DETAILED ACTION
Notice of Pre-AIA or AIA Status
The present application, filed on or after March 16, 2013, is being examined under the first inventor to file provisions of the AIA .
Status of Claims
This Office Action is in response to claims filed on 9/19/2024.
Claims 1-30 remain pending in the application.
Claim Objections
Claim 4 is objected to because of the following informalities: claim 4 is a dependent claim of claim 3, it does not further limit the claim 3. Appropriate correction is required.
Claim 10 is objected to because of the following informalities: claim 10 is a dependent claim of claim 9, it does not further limit the claim 9. Appropriate correction is required.
Claim 23 is objected to because of the following informalities: the phrase “…indicates that the sign of the gradient value is,” should be read as “…indicates that the sign of the gradient value is positive”. Appropriate correction is required.
Claim Rejections - 35 USC § 103
The following is a quotation of 35 U.S.C. 103 which forms the basis for all obviousness rejections set forth in this Office action:
A patent for a claimed invention may not be obtained, notwithstanding that the claimed invention is not identically disclosed as set forth in section 102, if the differences between the claimed invention and the prior art are such that the claimed invention as a whole would have been obvious before the effective filing date of the claimed invention to a person having ordinary skill in the art to which the claimed invention pertains. Patentability shall not be negated by the manner in which the invention was made.
The factual inquiries for establishing a background for determining obviousness under 35 U.S.C. 103 are summarized as follows:
1. Determining the scope and contents of the prior art.
2. Ascertaining the differences between the prior art and the claims at issue.
3. Resolving the level of ordinary skill in the pertinent art.
4. Considering objective evidence present in the application indicating obviousness or nonobviousness.
Claims 1-4, 6-7, 9-10, 12-19, 21-22, 27-28 and 30 are rejected under 35 U.S.C. 103 as being unpatentable over Sahin et al. (US 2022/0391696 A1); in view of Zhu et al. (Broadband Analog Aggregation for Low-Latency Federated Edge Learning).
Regarding claim 1; Sahin discloses an apparatus for wireless communication at a user equipment (UE), comprising: a memory (one or more non-transitory computer-readable media; see paragraph [0020]); and at least one processor coupled to the memory (one or more processors; see paragraph [0020]) and configured to:
calculate a gradient value relative to a parameter of the parameters of the machine learning model (the edge device calculates a gradient value based on the parameter vector w.sup.n in a machine learning model; see paragraphs [0057] and [0070]); the gradient value including a positive gradient value or a negative gradient value (the gradient values may be negative or positive; see paragraph [0066]).
and transmit, in one resource element (RE) of a pair of REs to the network node, an analog signal indicating a magnitude of the gradient value relative to the parameter of the parameters of the machine learning model, the pair of REs designated for indicating the gradient value relative to the parameter of the parameters of the machine learning model (symbols transmitted at a pair (m.sub.0, l.sub.0) and (m.sub.1, l.sub.1) are mapped based on gradient values; gradient estimation is calculated based on parameter vector w.sup.n with a machine learning model; the federal edge learning (FEEL) model can also be used for analog modulation over OFDM; see paragraphs [0057], [0061], [0063], [0068] and [0085]).
Sahin discloses an edge device calculates gradient values based on a parameter vector.
Sahin does not explicitly disclose the parameter vector is received from a network node.
Zhu discloses receive parameters of a machine learning model from a network node (in each communication round, the edge server broadcasts the current model under training w[n] to all edge devices; see last paragraph left column of page 4);
It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to combine the teachings of Sahin and Zhu to receive the parameter vector from a network node to implement federal edge learning (FEEL) model (see last paragraph left column of page 4 of Zhu).
Regarding claim 7; Sahin discloses an apparatus for wireless communication at a user equipment (UE), comprising: a memory (one or more non-transitory computer-readable media; see paragraph [0020]); and at least one processor coupled to the memory (one or more processors; see paragraph [0020]) and configured to:
calculate a gradient value relative to a parameter of the parameters of the machine learning model (the edge device calculates a gradient value based on the parameter vector w.sup.n; see paragraph [0057]); a sign of the gradient value being one of positive or negative (the gradient values may be negative or positive; the ED transmits signs of gradient values; see paragraphs [0060] and [0066])
and transmit or skip transmission of, in a single resource element (RE) to the network node, an analog signal based on the sign of the gradient value relative to the parameter of the parameters of the machine learning model, the single RE designated for indicating the sign of the gradient value relative to the parameter of the parameters of the machine learning model (the edge devices (ED) transmits the signs of their local gradient to the edge server (ES); EDs transmit the signs of local gradients by activating one of two orthogonal resources; gradient estimation is calculated from a parameter vector with a machine learning model; the federal edge learning (FEEL) model can also be used for analog modulation over OFDM; see paragraphs [0016], [0057], [0060], [0063], [0068] and [0085]).
Sahin discloses an edge device calculates a gradient value based on a parameter vector.
Sahin does not explicitly disclose the parameter vector is received from a network node.
Zhu discloses receive parameters of a machine learning model from a network node (in each communication round, the edge server broadcasts the current model under training w[n] to all edge devices; see last paragraph left column of page 4);
It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to combine the teachings of Sahin and Zhu to receive the parameter vector from a network node to implement federal edge learning (FEEL) model (see last paragraph left column of page 4 of Zhu).
Regarding claim 13; Sahin discloses an apparatus for wireless communication at a network node, comprising: a memory (one or more non-transitory computer-readable media; see paragraph [0020]);
and at least one processor (one or more processors; see paragraph [0020]) coupled to the memory and configured to:
obtain at least one aggregated analog signal in a pair of resource elements (REs) designated for the plurality of UEs to indicate a gradient value relative to a parameter of the parameters of the machine learning model of the plurality of UEs (the edge server observes the superposed symbols on the same subcarriers from a plurality of EDs; symbols received at a pair (m.sub.0, l.sub.0) and (m.sub.1, l.sub.1) are mapped based on gradient values; gradient estimation is calculated based on a parameter vector with a machine learning model; the federal edge learning (FEEL) model can also be used for analog modulation over OFDM; see paragraphs [0020], [0057], [0061], [0063], [0068], [0073] and [0085]),
the at least one aggregated analog signal obtained in the pair of REs representing a plurality of analog signals accumulatively obtained from the plurality of UEs at the network node (the edge server observes the superposed symbols on the same subcarriers from a plurality of EDs; see paragraph [0073]),
each analog signal of the plurality of analog signals indicating the gradient value relative to the parameter of the parameters of the machine learning model calculated at corresponding UE of the plurality of UEs (symbols received at a pair (m.sub.0, l.sub.0) and (m.sub.1, l.sub.1) are mapped based on gradient values; gradient estimation is calculated based on a parameter vector with a machine learning model; the federal edge learning (FEEL) model can also be used for analog modulation over OFDM; see paragraphs [0057], [0061], [0063], [0068], [0073] and [0085]),
the gradient value including a positive gradient value or a negative gradient value (the gradient values may be negative or positive; see paragraph [0066]);
and update the parameter of the parameters of the machine learning model based on the at least one aggregated analog signal obtained in the pair of REs (the ES distributes the global gradient estimate to the ED and the current model is updated based on a common update rule; the global gradient estimate is calculated based on gradient values transmitted from the ED; the federal edge learning (FEEL) model can also be used for analog modulation over OFDM; see paragraphs [0058] - [0059] and [0083]).
Sahin discloses an edge server receives superposed signals from a plurality of EDs.
Sahin does not explicitly disclose outputting parameters of a machine learning model for the plurality of EDs.
Zhu discloses output for transmission parameters of a machine learning model for a plurality of user equipments (UEs) (in each communication round, the edge server broadcasts the current model under training w[n] to all edge devices; see last paragraph left column of page 4);
It would have been obvious to one of the ordinary skills in the art before the effective filing date of the claimed invention to combine the teachings of Sahin and Zhu to output parameters of a machine learning model for UEs to implement federal edge learning (FEEL) model (see last paragraph left column of page 4 of Zhu).
Regarding claim 22; Sahin discloses an apparatus for wireless communication at a network node, comprising: a memory (one or more non-transitory computer-readable media; see paragraph [0020]); and at least one processor coupled to the memory (one or more processors; see paragraph [0020]) and configured to:
obtain an aggregated analog signal in a single resource element (RE) designated for the plurality of UEs to indicate a sign of a gradient value relative to a parameter of the parameters of the machine learning model of the plurality of UEs (the edge server observes the superposed symbols on the same subcarriers from a plurality of EDs; EDs transmit the signs of local gradients by activating one of two orthogonal resources; gradient estimation is calculated from data samples with a machine learning model; the federal edge learning (FEEL) model can also be used for analog modulation over OFDM; see paragraphs [0016], [0057], [0060], [0063], [0068], [0073] and [0085]),
the aggregated analog signal obtained in the single RE representing a plurality of analog signals accumulatively obtained from the plurality of UEs at the network node (the edge server observes the superposed symbols on the same subcarriers from a plurality of EDs; see paragraph [0073]),
each analog signal of the plurality of analog signals indicating that the sign of the gradient value relative to the parameter of the parameters of the machine learning model calculated at corresponding UE of the plurality of UEs is one of positive or negative (the edge devices (ED) transmits the signs of their local gradient to the edge server (ES); EDs transmit the signs of local gradients by activating one of two orthogonal resources; gradient estimation is calculated from parameter vector with a machine learning model; the federal edge learning (FEEL) model can also be used for analog modulation over OFDM; see paragraphs [0016], [0057], [0060], [0063], [0068] and [0085]);
and update the parameter of the parameters of the machine learning model based on the aggregated analog signal obtained in the single RE (the ES distributes the global gradient estimate to the ED and the current model is updated based on a common update rule; the global gradient estimate is calculated based on gradient values transmitted from the ED; the federal edge learning (FEEL) model can also be used for analog modulation over OFDM; see paragraphs [0058] - [0059] and [0083]).
Sahin discloses an edge server receives superimposed signals from a plurality of UEs.
Sahin does not explicitly disclose outputting parameters of a machine learning model for the plurality of UEs.
Zhu discloses output for transmission parameters of a machine learning model for a plurality of user equipments (UEs) (in each communication round, the edge server broadcasts the current model under training w[n] to all edge devices; see last paragraph left column of page 4);
It would have been obvious to one of ordinary skills in the art before the effective filing date of the claimed invention to combine the teachings of Sahin and Zhu to output parameters of a machine learning model for UEs to implement federal edge learning (FEEL) model (see last paragraph left column of page 4 of Zhu).
Regarding claim 2; Sahin discloses wherein the pair of REs designated for indicating the parameters of the machine learning model includes:
a first RE designated for indicating the positive gradient value (a positive gradient value maps to a symbol at (m.sub.0, l.sub.0); see paragraph [0066]);
and a second RE designated for indicating the negative gradient value (a negative gradient maps to a symbol at (m.sub.1, l.sub.1); see paragraph [0066]),
wherein the analog signal is transmitted in the first RE or the second RE based on the gradient value (symbols is transmitted at (m.sub.0, l.sub.0) or (m.sub.1, l.sub.1); the federal edge learning (FEEL) model can also be used for analog modulation over OFDM; see paragraphs [0057], [0061], [0063], [0068] and [0085]).
Regarding claims 3 and 9; Sahin discloses wherein the at least one processor is further configured to:
receive information associated with an updated parameter of the machine learning model from the network node, the updated parameter based at least in part on the analog signal transmitted to the network node (the ES distributes the global gradient estimate to the ED and the current model is updated based on a common update rule; the global gradient estimate is calculated based on gradient values transmitted from the ED; the federal edge learning (FEEL) model can also be used for analog modulation over OFDM; see paragraphs [0058] - [0059] and [0083]);
and update the parameter of the machine learning model based on the information associated with the updated parameter received from the network node (this process is repeated consecutively until a predetermined convergence criterion is achieved; see paragraph [0059]).
Regarding claims 4 and 10; Sahin discloses wherein the information associated with the updated parameter includes the updated parameter based at least in part on the analog signal transmitted to the network node (the ES distributes the global gradient estimate to the ED and the current model is updated based on a common update rule; the global gradient estimate is calculated based on gradient values transmitted from the ED; the federal edge learning (FEEL) model can also be used for analog modulation over OFDM; this process is repeated consecutively until a predetermined convergence criterion is achieved; see paragraphs [0058] - [0059] and [0083]).
Regarding claims 6, 12, 21 and 30; Sahin discloses further comprising a transceiver coupled to the at least one processor and configured to transmit the analog signal (transmitter and receiver for FEEL system; see paragraph [0073] and Fig. 1).
Regarding claim 14; Sahin discloses wherein the at least one aggregated analog signal includes a first aggregated analog signal and a second aggregated analog signal, and the pair of REs designated for indicating the parameter of the parameters of the machine learning model includes a first RE and a second RE (symbols received at a pair (m.sub.0, l.sub.0) and (m.sub.1, l.sub.1) are mapped based on gradient values; gradient estimation is calculated based on a parameter vector with a machine learning model; the federal edge learning (FEEL) model can also be used for analog modulation over OFDM; see paragraphs [0057], [0061], [0063], [0068], [0073] and [0085]),
wherein the first aggregated analog signal obtained in the first RE represents a first set of analog signals accumulatively obtained at the network node, each analog signal of the first set of analog signals indicating the positive gradient value (the ES receives superposed symbols from EDS; a positive gradient value maps to a symbol at (m.sub.0, l.sub.0); see paragraphs [0061] and [0066]),
and wherein the second aggregated analog signal obtained in the second RE represents a second set of analog signals accumulatively obtained at the network node, each analog signal of the second set of analog signals indicating the negative gradient value (the ES receives superposed symbols from EDs; a negative gradient value maps to a symbol at (m.sub.1, l.sub.1); see paragraphs [0061] and [0066]).
Regarding claim 15; Sahin discloses wherein, the at least one processor is further configured to compare a first magnitude of the first aggregated analog signal obtained in the first RE and a second magnitude of the second aggregated analog signal obtained in the second RE, and wherein the parameter is updated based on a comparison of the first magnitude and the second magnitude (the ES receives superposed local updates from EDs; compare the superposed symbols to determine the MV for each element of the update vector at the ES; see paragraph [0020], [0072] – [0073]).
Regarding claim 16; Sahin discloses wherein the at least one processor is further configured to generate an aggregated gradient of the parameter based on the comparison of the first magnitude and the second magnitude; wherein the parameter is updated based on the aggregated gradient of the parameter (the ES receives superposed local updates from EDs; compare the superposed symbols to determine the MV for each element of the update vector at the ES; aggregates the local updates at the ES using orthogonal resources and MV principle, and inputs the obtained data into the machine-learning model; see paragraphs [0020] – [0021] and [0072] – [0073]).
Regarding claim 17; Sahin discloses wherein the aggregated gradient of the parameter has a value of +p based on the first magnitude being greater than the second magnitude (if the energies on the first superposed symbol is greater than the second superposed symbol, the MV is 1; see paragraph [0072]),
wherein the aggregated gradient of the parameter has a value of -p based on the first magnitude being smaller than or equal to the second magnitude (if the energies on the first superposed symbol is smaller than the second superposed symbol, the MV is -1; see paragraph [0072]),
and wherein the p being a real number (1 is a real number; see paragraph [0072]).
Regarding claims 18 and 27; Sahin discloses wherein the at least one processor is further configured to: output for transmission information associated with the updated parameter of the machine learning model to plurality of UEs (the ES distributes the global gradient estimate to the EDs and the current model is updated based on a common update rule; the global gradient estimate is calculated based on gradient values transmitted from the ED; see paragraphs [0058] - [0059]).
Regarding claims 19 and 28; Sahin discloses wherein the information associated with the updated parameter includes the updated parameter (this process is repeated consecutively until a predetermined convergence criterion is achieved; see paragraph [0059]).
Claims 5, 11, 20 and 29 are rejected under 35 U.S.C. 103 as being unpatentable over Sahin; in view of Zhu; and in further view of Xiong et al. (WO 2022/155387 A1).
Regarding claim 5; the combination of Sahin and Zhu discloses an edge device transmits gradient values at a pair of resource elements to an edge server.
The combination of Sahin and Zhu does not explicitly disclose the network node transmits an indication of the resource elements for uplink transmission.
Xiong discloses wherein the at least one processor is further configured to receive an indication of the pair of REs from the network node, and wherein the analog signal is transmitted in the one RE of the pair of REs based on the indication of the pair of REs received from the network node (a radio node B device may transmit an indication of an allocation of one or more resource elements (REs) to be used for uplink transmission; see paragraph 6 of page 42).
It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to combine the teachings of Sahin, Zhu and Xiong to transmit an indication of the resource elements for uplink transmission to detect information received from the UE device (see Abstract of Xiong).
Regarding claim 11; the combination of Sahin and Zhu discloses an edge device transmits gradient values at a pair of resource elements to an edge server.
The combination of Sahin and Zhu does not explicitly disclose the network node transmits an indication of the resource element for uplink transmission.
Xiong discloses wherein the at least one processor is further configured to receive an indication of the single RE from the network node, wherein the analog signal is transmitted in the single RE based on the indication of the single RE received from the network node (a radio node B device may transmit an indication of an allocation of one or more resource elements (REs) to be used for uplink transmission; see paragraph 6 of page 42).
It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to combine the teachings of Sahin, Zhu and Xiong to transmit an indication of the resource element for uplink transmission to detect information received from the UE device (see Abstract of Xiong).
Regarding claims 20 and 29; The combination of Sahin and Zhu discloses an edge device transmits gradient values at a pair of resource elements to an edge server.
The combination of Sahin and Zhu does not explicitly disclose the network node transmits an indication of the resource elements for uplink transmission.
Xiong discloses wherein the at least one processor is further configured to output for transmission an indication of the pair of REs for the plurality of UEs, and wherein the at least one aggregated analog signal is obtained in the pair of REs based on the indication of the pair of REs (a radio node B device may transmit an indication of an allocation of one or more resource elements (REs) to be used for uplink transmission; see paragraph 6 of page 42).
It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to combine the teachings of Sahin, Zhu and Xiong to transmit an indication of the resource elements for uplink transmission to detect information received from the UE device (see Abstract of Xiong).
Claims 8 and 23 - 26 is rejected under 35 U.S.C. 103 as being unpatentable over Sahin; in view of Zhu; and in further view of Ferreira et al. (US 11790039 B2).
Regarding claims 8 and 23; the combination of Sahin and Zhu discloses an edge device transmits signs of gradient values in a RE to an edge server.
The combination of Sahin and Zhu does not explicitly disclose the content of the RE indicate a sign of a gradient value.
Ferreira discloses wherein a presence of the analog signal in the single RE indicates that the sign of the gradient value is, and an absence of the analog signal in the single RE indicates that the sign of the gradient value is negative (a vector of the signs of each gradient is transmitted, a single bit is used to represent the sign of the value was transmitted, for example, 0 for negative and 1 for positive; see lines 31-55, col. 5).
It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to combine the teachings of Sahin, Zhu and Ferreira to use content of the RE indicating a sign of a gradient value to reduce number of bits transmitted (see lines 31-55, col. 5).
Regarding claim 24; the combination of Sahin and Zhu discloses an edge device transmits signs of gradient values in a RE to an edge server.
The combination of Sahin and Zhu does not explicitly disclose comparing a magnitude and a threshold value.
Ferreira discloses wherein, the at least one processor is further configured to compare a magnitude of the aggregated analog signal obtained in the single RE and a threshold value, and wherein the parameter is updated based on a comparison of the magnitude and the threshold value (the model coordinator receives model updates as gradient sign vectors from client nodes; if slope of fitted curve above threshold, perform a compressor switch; see lines 11-30, col. 13 and see Fig. 2A and Fig. 2C).
It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to combine the teachings of Sahin, Zhu and Ferreira to compare a magnitude value and a threshold value to determine a new scheme (see lines 11-30, col. 13).
Regarding claim 25; Sahin discloses wherein the at least one processor is further configured to generate an aggregated gradient of the parameter based on the comparison of the first magnitude and the second magnitude; wherein the parameter is updated based on the aggregated gradient of the parameter (the ES receives superposed local updates from EDs; compare the superposed symbols to determine the MV for each element of the update vector at the ES; aggregates the local updates at the ES using orthogonal resources and MV principle, and inputs the obtained data into the machine-learning model; see paragraphs [0020] – [0021] and [0072] – [0073]).
Regarding claim 26; Sahin discloses wherein the aggregated gradient of the parameter has a value of +p based on the first magnitude being greater than the second magnitude (if the energies on the first superposed symbol is greater than the second superposed symbol, the MV is 1; see paragraph [0072]),
wherein the aggregated gradient of the parameter has a value of -p based on the first magnitude being smaller than or equal to the second magnitude (if the energies on the first superposed symbol is smaller than the second superposed symbol, the MV is -1; see paragraph [0072]),
and wherein the p being a real number (1 is a real number; see paragraph [0072]).
Conclusion
Any inquiry concerning this communication or earlier communications from the examiner should be directed to NING LI whose telephone number is (571)270-0624. The examiner can normally be reached Monday, Tuesday, Thursday 8:30am - 5:00pm.
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, Jeffrey Rutkowski can be reached at (571) 270-1215. 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.
/N.L/Examiner, Art Unit 2415
/MANSOUR OVEISSI/Primary Examiner, Art Unit 2415
Prosecution Insights