DETAILED ACTION
This Office Action is in response to the communication filed on 3/30/2026.
Claims 1-20 are pending.
Claims 1, 12 and 17 have been amended.
Claims 1-20 are rejected.
A request for continued examination under 37 CFR 1.114, including the fee set forth in 37 CFR 1.17(e), was filed in this application after final rejection. Since this application is eligible for continued examination under 37 CFR 1.114, and the fee set forth in 37 CFR 1.17(e) has been timely paid, the finality of the previous Office action has been withdrawn pursuant to 37 CFR 1.114. Applicant's submission filed on 03/30/2026 has been entered.
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 .
Note
Applicant redacted “randomly” on line 14 of claim 1 (and in similar claim 9). However this was already redacted in the previously entered claim set.
Double Patenting
The double patenting rejection was previously withdrawn.
Response to Arguments
On page 1 of Remarks filed on 3/30/2026 Applicant argues that the advisory action did not allege that Tsuchida taught offsetting or homomorphic encryption and amended claims 1 and 12 and claim 17 to recite the limitations. This argument is now moot because Tsuchida is no longer used to teach the limitations.
Applicant’s arguments with respect to claims 1-20 filed on 3/30/2026 have been considered but are moot because the new ground of rejection does not rely on any reference applied in the prior rejection of record for any teaching or matter specifically challenged in the argument.
On Page 6 Applicant asserts that the dependent claims are allowable based on allowable independent claims. However, the independent claims are not allowable thus the dependent claims are not allowable.
Claim Objections
Claim 3 is objected to as a repetitive. Claim 1 was amended to include “wherein at least one of the plurality of portions is generated via offsetting” however “wherein at least one of the plurality of portions is generated via offsetting” was not removed from claim 3.
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 for the rejection will not be considered a new ground of rejection if the prior art relied upon, and the rationale supporting the rejection, would be the same under either status.
The following is a quotation of 35 U.S.C. 103 which forms the basis for all obviousness rejections set forth in this Office action:
A patent for a claimed invention may not be obtained, notwithstanding that the claimed invention is not identically disclosed as set forth in section 102, if the differences between the claimed invention and the prior art are such that the claimed invention as a whole would have been obvious before the effective filing date of the claimed invention to a person having ordinary skill in the art to which the claimed invention pertains. Patentability shall not be negated by the manner in which the invention was made.
The factual inquiries for establishing a background for determining obviousness under 35 U.S.C. 103 are summarized as follows:
1. Determining the scope and contents of the prior art.
2. Ascertaining the differences between the prior art and the claims at issue.
3. Resolving the level of ordinary skill in the pertinent art.
4. Considering objective evidence present in the application indicating obviousness or nonobviousness.
Claims 1-5, 8-9, 12-18, and 20 are rejected under 35 U.S.C. 103 as being unpatentable over Shazeer (U.S. 20200042875), in view of Gu (U.S. 20220374763) and in further view of Buddhavarapu (U.S. 20220382908).
Regarding claim 1, 12 and 17,
Shazeer discloses: A method, comprising:
receiving, in a computing device, a tensor having elements specifying a computation of an artificial neural network, the tensor having a first dimension and a second dimension; (Shazeer [0083-0085 and 0102] describe receiving as input one or more respective input tensors each having one or more respective input dimensions; [0102] (i) receives as input one or more respective input tensors each having one or more respective input dimensions)
generating, by the computing device, a plurality of computing tasks via partitioning, along the first dimension and the second dimension, the tensor into a plurality of portions respectively, each of the plurality of computing tasks configured to operate a respective portion among the portions; (Shazeer [0032-0036; 0073-0077] Describe splitting tensors into slices across multiple dimensions; [0073] The computation system 100 can assign a respective tensor-slice to each computing device according to the computation layout, and execute the laid-out computation graph on the mesh 120, because the weight tensors are fully replicated but the data is split and distributed and operated on in parallel; [0085] the computation system can comprise a mesh execution engine that distributes tensors according to respective layouts for the tensors. Operations can be distributed across the plurality of computing devices by assigning slices of a tensor representing weights to each of one or more of the plurality of devices; [0102] (ii) generates as output one or more respective output tensors each having one or more respective output dimensions)
(Shazeer [0102] wherein the specification data specifies a respective layout for each input and output tensor that assigns each dimension of the input or output tensor to one or more of the plurality of computing devices; assigning, based on the layouts for the input and output tensors, respective device-local operations to each of the plurality of computing devices; [0007-0012, 0026-0032, 0037-0055, 0064-0086] teaches a layout (map) and specification data which is used to determine how tensors and operations should be distributed across the mesh and then uses the layout to combine the outputs distributed results from the various third party servers)
communicating, from the plurality of computing device, the portions to the external entities according to the computing tasks being (Shazeer [0021-0030]; [0102] wherein the specification data specifies a respective layout for each input and output tensor that assigns each dimension of the input or output tensor to one or more of the plurality of computing devices; assigning, based on the layouts for the input and output tensors, respective device-local operations to each of the plurality of computing devices; and causing the tensor computations to be executed by the plurality of computing devices by causing each of the plurality of computing devices to execute at least the respective-device local operations assigned to the computing devices.
receiving, by the computing device from the external entities, results of the computing tasks; and (Shazeer [0032-0038; 0050-0059] Describe receiving outsourced computational results [0054] computing device generates a respective output that the mesh execution engine 140 can combine to generate the output data 170, by executing communication primitive operations between the computing devices of the mesh 110.
generating, by the computing device (Shazeer [0003-0016]; [0085-0087] The system causes the tensor computations to be executed by the plurality of computing devices by causing each of the plurality of computing devices to execute at least the respective device-local operations assigned to the computing devices… combining individual outputs generated by executing device-local operations on the computing devices, to generate a combined output that can be an output for the tensor computations)
Although Shazeer discloses mapping using layouts and determining which device should process which portion of the computation, assigning that each device their portion of the computation and combining the results of those individual computation results into a combined output, Shazeer does not explicitly recite the shuffling using a map: shuffling, by the computing device and using a shuffling map
generating, by the computing device using the shuffling map… a result
However, in the same field of endeavor Gu discloses: shuffling, by the computing device and using a shuffling map
generating, by the computing device using the shuffling map… a result (Gu [Abstract, 0007-0009, 0087-0093, Claim 1] teaches mapping and shuffling partitions of data then reversely shuffling the partitions so they are in the correct order)
Shazeer and Gu are analogous art because they are from the same field of endeavor outsourcing computations while protecting the underlaying data.
Before the effective filing date of the claimed invention, it would have been obvious to one of ordinary skill in the art, having the teachings of Shazeer and Gu before him or her, to modify the method of Shazeer to include the shuffling and reversely shuffling of Gu because it will allow shuffled data to be placed in the correct order and facilitating aggregated obfuscation in federated learning.
The motivation for doing so would be [“The approach herein provides enhanced protection against information leakage by leveraging a multi-layered defense strategy that comprises several aspects, namely, trustworthy aggregation, decentralized aggregation with model partitioning, and dynamic permutation”] (Paragraph 0006 by Gu)].
Therefore, it would have been obvious to combine Shazeer and Gu to obtain the invention as specified in the instant claim.
While Shazeer and Gu teaches a randomized mapper that generates assignments which could be characterized as offsetting, Shazeer and Gu does not explicitly disclose: wherein at least one of the plurality of portions is generated via offsetting
However, in the same field of endeavor Buddhavarapu teaches: wherein at least one of the plurality of portions is generated via offsetting (Buddhavarapu [0086] teaches using random numbers as an offset to shuffle data records where that offset is removed for the data record to be matched using secret keys)
Shazeer in view of Gu and Buddhavarapu are analogous art because they are from the same field of endeavor secure multi-party computing.
Before the effective filing date of the claimed invention, it would have been obvious to one of ordinary skill in the art, having the teachings of Shazeer in view of Gu and Buddhavarapu before him or her, to modify the method of Shazeer in view of Gu to include the random numbers (i.e., an offset) of Buddhavarapu in the randomized mapper that generates assignments because the random numbers (i.e., an offset) may be utilized as additive shares in order to ensure cooperation between entities.
The motivation for doing so would be [“the outputted matching records may be encrypted (i.e., as “additive secret shares”) with each entity receiving only partial data and requiring another entity's cooperation to reveal any underlying data.”] (Paragraph 0029, 0086-0089 by Buddhavarapu)].
Therefore, it would have been obvious to combine Shazeer in view of Gu and Buddhavarapu to obtain the invention as specified in the instant claim.
Similar claim 12 recites “A computing device, comprising: memory; and at least one microprocessor coupled to the memory and configured via instructions to:” [see Claim 20 Shazeer] (One or more computer-readable storage media encoded with instructions that, when executed by one or more computers, cause the one or more computers to perform operations comprising), additionally claim 20 recites a matrix whereas claim 1 recites a tensor, there are being interpreted as synonymous.
Similar claim 17 recites “A non-transitory computer storage medium storing instructions which, when executed in a computing device, cause the computing device to perform a method, comprising” [see Claim 20 Shazeer] (One or more computer-readable storage media encoded with instructions that, when executed by one or more computers, cause the one or more computers to perform operations comprising); [0087] one or more modules of computer program instructions encoded on a tangible non transitory storage medium for execution by, or to control the operation of, data processing apparatus) additionally claim 20 recites a matrix whereas claim 1 recites a tensor, there are being interpreted as synonymous.
Additionally, similar claim 17 recites: wherein at least one of the plurality of computing tasks is generated via homomorphic encryption (Gu [0007-0009, 0071-0081] teaches encrypting the computing tasks [0004] teaches that the type of encryption can be homomorphic encryption); additionally/alternatively (Buddhavarapu [0025, 0086] teaches homomorphic encryption)
It would have been obvious to a person of ordinary skill in the art before the effective filing date of the claimed invention to modify with Gu or Buddhavarapu for similar reasons as cited in claim 1.
Regarding claim 2,
Shazeer in view of Gu and Buddhavarapu discloses: The method of claim 1, wherein the portions are configured with first subsets of the portions, the first subsets of the portions representing division of the tensor along the first dimension; (Shazeer [0032-0036; 0073-0077] Describe splitting tensors into slices across multiple dimensions; [0027-0050] describes different tensor slice (portion) dimensions and how they are laid out)
each computing task involving the first subsets of the portions is independent from one another; and (Shazeer [0026-0031, 0051-0021, 0073-0077] Describe parallelism where [0030] In some implementations, the layout engine 120 can generate and send layouts to the mesh execution engine 140 that in turn distributes respective tensors and operations to implement model parallelism where data is split and distributed and operated on in parallel)
the method further comprises: aggregating computing results of the first subsets of the portions along the first dimension to generate the result of the computation of the artificial neural network. (Shazeer [0003-0016]; [0085-0087] combining individual outputs generated by executing device-local operations on the computing devices, to generate a combined output that can be an output for the tensor computations or input for another computation).
Regarding claim 3,
Shazeer in view of Gu and Buddhavarapu discloses: The method of claim 2, wherein the portions are configured with second subsets of the portions, the second subsets of the portions representing division of the tensor along the second dimension; (Shazeer [0032-0036; 0073-0077] Describe splitting tensors into slices across multiple dimensions; [0027-0050] describes different tensor slice (portion) dimensions and how they are laid out; [0018] FIGS. 2A-D show example layouts for a two-dimensional tensor laid-out on a two-dimensional mesh; [0025-0055] Describe multiple tensor dimensions and how they interact along those dimensions)
each computing task associated with the second subsets of the portions is independent from one another; and (Shazeer [0026-0031, 0051-0021, 0073-0077] Describe parallelism where [0030] In some implementations, the layout engine 120 can generate and send layouts to the mesh execution engine 140 that in turn distributes respective tensors and operations to implement model parallelism where data is split and distributed and operated on in parallel)
the method further comprises summing computing results of the second subsets of the portions along the second dimension to generate the result of the computation of the artificial neural network. (Shazeer [0003-0016]; [0085-0087] combining individual outputs generated by executing device-local operations on the computing devices, to generate a combined output that can be an output for the tensor computations or input for another computation).
Regarding claim 4 and 16,
Shazeer in view of Gu and Buddhavarapu discloses: The method of claim 3, wherein at least one of the portions is generated via offsetting, bit-wise shifting, adding a constant, multiplying by a constant, or homomorphic encryption, or any combination thereof. (Gu [0004] teaches homomorphic encryption; although Gu states that a drawback of homomorphic encryption is that is computationally expensive, it does not “teach away” from homomorphic encryption)
It would have been obvious to a person of ordinary skill in the art before the effective filing date of the claimed invention to modify with Gu for similar reasons as cited in claim 1.
Regarding claim 5,
Shazeer in view of Gu and Buddhavarapu discloses: The method of claim 3, wherein each of the external entities is excluded from receiving a subset of the portions. (Shazeer [0084-0086]; Describes a mesh layout where entities receiving one portion cannot receive a portion meant for another; [0037] A layout has been referred to generally as specifying a format for how tensors and operations are assigned by the mesh execution engine 140 across the computing devices of the mesh 110. More specifically, the layout engine 120 can generate separate layouts for respective tensors, e.g., a tensor representing the input data 150, the output data 170, or any intermediate input and output data generated by processing the operations assigned to the mesh 110).
Regarding claim 8,
Shazeer in view of Gu and Buddhavarapu discloses: The method of claim 3, wherein the portions include a third subset; a sum of the third subset is equal to a corresponding portion in the tensor. (Shazeer [0053-0059] after the mesh execution engine 140 assigns sub-tensors of an input tensor to different computing devices across the mesh 110, the mesh execution engine 140 can execute communication operation primitives to combine output sub-tensors corresponding to assigned input sub-tensors, to form the output data 180 representing an output for the input data 150).
Regarding claim 9,
Shazeer in view of Gu and Buddhavarapu discloses: The method of claim 8, further comprising: summing results of the plurality of computing tasks, wherein each of the plurality of computing tasks are performed using one portion among the third subset, to generate the result of the computation of the artificial neural network. (Shazeer [0003-0016]; [0085-0087] combining individual outputs generated by executing device-local operations on the computing devices, to generate a combined output that can be an output for the tensor computations or input for another computation…As necessary, the system can also execute one or more communication primitives for combining individual outputs generated by executing device-local operations on the computing devices, to generate a combined output that can be an output for the tensor computations or input for another computation).
Regarding claim 13 and 18,
Shazeer in view of Gu and Buddhavarapu discloses: The computing device of claim 12, wherein the computing tasks are generated via further partitioning the matrix along a second dimension of the matrix. (Shazeer [0032-0036; 0073-0077] Describe splitting tensors into slices across multiple dimensions; [0073] The computation system 100 can assign a respective tensor-slice to each computing device according to the computation layout, and execute the laid-out computation graph on the mesh 120, because the weight tensors are fully replicated but the data is split and distributed and operated on in parallel; [0085] the computation system can comprise a mesh execution engine that distributes tensors according to respective layouts for the tensors. Operations can be distributed across the plurality of computing devices by assigning slices of a tensor representing weights to each of one or more of the plurality of devices; [0102] (ii) generates as output one or more respective output tensors each having one or more respective output dimensions).
Regarding claim 14,
Shazeer in view of Gu and Buddhavarapu discloses: The computing device of claim 13, wherein rows of the matrix are representative of division along the first dimension; (Shazeer [0032-0036; 0073-0077] Describe splitting tensors into slices across multiple dimensions; [0027-0050] describes different tensor slice (portion) dimensions and how they are laid out)
columns of the matrix are representative of division along the second dimension; and (Shazeer [0025] Dimension 0 is the “horizontal” dimension, and dimension 1 is the “vertical” dimension)
wherein the at least one microprocessor is further configured via the instructions to:
sum, based on the first results from the external entities, computing results corresponding portions of the matrix partitioned according to the second dimension; and (Shazeer [0061] Then, tensors assigned to each computing device and for each group can be summed or otherwise reduced and the resulting reduced value can be assigned to each computing device in the group)
aggregate, based on the first results from the external entities, computing results corresponding portions of the matrix partitioned according to the first dimension. (Shazeer [0003-0016]; [0085-0087] combining individual outputs generated by executing device-local operations on the computing devices, to generate a combined output that can be an output for the tensor computations or input for another computation).
Regarding claim 15,
Shazeer in view of Gu and Buddhavarapu discloses: The computing device of claim 14, wherein the at least one microprocessor is further configured via the instructions to: split a portion of the matrix as a sum of multiple portions operated upon in the computing tasks. (Shazeer [0032-0036; 0073-0077] Describe splitting tensors into slices across multiple dimensions; [0073] The computation system 100 can assign a respective tensor-slice to each computing device according to the computation layout, and execute the laid-out computation graph on the mesh 120, because the weight tensors are fully replicated but the data is split and distributed and operated on in parallel; [0085] the computation system can comprise a mesh execution engine that distributes tensors according to respective layouts for the tensors. Operations can be distributed across the plurality of computing devices by assigning slices of a tensor representing weights to each of one or more of the plurality of devices; [0102] (ii) generates as output one or more respective output tensors each having one or more respective output dimensions; [0003-0016]; [0085-0087] combining individual outputs generated by executing device-local operations on the computing devices, to generate a combined output that can be an output for the tensor computations or input for another computation).
Regarding claim 20,
Shazeer in view of Gu and Buddhavarapu discloses: The non-transitory computer storage medium of claim 17, wherein the computing tasks are generated via splitting a portion of the matrix into a sum of multiple randomized portions operated upon respectively in multiple tasks of the computing tasks. (Shazeer [0014]; Using the techniques described in this specification for specifying distribution schemes, schemes for parallelizing the training of very large machine learning models, i.e., models; [0113] multitasking and parallel processing.
Claims 6-7 and 19 are rejected under 35 U.S.C. 103 as being unpatentable over Shazeer (U.S. 20200042875), in view of Gu (U.S. 20220374763), view of Buddhavarapu (U.S. 20220382908) and in further view of Horne (U.S. 20230053566).
Regarding claim 6,
Shazeer in view of Gu and Buddhavarapu discloses: The method of claim 3, further comprising: shuffling rows of elements extending in the tensor along the first dimension to generate the portions. (Horne [0092] Shuffling may refer to a technique for randomizing the order of the rows so that the order in which rows of data appear in data set 408 cannot practically be determined based on the order of rows in cryptographically protected data set 420; [0125] In at least one embodiment, as part of a secure MPC protocol, first data set 702 is cryptographically protected using various techniques (e.g., HMAC, OPE, HE techniques), the rows are shuffled or randomized, and then transmitted to a server)
Although Shazeer in view of Gu and Buddhavarapu does not explicitly disclose: shuffling rows of elements extending in the tensor along the first dimension to generate the portions.
However, in the same field of endeavor Horne discloses: shuffling rows of elements extending in the tensor along the first dimension to generate the portions. (Horne [0048, 0092] As part of generating cryptographically protected second data set 216, the order of the cryptographically protected values may be shuffled or randomized.
Shazeer view of Gu and Buddhavarapu and Horne are analogous art because they are from the same field of endeavor outsourcing computations while protecting the underlaying data.
Before the effective filing date of the claimed invention, it would have been obvious to one of ordinary skill in the art, having the teachings of Shazeer view of Gu and Buddhavarapu and Horne before him or her, to modify the method of Shazeer view of Gu and Buddhavarapu to include the row and column shuffling of Horne because it will allow for data to be better protected while allowing for reading from and writing to the desired array element.
The motivation for doing so would be [“As part of generating first cryptographically protected data set 212, the order of the cryptographically protected values may, should, or must be shuffled or randomized, depending on use cases.” And “the secure computation system 100 according to the first example embodiment is able to read from and write to the desired array element [L.sub.idx] while keeping even which element in an array {[L.sub.j]}.sub.j=0.sup.m−1 is being accessed confidential.”] (Paragraph 0042 and 0053 by Horne)].
Therefore, it would have been obvious to combine Shazeer view of Gu and Buddhavarapu and Horne to obtain the invention as specified in the instant claim.
Regarding claim 7 and 19,
Shazeer in view of Gu and Buddhavarapu discloses: The method of claim 3, further comprising: shuffling columns of elements extending in the tensor along the second dimension to generate the portions. (Horne [0115] computing entity 604 may, accordingly, produce two or more columns of cryptographically protected data elements that are randomized/shuffled to produce a second cryptographically protected data set which is transmitted 622 to server 606)
It would have been obvious to a person of ordinary skill in the art before the effective filing date of the claimed invention to modify with Horne for similar reasons as cited in claim 6.
Claims 10-11 are rejected under 35 U.S.C. 103 as being unpatentable over Shazeer (U.S. 20200042875), in view of Gu (U.S. 20220374763), view of Buddhavarapu (U.S. 20220382908) and in further view of Li (U.S. 20210051007).
Regarding claim 10,
Shazeer in view of Gu and Buddhavarapu discloses: The method of claim 9, further comprising:
in at least one first portion among the third subset; and (Shazeer [0053-0059] after the mesh execution engine 140 assigns sub-tensors of an input tensor to different computing devices across the mesh 110, the mesh execution engine 140 can execute communication operation primitives to combine output sub-tensors corresponding to assigned input sub-tensors, to form the output data 180 representing an output for the input data 150)
Shazeer does not specifically disclose generating random numbers.
generating a second portion among the third subset from subtracting from the corresponding portion in the tensor a sum of the at least one first portion
However, in the same field of endeavor Li discloses: generating random numbers (Li [0010] obtaining a random number generated by the trusted random source)
generating a second portion among the third subset from subtracting from the corresponding portion in the tensor a sum of the at least one first portion. (Li [0091] Based on the foregoing expression, S.sub.d=S−(S.sub.1+S.sub.2+S.sub.3+ . . . +S.sub.d-1)% P. In this algorithm, S.sub.d may be obtained by subtracting a modulus computation result of a sum of the first d−1 shares and the value space P from the private data S).
Shazeer, Gu, Buddhavarapu and Li are analogous art because they are from the same field of endeavor Multi-party data protection.
Before the effective filing date of the claimed invention, it would have been obvious to one of ordinary skill in the art, having the teachings of Shazeer, Gu, Buddhavarapu and Li before him or her, to modify the method of Shazeer and Gu and Buddhavarapu to include the Random numbers of Li because it will better data protection.
The motivation for doing so would be providing data protection when private data is breached [“the secure multi-party algorithm is an algorithm designed for protecting private data, an erroneous or malicious algorithm implementation may cause a breach of private data”] (Paragraph 0127 by Li)].
Therefore, it would have been obvious to combine Shazeer in view of Gu and Buddhavarapu with Li to obtain the invention as specified in the instant claim.
Regarding claim 11,
Shazeer in view of Gu, Buddhavarapu and Li disclose: The method of claim 10, wherein each of the external entities receiving a first one in the third subset is excluded from receiving from the computing device a second one in the third subset. (Shazeer [0084-0086]; Describes a mesh layout where entities receiving one portion cannot receive a portion meant for another; [0037] A layout has been referred to generally as specifying a format for how tensors and operations are assigned by the mesh execution engine 140 across the computing devices of the mesh 110. More specifically, the layout engine 120 can generate separate layouts for respective tensors, e.g., a tensor representing the input data 150, the output data 170, or any intermediate input and output data generated by processing the operations assigned to the mesh 110).
Conclusion
The prior art made of record and not relied upon is considered pertinent to applicant's
disclosure.
Gentry 6/18/2019 (US 20200403781) teaches using homomorphic encryption at the bit level.
Klein 4/3/2021 (US 20220284114) teaches randomizing techniques for MPC security.
Any inquiry concerning this communication or earlier communications from the examiner should be directed to THOMAS A CARNES whose telephone number is (571)272-4378. The examiner can normally be reached Monday-Friday.
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, Shewaye Gelagay can be reached on (571) 272-4219. 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.
THOMAS A. CARNES
Examiner
Art Unit 2436
/THOMAS A CARNES/Examiner, Art Unit 2436
Prosecution Insights