DETAILED ACTION
Notice of Pre-AIA or AIA Status
The present application, filed on or after March 16, 2013, is being examined under the first inventor to file provisions of the AIA .
Amendment filed on 02/09/2026 has been entered. Claims 1-20 are pending. Claims 1, 5 and 15 have been amended.
Claim Rejections - 35 USC § 102
In the event the determination of the status of the application as subject to AIA 35 U.S.C. 102 and 103 (or as subject to pre-AIA 35 U.S.C. 102 and 103) is incorrect, any correction of the statutory basis (i.e., changing from AIA to pre-AIA ) for the rejection will not be considered a new ground of rejection if the prior art relied upon, and the rationale supporting the rejection, would be the same under either status.
The following is a quotation of the appropriate paragraphs of 35 U.S.C. 102 that form the basis for the rejections under this section made in this Office action:
A person shall be entitled to a patent unless –
(a)(1) the claimed invention was patented, described in a printed publication, or in public use, on sale, or otherwise available to the public before the effective filing date of the claimed invention.
Claims 1-3, 5-10 and 19 are rejected under 35 U.S.C. 102(a)(1) as being anticipated by Tianzhe Bao (Bao), NPL “Deep Learning Based Upper-limb Motion Estimation Using Surface Electromyography”, November, 2021.
As to independent claim 1, Bao teaches: "A system comprising: computer-readable memory storing executable instructions; and one or more processors programmed by the executable instructions to at least:" (preamble)
- "obtain a corpus of training data comprising a plurality of training data input vectors (Bao 41, "a matrix is firstly constructed from Xt sEMG signals", The transformed signals are the training data.) and a plurality of reference data output vectors, wherein a reference data output vector of the plurality of reference data output vectors represents a desired output (Bao 41, "Herein G is denoted as the hand gesture", G is the labeled output vector) generated by an artificial neural network from a corresponding training data input vector of the plurality of training data input vectors;" (Bao Table 3.1 and Equation 3.1, The table depicts the neural network architecture and the equation depicts the distribution being learned where G-hat is the networks inferred class for a given input.)
PNG
media_image1.png
658
673
media_image1.png
Greyscale
- "train the artificial neural network using the corpus of training data to generate classification determinations;" (Bao 52, "the network was trained in a 128-sized mini batch using stochastic gradient descent with momentum (SDGM)."; Bao Equation 3.1, The equation teaches the classification being performed wherein G-hat is what Xt is classified as)
- "generate, using the artificial neural network and the corpus of training data, a generative confidence model of training data support for points in a feature space," (Bao 43 and Equation 3.2, "To indicate how confident the CNN classifier is about its prediction, a confidence estimation model is proposed by applying a zero-order smooth-step function to the weighted posterior probability distribution of CNN. The mathematical expression of this confidence model is [Equation 3.2]", Bao Figure 5.1, "LSTM-KF for
Sequential Regression", Examiner notes that the generative confidence model comprises both the ConfScore computation and the LSTM-KF module)
PNG
media_image2.png
212
832
media_image2.png
Greyscale
PNG
media_image3.png
742
727
media_image3.png
Greyscale
- "wherein the artificial neural network is configured to generate a point in the feature space during generation of a classification determination;" (Bao Figure 5.1, "Feature Vectors", f1...fk are generated by the CNN)
- "and evaluate a plurality of operational input vectors using the artificial neural network and the generative confidence model, wherein to evaluate each operational input vector of the plurality of operational input vectors, the one or more processors are programed by the executable instructions to: classify the operational input vector using the artificial neural network;" (Bao Equation 3.1, The equation teaches the classification being performed wherein G-hat is what Xt is classified as)
- "generate a confidence value for the classification determination using the generative confidence model;" (Bao Equation 3.2, "ConfScore")
- "and modify the generative confidence model based at least partly on the operational input vector and application of adaptive filtering to one or more variables of the generative confidence model (Bao Figure 5.2, "Kalman Update"), wherein the generative confidence model, after modification, is configured to generate a different confidence value for the classification determination (Bao Equation 3.2, The parameter Beta may be modified to modulate the confidence value)."
PNG
media_image4.png
433
720
media_image4.png
Greyscale
In addition, Bao discloses on pages iv and v, Abstract: a confidence model was designed to predict the possibility of correctness with regard to each classification of convolutional neural networks ((CNN), a hybrid framework using CNN for deep feature (vector) extraction and long short-term memory neural network (LSTM) was constructed to conduct sequence regression, wherein the hybrid framework was further extended by integrating Kalman filter with LSTM units in the recursive learning process, obtaining a deep Kalman filter network (DKFN) to perform kinematics estimation more effectively. Bao further discloses in Chapter 5, pages 93-94: LSTM-based Kalman Filter for sequence regression, wherein Kalman filters (KF) is a model-based method that recursively provides current estimate using the previous estimate and the most recent observation, wherein LSTM-KF is proposed to further enhance the sequence regression on features extracted by CNN. Bao further discloses on page 5, 2nd paragraph: a confidence estimation model is established to generate confidence scores (ConfScore) based on posterior probabilities of CNN, and an objective function is specially designed to train the parameters of this model. Bao further discloses on pages 61-62, 3.5 Discussion: a novel confidence estimation model ConfScore (p*, β) is proposed to generate confidence scores based on posterior probabilities of CNN, wherein p* =p*1, p*2, …p*m is obtained by sorting the posterior probability vector p, and β = β1, β2, … βm is a coefficient vector, and β is designed as a learnable parameter that can be tuned in a supervised manner (pages 43-44). Bao further discloses on page 71, 4.2 CNN-LSTM Hybrid Model: CNN is utilized to extract deep feature vectors which are rearranged into a series of feature sequences. Bao further discloses on pages 139-140, 7.1 Conclusion: Confidence estimation for CNN-based gesture recognition: Chapter 3 demonstrated a confidence estimation model for CNN-based gesture recognition. Specifically, the proposed model worked to predict the probability of correctness, i.e., ConfScore, for each classification of CNN. A novel objective function was designed to train parameters of the model.
As to independent claim 2, Bao teaches:
- "2. The system of claim 1, wherein to apply adaptive filtering to one or more variables of the generative confidence model, the one or more processors are further programmed by the executable instructions to apply a Kalman filter to a variable of the generative confidence model." (Bao Figure 5.2, "Kalman Update")
As to independent claim 3, Bao teaches:
"3. The system of claim 1, wherein the generative confidence model that is modified is configured to determine a second confidence value for a feature space point that corresponds to the operational input vector, wherein the second confidence value is higher than the confidence value, and wherein the feature space point is different from each of the points in the feature space observed during generation of the generative confidence model." (Bao 43 and Equation 3.2, The ConfScore function is configured to
do what is being claimed as it can generate scores between 0 and 1 inclusive.)
As to independent claim 5, Bao teaches:
"5. A computer-implemented method comprising: under control of a computing system comprising one or more processors configured to execute specific instructions," (preamble)
"obtaining a machine learning model trained to generate prediction outputs;" (Bao 52, "the network was trained in a 128-sized mini batch using stochastic gradient descent with momentum (SDGM)."; Bao Equation 3.1, The Equation teaches the classification being performed wherein G-hat is what Xt is classified as)
"obtaining a confidence model of training data support for points in a feature space," (Bao 43 and Equation 3.2, "To indicate how confident the CNN classifier is about its prediction, a confidence estimation model is proposed by applying a zero-order smooth-step function to the weighted posterior probability distribution of CNN. The mathematical expression of this confidence model is [Equation 3.2]", Bao Figure 5.1, "LSTM-KF for Sequential Regression", Examiner notes that the generative confidence model comprises both the ConfScore computation and the LSTM-KF module)
"wherein the machine learning model is configured to generate a point in the feature space during generation of a prediction output;" (Bao Figure 5.1, "Feature Vectors", f1...fk are generated by the CNN)
"and evaluating a plurality of input data items using the machine learning model and the confidence model, wherein evaluating an input data item of the plurality of input data items comprises: performing inference on the input data item using the machine learning model to generate prediction output data;" (Bao Equation 3.1, The Equation teaches the classification being performed wherein G-hat is what Xt is classified as)
"generating, using the confidence model, a confidence value for the prediction output data generated from the input data item;" (Bao Equation 3.2, "ConfScore")
"and modifying the confidence model based at least partly the input data item and application of adaptive filtering to one or more variables of the confidence model." (Bao Figure 5.2, "Kalman Update").
In addition, Bao discloses on pages iv and v, Abstract: a confidence model was designed to predict the possibility of correctness with regard to each classification of convolutional neural networks ((CNN), a hybrid framework using CNN for deep feature (vector) extraction and long short-term memory neural network (LSTM) was constructed to conduct sequence regression, wherein the hybrid framework was further extended by integrating Kalman filter with LSTM units in the recursive learning process, obtaining a deep Kalman filter network (DKFN) to perform kinematics estimation more effectively. Bao further discloses in Chapter 5, pages 93-94: LSTM-based Kalman Filter for sequence regression, wherein Kalman filters (KF) is a model-based method that recursively provides current estimate using the previous estimate and the most recent observation, wherein LSTM-KF is proposed to further enhance the sequence regression on features extracted by CNN. Bao further discloses on page 5, 2nd paragraph: a confidence estimation model is established to generate confidence scores (ConfScore) based on posterior probabilities of CNN, and an objective function is specially designed to train the parameters of this model. Bao further discloses on pages 61-62, 3.5 Discussion: a novel confidence estimation model ConfScore (p*, β) is proposed to generate confidence scores based on posterior probabilities of CNN, wherein p* =p*1, p*2, …p*m is obtained by sorting the posterior probability vector p, and β = β1, β2, … βm is a coefficient vector, and β is designed as a learnable parameter that can be tuned in a supervised manner (pages 43-44). Bao further discloses on page 71, 4.2 CNN-LSTM Hybrid Model: CNN is utilized to extract deep feature vectors which are rearranged into a series of feature sequences. Bao further discloses on pages 139-140, 7.1 Conclusion: Confidence estimation for CNN-based gesture recognition: Chapter 3 demonstrated a confidence estimation model for CNN-based gesture recognition. Specifically, the proposed model worked to predict the probability of correctness, i.e., ConfScore, for each classification of CNN. A novel objective function was designed to train parameters of the model.
As to independent claim 6, Bao teaches:
"6. The computer-implemented method of claim 5, wherein modifying the confidence model comprises configuring the confidence model to generate a different confidence value for a classification of the input data item into a class." (Bao Equation 3.2, The parameter beta may be modified to modulate the confidence value)
As to independent claim 7, Bao teaches:
"7. The computer-implemented method of claim 5, wherein obtaining the confidence model comprises generating one of:" (preamble)
"Gaussian mixture model to represent at least a subset of the points in a training data feature space (Bao 21-22, "Fukuda et al. suggested to calculate the entropy of a log-linearised Gaussian mixture network to indicate the risk of incorrect discrimination. If the entropy exceeded a pre-specified threshold, meaning that the network output is ambiguous, the associated motor control should be suspended."), or a clustering model to represent at least a subset of the points in the training data feature space (Bao 79,
"From Figure 4.5 we can see that in each dataset the clustering of scatters projected from CNN features is significantly better than that of hand-crafted features.")."
As to independent claim 8, Bao teaches:
- "8. The computer-implemented method of claim 5, wherein obtaining the machine learning model comprises:" (preamble)
- "obtaining a corpus of training data comprising a plurality of training data input vectors (Bao 41, "a matrix is firstly constructed from Xt sEMG signals") and a plurality of reference data output vectors, wherein a reference data output vector of the plurality of reference data output vectors represents a desired output (Bao 41, "Herein G is denoted as the hand gesture", G is the labeled output vector)
generated by an artificial neural network from a corresponding training data input vector of the plurality of training data input vectors;" (Bao Table 3.1 and Equation 3.1, The table depicts the neural network architecture and the equation depicts the distribution being learned where G-hat is the networks inferred class for a given input.)
- "and training the artificial neural network using the corpus of training data to generate classification determinations." (Bao 52, "the network was trained in a 128-sized mini batch using stochastic gradient descent with momentum (SDGM)."; Bao Equation 3.1, The equation teaches the classification being performed wherein G-hat is what Xt is classified as)
As to independent claim 9, Bao teaches:
- "9. The computer-implemented method of claim 5, wherein modifying the confidence model comprises applying a Kalman filter to a variable of the confidence model." (Bao Figure 5.2, "Kalman Update")
As to independent claim 10, Bao teaches:
- "10. The computer-implemented method of claim 9, wherein modifying the confidence model further comprises applying a second Kalman filter to a second variable of the confidence model." (Bao Figure 5.1, "LSTM-KF", Multiple LSTM Kalman filters are applied to each feature vector of the CNN)
Claim Rejections - 35 USC § 103
In the event the determination of the status of the application as subject to AIA 35 U.S.C. 102 and 103 (or as subject to pre-AIA 35 U.S.C. 102 and 103) is incorrect, any correction of the statutory basis (i.e., changing from AIA to pre-AIA ) for the rejection will not be considered a new ground of rejection if the prior art relied upon, and the rationale supporting the rejection, would be the same under either status.
The following is a quotation of 35 U.S.C. 103 which forms the basis for all obviousness rejections set forth in this Office action:
A patent for a claimed invention may not be obtained, notwithstanding that the claimed invention is not identically disclosed as set forth in section 102, if the differences between the claimed invention and the prior art are such that the claimed invention as a whole would have been obvious before the effective filing date of the claimed invention to a person having ordinary skill in the art to which the claimed invention pertains. Patentability shall not be negated by the manner in which the invention was made.
The text of those sections of Title 35, U.S. Code not included in this action can be found in a prior Office action.
The factual inquiries for establishing a background for determining obviousness under 35 U.S.C. 103 are summarized as follows:
1. Determining the scope and contents of the prior art.
2. Ascertaining the differences between the prior art and the claims at issue.
3. Resolving the level of ordinary skill in the pertinent art.
4. Considering objective evidence present in the application indicating obviousness or nonobviousness.
Claims 4, 11-13, 15-16, 19-20 are rejected under 35 U.S.C. 103 as being unpatentable over Bao as applied to claims 1-3 and 5-10 above, and further in view of Fearn et al. (Fearn), NPL “Classifying with confidence using Bayes rule and kernel density estimation”, April 15, 2019.
As to dependent claim 4, Bao does not teach, but Fearn teaches:
"4. The system of claim 1, wherein the one or more processors are further programmed by the executable instructions to add a kernel to the generative confidence model based at least partly on results of evaluating the plurality of operational input vectors, wherein the kernel defines a region of the feature space associated with higher confidence than generated using the generative confidence model without the kernel (Fearn 83, "The probability distribution within a class is estimated by placing a spherical multivariate normal distribution, known as a kernel, at the location of each training sample in
the class and then averaging these distributions.", Bao establishes that the neural network feature space is being analyzed to determine confidence whereas Fearn teaches a specific analytical tool (kernel density estimation) to carry out that task. Because the kernel is being used to estimate the probability distribution, it is associated with a higher confidence value.)."
It would have been obvious to one of ordinary skill in the art before the effective filing date of the present application to combine Bao and Fearn. Bao discloses a system of feature extraction and upper limb motion estimation that includes confidence modeling for said motion estimation. Fearn discloses confidence classification using kernel density estimation. One would be motivated to combine these references because the system of Fearn could be combined with the movement and confidence estimation
system of Bao in order to provide an improved or alternate confidence estimation. Further, MPEP 2143 sets forth the Supreme Court rationales for obviousness including: (D) Applying a known technique to a known device (method, or product) ready for improvement to yield predictable results.
As to dependent claim 11, Bao does not teach, but Fearn teaches:
- "11. The computer-implemented method of claim 5, further comprising adding a kernel to the confidence model based at least partly on results of evaluating the plurality of input data items, wherein the kernel defines a region of the feature space associated with a higher confidence value than generated using the confidence model without the kernel (Fearn 83, "The probability distribution within a class is estimated by placing a spherical multivariate normal distribution, known as a kernel, at the location of each training sample in the class and then averaging these distributions.", Bao establishes that the neural network feature space is being analyzed to determine confidence whereas Fearn teaches a specific analytical tool (kernel density estimation) to carry out that task. Because the kernel is being used to estimate the probability distribution, it is associated with a higher confidence value.
As to dependent claim 12, Bao does not teach, but Fearn teaches:
- "12. The computer-implemented method of claim 11, further comprising: determining that the region of the feature space is associated with a second region of the feature space, wherein the region of the feature space is associated with a first degree of confidence that is lower than a second degree of confidence associated with the second region of the feature space; and determining, based on the region of the feature space being associated with the second region of the feature space, to add the kernel defining the region of the feature space to the confidence model (Fearn 83, "The probability
distribution within a class is estimated by placing a spherical multivariate normal distribution, known as a kernel, at the location of each training sample in the class and then averaging these distributions.", Bao establishes that the neural network feature space is being analyzed to determine confidence whereas Fearn teaches a specific analytical tool (kernel density estimation) to carry out that task. Because the kernel
is being used to estimate the probability distribution, it is associated with a region indicative of a lower confidence value, i.e. outside of the estimated distribution)."
As to dependent claim 13, Bao does not teach, but Fearn teaches:
- "13. The computer-implemented method of claim 5, further comprising adding a kernel to the confidence model based at least partly on results of evaluating the plurality of input data items, wherein the kernel defines a region of the feature space associate with a lower confidence value than generated using the confidence model without the kernel (Fearn 83, "The probability distribution within a class is estimated by placing a spherical multivariate normal distribution, known as a kernel, at the location of each training sample in the class and then averaging these distributions.", Bao establishes that the neural network feature space is being analyzed to determine confidence whereas Fearn teaches a specific analytical tool (kernel density estimation) to carry out that task. Because the kernel is being used to estimate the probability distribution, it is associated with a region indicative of a lower confidence value, i.e. outside of the estimated distribution)."
As to independent claim 15, Bao teaches:
- "15. A system comprising: computer-readable memory storing executable instructions; and one or more processors programmed by the executable instructions to at least:" (preamble)
- "obtain a machine learning model trained to generate prediction outputs;" (Bao 52, "the network was trained in a 128-sized mini batch using stochastic gradient descent with momentum (SDGM)."; Bao Equation 3.1, The equation teaches the classification being performed wherein G-hat is what Xt is classified as)
- "obtain a confidence model of training data support for points in a feature space," (Bao 43 and Equation 3.2, "To indicate how confident the CNN classifier is about its prediction, a confidence estimation model is proposed by applying a zero-order smooth-step function to the weighted posterior probability distribution of CNN. The mathematical expression of this confidence model is", Bao Figure 5.1, "LSTM-KF for Sequential Regression", Examiner notes that the generative confidence model comprises both the
ConfScore computation and the LSTM-KF module)
- "wherein the machine learning model is configured to generate a point in the feature space during generation of a prediction output;" (Bao Figure 5.1, "Feature Vectors", f1...fk are generated by the CNN)
- "and evaluate a plurality of input data items using the machine learning model and the confidence model, wherein to evaluate an input data item of the plurality of input data items, the one or more processors are programed to: perform inference on the input data item using the machine learning model to generate prediction output data;" (Bao Equation 3.1, The Equation teaches the classification being performed wherein G-hat is what Xt is classified as)
- "generate, using the confidence model, a confidence value for the prediction output data generated from the input data item;" (Bao Equation 3.2, "ConfScore"); and
-“modify the confidence model based at least in part on the input data item” (Bao Figure 5.2, "Kalman Update").
In addition, Bao discloses on pages iv and v, Abstract: a confidence model was designed to predict the possibility of correctness with regard to each classification of convolutional neural networks ((CNN), a hybrid framework using CNN for deep feature (vector) extraction and long short-term memory neural network (LSTM) was constructed to conduct sequence regression, wherein the hybrid framework was further extended by integrating Kalman filter with LSTM units in the recursive learning process, obtaining a deep Kalman filter network (DKFN) to perform kinematics estimation more effectively. Bao further discloses in Chapter 5, pages 93-94: LSTM-based Kalman Filter for sequence regression, wherein Kalman filters (KF) is a model-based method that recursively provides current estimate using the previous estimate and the most recent observation, wherein LSTM-KF is proposed to further enhance the sequence regression on features extracted by CNN. Bao further discloses on page 5, 2nd paragraph: a confidence estimation model is established to generate confidence scores (ConfScore) based on posterior probabilities of CNN, and an objective function is specially designed to train the parameters of this model. Bao further discloses on pages 61-62, 3.5 Discussion: a novel confidence estimation model ConfScore (p*, β) is proposed to generate confidence scores based on posterior probabilities of CNN, wherein p* =p*1, p*2, …p*m is obtained by sorting the posterior probability vector p, and β = β1, β2, … βm is a coefficient vector, and β is designed as a learnable parameter that can be tuned in a supervised manner (pages 43-44). Bao further discloses on page 71, 4.2 CNN-LSTM Hybrid Model: CNN is utilized to extract deep feature vectors which are rearranged into a series of feature sequences. Bao further discloses on pages 139-140, 7.1 Conclusion: Confidence estimation for CNN-based gesture recognition: Chapter 3 demonstrated a confidence estimation model for CNN-based gesture recognition. Specifically, the proposed model worked to predict the probability of correctness, i.e., ConfScore, for each classification of CNN. A novel objective function was designed to train parameters of the model.
Bao, however, does not teach, but Fearn teaches:
- "and add a kernel to the confidence model based at least partly on results of evaluating the plurality of input data items, wherein the kernel defines a region of the feature space associated with a different confidence value than generated using the confidence model without the kernel (Fearn 83, "The probability distribution within a class is estimated by placing a spherical multivariate normal distribution, known as a kernel, at the location of each training sample in the class and then averaging these
distributions.", Bao establishes that the neural network feature space is being analyzed to determine confidence whereas Fearn teaches a specific analytical tool (kernel density estimation) to carry out that task. Because the kernel is being used to estimate the probability distribution, it is associated with a higher confidence value.)."
As to dependent claim 16, Bao does not teach, but Fearn teaches:
- "16. The system of claim 15, wherein the one or more processors are further programmed by the executable instructions to: determine that the region of the feature space is associated with a second region of the feature space, wherein the region of the feature space is associated with a first degree of confidence that is lower than a second degree of confidence associated with the second region of the feature space; and determine, based on the region of the feature space being associated with the second
region of the feature space, to add the kernel defining the region of the feature space to the confidence model (Fearn 83, "The probability distribution within a class is estimated by placing a spherical multivariate normal distribution, known as a kernel, at the location of each training sample in the class and then averaging these distributions.", Bao establishes that the neural network feature space is being analyzed to determine confidence whereas Fearn teaches a specific analytical tool (kernel density
estimation) to carry out that task. Because the kernel is being used to estimate the probability distribution, it is associated with a region indicative of a lower confidence value, i.e. outside of the estimated distribution)."
As to dependent to claim 19, Bao teaches:
- "19. The system of claim 15, wherein to evaluate the plurality of input data items, the one or more processors are further programmed by the executable instructions to modify the confidence model based at least partly on application of adaptive filtering to one or more variables of the confidence model." (Bao Figure 5.2, "Kalman Update")
As to dependent to claim 20, Bao teaches:
- "20. The system of claim 15, wherein the confidence model comprises one of:" (preamble)
- "a Gaussian mixture model to represent at least a subset of the points in a training data feature space (Bao 21-22, "Fukuda et al. suggested to calculate the entropy of a log-linearised Gaussian mixture network to indicate the risk of incorrect discrimination. If the entropy exceeded a pre-specified threshold, meaning that the network output is ambiguous, the associated motor control should be suspended."), or a
clustering model to represent at least a subset of the points in the training data feature space (Bao 79, "From Figure 4.5 we can see that in each dataset the clustering of scatters projected from CNN features is significantly better than that of hand-crafted features.")."
Claims 14, 17 and 18 are rejected under 35 U.S.C. 103 as being unpatentable over Bao and Fearn as applied to claims 4, 11-13, 15-16 and 19-20 above, and further in view of Virkar, US Patent Application Publication No. US 20100063948-A1.
As to dependent claim 14, Bao does not teach, but Fearn teaches .
- "[determining that the region of the feature space is associated with a training data bias; and determining, based on the region of the feature space being associated with the training data bias, to] add the kernel [defining the region of the feature space to the confidence model]." (Fearn 83, "The probability distribution within a class is estimated by placing a spherical multivariate normal distribution, known as a kernel, at the location of each training sample in the class and then averaging these distributions."
The combination of Bao and Fearn does not teach, but Virkar teaches:
- "determining that the region of the feature space is associated with a training data bias; and determining, based on the region of the feature space being associated with the training data bias, to [add the kernel] defining the region of the feature space to the confidence model." (Virkar [0249], "In either case, this method allows the decision line to find equilibrium between the classes, which would otherwise create an unwanted bias in favor of the larger class. This approach is advantageous when positive samples are outnumbered by negative samples. Given an experiment which lacks a negative class
altogether, we have proposed a novel method of employing negative examples from a random sampling over the entire feature space. By providing an abundance of random samples, each with an extremely weak bound in comparison with the positive samples, a negative region can be defined, and a decision boundary for membership in the positive class can be obtained.", The reference describes a dataset bias with the negative and positive samples. The reference defines a region using the decision boundary.)
Motivation to combiner Bao, Fearn, Virkar.
It would have been obvious to one of ordinary skill in the art before the effective filing date of the present application to combine Bao, Fearn and Virkar. Bao, Fearn discloses a system of feature extraction and upper limb motion estimation that includes confidence modeling for said motion estimation. Virkar discloses a method for identifying and handling bias in the training of machine learning algorithms. One
would be motivated to combine these references because the method of Virkar could be used to augment the system of Bao, Fearn in order to increase the overall confidence estimates (i.e. confidence goes up) or better estimate the confidence estimate (i.e. accuracy of estimate goes up). Further, MPEP 2143 sets forth
the Supreme Court rationales for obviousness including: (D) Applying a known technique to a known device (method, or product) ready for improvement to yield predictable results.
As to dependent claim 17, Bao does not teach, but Fearn teaches:
- "17. The system of claim 15, wherein the one or more processors are further programmed by the executable instructions to:" (preamble)
- "[determine that the region of the feature space is associated with a training data bias; and determine, based on the region of the feature space being associated with the training data bias, to] add the kernel [defining the region of the feature space to the confidence model]." (Fearn 83, "The probability distribution within a class is estimated by placing a spherical multivariate normal distribution, known as a kernel, at the location of each training sample in the class and then averaging these distributions."
The combination of Bao and Fearn does not teach, but Virkar teaches:
- "determine that the region of the feature space is associated with a training data bias; and determine, based on the region of the feature space being associated with the training data bias, to [add the kernel] defining the region of the feature space to the confidence model." (Virkar [0249], "In either case, this method allows the decision line to find equilibrium between the classes, which would otherwise create an unwanted bias in favor of the larger class. This approach is advantageous when positive samples are
outnumbered by negative samples. Given an experiment which lacks a negative class altogether, we have proposed a novel method of employing negative examples from a random sampling over the entire feature space. By providing an abundance of random samples, each with an extremely weak bound in comparison with the positive samples, a negative region can be defined, and a decision boundary for membership in the positive class can be obtained.", The reference describes a dataset bias with the negative and positive samples. The reference defines a region using the decision boundary.)
As to dependent claim 18, Bao teaches:
- "18. The system of claim 15, wherein the one or more processors are further programmed by the executable instructions to" (preamble)
- "receive, from a management computing system, [bias data indicating the region of the feature space is associated with a training data bias, wherein the kernel is added in response to receiving the bias data]." (Bao Figure 5.1, "Feature Vectors", f1...fk are generated by the CNN and received by the LSTM-KF system.)
Bao, however, does not teach, but Fearn teaches:
- "[receive, from a management computing system, bias data indicating the region of the feature space is associated with a training data bias], wherein the kernel is added in response to receiving the bias data." (Fearn 83, "The probability distribution within a class is estimated by placing a spherical multivariate normal distribution, known as a kernel, at the location of each training sample in the class and then averaging these distributions."
The combination of Bao and Fearn does not teach, but Virkar teaches:
- "[receive, from a management computing system], bias data indicating the region of the feature space is associated with a training data bias, [wherein the kernel is added in response to receiving the bias data]." (Virkar [0249], "In either case, this method allows the decision line to find equilibrium between the classes, which would otherwise create an unwanted bias in favor of the larger class. This approach is advantageous when positive samples are outnumbered by negative samples. Given an experiment which
lacks a negative class altogether, we have proposed a novel method of employing negative examples from a random sampling over the entire feature space. By providing an abundance of random samples, each with an extremely weak bound in comparison with the positive samples, a negative region can be defined, and a decision boundary for membership in the positive class can be obtained.” The reference describes a dataset bias with the negative and positive samples. The reference defines a region using the decision boundary.)
Response to Arguments
In the Remarks, Applicant argues in substance that:
Applicant’s arguments regarding claims rejections under 35 U.S.C. § 101 (see pages 9-17 of Remarks) are persuasive. Therefore, the rejection under 35 U.S.C. § 101 is hereby withdrawn.
Bao does not describe any combination of the LSTM-KF model and the ConfScore system.
In reply to this argument, Bao discloses on pages iv and v, Abstract: a confidence model was designed to predict the possibility of correctness with regard to each classification of convolutional neural networks ((CNN), a hybrid framework using CNN for deep feature (vector) extraction and long short-term memory neural network (LSTM) was constructed to conduct sequence regression, wherein the hybrid framework was further extended by integrating Kalman filter with LSTM units in the recursive learning process, obtaining a deep Kalman filter network (DKFN) to perform kinematics estimation more effectively. Bao further discloses in Chapter 5, pages 93-94: LSTM-based Kalman Filter for sequence regression, wherein Kalman filters (KF) is a model-based method that recursively provides current estimate using the previous estimate and the most recent observation, wherein LSTM-KF is proposed to further enhance the sequence regression on features extracted by CNN. Bao further discloses on page 5, 2nd paragraph: a confidence estimation model is established to generate confidence scores (ConfScore) based on posterior probabilities of CNN, and an objective function is specially designed to train the parameters of this model. Bao further discloses on pages 61-62, 3.5 Discussion: a novel confidence estimation model ConfScore (p*, β) is proposed to generate confidence scores based on posterior probabilities of CNN, wherein p* =p*1, p*2, …p*m is obtained by sorting the posterior probability vector p, and β = β1, β2, … βm is a coefficient vector, and β is designed as a learnable parameter that can be tuned in a supervised manner (pages 43-44). Bao further discloses on page 71, 4.2 CNN-LSTM Hybrid Model: CNN is utilized to extract deep feature vectors which are rearranged into a series of feature sequences. Bao further discloses on pages 139-140, 7.1 Conclusion: Confidence estimation for CNN-based gesture recognition: Chapter 3 demonstrated a confidence estimation model for CNN-based gesture recognition. Specifically, the proposed model worked to predict the probability of correctness, i.e., ConfScore, for each classification of CNN. A novel objective function was designed to train parameters of the model.
Conclusion
THIS ACTION IS MADE FINAL. Applicant is reminded of the extension of time policy as set forth in 37CFR 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 extension fee 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 CHAU T NGUYEN whose telephone number is (571)272-4092. The examiner can normally be reached on Monday-Friday from 8am to 5pm.
If attempts to reach the examiner by telephone are unsuccessful, the examiner’s supervisor, Cesar Paula, can be reached at telephone number 5712724128. The fax phone number for the organization where this application or proceeding is assigned is 571-273-8300.
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) Form at https://www.uspto.gov/patents/uspto-automated-interview-request-air-form.
Information regarding the status of an application may be obtained from the Patent Application Information Retrieval (PAIR) system. Status information for published applications may be obtained from either Private PAIR or Public PAIR. Status information for unpublished applications is available through Private PAIR only. For more information about the PAIR system, see http://pair-direct.uspto.gov. Should you have questions on access to the Private PAIR system, contact the Electronic Business Center (EBC) at 866-217-9197 (toll-free).
/CHAU T NGUYEN/Primary Examiner, Art Unit 2145
Prosecution Insights