DETAILED ACTION
This Office Action is responsive to the claims filed on: 03/05/2024.
Claims 1-17 are pending for Examination.
Notice of Pre-AIA or AIA Status
The present application, filed on or after March 16, 2013, is being examined under the first inventor to file provisions of the AIA .
Information Disclosure Statements
The information disclosure statement (IDS) submitted on: 03/05/2024 is determined to be compliance with the provisions of 37 CFR 1.97. Accordingly, this IDS is being considered by the Examiner.
Claim Interpretation – Alternative Claim Language
The claims of the instant application are given their Broadest Reasonable Interpretation (BRI) using the plain meaning of the claim language in light of the specification, as it would be understood by one of ordinary skill in the art. Accordingly, the BRI of an alternative claim limitation or term can be determined to be the least-limiting interpretation, consistent with the specification. In this context, the term “or” by plain meaning can be interpreted to alternatively be: one or the other (i.e., A or B), but not both (i.e., not A and B). The term “and/or” by plain meaning can be interpreted to be: “and” or alternatively “or,” but not both, as this would not make sense. In this context, the forward-slash “/” is equivalent to the alternative “or.” Likewise, the alternative terms “at least one of,” “one or more of,” and the like, followed by multiple alternative claim limitations can be reasonably interpreted to be only “one of” a group of alternative claim limitations.
Prior art disclosing any one of multiple alternative claim limitations discloses matter within the scope of the claimed invention. "When a claim covers several structures or compositions, either generically or as alternatives, the claim is deemed anticipated if any of the structures or compositions within the scope of the claim is known in the prior art." Brown v. 3M, 265 F.3d 1349, 1351, 60 USPQ2d 1375, 1376 (Fed. Cir. 2001) (claim to a system for setting a computer clock to an offset time to address the Year 2000 (Y2K) problem, applicable to records with year date data in "at least one of two-digit, three-digit, or four-digit" representations, was held anticipated by a system that offsets year dates in only two-digit formats). See MPEP 2131.
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.
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.
Claims 1-2, 5, 7-8, and 14-17 are rejected under 35 U.S.C. 103 as being unpatentable over US PG Pub. 2022/0200669 A1, Banuli Nanje Gowda et al. (hereinafter “Gowda”), in view of WO 2021/112732 A1, Girnyk.
With respect to claim 1, Gowda teaches:
A radio transmitter device (para. [0059]-[0062] and [0071]; and RRH/BS of Figs. 1, 17, 34, and 39), comprising:
a transmit antenna array comprising at least two transmit antennas (paras. [0065]-[0067]; Tx antenna array 202/208 having multiple transmit antennas of Fig. 2 and antenna 3912 of Fig. 39);
at least one processor (paras. [0448]-[0452]; controller/processor of Fig. 34, and baseband CPU 3908E of Figs. 39-40 —a BS controller can be a parallel-processor capable of performing precoding); and
at least one non-transitory memory storing instructions that, when executed with the at least one processor (paras. [0448]-[0452]; and memory 3908G of Figs. 39-40—a BS baseband circuitry can include memory coupled to the CPU), cause the radio transmitter device to at least perform:
receiving uplink channel information (paras. [0062], [0080], [0486]-[0487], and [0535]; and UEs 318 of Fig. 3; and block 502 of Fig. 5 —a RRH/BS can receive UL channel information from a UE device or based on sensed UL channel, i.e., via a received SRS);
determining resource element specific precoding matrices for a downlink channel based on the received uplink channel information (paras. [0062]-[0064], [0080]-[0082], [0493], and [0535]; blocks 504 and 506 of Fig. 5, and blocks 602 and 604 of Fig. 6 —RE precoding matrices can be determined for a DL channel transmission(s) based on weights determined from received/sensed UCI and/or plink SRS’); and
generating transmit antenna specific output signals for the transmit antenna array based on the determined resource element specific precoding matrices and symbols to be transmitted (paras. [0062]-[0065], [0080]-[0082], [0113], and [0535]; and DL antenna array of Fig. 2 —a RRH/BS can use its determined RE precoding matrices to generate and then transmit DL signaling via its Tx antenna array, as depicted in Fig. 2);
wherein the determining of the resource element specific precoding matrices for the downlink channel based on the received uplink channel information is performed. (paras. [0080]-[0082], [0418], [0412], [0429]-[0430], and [0474]; and 3722 of Fig. 37 —received UL channel information can be used to determine RE precoding matrices for DL signaling),
However, Gowda does not explicitly teach:
applying a neural network to the received uplink channel information; and
the neural network comprising at least one neural network layer executable to process the received uplink channel information to output the resource element specific precoding matrices for the downlink channel.
Girnyk does teach:
applying a neural network to received uplink channel information, the neural network comprising at least one neural network layer executable to process the received uplink channel information to output resource element specific precoding matrices for the downlink channel (p. 3, lines 8-27; p. 10, lines 23-28; p. 12, ln. 27 to p. 13, ln. 11; and p. 13, ln. 29, to p. 14, ln. 10 —for TDD operation w/reciprocity between DL-UL transmission of a channel pair, a CSI learned in the UL, can be applied for DL estimation of channel state, relating to matrix H —the precoder matrix G can then be optimized/output for the DL channel, by applying a neural network layer to the channel realization H input).
It would have been prima-facie obvious to one having ordinary skill in the art before the effective filing date of the claimed invention to have modified Gowda’s neural network solution for optimizing DL antenna precoder matrices, with application of a neural network to UL CSI input, as taught by Girnyk.
The motivation for doing so would have been to improve DL precoder matrices determination by considering DL-UL CSI reciprocity, when applying neural network solutions, as recognized by Girnyk (p. 3, lines 8-27; p. 10, lines 23-28; p. 12, ln. 27 to p. 13, ln. 11; and p. 13, ln. 29, to p. 14, ln. 10).
With respect to claim 2, Gowda in view of Girnyk teaches:
The radio transmitter device according to claim 1, wherein the neural network comprises at least one of a convolutional neural network, a transformer neural network, or a combination thereof (Gowda: paras. [0126]-[0127] and [0429] —the neural network can be a convolutional neural network (CNN) —the alternative term “or” only requires examination on-the-merits of a single claimed alternative for the reasons explained above in the Claim Interpretation — Alternative Claim Language section).
With respect to claim 5, Gowda in view of Girnyk teaches:
The radio transmitter device according to claim 1, wherein the instructions, when executed with the at least one processor, cause the radio transmitter device to perform the determining of the resource element specific precoding matrices further with
applying a zero-forcing transformation or an approximation of the zero-forcing transformation to the output of the neural network (Gowda: paras. [0063]-[0064], [0077]-[0078], [0193], and [0474] —a regularized zero forcing (RZF) transform can be applied to the precoding matrices output of the NN for estimating the DL channel—the alternative term “or” only requires examination on-the-merits of a single claimed alternative for the reasons explained above in the Claim Interpretation — Alternative Claim Language section).
With respect to claim 7, Gowda in view of Girnyk teaches:
The radio transmitter device according to claim 1,wherein the uplink channel information comprises uplink channel estimate information provided with a channel estimator (Gowda: paras. [0062]-[0063], [0072]-[0073], and [0484]-[0486]; and 116 of Fig. 1, 333 of Fig. 3 —an uplink channel estimate information can be determined explicitly or via DL-UL channel reciprocity).
With respect to claim 8, Gowda in view of Girnyk teaches:
The radio transmitter device according to claim 1, wherein the uplink channel information comprises uplink channel estimate information provided with a radio receiver device utilizing an iterative neural network to generate the uplink channel estimate information (Gowda: paras. [0059]-[0060], [0062]-[0063], [0074]-[0078]; and SRS channel estimator 116 of Fig. 1 —uplink channel information can be estimated from the SRS’ received via a RRH, which can employ a NN to generate channel estimate information for determining the channel state; Girnyk: p. 3, lines 7-33, p. 5, and pp. 12-14, also describes an iterative approach to UL channel estimation that can be processed by a NN.).
With respect to claim 14, Gowda in view of Girnyk teaches:
The radio transmitter device according to claim 1, wherein the radio transmitter device comprises a time division duplexing capable radio transmitter device (Gowda: paras. [0478] and [0486]; and Figs. 1-2 and 39-40 —the RRH can be TDD-capable device, where channel reciprocity operations may be assumed).
With respect to claim 15, Gowda in view of Girnyk teaches:
The radio transmitter device according to claim 1, wherein the radio transmitter device comprises a multiple-input and multiple-output capable radio transmitter device (Gowda: paras. [0064]-[0067], [0070], and [0482]-[0486]; and Figs. 1-2 and 39-40 —the RRH can be a MIMO-capable NR device).
With respect to claim 16, this claim recites similar features to independent claim 1, except claim 16 is directed to a method. As such, claim 16 is likewise rejected under §103 based on Gowda in view of Girnyk, for the same reasons explained above for independent claim 1.
With respect to claim 17, this claim recites similar features to independent claim 1, except claim 17 is directed to non-transitory program storage device (Gowda: paras. [0448]-[0452]; and memory 3908G of Figs. 39-40 —a BS baseband circuitry can include memory coupled to the CPU). As such, claim 17 is likewise rejected under §103 based on Gowda in view of Girnyk, for the same reasons explained above for independent claim 1.
Claim 3 is rejected under 35 U.S.C. 103 as being unpatentable over Gowda in view of Girnyk, in further view of US PG Pub. 2024/0137082 A1, Shi et al. (hereinafter “Shi”).
With respect to claim 3, Gowda in view of Girnyk teaches he radio transmitter device according to claim 1.
However, Gowda in view of Girnyk does not explicitly teach:
wherein the neural network utilizes residual connections.
Shi does teach:
wherein the neural network utilizes residual connections (para. [0594] —the RefineNet neural network can be configured to use residual connections).
It would have been prima-facie obvious to one having ordinary skill in the art before the effective filing date of the claimed invention to have modified Gowda in view of Girnyk’s neural network solution for optimizing DL antenna precoder matrices, to include residual connections, such as in the RefineNet solution, taught by Shi.
The motivation for doing so would have been to improve neural network determination through the use of a large quantity of residual connections within a network model, as recognized by Shi (para. [0594]).
Claim 4 is rejected under 35 U.S.C. 103 as being unpatentable over Gowda in view of Girnyk, in further view of US PG Pub. 2023/0004350 A1, Li.
With respect to claim 4, Gowda in view of Girnyk teaches the radio transmitter device according to claim 2.
However, Gowda in view of Girnyk does not explicitly teach:
wherein at least one of the at least one neural network layers utilizes depthwise separable convolution.
Li does teach:
a neural network layer that utilizes depthwise separable convolution (para. [0127] —weights for a depthwise portion of a depthwise separable convolution (e.g., for a convolutional neural network layer) can be processed for one or more NN determinations).
It would have been prima-facie obvious to one having ordinary skill in the art before the effective filing date of the claimed invention to have modified Gowda in view of Girnyk’s neural network solution to utilizes depthwise separable convolution, taught by Li.
The motivation for doing so would have been to improve neural network determination through the use known depthwise separable convolution techniques, as recognized by Li (para. [0127]).
Claim 6 is rejected under 35 U.S.C. 103 as being unpatentable over Gowda in view of Girnyk, in further view of US PG Pub. 2023/0186079 A1, Lee et al. (hereinafter “Lee”).
With respect to claim 6, Gowda in view of Girnyk teaches the radio transmitter device according to claim 1.
However, Gowda in view of Girnyk does not explicitly teach:
wherein the radio transmitter device performs the determining of the resource element specific precoding matrices further based on a prediction length.
Lee does teach:
determining of resource element specific precoding matrices based on a prediction length (paras. [0007], [0013]-[0014], [0055], [0060], [0067], and [0073] —RE matrices can be determined based on a set time period(s)/episode(s), i.e., a prediction length, of an iterative NN training procedure).
It would have been prima-facie obvious to one having ordinary skill in the art before the effective filing date of the claimed invention to have modified Gowda in view of Girnyk’s neural network solution for determining RE specific matrices, with determining the matrices based on a prediction time(s)/period(s), taught by Lee.
The motivation for doing so would have been to improve NN determination of RE matrices using a set prediction time(s)/length(s) for iterative training procedures, as recognized by Lee (paras. [0007], [0013]-[0014], [0055], [0060], [0067], and [0073]).
Claims 9-12 are rejected under 35 U.S.C. 103 as being unpatentable over Gowda in view of Girnyk, in further view of US PG Pub. 2019/0274108 A1, O’Shea et al. (hereinafter “O’Shea”).
With respect to claim 9, Gowda in view of Girnyk teaches the radio transmitter device according to claim 1.
However, Gowda in view of Girnyk does not explicitly teach:
the radio transmitter device performing training the neural network with differentiating through a simulated channel.
O’Shea does teach:
performing training of a NN by differentiating through a simulated channel (paras. [0207]-[0208], [0211]-[0212], and [0223]; and 510 of Fig. 5, 706/712 of Figs. 7-8, and 922 of Fig. 9 —a NN can be trained via differentiating a simulated channel from a real world channel).
It would have been prima-facie obvious to one having ordinary skill in the art before the effective filing date of the claimed invention to have modified Gowda in view of Girnyk’s neural network solution for determining RE specific matrices, with training of a NN by differentiating via a simulated channel, taught by O’Shea.
The motivation for doing so would have been to improve NN determination of encoder/decoder mappings by differentiating simulated from real-world channel signaling, as recognized by O’Shea (paras. [0207]-[0208], [0211]-[0212], and [0223]; and 510 of Fig. 5, 706/712 of Figs. 7-8, and 922 of Fig. 9).
With respect to claim 10, Gowda in view of Girnyk and O’Shea teaches the radio transmitter device according to claim 9.
However, Gowda does not explicitly teach:
a simulated channel that is based on at least one of a statistically simulated channel, a raytraced channel, or a captured channel.
O’Shea does teach:
a simulated channel that is based on at least one of a statistically simulated channel, a raytraced channel, or a captured channel (paras. [0207]-[0208], [0211]-[0212], and [0223]; and 510 of Fig. 5, 706/712 of Figs. 7-8, and 922 of Fig. 9 —a simulated channel can be based on captured/measured real-world channel and/or a statistical representation thereof in terms of percentage of similarity thereto —the alternative terms “at least one of” and “or” only require examination on-the-merits of a single claimed alternative, for the reasons explained above in the Claim Interpretation — Alternative Claim Limitations section).
It would have been prima-facie obvious to one having ordinary skill in the art before the effective filing date of the claimed invention to have modified Gowda’s neural network solution for determining RE specific matrices, with training of a NN by differentiating via a simulated channel, taught by O’Shea.
The motivation for doing so would have been to improve NN determination of encoder/decoder mappings by differentiating simulated from real-world channel signaling, as recognized by O’Shea (paras. [0207]-[0208], [0211]-[0212], and [0223]; and 510 of Fig. 5, 706/712 of Figs. 7-8, and 922 of Fig. 9).
With respect to claim 11, Gowda in view of Girnyk and O’Shea teaches the radio transmitter device according to claim 9.
However, Gowda does not explicitly teach:
training the neural network with applying a loss.
O’Shea does teach:
training a neural network by applying a loss (paras. [0102], [0111]-[0112], [0139], [0212] and [0217] —the neural network can be trained applying one or more loss functions, i.e., to minimize loss between a simulated/approximated channel and a real world channel).
It would have been prima-facie obvious to one having ordinary skill in the art before the effective filing date of the claimed invention to have modified Gowda’s neural network solution for determining RE specific matrices, with training of a NN by differentiating via a simulated channel, taught by O’Shea.
The motivation for doing so would have been to improve NN determination of encoder/decoder mappings by differentiating simulated from real-world channel signaling, as recognized by O’Shea (paras. [0207]-[0208], [0211]-[0212], and [0223]; and 510 of Fig. 5, 706/712 of Figs. 7-8, and 922 of Fig. 9).
With respect to claim 12, Gowda in view of Girnyk and O’Shea teaches the radio transmitter device according to claim 11.
However, Gowda does not explicitly teach:
wherein the loss comprises a sum of one or more cross-entropy losses.
O’Shea does teach:
wherein the loss comprises a sum of one or more cross-entropy losses (paras. [0143]-[0144], [0198], [0237], and [0257] —the neural network can be trained applying entropy-loss functions to multiple parameters, collectively, i.e., to minimize losses between a simulated channel and a real world channel).
It would have been prima-facie obvious to one having ordinary skill in the art before the effective filing date of the claimed invention to have modified Gowda’s neural network solution for determining RE specific matrices, with training of a NN by differentiating via a simulated channel, taught by O’Shea.
The motivation for doing so would have been to improve NN determination of encoder/decoder mappings by differentiating simulated from real-world channel signaling, as recognized by O’Shea (paras. [0207]-[0208], [0211]-[0212], and [0223]; and 510 of Fig. 5, 706/712 of Figs. 7-8, and 922 of Fig. 9).
Claim 13 is rejected under 35 U.S.C. 103 as being unpatentable over Gowda in view of Girnyk and O’Shea, in further view of US PG Pub. 2021/0304891 A1, Kozloski et al. (hereinafter “Kozloski”).
With respect to claim 9, Gowda in view of Girnyk and O’Shea teaches the radio transmitter device according to claim 11.
However, Gowda in view of Girnyk and O’Shea does not explicitly teach:
training the neural network with optimizing the loss based on stochastic gradient descent and backpropagation.
Kozloski does teach:
training a neural network by optimizing a loss based on stochastic gradient descent and backpropagation (paras. [0029]-[0031]; and Fig. 4 —a NN model/training can employ known optimization solutions, such as stochastic gradient descent and backpropagation algorithms, to minimize one or more loss functions).
It would have been prima-facie obvious to one having ordinary skill in the art before the effective filing date of the claimed invention to have modified Gowda in view of Girnyk and O’Shea’s neural network solution for determining RE specific matrices, with NN training optimization solutions, such as stochastic gradient descent and backpropagation, to minimize loss functions, as taught by Kozloski.
The motivation for doing so would have been to improve NN training through improved minimization of utilized loss functions, as recognized by Kozloski (paras. [0029]-[0031]).
Conclusion
The prior art made of record and not relied upon is considered pertinent to Applicant's disclosure is as follows:
US PG Pub. 2022/0374685 A1, Lee et al.: teaches various neural network solutions including training antenna precoder matrices related to the present invention.
US Patent No. 11,968,005 B2, Lee et al.: teaches various neural network solutions including training antenna precoder matrices related to the present invention.
US PG Pub. 2020/0293896 A1, Kwon et al.: teaches various neural network solutions including training antenna precoder matrices related to the present invention.
US PG Pub. 2024/0356591 A1, Chi et al.: teaches various neural network solutions including training antenna precoder matrices related to the present invention.
Any inquiry concerning this communication or earlier communications from the Examiner should be directed to Scott Schlack whose telephone number is (571)272-2332. The Examiner can normally be reached Mon. through Fri., from 11am-6pm EST.
Examiner interviews are available via telephone, in-person, and video conferencing using a USPTO supplied web-based collaboration tool. To schedule an interview, applicant is encouraged to use the USPTO Automated Interview Request (AIR) at http://www.uspto.gov/interviewpractice.
If attempts to reach the Examiner by telephone are unsuccessful, the Examiner’s supervisor, Moo Jeong can be reached at (571)272-9617. 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.
/Scott A. Schlack/Examiner, Art Unit 2418
/Moo Jeong/Supervisory Patent Examiner, Art Unit 2418