] "We have demonstrated that domain randomization (DR) is an effective technique to bridge the reality gap. Using synthetic DR data alone, we have trained a neural network to accomplish complex tasks like object detection with performance comparable to more labor-intensive (and therefore more expensive) datasets.")
It would have been obvious to a person of ordinary skill in the art before the effective filing date of the invention to modify the invention of Keeler as modified by Green to acquire the training data including data obtained by adding noise to data of the simulation result, as taught by Tremblay, because adding noise and other randomized augmentations to synthetic training data is "an effective technique to bridge the reality gap," enabling a neural network trained on synthetic data alone to achieve "performance comparable to more labor-intensive (and therefore more expensive) datasets" (Tremblay section 5), thereby improving the trained model's performance on real measurement data while reducing the cost of training data collection.
2. (canceled)
3. (canceled)
As to claim 4, Keeler as modified by Green and Tremblay teaches the information processing device according to claim 1,
wherein the processor is further configured to execute the instructions to acquire the training data including actual measurement data of the measurement item in addition to data based on the simulation result. (see Green paragraph [0069] "As an example, a circular saw may be used to rip half-way down the length of a board, then pinched in a material using a clamp. The trigger is fastened in the on position such that when power is applied, the back edge of the blade immediately gains purchase on the board and causes the tool to jump from the board. Such a scenario is not 100% representative of a real-world event, but sufficient to train the neural network for the machine learning model")
As to claim 5, Keeler as modified by Green and Tremblay teaches the information processing device according to claim 1,
wherein the processor is further configured to execute the instructions to receive designation of a non-estimation target item among items influencing the target, and estimates a value of an item that is not the non-estimation target. (see Keeler [Col 5 L 7] "Referring now to FIG. 1, there is illustrated a diagrammatic view of a predicted model 10 of a plant 12. The plant 12 is any type of physical, chemical, biological, electronic or economic process with inputs and outputs.", and see Keeler [Col 14 L 39] "The output variables y(t) are functions of the control variables c(t), the measured state variables s(t) and the external influences E(t)", and see Keeler [Col 15 L 34] "useful information that is captured by the measured state variables, and that implicitly contains the external disturbances, is not discarded. Note that since the neural network learning state variable predictions can learn non-linear functions, this is a fully general non-linear projection to f(c(t)). Furthermore, by calculating the residuals, an excellent estimation of the external variations has been provided.")
As to claim 6, Keeler as modified by Green and Tremblay teaches the information processing device according to claim 5,
wherein the processor is further configured to execute the instructions to estimate the value of the item that is not the non-estimation target, using the inverse model that is adjusted so that the value of the non-estimation target item is a predetermined value for the item. (see Keeler [Col 6 L 33] "Referring now to FIG. 3.... The time series represents the actual output of a plant, which is referred to as y(t). As will be described in more detail hereinbelow, a first network is provided for making a first prediction, and then the difference between that prediction and the actual output y(t) is then determined to define a second time series representing the residual.", and see Keeler [Col 14 L 39] "The output variables y(t) are functions of the control variables c(t), the measured state variables s(t) and the external influences E(t)", and see Keeler [Col 15 L 34] "useful information that is captured by the measured state variables, and that implicitly contains the external disturbances, is not discarded. ... Furthermore, by calculating the residuals, an excellent estimation of the external variations has been provided.")
As to claim 7, Keeler as modified by Green and Tremblay teaches the information processing device according to claim 6,
wherein the inverse model is configured using a neural network, and
wherein the processor is further configured to execute the instructions to use the inverse model that is adjusted by performing at least any one of: switching an input value to one or more nodes of the neural network to a constant value; biasing an input to the one or more nodes of the neural network; rewriting a weight coefficient of one or more edges of the neural network; and switching part or all of the neural network to another neural network. (see Keeler [Co 13 L 5] "The residual states s.sup.r(t) in layer 102 are calculated after the weights in the network labelled NET 1 are frozen. This network is referred to as the "state prediction" net. The values in the residual layer 102 are referred to as the "residual activation" of the state variables.", and see Keeler [Col 13 L 18] "Referring now to FIG. 12, there is illustrated the next step in building the network, wherein the overall residual network is built.... until the output reaches a desired output or until a given number of BPA iterations has been achieved.")
As to claim 8, Keeler as modified by Green and Tremblay teaches the information processing device according to claim 6,
wherein the processor is further configured to execute the instructions to learn a method of adjusting the inverse model for each item that has a possibility of being designated as the non-estimation target item. (see Keeler [Col 13 L 5] "The residual states s.sup.r(t) in layer 102 are calculated after the weights in the network labelled NET 1 are frozen. This network is referred to as the "state prediction" net. The values in the residual layer 102 are referred to as the "residual activation" of the state variables. These residuals represent a good estimation of the external variables that affect the plant operation. This is important additional information for the network as a whole, and it is somewhat analogous to noise estimation in Weiner and Kahlman filtering, wherein the external perturbations can be viewed as noise and the residuals are the optimal (non-linear) estimate of this noise.")
As to claim 9, Keeler as modified by Green and Tremblay teaches the information processing device according to claim 1,
wherein the processor is further configured to execute the instructions to determine validity of output data of the inverse model, based on input data to the inverse model. (see Keeler [Col 5 L 20] "The input of the model 10 is comprised of an input vector 20 of known plant inputs, which inputs comprise in part manipulated variables referred to as "control" variables, and in part measured or non-manipulated variables referred to as "state" variables. The control variables are the input to the plant 12. When the inputs are applied to the plant 12, an actual output results. By comparison, the output of the model 10 is a predicted output. To the extent that the model 10 is an accurate model, the actual output and the predicted output will be essentially identical. However, whenever the actual output is to be varied to a set point, the plant control inputs must be varied.")
As to claim 10, Keeler as modified by Green and Tremblay teaches the information processing device according to claim 9,
wherein the processor is further configured to execute the instructions to determine the validity of the output data of the inverse model, based on a consistency between an inference result obtained by qualitative inference using a qualitative expression of the output data of the inverse model, and the input data to the inverse model. (see Green paragraph [0070] "The time-series data representing abnormal motion events are labeled and manually annotated with a timestamp of the exact event. For example, a time-series that includes a kickback event may be 250 ms with the actual kickback event occurring 150 ms into the time series. The entire time-series data of 250 ms can be labeled as representing abnormal motion with the actual event at 150 ms being labeled as the kickback event. This labelling process is done by manually inspecting the data and determining when the event occurred.", and see Green paragraph [0089] "The system then compares extracted time series data with known, labelled data from power tools to determine whether the sensed motion is normal or abnormal for the given power tool and user. The machine learning model outputs a decimal value between 0 and 1, representing a probability of abnormal motion.")
As to claim 12, Keeler as modified by Green and Tremblay teaches the information processing device according to claim 1,
wherein the processor is further configured to execute the instructions to update the inverse model, using actual measurement data of the measurement item. (see Green paragraph [0080] "Neural network models can then be continually refined by retraining on additional real-world data as it is captured from the field and added to the repository. For example, anonymous time series data collected through use of the power tool can be transmitted using the mobile device running the mobile application for power tool safety to the central repository or directly to a remote deep-learning machine learning model, e.g., a cloud-based deep-learning model, and used to improve detection of future hazardous conditions. New versions of the neural network can be deployed to the system 100 in various means")
As to claim 13, Keeler as modified by Green and Tremblay teaches the information processing device according to claim 1,
wherein the processor is further configured to execute the instructions to estimate an abnormality related to the target, based on output data calculated by inputting actual measurement data of the measurement item into the inverse model. (see Green paragraph [0060] "sensor data is collected in real-time and processed by processing device 114, e.g., a microcontroller. The processing device 114 can include one or more of the following components: a memory 110 for storing sensor data, an environment sensor 106d for collecting information about the workspace 110 environment, wireless connectivity 117, a general purpose input/output 118, a battery 119, a battery charger 120, and a machine learning model 112 that is trained to determine abnormal motion of a power tool.")
As to claim 14 is directed to a method that corresponds to the device of claim 1. See the rejection for claim 1 above, which also applies for claim 14.
As to claim 15 is directed to a computer-program embodiment that corresponds to device claim 1. See the rejection for claim 1 above, which also applies for claim 15.
Claim(s) 11 is/are rejected under 35 U.S.C. 103 as being unpatentable over Keeler et al. (US 6985781 B2) in view of Green et al. (US 20200276680 A1) and Tremblay et al. (Training Deep Networks with Synthetic Data: Bridging the Reality Gap by Domain Randomization, 23 April 2018), and further in view of Nakaya et al. (US 20050240382 A1).
As to claim 11, Keeler as modified by Green and Tremblay teaches the information processing device according to claim 9,
Keeler as modified by Green and Tremblay does not explicitly teach "wherein the processor is further configured to execute the instructions to determine the validity of the output data of the inverse model by comparing the simulation result with the input data to the inverse model"
However, Nakaya teaches
wherein the processor is further configured to execute the instructions to determine the validity of the output data of the inverse model by comparing the simulation result with the input data to the inverse model. (see Nakaya paragraph [0040] "plant diagnosis means 110 checks consistency between data calculated by process simulation means 102 and the actual data obtained from actual plant 30 and, if the difference between the two exceeds a predetermined tolerance, displays the difference as a plant abnormality on display means 107.")
It would have been obvious to a person of ordinary skill in the art before the effective filing date of the invention to modify the invention of Keeler as modified by Green and Tremblay to determine the validity of the output data of the model by comparing the simulation result with the input data to the model, as taught by Nakaya, in order to reflect the status of the actual plant in the simulation model consecutively "so that operations of the actual plant can be predicted in a highly accurate manner" (Nakaya [0066]) and to enable "early discovery of plant abnormalities" (Nakaya [0044]).
Conclusion
THIS ACTION IS MADE FINAL. Applicant is reminded of the extension of time policy as set forth in 37 CFR 1.136(a).
A shortened statutory period for reply to this final action is set to expire THREE MONTHS from the mailing date of this action. In the event a first reply is filed within TWO MONTHS of the mailing date of this final action and the advisory action is not mailed until after the end of the THREE-MONTH shortened statutory period, then the shortened statutory period will expire on the date the advisory action is mailed, and any nonprovisional extension fee (37 CFR 1.17(a)) pursuant to 37 CFR 1.136(a) will be calculated from the mailing date of the advisory action. In no event, however, will the statutory period for reply expire later than SIX MONTHS from the mailing date of this final action.
Any inquiry concerning this communication or earlier communications from the examiner should be directed to ABDULLAH K ABOUD whose telephone number is (571)272-0025. The examiner can normally be reached Mon-Fri 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, Li B Zhen, can be reached at (571) 272-3768. 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.
/ABDULLAH KHALED ABOUD/ Examiner, Art Unit 2121
/Li B. Zhen/ Supervisory Patent Examiner, Art Unit 2121
Prosecution Insights