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 .
Specification
The abstract of the disclosure is objected to because it exceeds 150 words, and additionally because it exceeds 15 lines. A corrected abstract of the disclosure is required and must be presented on a separate sheet, apart from any other text. See MPEP § 608.01(b).
Claim Objections
Claims 1, 10, and 13 objected to because of the following informality: optimizing a first hyper-parameter (λ) and a second hyper parameter (α) should read “optimizing a first hyper-parameter (λ) and a second hyper-parameter (α)”. Appropriate correction is required.
Claims 2, 16, and 19 objected to because of the following informality: wherein the training data set (D) is generated by randomly generation on the noisy observation should read “wherein the training data set (D) is generated by random generation of the noisy observation”. Appropriate correction is required.
Claim 9 objected to because of the following informality: and processing unit (PU) is a user equipment (UE) should read “and the processing unit (PU) is a user equipment (UE)”. Appropriate correction is required.
Claim 13 objected to because of the following informalities: the iterations (T) the subprocess should read “the iterations (T) of the subprocess”, and “the subprocess the hyper-parameter node” should read “the distinct subprocesses, the at least one hyper-parameter node”. Appropriate correction is required.
Claim Interpretation
Various mathematical symbols or terms such as H, σ, IN, -1, T, diag, ci, and M are used within the equations within claims 5-8, without being defined within those claims or any parent claims. To the extent possible, definitions for these symbols or terms will be taken from the following parts of the specification: [0009] “The transpose, conjugate transpose, inverse, and diagonalize operations is represented as T, H, -1, and diag(·), respectively. The M x M identity matrix, and the complex Gaussian distributions with mean µ and variance σ is denoted by IM, and CN(µ, σ), respectively” and [0038] “a finite discrete alphabet set C = {c1, c2, …, c|C|}”.
Claim Rejections - 35 USC § 112b
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 1-10, 13, and 16-21 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 claims are generally narrative and indefinite, failing to conform with current U.S. practice. They appear to be a literal translation into English from a foreign document and are replete with grammatical and idiomatic errors.
Regarding claim 1,
Claim 1 recites the limitation - recovering the unknown signal vector (x) from the noisy observation vector (y) by, followed by a list of four limitations without a conjunction or other indication of how many limitations from the list are required by the claim scope. Therefore, the scope of the claim is indefinite. For examination purposes, this limitation will be interpreted as reading “- recovering the unknown signal vector (x) from the noisy observation vector (y) by performing all of the following steps:”.
Additionally, claim 1 recites the limitation - initializing a first function (X-hatIDLS-Net) computing a second function (βi,j(t)) computing a third function (b(t)), fourth function (B(t)) for every layer (t), calculating the first function (X-hatIDLS-Net) based on the result of the second function (βi,j(t)), third function (b(t)), fourth function (B(t)). It is unclear how many of the second, third, and fourth functions are computed by the first function and for every layer, or if the fourth function is initialized or calculated at all. Therefore, the scope of the claim is indefinite. For examination purposes, this limitation will be interpreted as reading “- initializing a first function (X-hatIDLS-Net), and calculating the first function (X-hatIDLS-Net) based on the results of calculating a second function (βi,j(t)), a third function (b(t)), and a fourth function (B(t)), wherein all of the functions are calculated for every layer (t)”.
Additionally, claim 1 recites the phrase the max operation is taken over the mini-batch (Dl). The term “the max operation” does not have antecedent basis as it is not previously recited. Therefore, the scope of the claim is indefinite. For examination purposes, this phrase will be interpreted as reading “a max operation”.
Additionally, claim 1 recites the phrase “every layer (t')”. However, claim 1 earlier recites “every layer (t)”. Therefore it is unclear whether each layer is represented as the symbol “t” or “t’”. Therefore, the scope of the claim is indefinite. For examination purposes, the phrase “every layer (t')” will be interpreted as reading “every layer (t)”.
Additionally, claim 1 recites the phrase “performing a standard supervised mini-batch training”. The term “standard” appears to be relative terminology, as no objective boundary is provided in the claim or the specification as to what constitutes the standard for supervised mini-batch training. Therefore, the scope of the claim is indefinite. For examination purposes, the phrase “performing a standard supervised mini-batch training” will be interpreted as reading “performing a supervised mini-batch training”.
In reference to dependent claims 2-9, claims 2-9 do not cure the deficiencies noted in the rejection of claim 1. Therefore, claims 2-9 are rejected under the same rationale as claim 1.
Regarding claim 2,
Claim 2 recites the term “the noisy observation (y)”. This term does not have antecedent basis as it is not previously recited, although claim 1 recites the similar term “the noisy observation (y) vector”. Therefore, the scope of the claim is indefinite. For examination purposes, the term “the noisy observation (y)” within claim 2 will be interpreted as reading “the noisy observation vector (y)”.
Additionally, claim 2 recites the term “the noise vector (n)”. This term does not have antecedent basis as it is not previously recited. Therefore, the scope of the claim is indefinite. For examination purposes, the term “the noise vector (n)” within claim 2 will be interpreted as reading “a noise vector (n)”.
Regarding claim 5,
Claim 5 recites the limitation wherein the first function is defined by X-hatIDLS = (AHA + σ2IN + λB)-1(AHy + λb). However, parent claim 1 recites “a first function (X-hatIDLS-Net)”. Therefore it is unclear whether the first function is “X-hatIDLS” or “X-hatIDLS-Net”. Therefore, the scope of the claim is indefinite. For examination purposes, this limitation will be interpreted as reciting “wherein the first function is defined by X-hatIDLS-Net = (AHA + σ2IN + λB)-1(AHy + λb)”.
Additionally, claim 5 recites the term “σ” within the equation “X-hatIDLS = (AHA + σ2IN + λB)-1(AHy + λb)”. This term is not previously defined within the claims. Within the specification, it is recited at [0009] “The M x M identity matrix, and the complex Gaussian distributions with mean µ and variance σ is denoted by IM, and CN(µ, σ), respectively”. However, a variance is a statistical measure of a distribution, and the specification does not state what distribution the variance in this equation is for. Therefore, the scope of the claim is indefinite. For examination purposes, it is unclear how to interpret this equation in a meaningful way given the undefined term.
Additionally, claim 5 recites the term “IN” within the equation “X-hatIDLS = (AHA + σ2IN + λB)-1(AHy + λb)”. This term is not previously defined within the claims. Within the specification, it is recited at [0009] “The M x M identity matrix, and the complex Gaussian distributions with mean µ and variance σ is denoted by IM, and CN(µ, σ), respectively”. However, while this defines “IM”, the equation recites “IN”, and it is not clear what “N” is supposed to be in this context. Therefore, the scope of the claim is indefinite. For examination purposes, it is unclear how to interpret this equation in a meaningful way given the undefined term.
Regarding claim 6,
Claim 6 recites the limitation wherein the second function is defined by βi,j = √α / |x-hatj – ci|2 + α. This recites the terms “x-hatj” and “ci”, which are not defined within the claim, the parent claim, or in the specification. Although it is clear from parent claim 1 and the specification that “x-hat” is a vector and “c” is a term within a finite discrete alphabet set, it is unclear what values result from applying the subscript “j” to x-hat and the subscript “i” to c. Therefore it is unclear what the terms “x-hatj” and “ci” represent, and therefore the scope of the claim is indefinite. For examination purposes, it is unclear how to interpret this equation in a meaningful way given the undefined terms.
Regarding claim 7,
Claim 7 recites the limitation wherein the third function is defined by b = Σ|C|i=1 ci[β2i,1, β2i,2, …, β2i,M]T. However, parent claim 1 recites “a third function (b(t))”. Therefore it is unclear whether the third function is “b” or “b(t)”. Therefore, the scope of the claim is indefinite. For examination purposes, this limitation will be interpreted as reciting “wherein the third function is defined by b(t) = Σ|C|i=1 ci[β2i,1, β2i,2, …, β2i,M]T”.
Additionally, claim 7 recites the terms “ci”, “β2i,1”, “β2i,2”, and “β2i,M”. It is unclear what values result from subscripting c and β in this way as no definitions for these terms are recited in the claim, the parent claim, or the specification. Therefore it is unclear what the terms “ci”, “β2i,1”, “β2i,2”, and “β2i,M” represent, and therefore the scope of the claim is indefinite. For examination purposes, it is unclear how to interpret this equation in a meaningful way given the undefined terms.
Regarding claim 8,
Claim 8 recites the limitation wherein the fourth function is defined by B = Σ|C|i=1 diag(β2i,1, β2i,2, …, β2i,M). However, parent claim 1 recites “fourth function (B(t))”. Therefore it is unclear whether the third function is “B” or “B(t)”. Therefore, the scope of the claim is indefinite. For examination purposes, this limitation will be interpreted as reciting “wherein the fourth function is defined by B(t) = Σ|C|i=1 diag(β2i,1, β2i,2, …, β2i,M)”.
Additionally, claim 7 recites the terms “β2i,1”, “β2i,2”, and “β2i,M”. It is unclear what values result from subscripting β in this way as no definitions for these terms are recited in the claim, the parent claim, or the specification. Therefore it is unclear what the terms “β2i,1”, “β2i,2”, and “β2i,M” represent, and therefore the scope of the claim is indefinite. For examination purposes, it is unclear how to interpret this equation in a meaningful way given the undefined terms.
Regarding claim 10,
Claim 10 recites a receiver that performs the function of the method of claim 1. All limitations of claim 1 have substantial equivalents within claim 10, therefore claim 10 is considered indefinite with an equivalent rationale and is interpreted for examination purposes in the same way.
Additionally, claim 10 recites A receiver (R) of a communication system having a processor, volatile and/or non-volatile memory, at least one interface adapted to receive a signal in a communication channel, wherein the non-volatile memory stores computer program instructions which, when executed by the microprocessor, configure the receiver to [perform the method of claim 1]. The term “the microprocessor” is recited here, but this term does not have antecedent basis as it is not previously recited, though the similar term “a processor” is. Therefore, the scope of the claim is indefinite. For examination purposes, the term “the microprocessor” within claim 10 will be interpreted as reading “the processor”.
In reference to dependent claims 16-18, claims 16-18 do not cure the deficiencies noted in the rejection of claim 10. Therefore, claims 16-18 are rejected under the same rationale as claim 10.
Regarding claim 13,
Claim 13 recites a system for performing the function of the method of claim 1. All limitations of claim 1 have substantial equivalents within claim 13, therefore claim 13 is considered indefinite with an equivalent rationale and is interpreted for examination purposes in the same way.
Additionally, claim 13 recites A data processing system characterized by having at least one hyper-parameter node, where computer-implemented operations to optimize hyper-parameters of discrete digital signal recovery for the data processing system generate distinct subprocesses and hyper-parameters λ(t) and α(t) for every layer (t) and after processing the maximum numbers of the iterations (T) the subprocess the hyper-parameter node are optimized layerwise within the data processing system, the computer-implemented operations comprising: [the method of claim 1].
The phrase “the maximum numbers of the iterations (T)” recites the term “the maximum numbers”, however this term does not have antecedent basis as it is not previously recited. Additionally, “maximum numbers of the iterations (T)” appears to be relative terminology, as the term “maximum numbers” is used without an apparent objective boundary for what causes the numbers to be maximized, in the claim or in the specification. Therefore, the scope of the claim is indefinite. For examination purposes, the phrase “maximum numbers of the iterations (T)” within claim 13 will be interpreted as reading “all of the iterations (T)”.
Additionally, the term “the subprocess” is recited in claim 13. However, only the plural “distinct subprocesses” is earlier recited in claim 13. Therefore it is unclear which subprocess of the plurality of distinct subprocesses is being referred to here. Therefore, the scope of the claim is indefinite. For examination purposes, the term “the subprocess” within claim 13 will be interpreted as reading “the distinct subprocesses”.
In reference to dependent claims 19-21, claims 19-21 do not cure the deficiencies noted in the rejection of claim 13. Therefore, claims 19-21 are rejected under the same rationale as claim 13.
Regarding claims 16 and 19,
Claims 16 and 19 recite a receiver and a system, respectively, that perform the function of the method of claim 2. All limitations of claim 2 have substantial equivalents within both claims 16 and 19, therefore claims 16 and 19 are considered indefinite with an equivalent rationale and are interpreted for examination purposes in the same way.
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 No therefor, subject to the conditions and requirements of this title.
Claims 13 and 19-21 are rejected under 35 U.S.C. 101 because the claimed invention is not directed towards one of the four statutory categories of eligible subject matter. Claims 13 and 19-21 recite a data processing system without any tangible hardware components, which is software per se. See MPEP 2106.03.
Additionally, claims 1-10, 13, and 16-21 are rejected under 35 U.S.C. 101 because the claimed invention is directed to abstract ideas without significantly more.
Regarding claim 1,
Step 1 - “Is the claim to a process, machine, manufacture or composition of matter?”
Yes, the claim is directed towards a process.
Step 2A, Prong 1 - “Is the claim directed to a law of nature, a natural phenomenon (product of nature) or an abstract idea?”:
The limitation of where the unknown signal vector (x) is composed of quantities randomly sampled from a finite discrete alphabet set (C) recites a mathematical calculation of random sampling from a set of quantities, which is a mathematical concept, which is an abstract idea.
The limitation of - recovering the unknown signal vector (x) from the noisy observation vector (y) by recites an evaluation of one quantity based on another quantity, which is a mental process, which is an abstract idea, regardless of if it’s performed on a generic computer or using a generic machine learning model.
The limitation of whereby a training data set (D) is split (L) into mini-batches (Dl) and every mini-batch (Dl) is composed of many pairs of signal vectors (x) and noisy observation vectors (y), recites an evaluation of a data set to split the data into portions, which is a mental process, which is an abstract idea, regardless of if it’s performed on a generic computer or using a generic machine learning model.
The limitation of - initializing a first function (X-hatIDLS-Net) computing a second function (βi,j(t)) computing a third function (b(t)), fourth function (B(t)) for every layer (t), calculating the first function (X-hatIDLS-Net) based on the result of the second function (βi,j(t)), third function (b(t)), fourth function (B(t)) recites a mathematical calculation of functions, which is a mathematical concept, which is an abstract idea.
The limitation of whereby a loss function (Loss (t)) is determined at the end of every layer (t') and the max operation is taken over the mini-batch (Dl) recites a mathematical calculation of a loss function and a max operation, which is a mathematical concept, which is an abstract idea.
The limitation of - updating the first hyper-parameter (λ) and the second hyper parameter (α) recites an evaluation of hyperparameters to update them, which is a mental process, which is an abstract idea, regardless of if it’s performed on a generic computer or using a generic machine learning model.
Step 2A, Prong 2 - “Does the claim recite additional elements that integrate the judicial exception into a practical application?”:
The limitation of - a processing unit (PU) receiving a noisy observation vector (y) of scalar measurements (N) from an unknown signal vector (x) linearly transformed by the measurement matrix (A), recites the mere extra-solution activity of data gathering, which does not integrate the exception into a practical application, MPEP 2106.05(d) and 2106.05(g).
The limitation of - optimizing a first hyper-parameter (λ) and a second hyper parameter (α) by performing a standard supervised mini-batch training, recites mere instructions to apply supervised mini-batch training to optimizing hyperparameters, which does not integrate the exceptions into a practical application, MPEP 2106.05(d) and 2106.05(f).
The limitation of - appending the next layer (t), recites mere instructions to apply appending of a layer, which does not integrate the exceptions into a practical application, MPEP 2106.05(d) and 2106.05(f).
The limitation of - returning the first hyper-parameter (λ) and the second hyper parameter (α) recites the mere extra-solution activity of data outputting, which does not integrate the exception into a practical application, MPEP 2106.05(d) and 2106.05(g).
Step 2B - “Does the claim recite additional elements that amount to significantly more than the judicial exception?”:
The limitation of - a processing unit (PU) receiving a noisy observation vector (y) of scalar measurements (N) from an unknown signal vector (x) linearly transformed by the measurement matrix (A), recites receiving data over a network, which is well-understood, routine, and conventional, MPEP 2106.05(d).II., example (i) of WURC computer functions.
The limitation of - optimizing a first hyper-parameter (λ) and a second hyper parameter (α) by performing a standard supervised mini-batch training, recites mere instructions to apply supervised mini-batch training to optimizing hyperparameters, which is not significantly more than any recited judicial exceptions, MPEP 2106.05(f).
The limitation of - appending the next layer (t), recites mere instructions to apply appending of a layer, which is not significantly more than any recited judicial exceptions, MPEP 2106.05(f).
The limitation of - returning the first hyper-parameter (λ) and the second hyper parameter (α) recites transmitting data over a network, which is well-understood, routine, and conventional, MPEP 2106.05(d).II., example (i) of WURC computer functions.
Therefore, claim 1 is found to be ineligible subject matter under 35 U.S.C. 101.
Regarding claim 2,
Claim 2 adds the additional limitations to claim 1:
wherein the training data set (D) is generated by randomly generation on the noisy observation (y), the measurement matrix (A), signal vector (x) and the noise vector (n) according to the relationship y = Ax + n recites a mathematical calculation of random generation based on a mathematical equation, which is a mathematical concept, which is an abstract idea.
Therefore, claim 2 is found to be ineligible subject matter under 35 U.S.C. 101.
Regarding claim 3,
Claim 3 adds the additional limitations to claim 1:
wherein the optimizing is done by deep learning techniques recites mere instructions to apply generic deep learning techniques to optimization, which does not integrate the exceptions into a practical application, and is not significantly more than any recited judicial exceptions, MPEP 2106.05(d) and 2106.05(f).
Therefore, claim 3 is found to be ineligible subject matter under 35 U.S.C. 101.
Regarding claim 4,
Claim 4 adds the additional limitations to claim 3:
wherein the optimizing is done by stochastic gradient descent and back-propagation recites mere instructions to apply stochastic gradient descent and backpropagation to optimization, which does not integrate the exceptions into a practical application, and is not significantly more than any recited judicial exceptions, MPEP 2106.05(d) and 2106.05(f).
Therefore, claim 4 is found to be ineligible subject matter under 35 U.S.C. 101.
Regarding claim 5,
Claim 5 adds the additional limitations to claim 1:
wherein the first function is defined by X-hatIDLS = (AHA + σ2IN + λB)-1(AHy + λb) recites a mathematical equation, which is a mathematical concept, which is an abstract idea.
Therefore, claim 5 is found to be ineligible subject matter under 35 U.S.C. 101.
Regarding claim 6,
Claim 6 adds the additional limitations to claim 1:
wherein the second function is defined by βi,j = √α / |x-hatj – ci|2 + α recites a mathematical equation, which is a mathematical concept, which is an abstract idea.
Therefore, claim 6 is found to be ineligible subject matter under 35 U.S.C. 101.
Regarding claim 7,
Claim 7 adds the additional limitations to claim 1:
wherein the third function is defined by b = Σ|C|i=1 ci[β2i,1, β2i,2, …, β2i,M]T recites a mathematical equation, which is a mathematical concept, which is an abstract idea.
Therefore, claim 7 is found to be ineligible subject matter under 35 U.S.C. 101.
Regarding claim 8,
Claim 8 adds the additional limitations to claim 1:
wherein the fourth function is defined by B = Σ|C|i=1 diag(β2i,1, β2i,2, …, β2i,M) recites a mathematical equation, which is a mathematical concept, which is an abstract idea.
Therefore, claim 8 is found to be ineligible subject matter under 35 U.S.C. 101.
Regarding claim 9,
Claim 9 adds the additional limitations to claim 1:
wherein the data processing system is a communication system … recites a mere field of use of communications for the data processing system, which does not integrate the exceptions into a practical application, and is not significantly more than any recited judicial exceptions, MPEP 2106.05(d) and 2106.05(h).
…and processing unit (PU) is a user equipment (UE), in view of the specification at [0074] “A user equipment means the equipment designed for consumer use. It is any device used by an end user such as a smart phone or other mobile device, laptop, or tablet equipped with at least one wired or wireless broadband adapter”, recites mere instructions to apply the exceptions with generic computers, which does not integrate the exceptions into a practical application, and is not significantly more than any recited judicial exceptions, MPEP 2106.05(d) and 2106.05(f).
Therefore, claim 9 is found to be ineligible subject matter under 35 U.S.C. 101.
Regarding claims 10 and 16-18,
Claims 10 and 16-18 recite a machine, specifically a receiver for a communication system comprising a processor and a memory with instructions that implements the function of the method of claims 1-4, respectively. Therefore the same analysis and rejection applied to claims 1-4 applies to claims 10 and 16-18.
Therefore, claims 10 and 16-18 are found to be ineligible subject matter under 35 U.S.C. 101.
Regarding claims 13 and 19-21,
Claims 13 and 19-21 recite software per se, specifically a data processing system that implements the function of the method of claims 1-4, respectively. Therefore the same analysis and rejection applied to claims 1-4 applies to claims 13 and 19-21.
Therefore, claims 13 and 19-21 are found to be ineligible subject matter under 35 U.S.C. 101.
Prior Art
The following references are used for prior art claim rejections:
Takabe et al. “Deep-Unfolded Sparse CDMA: Multiuser Detector and Sparse Signature Design”, hereinafter Takabe
Iimori et al. “Discreteness-aware Receivers for Overloaded MIMO Systems”, hereinafter Iimori_1
Jin and Kim “Deep Learning Detection Networks in MIMO Decode-Forward Relay Channels”, hereinafter Jin
Iimori et al. “Robust Symbol Detection in Overloaded NOMA Systems”, hereinafter Iimori_2
Balatsoukas-Stimming et al. “Neural-Network Optimized 1-bit Precoding for Massive MU-MIMO”, hereinafter Balatsoukas-Stimming
Claim Rejections - 35 USC § 103
The following is a quotation of 35 U.S.C. 103 which forms the basis for all obviousness rejections set forth in this Office action:
A patent for a claimed invention may not be obtained, notwithstanding that the claimed invention is not identically disclosed as set forth in section 102, if the differences between the claimed invention and the prior art are such that the claimed invention as a whole would have been obvious before the effective filing date of the claimed invention to a person having ordinary skill in the art to which the claimed invention pertains. Patentability shall not be negated by the manner in which the invention was made.
Claims 1-4, 6-8, 10, 13, and 16-21 are rejected under 35 U.S.C. 103 as being unpatentable over Takabe in view of Iimori_1, further in view of Jin.
Regarding claim 1,
Takabe teaches A computer-implemented method to optimize hyperparameters of discrete digital signal recovery for a data processing system that is characterized by a measurement matrix (A), ((Takabe Pg. 8) “We can modify the training process of the STPG and C-STPG detector such that not only the trainable parameters of the detector but also the signature matrix A are updated by an optimizer”) the method comprising:
- a processing unit (PU) receiving a noisy observation vector (y) of scalar measurements (N) from an unknown signal vector (x) linearly transformed by the measurement matrix (A), ((Takabe Pg. 3) “The SCDMA system discussed herein is defined as follows…Under these assumptions, the received signal y = (y1, ..., yM)T is given by [Equation 1] where SNR represents the signal-to-noise ratio (SNR) of the system and w ∈ CM is an additive white Gaussian noise vector with zero mean and unit variance. Equivalently, using the signature matrix A, the system can be concisely represented by [Equation 2]”)
PNG
media_image1.png
57
302
media_image1.png
Greyscale
PNG
media_image2.png
52
295
media_image2.png
Greyscale
- recovering the unknown signal vector (x) from the noisy observation vector (y) by ((Takabe Abstract) “In SCDMA, transmitted symbols from multiple users are coded by their own sparse signature sequences, and a base station attempts to detect those symbols using the signature sequences”, (Takabe Pg. 3) “The SCDMA system discussed herein is defined as follows. Assume that N users try to send their own symbol xi ∈ X (i = 1, ..., N) in the signal constellation X ⊂ C to a BS…Under these assumptions, the received signal y = (y1, ..., yM)T is given by [Equation 1]”)
- optimizing a first hyper-parameter (λ) and a second hyper parameter (α) by performing a standard supervised mini-batch training, ((Takabe Pg. 4) “The update rule of an STPG detector is given as follows:…where {γt}Tt=1 and α are trainable parameters…These parameters are learned by using randomly generated supervised data and standard deep-learning techniques such as back propagation and SGD”, (Takabe Pg. 5) “For the STPG detector, the initial values of the trainable parameters are set to γ2t =0.01 (t = 1, ..., T), and α = 2. In each generation, we prepare 100 mini batches of size 200 containing a pairing of a transmit signal x-tilde and received signal y-tilde, and train parameters using the Adam optimizer”)
whereby a training data set (D) is split (L) into mini-batches (Dl) and every mini-batch (Dl) is composed of many pairs of signal vectors (x) and noisy observation vectors (y), ((Takabe Pg. 5) “For the STPG detector, the initial values of the trainable parameters are set to γ2t =0.01 (t = 1, ..., T), and α = 2. In each generation, we prepare 100 mini batches of size 200 containing a pairing of a transmit signal x-tilde and received signal y-tilde, and train parameters using the Adam optimizer”)
whereby a loss function (Loss (t)) is determined at the end of every layer (t') … (Takabe Pg. 9, Algorithm 1 shows that a loss, using the MSE (mean squared error) function, is determined at line 9)
PNG
media_image3.png
657
457
media_image3.png
Greyscale
- updating the first hyper-parameter (λ) and the second hyper parameter (α) (Takabe Pg. 9, Algorithm 1 shows that the parameters {γt}, α, and A are updated at line 10)
- appending the next layer (t), (Takabe Pg. 9, Algorithm 1 shows that t is incremented from 1 until it reaches T every loop at line 2, (Takabe Pg. 6) “Figure 3 shows the block diagram of the C-STPG detector in the Tth iteration (layer)”)
PNG
media_image4.png
198
405
media_image4.png
Greyscale
- returning the first hyper-parameter (λ) and the second hyper parameter (α) (Takabe Pg. 9, Algorithm 1 shows that the output of the algorithm consists of trained parameters {γt}, α, and A)
Iimori_1 teaches the following further limitations that Takabe doesn’t teach or more explicitly than Takabe teaches:
where the unknown signal vector (x) is composed of quantities randomly sampled from a finite discrete alphabet set (C) ((Iimori_1 Pg. 2) “where the transmit symbol vector s = [s1, …, sNt]T ∈ CNt x 1 is normalized…such that each of its elements is sampled from the same discrete and regular quadrature amplitude modulation (QAM) constellation set C = {c1, …, c2b} of cardinality 2b, with b denoting the number of bits per symbol”)
- initializing a first function (X-hatIDLS-Net) computing a second function (βi,j(t)) computing a third function (b(t)), fourth function (B(t)) for every layer (t), ((Iimori_1 Pg. 9) “thanks to the quadratic shape of the function q(s), equation (28) can be solved in closed form by setting its Wirtinger derivative with respect to s equal to 0, that is, [Equation 29]”, Equation 29 depends on values calculated in equations 24, 25, and 26)
PNG
media_image5.png
32
346
media_image5.png
Greyscale
PNG
media_image6.png
65
428
media_image6.png
Greyscale
PNG
media_image7.png
61
351
media_image7.png
Greyscale
PNG
media_image8.png
63
368
media_image8.png
Greyscale
calculating the first function (X-hatIDLS-Net) based on the result of the second function (βi,j(t)), third function (b(t)), fourth function (B(t)) (Iimori_1 Pg. 5, Algorithm 1 computes sk based on equation 29, which depends on equations 24, 25, and 26)
PNG
media_image9.png
396
445
media_image9.png
Greyscale
At the time of filing, one of ordinary skill in the art would have motivation to combine Takabe and Iimori_1 by taking the method for optimizing hyperparameters for discrete digital signal recovery, including a signal vector, taught by Takabe, and including the signal vector being sampled from a finite discrete set, and computing a first function based on a second function (βi,j(t)), a third function (b(t)), and a fourth function (B(t)), for every layer (t), taught by Iimori_1, as Iimori_1 teaches (Iimori_1 Pg. 6) “the overloaded multi-user signal detection scheme derived in this subsection, expressed compactly in equation (29) and summarized in the pseudocode offered in Algorithm 1 is referred to as the discreteness-aware penalized zero-forcing (DAPZF) detector” and “the proposed DAPZF method outperforms the SotA methods in all cases”, that is, the detector function of Iimori_1 outperforms other state of the art methods for signal detection, and it is thus a natural choice to use it with the optimized hyperparameters produced by Takabe’s method, substituting these optimized hyperparameters for the hyperparameters used within Iimori_1’s Algorithm 1. Such a combination would be obvious.
Jin teaches the following further limitations that neither Takabe, nor Iimori_1 teach:
… and the max operation is taken over the mini-batch (Dl) (Jin Pg. 7, Algorithm 1 shows generation of batches at line 3 and computation using the samples in the batch for every layer at lines 5 and 6, (Jin Pg. 4) “The L iterations are unfolded to the L neural-network layers (NNLs)…the main parts in the kth NNL include uk = Relu(W1kik + b1k)…where Relu(x) = max(0,x)”, corresponding to a max operation being taken over each batch at every layer during training)
PNG
media_image10.png
665
427
media_image10.png
Greyscale
At the time of filing, one of ordinary skill in the art would have motivation to combine Takabe, Iimori_1, and Jin by taking the method for optimizing hyperparameters for discrete digital signal recovery, including training the optimizing method with mini-batches over a number of layers, jointly taught by Takabe and Iimori_1, and including performing a max operation over the mini-batch, as part of a Relu layer, taught by Jin, as Relu layers are a well-known and popular activation function for layers of neural networks in the art, as it is relatively easy to calculate and is not as prone to vanishing gradient problems compared to other common activation functions. Such a combination would be obvious.
Regarding claim 2,
Takabe, Iimori_1, and Jin jointly teach The method of claim 1,
Takabe further teaches:
wherein the training data set (D) is generated by randomly generation ((Takabe Pg. 4) “These parameters are learned by using randomly generated supervised data”, randomly generated supervised data for learning parameters corresponds to a randomly generated training data set) on the noisy observation (y), (Takabe Pg. 9, Algorithm 1 shows at line 10 that y is generated from an equation using variables that are randomly generated) the measurement matrix (A), ((Takabe Pg. 4) “In the numerical simulations…A signature matrix A is randomly generated”) signal vector (x) (Takabe Pg. 9, Algorithm 1 shows at line 9 that x is randomly generated) and the noise vector (n) ((Takabe Pg. 3) “w is a complex Gaussian random vector with zero mean and unit variance”) according to the relationship y = Ax + n ((Takabe Pg. 3) “using the signature matrix A, the system can be concisely represented by [Equation 2]”, Iimori_1 but not Takabe teaches the relationship y = Ax + n verbatim)
At the time of filing, one of ordinary skill in the art would have motivation to combine the method jointly taught by Takabe, Iimori_1, and Jin for the parent claim of claim 2, claim 1. No new embodiments are introduced, so the reason to combine is the same as for the parent claim.
Regarding claim 3,
Takabe, Iimori_1, and Jin jointly teach The method of claim 1,
Takabe further teaches:
wherein the optimizing is done by deep learning techniques ((Takabe Pg. 4) “The update rule of an STPG detector is given as follows:…where {γt}Tt=1 and α are trainable parameters…These parameters are learned by using randomly generated supervised data and standard deep-learning techniques such as back propagation and SGD”)
At the time of filing, one of ordinary skill in the art would have motivation to combine the method jointly taught by Takabe, Iimori_1, and Jin for the parent claim of claim 3, claim 1. No new embodiments are introduced, so the reason to combine is the same as for the parent claim.
Regarding claim 4,
Takabe, Iimori_1, and Jin jointly teach The method of claim 3,
Takabe further teaches:
wherein the optimizing is done by stochastic gradient descent and back-propagation ((Takabe Pg. 4) “The update rule of an STPG detector is given as follows:…where {γt}Tt=1 and α are trainable parameters…These parameters are learned by using randomly generated supervised data and standard deep-learning techniques such as back propagation and SGD”)
At the time of filing, one of ordinary skill in the art would have motivation to combine the method jointly taught by Takabe, Iimori_1, and Jin for the parent claim of claim 4, claim 3. No new embodiments are introduced, so the reason to combine is the same as for the parent claim.
Regarding claim 6,
Takabe, Iimori_1, and Jin jointly teach The method of claim 1,
Takabe further teaches:
wherein the second function is defined by βi,j = √α / |x-hatj – ci|2 + α (Iimori_1 Pg. 5, Equation 24 defines an analogous function)
PNG
media_image6.png
65
428
media_image6.png
Greyscale
At the time of filing, one of ordinary skill in the art would have motivation to combine the method jointly taught by Takabe, Iimori_1, and Jin for the parent claim of claim 6, claim 1. No new embodiments are introduced, so the reason to combine is the same as for the parent claim.
Regarding claim 7,
Takabe, Iimori_1, and Jin jointly teach The method of claim 1,
Takabe further teaches:
wherein the third function is defined by b = Σ|C|i=1 ci[β2i,1, β2i,2, …, β2i,M]T (Iimori_1 Pg. 5, Equation 25 defines an analogous function)
PNG
media_image7.png
61
351
media_image7.png
Greyscale
At the time of filing, one of ordinary skill in the art would have motivation to combine the method jointly taught by Takabe, Iimori_1, and Jin for the parent claim of claim 7, claim 1. No new embodiments are introduced, so the reason to combine is the same as for the parent claim.
Regarding claim 8,
Takabe, Iimori_1, and Jin jointly teach The method of claim 1,
Takabe further teaches:
wherein the fourth function is defined by B = Σ|C|i=1 diag(β2i,1, β2i,2, …, β2i,M) (Iimori_1 Pg. 5, Equation 26 defines an analogous function)
PNG
media_image8.png
63
368
media_image8.png
Greyscale
At the time of filing, one of ordinary skill in the art would have motivation to combine the method jointly taught by Takabe, Iimori_1, and Jin for the parent claim of claim 8, claim 1. No new embodiments are introduced, so the reason to combine is the same as for the parent claim.
Regarding claims 10 and 16-18,
Claims 10 and 16-18 recite a receiver for a communication system comprising a processor and a memory with instructions for performing the function of the method of claims 1-4, respectively. Specifically, claim 10 recites A receiver (R) of a communication system having a processor, volatile and/or non-volatile memory, at least one interface adapted to receive a signal in a communication channel, wherein the non-volatile memory stores computer program instructions which, when executed by the microprocessor, configure the receiver to [perform the method of claim 1]. Iimori_1 recites: (Iimori_1 Abstract) “We describe three new high-performance receivers suitable for symbol detection of large-scaled and overloaded multidimensional wireless communication systems”,
and Takabe recites: (Takabe Pg. 4) “the BP decoder with 30 iterations executes in about 2 seconds on a PC with Intel Core i9-10850K CPU”, with a CPU comprising both a processor and memory.
All other limitations in claims 10 and 16-18 are substantially the same as those in claims 1-4, respectively, therefore the same rationale for rejection applies.
Regarding claims 13 and 19-21,
Claims 13 and 19-21 recite a data processing system for performing the function of the method of claims 1-4, respectively. Specifically, claim 13 recites A data processing system characterized by having at least one hyper-parameter node, where computer-implemented operations to optimize hyper-parameters of discrete digital signal recovery for the data processing system generate distinct subprocesses and hyper-parameters λ(t) and α(t) for every layer (t) and after processing the maximum numbers of the iterations (T) the subprocess the hyper-parameter node are optimized layerwise within the data processing system, the computer-implemented operations comprising: [the method of claim 1]. Takabe recites: (Takabe Pg. 6) “Figure 3 shows the block diagram of the C-STPG detector in the Tth iteration (layer)”, Takabe Pg. 6, Fig. 3 shows nodes including hyperparameters γt and α, and Takabe Pg. 9, Algorithm 1 shows that different loop iterations are done for every layer t at line 2, corresponding to distinct subprocesses, and that the hyperparameters are optimized at line 10.
All other limitations in claims 13 and 19-21 are substantially the same as those in claims 1-4, respectively, therefore the same rationale for rejection applies.
Claim 5 is rejected under 35 U.S.C. 103 as being unpatentable over Takabe in view of Iimori_1, further in view of Jin, further in view of Iimori_2.
Regarding claim 5,
Takabe, Iimori_1, and Jin jointly teach The method of claim 1,
Iimori_2 teaches the following further limitations that neither Takabe, nor Iimori_1, nor Jin teach:
wherein the first function is defined by X-hatIDLS = (AHA + σ2IN + λB)-1(AHy + λb) (Iimori_2 Pg. 26, Equation 49a defines an analogous function)
PNG
media_image11.png
43
595
media_image11.png
Greyscale
At the time of filing, one of ordinary skill in the art would have motivation to combine Takabe, Iimori_1, and Jin by taking the method for optimizing hyperparameters for discrete digital signal recovery of claim 1, including a function to detect a received signal, jointly taught by Takabe, Iimori_1, and Jin, and defining the function as X-hatIDLS = (AHA + σ2IN + λB)-1(AHy + λb), taught by Iimori_2, as doing so is a simple substitution of one signal computation function for another, yielding the predictable results of also computing the received signal. Such a combination would be obvious.
Claim 9 is rejected under 35 U.S.C. 103 as being unpatentable over Takabe in view of Iimori_1, further in view of Jin, further in view of Balatsoukas-Stimming.
Regarding claim 9,
Takabe, Iimori_1, and Jin jointly teach The method of claim 1,
Takabe further teaches:
wherein the data processing system is a communication system … ((Iimori_1 Abstract) “We describe three new high-performance receivers suitable for symbol detection of large-scaled and overloaded multidimensional wireless communication systems”)
Balatsoukas-Stimming teaches the following further limitations that neither Takabe, nor Iimori_1, nor Jin explicitly teach:
wherein the data processing system is a communication system and processing unit (PU) is a user equipment (UE) ((Balatsoukas-Stimming Pg. 2) “We focus on the downlink of a single-cell, narrowband massive MU-MIMO system as shown in Figure 1, where a BS equipped with B antennas serves U << B UEs with one antenna each. This system is modeled as y = Hx+n, where y…comprises the signals received by each UE”, (Balatsoukas-Stimming Pg. 1) “By equipping the base station (BS) with hundreds of antennas, massive MU-MIMO permits communication with tens of user equipments (UEs)”)
At the time of filing, one of ordinary skill in the art would have motivation to combine Takabe, Iimori_1, Jin, and Balatsoukas-Stimming by taking the method for optimizing hyperparameters for discrete digital signal recovery of claim 1, including a communication system, jointly taught by Takabe, Iimori_1, and Jin, and having the signal processing take place on user equipment, taught by Balatsoukas-Stimming, as communication systems incorporating the devices of users is well-known within the art, and confers the predictable benefit of allowing the communication system to provide service to users. Such a combination would be obvious.
Conclusion
The prior art made of record and not relied upon is considered pertinent to applicant's disclosure.
Muraoka et al. (U.S. Patent Application Publication No. 2022/0337341) teaches a method for unfolding an iterative algorithm and producing a set of learned parameters of the algorithm.
Balatsoukas-Stimming and Studer “Deep Unfolding for Communications Systems: A Survey and Some New Directions” teaches an overview of deep unfolding techniques applied to communication systems.
Any inquiry concerning this communication or earlier communications from the examiner should be directed to VICTOR A NAULT whose telephone number is (703) 756-5745. The examiner can normally be reached M - F, 12 - 8.
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.
/V.A.N./Examiner, Art Unit 2124
/Kevin W Figueroa/Primary Examiner, Art Unit 2124
Prosecution Insights