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 .
This office action is in response to submission of application on 5/18/2023.
Claims 1-19 are presented for examination.
Specification
The specification is objected to as failing to provide proper antecedent basis for the claimed subject matter. See 37 CFR 1.75(d)(1) and MPEP § 608.01(o). Correction of the following is required: The instant specification does not define or provide support for a "pixel value change rate" or "pixel change ratio"
Claim Rejections - 35 USC § 112
The following is a quotation of 35 U.S.C. 112(b):
(b) CONCLUSION.—The specification shall conclude with one or more claims particularly pointing out and distinctly claiming the subject matter which the inventor or a joint inventor regards as the invention.
The following is a quotation of 35 U.S.C. 112 (pre-AIA ), second paragraph:
The specification shall conclude with one or more claims particularly pointing out and distinctly claiming the subject matter which the applicant regards as his invention.
Claims 6 and 15 are rejected under 35 U.S.C. 112(b) or 35 U.S.C. 112 (pre-AIA ), second paragraph, as being indefinite for failing to particularly point out and distinctly claim the subject matter which the inventor or a joint inventor (or for applications subject to pre-AIA 35 U.S.C. 112, the applicant), regards as the invention.
The term “low-capacity device” in claim 6 and 15 is a relative term which renders the claim indefinite. The term “low-capacity device” is not defined by the claim, the specification does not provide a standard for ascertaining the requisite degree, and one of ordinary skill in the art would not be reasonably apprised of the scope of the invention. For purposes of examination, "low-capacity device" is interpreted under BRI as referring to being able to perform lower computations.
Appropriate correction is required.
Claim Rejections - 35 USC § 101
35 U.S.C. 101 reads as follows:
Whoever invents or discovers any new and useful process, machine, manufacture, or composition of matter, or any new and useful improvement thereof, may obtain a patent therefor, subject to the conditions and requirements of this title.
Claims 1-19 are rejected under 35 U.S.C. 101 because the claimed invention is directed to an abstract idea without significantly more.
Step 1: Is the claim to a process, machine, manufacture or composition of matter?
Claims 1-9 are directed to an apparatus (i.e., a machine/apparatus); claims 10-18 are directed to a method (i.e., a process); and claim 19 is directed to an article of manufacture (i.e., a product); therefore, all pending claims are directed to one of the four categories of invention.
Step 2A: Prong 1: Does the claim recite an abstract idea, law of nature, or natural
phenomenon?
Claim 1 recites the limitations of:
identify a first activation function used in at least one layer of the artificial intelligence model - mental process (observation, evaluation, judgement) as a human mind can identify a function.
obtain a second activation function by adding a first periodic function corresponding to a first time interval to the first activation function - mental process (observation, evaluation, judgement) as a human mind can add a function to function.
apply the second activation function to an output layer during the first time
interval - mental process (observation, evaluation, judgement) as a human mind can apply a function to an output.
obtain a third activation function by adding a second periodic function corresponding to a second time interval to the first activation function, and -mental process (observation, evaluation, judgement) as a human mind can add a function to function.
update the artificial intelligence model by applying the third activation function to the output layer during the second time interval - mental process (observation, evaluation, judgement) as a human mind can apply a function to an output.
which are abstract ideas.
Step 2A: Prong 2 & Step 2B: There are no additional elements recited so the claim does not provide a practical application and is not considered to be significantly more. As such, the claim is patent ineligible.
Independent claims 10 and 19 recite the same relevant limitations and a similar analysis applies. Claim 10 recites the additional elements of "A control method of an electronic apparatus, the method comprising:" - components recited at a high level are construed as generic computer components used to implement the abstract idea. See MPEP 2106.05(f)(2). As such, the limitations do not integrate the abstract idea into a practical application. Nor to do they amount to significantly more. Claim 19 recites the additional limitations of "A non-transitory computer-readable storage medium storing instructions that, when executed by a processor of an electronic apparatus, cause the processor to:" - components recited at a high level are construed as generic computer components used to implement the abstract idea. See MPEP 2106.05(f)(2). As such, the limitations do not integrate the abstract idea into a practical application. Nor to do they amount to significantly more.
Therefore, the independent claims are not patent eligible.
Claim 2 and 11 recites the additional elements of “each of the first periodic function and the second periodic function comprises functions added to a plurality of periodic functions,” – description of a limitation on the data merely identifies a field of use. See MPEP 2106.05(h); and “wherein the first periodic function comprises at least one of a period of a function, an amplitude of a function, and a number of periodic functions, and” - description of a limitation on the data merely identifies a field of use. See MPEP 2106.05(h); and “wherein the second periodic function comprises at least one of a period of a function, an amplitude of a function, and a number of periodic functions that is different from the first periodic function.” - description of a limitation on the data merely identifies a field of use. See MPEP 2106.05(h).
Claim 3 recites the additional elements of “a communication interface,” - sending data is insignificant, extra-solution activity. See MPEP 2106.05(g). Transmitting data is well-understood, routine, and conventional. See MPEP 2106.05(d)(II)(i); and “wherein the at least one processor is further configured to: obtain the first periodic function corresponding to the first time interval from an external device through the communication interface, and” - sending data is insignificant, extra-solution activity. See MPEP 2106.05(g). Transmitting data is well-understood, routine, and conventional. See MPEP 2106.05(d)(II)(i); and “obtain the second periodic function corresponding to the second time interval from the external device through the communication interface.” - sending data is insignificant, extra-solution activity. See MPEP 2106.05(g). Transmitting data is well-understood, routine, and conventional. See MPEP 2106.05(d)(II)(i).
Claim 4 and 13 recites the additional elements of “identify whether there is an attack on the updated artificial intelligence model based on a pixel value change rate of an image output from the updated artificial intelligence model.” - mental process (observation, evaluation, judgement) as a human mind can identify a change in pixel value.
Claim 5 recites the additional elements of “training the at least one layer having the second activation function or the third activation function applied thereto, or” - general class of computer algorithms recited at a high level are construed as generic computer components used to implement the abstract idea. See MPEP 2106.05(f)(2); and “additionally training all layers of the artificial intelligence model.” - general class of computer algorithms recited at a high level are construed as generic computer components used to implement the abstract idea. See MPEP 2106.05(f)(2).
Claim 6 and 15 recites the additional elements of “the electronic apparatus is implemented as a low-capacity device in which the at least one layer of the artificial intelligence model is not encrypted.” - components recited at a high level are construed as generic computer components used to implement the abstract idea. See MPEP 2106.05(f)(2).
Claim 7 and 16 recites the additional elements of “obtain the first periodic function and the second periodic function based on a type of the artificial intelligence model, a type of the at least one layer, a type of the electronic apparatus, a user of the electronic apparatus, or a type of the first activation function.” - mental process (observation, evaluation, judgement) as a human mind can obtain a function.
Claim 8 and 17 recites the additional elements of “the at least one layer comprises the output layer.” - description of a limitation on the data merely identifies a field of use. See MPEP 2106.05(h).
Claim 9 and 18 recites the additional elements of “the first activation function comprises at least one of a continuous function and a discontinuous function.” - description of a limitation on the data merely identifies a field of use. See MPEP 2106.05(h).
Claim 12 recites the additional elements of “obtaining the first periodic function corresponding to the first time interval from an external device through a communication interface, and” - sending data is insignificant, extra-solution activity. See MPEP 2106.05(g). Transmitting data is well-understood, routine, and conventional. See MPEP 2106.05(d)(II)(i); and “obtaining the second periodic function corresponding to the second time interval from the external device through the communication interface.” - sending data is insignificant, extra-solution activity. See MPEP 2106.05(g). Transmitting data is well-understood, routine, and conventional. See MPEP 2106.05(d)(II)(i).
Claim 14 recites the additional elements of “based on the artificial intelligence model being updated, training the at least one layer having the second activation function or the third activation function applied thereto, or” - general class of computer algorithms recited at a high level are construed as generic computer components used to implement the abstract idea. See MPEP 2106.05(f)(2); and “additionally training all layers of the artificial intelligence model.” - general class of computer algorithms recited at a high level are construed as generic computer components used to implement the abstract idea. See MPEP 2106.05(f)(2).
Claim Rejections - 35 USC § 102
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, 5-12, and 14-19 are rejected under 35 U.S.C. 102(a)(1) as being anticipated by Neural Networks Fail to Learn Periodic Functions and How to Fix It to Ziyin et al. (hereinafter Ziyin).
Per claim 1, Ziyin discloses An electronic apparatus, comprising: a memory storing an artificial intelligence model; and at least one processor configured to [pg. 6, Section 6 "Applications…demonstrate the wide applicability of Snake...image classification task...temperature and financial data prediction...". (note: section 6 discusses different applications of Snake, which requires running on a computer system); pg. 9 "the training is done with gradient descent on the full batch, and the computation time remains low and does not increase visibly as long as the GPU memory is not overloaded"]:
identify a first activation function used in at least one layer of the artificial intelligence model [pg. 1 "...the standard activations functions, such as ReLU, tanh, sigmoid, along with their variants, all fail to learn to extrapolate simple periodic functions. We hypothesize that this is due to their lack of a “periodic” inductive bias. As a fix of this problem, we propose a new activation, namely, x + sin2(x), which achieves the desired periodic inductive bias...". (note: the new, or second, activation function proposed is 'x + sin2(x)', which makes it different than the standard, or first, activation function because of its periodic inductive bias, which is achieved by the addition of 'sin2(x)'. 'x' by itself is not a periodic function. The standard activation functions referred to do not have a periodic inductive bias. Therefore, the difference between the first and second activation function, is the addition of 'sin2(x)', which is the periodic inductive bias. Therefore, the first activation function is simply 'x'. f(x)=x is the ‘identity activation function’, which is a type of known activation function in the art) (note: activation functions are applied in the hidden layers and the output layer just before passing the data to the next layer. This is how activation functions work. Ziyin says that standard activation functions like ReLU are replaced with Snake activation function, x + sin2(x), so Snake is being applied as an activation function throughout all the layers.); pg. 5 "As shown in [16], different activation functions actually require different initialization schemes (in terms of the sampling variance) to make the output of each layer unit variance, thus avoiding divergence or vanishing of the forward signal." (note: this is under 'Initializtion for Snake', Ziyin is initializing the Snake function to maintain unit variance for each layer, thus the activation function is to be applied in all the layers, which includes the output layer)],
obtain a second activation function by adding a first periodic function corresponding to a first time interval to the first activation function [pg. 3 "we propose to use x + sin2(x) as an activation function, which we call the “Snake” function.". (note: The first periodic function sin2(x) has an intrinsic corresponding time interval being its cyclical period, which is the first time interval, that is associated with the sinusoidal function and this being added to the first activation function ‘x’ obtains the second activation function ‘x + sin2(x)’)],
apply the second activation function to an output layer during the first time interval [pg. 1 "...standard activations functions, such as ReLU, tanh, sigmoid, along with their variants, all fail to learn to extrapolate simple periodic functions. We hypothesize that this is due to their lack of a “periodic” inductive bias. As a fix of this problem, we propose a new activation, namely, x + sin2(x),". (note: activation functions are applied in the hidden layers and the output layer just before passing the data to the next layer. This is how activation functions work. Ziyin says that standard activation functions like ReLU are replaced with the Snake activation function, ‘x + sin2(x)’, so Snake is being applied as an activation function throughout all the layers.); pg. 5 "As shown in [16], different activation functions actually require different initialization schemes (in terms of the sampling variance) to make the output of each layer unit variance, thus avoiding divergence or vanishing of the forward signal." (note: this is under 'Initializtion for Snake', Ziyin is initializing the Snake function to maintain unit variance for each layer, therefore the activation function is to be applied in all the layers, which includes the output layer.)],
obtain a third activation function by adding a second periodic function corresponding to a second time interval to the first activation function, and [pg. 4 "We also propose two other alternatives, x + sin(x) and x + cos(x)." (note: the second periodic function, 'sin x' or 'cos x', has an intrinsic corresponding time interval being its cyclical period, which is the second time interval, that is associated with the sinusoidal function and this being added to the first activation function ‘x’ obtains the third activation function)],
update the artificial intelligence model by applying the third activation function to the output layer during the second time interval [pg. 1 "...standard activations functions, such as ReLU, tanh, sigmoid, along with their variants, all fail to learn to extrapolate simple periodic functions. We hypothesize that this is due to their lack of a “periodic” inductive bias. As a fix of this problem, we propose a new activation, namely, x + sin2(x),"; pg. 4 "We also propose two other alternatives, x + sin(x) and x + cos(x).". (note: activation functions are applied in the hidden layers and the output layer just before passing the data to the next layer. This is how activation functions work. Ziyin says that standard activation functions like ReLU are replaced with periodic activation functions, ‘x + sin(x)’ and ‘x + cos(x)’, which are therefore being applied as an activation function throughout all the layers.); pg. 5 "As shown in [16], different activation functions actually require different initialization schemes (in terms of the sampling variance) to make the output of each layer unit variance, thus avoiding divergence or vanishing of the forward signal." (note: this is under 'Initializtion for Snake', Ziyin is initializing the Snake function to maintain unit variance for each layer, therefore the activation function is to be applied in all the layers, which includes the output layer. The same is true for its two alternative activation functions, ‘x + sin(x)’ and ‘x + cos(x)’)].
Per claim 2, Ziyin discloses claim 1, further disclosing each of the first periodic function and the second periodic function comprises functions added to a plurality of periodic functions [pg. 3 "One is the sin function, which has been proposed in [30], along with cos and their linear combinations as proposed in Fourier neural networks". (note: Fourier combinations show that both sine and cosine can be expressed as linear combinations, which are functions added to a plurality of periodic functions ); pg. 21 "It suffices to show that a neural network with sin as activation function can represent a Fourier series to arbitrary order"],
wherein the first periodic function comprises at least one of a period of a function, an amplitude of a function, and a number of periodic functions [pg. 3 "One can augment it with a factor a to control the frequency of the periodic part. Thus propose the Snake activation with frequency a." (note: frequency a is a period of a function); pg. 3, Equation 3. (note: (1/a) in the equation is the amplitude of a function))], and
wherein the second periodic function comprises at least one of a period of a function, an amplitude of a function, and a number of periodic functions that is different from the first periodic function [pg. 4 "We also propose two other alternatives, x + sin(x) and x + cos(x)." (note: since both possible functions are either sine or cosine, their period can be adjusted and their amplitude. Both possibilities are also different from the first periodic function)].
Per claim 3, Ziyin discloses claim 1, further disclosing a communication interface, wherein the at least one processor is further configured to: obtain the first periodic function corresponding to the first time interval from an external device through the communication interface [pg. 9 "the training is done with gradient descent on the full batch, and the computation time remains low and does not increase visibly as long as the GPU memory is not overloaded" (note: the use of a GPU, which is an external device from the main processor, requires a communication interface, such as a PCIe)], and
obtain the second periodic function corresponding to the second time interval from the external device through the communication interface [pg. 9 "the training is done with gradient descent on the full batch, and the computation time remains low and does not increase visibly as long as the GPU memory is not overloaded" (note: the use of a GPU, which is an external device from the main processor, requires a communication interface, such as a PCIe)].
Per claim 5, Ziyin discloses claim 1, further disclosing training the at least one layer having the second activation function or the third activation function applied thereto, or additionally training all layers of the artificial intelligence model [pg. 4 "…we regress a simple 1−d periodic function, sin(x), with the proposed activation function. See Figure 4 and compare with the related experiments on tanh and ReLU in Figure 1. As expected, all three activation functions learn to regress the training points.", (note: here Snake specifically learns to capture the correct frequency and periodic nature of the function at all training points, thus training across all layers); pg. 5 “Initialization for Snake…different activation functions actually require different initialization schemes (in terms of the sampling variance) to make the output of each layer unit variance…using the correction leads to better training speed…”; pg. 6 “Applications…We start with a standard image classification task, where Snake is shown to perform competitively against the popular activation functions…Image Classification…We train ResNet-18 [17], with roughly 10M parameters, on the standard CIFAR-10 dataset. We simply replace the activation functions in ReLU with the specified ones for comparison. CIFAR-10 is a 10-class image classification task…”. (note: for standard tasks like CIFAR-10, the activation function is applied across the specified layers and the entire network is trained to achieve unit variance and competitive accuracy)].
Per claim 6, Ziyin discloses claim 1, further disclosing the electronic apparatus is implemented as a low-capacity device in which the at least one layer of the artificial intelligence model is not encrypted. [pg. 6, Section 6 "Applications…demonstrate the wide applicability of Snake...image classification task...temperature and financial data prediction...". (note: section 6 discusses different applications of Snake, which requires running on a computer device); pg. 9 "the training is done with gradient descent on the full batch, and the computation time remains low and does not increase visibly as long as the GPU memory is not overloaded". (note: maintaining low computation time by operating within available GPU memory limits, is the equivalent of implementing on a low-capacity device, such constraints and optimizations are necessitated only in a resource-limited, low capacity, environment)].
Per claim 7, Ziyin discloses claim 1, further disclosing obtain the first periodic function and the second periodic function based on a type of the artificial intelligence model, a type of the at least one layer, a type of the electronic apparatus, a user of the electronic apparatus, or a type of the first activation function. [pg. 3 "we propose to use x + sin2(x) as an activation function, which we call the “Snake” function."; pg. 4 "We also propose two other alternatives, x + sin(x) and x + cos(x)." (note: the first periodic function, ‘sin2(x)’, and the second periodic function, ‘sin(x)’ or ‘cos(x)’, are periodic type activation functions.)]:
Per claim 8, Ziyin discloses claim 1, further disclosing the at least one layer comprises the output layer [pg. 1 "...the standard activations functions, such as ReLU, tanh, sigmoid, along with their variants, all fail to learn to extrapolate simple periodic functions. We hypothesize that this is due to their lack of a “periodic” inductive bias. As a fix of this problem, we propose a new activation, namely, x + sin2(x), which achieves the desired periodic inductive bias...". (note: activation functions are applied in the hidden layers and the output layer just before passing the data to the next layer. This is how activation functions work. Ziyin says that standard activation functions like ReLU are replaced with Snake activation function, x + sin2(x), so Snake is being applied as an activation function throughout all the layers.); pg. 5 "As shown in [16], different activation functions actually require different initialization schemes (in terms of the sampling variance) to make the output of each layer unit variance, thus avoiding divergence or vanishing of the forward signal." (note: this is under 'Initializtion for Snake', Ziyin is initializing the Snake function to maintain unit variance for each layer, thus the activation function is to be applied in all the layers, which includes the output layer)].
Per claim 9, Ziyin discloses claim 1, further disclosing the first activation function comprises at least one of a continuous function and a discontinuous function [pg. 1 "the standard activations functions, such as ReLU, tanh, sigmoid, along with their variants,...". (note: the initial activation function 'x', is one of the listed functions here, and ReLU, tanh, and sigmoid are all continuous functions)].
Claims 10 and 19 are substantially similar in scope and spirit to claim 1. Therefore, the rejection of claim 1 is applied accordingly. Regarding claim 10, claim 10 is directed towards the method performed by the apparatus of claim 1. Therefore, the rejection applied to claim 1 also applies to claim 10. Regarding claim 19, Ziyin also shows the apparatus being implemented by an article of manufacture (pg. 9 "… as long as the GPU memory is not overloaded").
Dependent claims 11-12 and 14-18 are directed towards the method performed by the apparatus of claims 2-3 and 5-9, respectively. Therefore, the rejections applied to claims 2-3 and 5-9 also apply to claims 11-12 and 14-18.
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.
The factual inquiries for establishing a background for determining obviousness under 35 U.S.C. 103 are summarized as follows:
1. Determining the scope and contents of the prior art.
2. Ascertaining the differences between the prior art and the claims at issue.
3. Resolving the level of ordinary skill in the pertinent art.
4. Considering objective evidence present in the application indicating obviousness or nonobviousness.
This application currently names joint inventors. In considering patentability of the claims the examiner presumes that the subject matter of the various claims was commonly owned as of the effective filing date of the claimed invention(s) absent any evidence to the contrary. Applicant is advised of the obligation under 37 CFR 1.56 to point out the inventor and effective filing dates of each claim that was not commonly owned as of the effective filing date of the later invention in order for the examiner to consider the applicability of 35 U.S.C. 102(b)(2)(C) for any potential 35 U.S.C. 102(a)(2) prior art against the later invention.
Claim 4 is rejected under 35 U.S.C. 103 as being unpatentable over Ziyin in view of Feature Squeezing: Detecting Adversarial Examples in Deep Neural Networks to Xu et al. (hereinafter Xu).
Per claim 4, Ziyin discloses claim 1.
Ziyin does not fully disclose, but with Xu does teach: identify whether there is an attack [pg. 1 “compare the model’s prediction on the original sample with its prediction on the sample after squeezing, as depicted in Figure 1. If the original and squeezed inputs produce substantially different outputs from the model, the input is likely to be adversarial.” (note: this comparison is the identification); pg. 2 “The goal of this work is to harden DNN systems against adversarial examples by detecting adversarial inputs”] on the updated artificial intelligence model based on a pixel value change rate of an image output [pg. 1 “reducing the color bit depth of each pixel” (note: color bit depth is the number of bits used to represent the color of a single pixel. Pixel value is the actual numerical data stored within those bits. A color pixel necessarily involves a color bit depth. The reduction of the color bit depth is the change rate.)] from the updated artificial intelligence model [pg. 1 “comparing a DNN model’s prediction on the original input with that on squeezed inputs,” (note: the squeezed inputs correspond to the updated model.)].
Ziyin and Xu are analogous art because they are from the same field of endeavor encompassing neural networks.
Before the effective filing date of the claimed invention, it would have been obvious to a person of ordinary skill in the art to incorporate the identification of an attack based on a pixel value change rate as taught by Xu in Ziyin.
The suggestion/motivation for doing so would have been to defend against adversarial examples. Xu states that its defense strategy is simple, low-cost, and does not change the model. Xu also states the benefit of its feature squeezing significantly enhances the robustness of a model. [Xu, pg. 1 “…deep neural networks (DNNs)…can often be fooled by adversarial examples”, “We propose a new strategy, feature squeezing, that can be used to harden DNN models by detecting adversarial examples”, “Detecting an attempted attack may be as important as predicting correct outputs”, “In contrast, our work focuses on finding simple and low-cost defensive strategies that alter the input samples but leave the model unchanged”; pg. 2 “…feature squeezing significantly enhances the robustness of a model…”].
Regarding claim 13, claim 13 is directed towards a method performed by the apparatus of claim 4. Therefore, the rejection applied to claim 4 also applies to claim 13.
Conclusion
The prior art made of record and not relied upon is considered pertinent to applicant's disclosure. Patents and/or related publications are cited in the Notice of References Cited (Form PTO-892) attached to this action to further show the state of the art with respect to reinforcing security of an artificial intelligence model.
Sengupta discloses adversarial attack defense based on corresponding time intervals.
Any inquiry concerning this communication or earlier communications from the examiner should be directed to Sayed M Shah whose telephone number is (571)272-9406. The examiner can normally be reached Monday-Friday 6:00 am - 2:00 pm.
Examiner interviews are available via telephone, in-person, and video conferencing using a USPTO supplied web-based collaboration tool. To schedule an interview, applicant is encouraged to use the USPTO Automated Interview Request (AIR) at http://www.uspto.gov/interviewpractice.
If attempts to reach the examiner by telephone are unsuccessful, the examiner’s supervisor, Miranda Huang can be reached at (571) 270-7092. 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.
/SAYED MUNEER SHAH/Examiner, Art Unit 2124
/MIRANDA M HUANG/Supervisory Patent Examiner, Art Unit 2124
Prosecution Insights