DETAILED ACTION
This communication is in response to Application No. 18/935,381 filed on November 1st, 2024, in which claims 1-20 are presented 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 .
Priority
Acknowledgment is made of applicant's claim for foreign priority based on an application filed in the Republic of Korea on 11/03/2023. It is noted, however, that applicant has not filed a certified copy of the 10-2023-0150745 application as required by 37 CFR 1.55.
Acknowledgment is also made of applicant's claim for foreign priority based on the following applications: Application No. 10-2023-0166075, filed in the Republic of Korea on 11/24/2023; Application No. 10-2024-0041278, filed in the Republic of Korea on 03/26/2024; Application No. 10-2024-0061937, filed in the Republic of Korea on 05/10/2024; and Application No. 10-2024-0153518, filed in the Republic of Korea on 11/01/2024. Receipt is acknowledged of certified copies of papers required by 37 CFR 1.55.
Information Disclosure Statement
The information disclosure statement submitted on 11/01/2024 is in compliance with the provisions of 37 CFR 1.97. Accordingly, the information disclosure statement was considered by the examiner.
Specification
The contents of the specification are sufficient for examination purposes.
Claim Rejections - 35 USC § 101
35 U.S.C. 101 reads as follows:
Whoever invents or discovers any new and useful process, machine, manufacture, or composition of matter, or any new and useful improvement thereof, may obtain a patent therefor, subject to the conditions and requirements of this title.
Claims 1-20 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: Claim 1 is a process claim. Therefore, claims 1-10 are directed to a statutory category of eligible subject matter.
Step 2A Prong 1: If a claim limitation, under its broadest reasonable interpretation, covers performance of the limitation in the mind but for the recitation of generic computer components, then it falls within the "Mental Processes" grouping of abstract ideas. Here, elements of the claimed subject matter are mental processes. Specifically, the claim recites
“A method . . . comprising . . . configuring at least one intelligent functionality among a plurality of intelligent functionalities based on functionality configuration information” (mental process – amounts to exercising judgment to form on opinion on a state of mind or mental process, with reference to known or observed information, which may be aided by pen and paper) and
“performing intelligent operations according to the at least one intelligent functionality” (mental process – amounts to exercising judgment to perform evaluations and form opinions, with reference to the above discussed state of mind or the mental process, which may be aided by pen and paper).
Step 2A Prong 2: This judicial exception is not integrated into a practical application.
The claim recites the additional elements:
“of a first communication node . . . from a second communication node . . . corresponding to the at least one intelligent functionality based on an associated identifier (ID) . . . from the second communication node . . . to perform the at least one intelligent functionality” (amounts to merely generally linking the use of the judicial exception to a particular technological environment or field of use, which do not impose any meaningful limits on practicing the abstract idea);
“information received . . . obtaining a training dataset . . . identifier (ID) received” (amounts to extra-solution activity because it is the transmission of data, which is incidental to the claimed subject matter); and
“training an intelligent model . . . using the training dataset . . . using the trained intelligent model” (amounts to mere instructions to apply the judicial exception on generic and unspecialized computer components, which do not impose any meaningful limits on practicing the abstract idea).
Step 2B: The claim does not include additional elements considered individually and in combination that are sufficient to amount to significantly more than the judicial exception.
The claim recites the additional elements:
“of a first communication node . . . from a second communication node . . . corresponding to the at least one intelligent functionality based on an associated identifier (ID) . . . from the second communication node . . . to perform the at least one intelligent functionality” (merely generally linking the use of the judicial exception to a particular technological environment or field of use does not provide an inventive concept);
“information received . . . obtaining a training dataset . . . identifier (ID) received” (transmitting data is well‐understood, routine, and conventional, see Symantec, 838 F.3d at 1321, 120 USPQ2d at 1362; see also buySAFE, Inc. v. Google, Inc., 765 F.3d 1350, 1355, 112 USPQ2d 1093, 1096 (Fed. Cir. 2014); therefore the limitation, which is recited with a high level of generality, remains insignificant extra-solution activity even upon reconsideration); and
“training an intelligent model . . . using the training dataset . . . using the trained intelligent model” (mere instructions to apply the exception using generic computer components does not provide an inventive concept).
For the reasons above, Claim 1 is rejected as being directed to an abstract idea without significantly more. This rejection applies equally to dependent claims 2-10. The additional limitations of the dependent claims are addressed below.
Regarding Claim 2:
Step 2A Prong 1: See the rejection of Claim 1 above, which Claim 2 depends on. Here, the claim recites additional limitations that are mental processes. Specifically, the claim recites
“wherein the configuring of the at least one intelligent functionality comprises: . . . in response to a capability information . . . including information on each of the plurality of intelligent functionalities supportable . . . configuring the at least one intelligent functionality corresponding to the functionality configuration information among the plurality of intelligent functionalities” (mental process – amounts to exercising judgment to form on opinion on mental states or mental processes, with reference to known or observed information, which may be aided by pen and paper).
Step 2A Prong 2: This judicial exception is not integrated into a practical application.
The claim recites the additional elements:
“from the second communication node . . . to the second communication node . . . the first communication node; in response to the functionality configuration . . . from the second communication node based on the capability report . . . to the second communication node upon completion of the configuring of the at least one intelligent functionality” (amounts to merely generally linking the use of the judicial exception to a particular technological environment or field of use, which do not impose any meaningful limits on practicing the abstract idea) and
“request received . . . transmitting . . . information received . . . transmitting a completion report” (amounts to extra-solution activity because it is the transmission of data, which is incidental to the claimed subject matter).
Step 2B: The claim does not include additional elements considered individually and in combination that are sufficient to amount to significantly more than the judicial exception.
The claim recites the additional elements:
“from the second communication node . . . to the second communication node . . . the first communication node; in response to the functionality configuration . . . from the second communication node based on the capability report . . . to the second communication node upon completion of the configuring of the at least one intelligent functionality” (merely generally linking the use of the judicial exception to a particular technological environment or field of use does not provide an inventive concept) and
“request received . . . transmitting . . . information received . . . transmitting a completion report” (transmitting data is well‐understood, routine, and conventional, see Symantec, 838 F.3d at 1321, 120 USPQ2d at 1362; see also buySAFE, Inc. v. Google, Inc., 765 F.3d 1350, 1355, 112 USPQ2d 1093, 1096 (Fed. Cir. 2014); therefore the limitation, which is recited with a high level of generality, remains insignificant extra-solution activity even upon reconsideration).
Accordingly, Claim 2 is rejected as being directed to an abstract idea without significantly more.
Regarding Claim 3:
Step 2A Prong 1: See the rejection of Claim 1 above, which Claim 3 depends on. Here, the claim recites additional limitations that are mental processes. Specifically, the claim recites
“reconfiguring the at least one intelligent functionality by incorporating the additional information based on the functionality reconfiguration information” (mental process – amounts to exercising judgment to form on opinion on an adjustment to a state of mind or mental process, with reference to known or observed information, which may be aided by pen and paper).
Step 2A Prong 2: This judicial exception is not integrated into a practical application.
The claim recites the additional elements:
“after configuring the at least one intelligent functionality . . . to the second communication node . . . from the second communication node based on the additional information” (amounts to merely generally linking the use of the judicial exception to a particular technological environment or field of use, which do not impose any meaningful limits on practicing the abstract idea) and
“transmitting additional information . . . receiving functionality reconfiguration information” (amounts to extra-solution activity because it is the transmission of data, which is incidental to the claimed subject matter).
Step 2B: The claim does not include additional elements considered individually and in combination that are sufficient to amount to significantly more than the judicial exception.
The claim recites the additional elements:
“after configuring the at least one intelligent functionality . . . to the second communication node . . . from the second communication node based on the additional information” (merely generally linking the use of the judicial exception to a particular technological environment or field of use does not provide an inventive concept) and
“transmitting additional information . . . receiving functionality reconfiguration information” (transmitting data is well‐understood, routine, and conventional, see Symantec, 838 F.3d at 1321, 120 USPQ2d at 1362; see also buySAFE, Inc. v. Google, Inc., 765 F.3d 1350, 1355, 112 USPQ2d 1093, 1096 (Fed. Cir. 2014); therefore the limitation, which is recited with a high level of generality, remains insignificant extra-solution activity even upon reconsideration).
Accordingly, Claim 3 is rejected as being directed to an abstract idea without significantly more.
Regarding Claim 4:
Step 2A Prong 1: See the rejection of Claim 1 above, which Claim 4 depends on. Here, the claim recites additional limitations that are mental processes. Specifically, the claim recites
“reconfiguring the at least one intelligent functionality by incorporating the additional information based on the functionality reconfiguration information” (mental process – amounts to exercising judgment to form on opinion on an adjustment to a state of mind or mental process, with reference to known or observed information, which may be aided by pen and paper).
Step 2A Prong 2: This judicial exception is not integrated into a practical application.
The claim recites the additional elements:
“after configuring the at least one intelligent functionality . . . from the second communication node” (amounts to merely generally linking the use of the judicial exception to a particular technological environment or field of use, which do not impose any meaningful limits on practicing the abstract idea) and
“receiving additional information and functionality reconfiguration information” (amounts to extra-solution activity because it is the transmission of data, which is incidental to the claimed subject matter).
Step 2B: The claim does not include additional elements considered individually and in combination that are sufficient to amount to significantly more than the judicial exception.
The claim recites the additional elements:
“after configuring the at least one intelligent functionality . . . from the second communication node” (merely generally linking the use of the judicial exception to a particular technological environment or field of use does not provide an inventive concept) and
“receiving additional information and functionality reconfiguration information” (transmitting data is well‐understood, routine, and conventional, see Symantec, 838 F.3d at 1321, 120 USPQ2d at 1362; see also buySAFE, Inc. v. Google, Inc., 765 F.3d 1350, 1355, 112 USPQ2d 1093, 1096 (Fed. Cir. 2014); therefore the limitation, which is recited with a high level of generality, remains insignificant extra-solution activity even upon reconsideration).
Accordingly, Claim 4 is rejected as being directed to an abstract idea without significantly more.
Regarding Claim 5:
Step 2A Prong 1: See the rejection of Claim 1 above, which Claim 5 depends on.
Step 2A Prong 2: This judicial exception is not integrated into a practical application.
The claim recites the additional elements:
“corresponding to the at least one intelligent functionality from the second communication node . . . corresponding to the dataset information from the second communication node and acquiring the dataset as the training dataset” (amounts to merely generally linking the use of the judicial exception to a particular technological environment or field of use, which do not impose any meaningful limits on practicing the abstract idea) and
“wherein the obtaining of the training dataset comprises: receiving dataset information . . . receiving a dataset” (amounts to extra-solution activity because it is the transmission of data, which is incidental to the claimed subject matter).
Step 2B: The claim does not include additional elements considered individually and in combination that are sufficient to amount to significantly more than the judicial exception.
The claim recites the additional elements:
“corresponding to the at least one intelligent functionality from the second communication node . . . corresponding to the dataset information from the second communication node and acquiring the dataset as the training dataset” (merely generally linking the use of the judicial exception to a particular technological environment or field of use does not provide an inventive concept) and
“wherein the obtaining of the training dataset comprises: receiving dataset information . . . receiving a dataset” (transmitting data is well‐understood, routine, and conventional, see Symantec, 838 F.3d at 1321, 120 USPQ2d at 1362; see also buySAFE, Inc. v. Google, Inc., 765 F.3d 1350, 1355, 112 USPQ2d 1093, 1096 (Fed. Cir. 2014); therefore the limitation, which is recited with a high level of generality, remains insignificant extra-solution activity even upon reconsideration).
Accordingly, Claim 5 is rejected as being directed to an abstract idea without significantly more.
Regarding Claim 6:
Step 2A Prong 1: See the rejection of Claim 1 above, which Claim 6 depends on.
Step 2A Prong 2: This judicial exception is not integrated into a practical application.
The claim recites the additional elements:
“based on data collection configuration . . . from the second communication node . . . in response to . . . request signal from the second communication node . . . to the second communication node . . . including one or more data units corresponding to the at least one intelligent functionality among the plurality of data units from the second communication node . . . as the training dataset” (amounts to merely generally linking the use of the judicial exception to a particular technological environment or field of use, which do not impose any meaningful limits on practicing the abstract idea) and
“wherein the obtaining of the training dataset comprises: collecting a plurality of data units . . . information received . . . receipt of a data . . . transmitting the plurality of data units . . . receiving a dataset . . . and acquiring the dataset” (amounts to extra-solution activity because it is the transmission of data, which is incidental to the claimed subject matter).
Step 2B: The claim does not include additional elements considered individually and in combination that are sufficient to amount to significantly more than the judicial exception.
The claim recites the additional elements:
“based on data collection configuration . . . from the second communication node . . . in response to . . . request signal from the second communication node . . . to the second communication node . . . including one or more data units corresponding to the at least one intelligent functionality among the plurality of data units from the second communication node . . . as the training dataset” (merely generally linking the use of the judicial exception to a particular technological environment or field of use does not provide an inventive concept) and
“wherein the obtaining of the training dataset comprises: collecting a plurality of data units . . . information received . . . receipt of a data . . . transmitting the plurality of data units . . . receiving a dataset . . . and acquiring the dataset” (transmitting data is well‐understood, routine, and conventional, see Symantec, 838 F.3d at 1321, 120 USPQ2d at 1362; see also buySAFE, Inc. v. Google, Inc., 765 F.3d 1350, 1355, 112 USPQ2d 1093, 1096 (Fed. Cir. 2014); therefore the limitation, which is recited with a high level of generality, remains insignificant extra-solution activity even upon reconsideration).
Accordingly, Claim 6 is rejected as being directed to an abstract idea without significantly more.
Regarding Claim 7:
Step 2A Prong 1: See the rejection of Claim 6 above, which Claim 7 depends on.
Step 2A Prong 2: This judicial exception is not integrated into a practical application.
The claim recites the additional elements:
“of the first communication node to the second communication node . . . to the second communication node . . . collected up to a time of changing configuration of the first communication node to the second communication node upon receipt of . . . from the second communication node based on the configuration change information” (amounts to merely generally linking the use of the judicial exception to a particular technological environment or field of use, which do not impose any meaningful limits on practicing the abstract idea) and
“wherein the obtaining of the training dataset comprises: transmitting configuration change information . . . and the transmitting of the plurality of data units . . . comprises: transmitting the plurality of data units . . . a data request signal transmitted” (amounts to extra-solution activity because it is the transmission of data, which is incidental to the claimed subject matter).
Step 2B: The claim does not include additional elements considered individually and in combination that are sufficient to amount to significantly more than the judicial exception.
The claim recites the additional elements:
“of the first communication node to the second communication node . . . to the second communication node . . . collected up to a time of changing configuration of the first communication node to the second communication node upon receipt of . . . from the second communication node based on the configuration change information” (merely generally linking the use of the judicial exception to a particular technological environment or field of use does not provide an inventive concept) and
“wherein the obtaining of the training dataset comprises: transmitting configuration change information . . . and the transmitting of the plurality of data units . . . comprises: transmitting the plurality of data units . . . a data request signal transmitted” (transmitting data is well‐understood, routine, and conventional, see Symantec, 838 F.3d at 1321, 120 USPQ2d at 1362; see also buySAFE, Inc. v. Google, Inc., 765 F.3d 1350, 1355, 112 USPQ2d 1093, 1096 (Fed. Cir. 2014); therefore the limitation, which is recited with a high level of generality, remains insignificant extra-solution activity even upon reconsideration).
Accordingly, Claim 7 is rejected as being directed to an abstract idea without significantly more.
Regarding Claim 8:
Step 2A Prong 1: See the rejection of Claim 1 above, which Claim 8 depends on.
Step 2A Prong 2: This judicial exception is not integrated into a practical application.
The claim recites the additional elements:
“from the second communication node . . . corresponding to the at least one intelligent functionality based on the RSs and the associated ID and . . . as the training data” (amounts to merely generally linking the use of the judicial exception to a particular technological environment or field of use, which do not impose any meaningful limits on practicing the abstract idea) and
“wherein the obtaining of the training dataset comprises: receiving the associated ID and reference signals (RSs) . . . and collecting a plurality of data units . . . acquiring the plurality of data units” (amounts to extra-solution activity because it is the transmission of data, which is incidental to the claimed subject matter).
Step 2B: The claim does not include additional elements considered individually and in combination that are sufficient to amount to significantly more than the judicial exception.
The claim recites the additional elements:
“from the second communication node . . . corresponding to the at least one intelligent functionality based on the RSs and the associated ID and . . . as the training data” (merely generally linking the use of the judicial exception to a particular technological environment or field of use does not provide an inventive concept) and
“wherein the obtaining of the training dataset comprises: receiving the associated ID and reference signals (RSs) . . . and collecting a plurality of data units . . . acquiring the plurality of data units” (transmitting data is well‐understood, routine, and conventional, see Symantec, 838 F.3d at 1321, 120 USPQ2d at 1362; see also buySAFE, Inc. v. Google, Inc., 765 F.3d 1350, 1355, 112 USPQ2d 1093, 1096 (Fed. Cir. 2014); therefore the limitation, which is recited with a high level of generality, remains insignificant extra-solution activity even upon reconsideration).
Accordingly, Claim 8 is rejected as being directed to an abstract idea without significantly more.
Regarding Claim 9:
Step 2A Prong 1: See the rejection of Claim 1 above, which Claim 9 depends on.
Step 2A Prong 2: This judicial exception is not integrated into a practical application.
The claim recites the additional elements:
“after the training of the intelligent model . . . for the intelligent model to the second communication node” (amounts to merely generally linking the use of the judicial exception to a particular technological environment or field of use, which do not impose any meaningful limits on practicing the abstract idea) and
“transmitting model information” (amounts to extra-solution activity because it is the transmission of data, which is incidental to the claimed subject matter).
Step 2B: The claim does not include additional elements considered individually and in combination that are sufficient to amount to significantly more than the judicial exception.
The claim recites the additional elements:
“after the training of the intelligent model . . . for the intelligent model to the second communication node” (merely generally linking the use of the judicial exception to a particular technological environment or field of use does not provide an inventive concept) and
“transmitting model information” (transmitting data is well‐understood, routine, and conventional, see Symantec, 838 F.3d at 1321, 120 USPQ2d at 1362; see also buySAFE, Inc. v. Google, Inc., 765 F.3d 1350, 1355, 112 USPQ2d 1093, 1096 (Fed. Cir. 2014); therefore the limitation, which is recited with a high level of generality, remains insignificant extra-solution activity even upon reconsideration).
Accordingly, Claim 9 is rejected as being directed to an abstract idea without significantly more.
Regarding Claim 10:
Step 2A Prong 1: See the rejection of Claim 1 above, which Claim 10 depends on. Here, the claim recites additional limitations that are mental processes. Specifically, the claim recites
“wherein the performing of the intelligent operations comprises . . . perform the at least one intelligent functionality based on a function activation signal . . . and deactivating operations . . . based on a function deactivation signal” (mental process – amounts to exercising judgment to perform evaluations and form opinions, with reference to the above discussed state of mind or the mental process where specific processes can be performed or discarded based on observed signals, which may be aided by pen and paper).
Step 2A Prong 2: This judicial exception is not integrated into a practical application.
The claim recites the additional elements:
“operating the intelligent model to” (amounts to mere instructions to apply the judicial exception on generic and unspecialized computer components, which do not impose any meaningful limits on practicing the abstract idea);
“from the second communication node . . . of the intelligent model . . . from the second communication node” (amounts to merely generally linking the use of the judicial exception to a particular technological environment or field of use, which do not impose any meaningful limits on practicing the abstract idea); and
“signal received . . . signal received” (amounts to extra-solution activity because it is the transmission of data, which is incidental to the claimed subject matter).
Step 2B: The claim does not include additional elements considered individually and in combination that are sufficient to amount to significantly more than the judicial exception.
The claim recites the additional elements:
“operating the intelligent model to” (mere instructions to apply the exception using generic computer components does not provide an inventive concept);
“from the second communication node . . . of the intelligent model . . . from the second communication node” (merely generally linking the use of the judicial exception to a particular technological environment or field of use does not provide an inventive concept); and
“signal received . . . signal received” (transmitting data is well‐understood, routine, and conventional, see Symantec, 838 F.3d at 1321, 120 USPQ2d at 1362; see also buySAFE, Inc. v. Google, Inc., 765 F.3d 1350, 1355, 112 USPQ2d 1093, 1096 (Fed. Cir. 2014); therefore the limitation, which is recited with a high level of generality, remains insignificant extra-solution activity even upon reconsideration).
Accordingly, Claim 10 is rejected as being directed to an abstract idea without significantly more.
Regarding Claim 11, the claim recites limitations that are all substantially the same as limitations of Claim 1, in the form of a machine with generic computer components. For substantially the same reasoning, the claim is also directed to performing mental processes without integration into a practical application or amounting to either significantly more or an inventive concept.
Accordingly, Claim 11 is rejected under the same rationale.
Regarding Claim 12, the claim recites limitations that are all substantially the same as limitations of Claim 2, in the form of a machine with generic computer components. For substantially the same reasoning, the claim is also directed to performing mental processes without integration into a practical application or amounting to either significantly more or an inventive concept.
Accordingly, Claim 12 is rejected under the same rationale.
Regarding Claim 13, the claim recites limitations that are all substantially the same as limitations of Claim 3, in the form of a machine with generic computer components. For substantially the same reasoning, the claim is also directed to performing mental processes without integration into a practical application or amounting to either significantly more or an inventive concept.
Accordingly, Claim 13 is rejected under the same rationale.
Regarding Claim 14, the claim recites limitations that are all substantially the same as limitations of Claim 4, in the form of a machine with generic computer components. For substantially the same reasoning, the claim is also directed to performing mental processes without integration into a practical application or amounting to either significantly more or an inventive concept.
Accordingly, Claim 14 is rejected under the same rationale.
Regarding Claim 15, the claim recites limitations that are all substantially the same as limitations of Claim 5, in the form of a machine with generic computer components. For substantially the same reasoning, the claim is also directed to performing mental processes without integration into a practical application or amounting to either significantly more or an inventive concept.
Accordingly, Claim 15 is rejected under the same rationale.
Regarding Claim 16, the claim recites limitations that are all substantially the same as limitations of Claim 6, in the form of a machine with generic computer components. For substantially the same reasoning, the claim is also directed to performing mental processes without integration into a practical application or amounting to either significantly more or an inventive concept.
Accordingly, Claim 16 is rejected under the same rationale.
Regarding Claim 17, the claim recites limitations that are all substantially the same as limitations of Claim 7, in the form of a machine with generic computer components. For substantially the same reasoning, the claim is also directed to performing mental processes without integration into a practical application or amounting to either significantly more or an inventive concept.
Accordingly, Claim 17 is rejected under the same rationale.
Regarding Claim 18, the claim recites limitations that are all substantially the same as limitations of Claim 8, in the form of a machine with generic computer components. For substantially the same reasoning, the claim is also directed to performing mental processes without integration into a practical application or amounting to either significantly more or an inventive concept.
Accordingly, Claim 18 is rejected under the same rationale.
Regarding Claim 19, the claim recites limitations that are all substantially the same as limitations of Claim 9, in the form of a machine with generic computer components. For substantially the same reasoning, the claim is also directed to performing mental processes without integration into a practical application or amounting to either significantly more or an inventive concept.
Accordingly, Claim 19 is rejected under the same rationale.
Regarding Claim 20, the claim recites limitations that are all substantially the same as limitations of Claim 10, in the form of a machine with generic computer components. For substantially the same reasoning, the claim is also directed to performing mental processes without integration into a practical application or amounting to either significantly more or an inventive concept.
Accordingly, Claim 20 is rejected under the same rationale.
Claim Rejections - 35 USC § 102
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 the appropriate paragraphs of 35 U.S.C. 102 that form the basis for the rejections under this section made in this Office action:
A person shall be entitled to a patent unless –
(a)(1) the claimed invention was patented, described in a printed publication, or in public use, on sale, or otherwise available to the public before the effective filing date of the claimed invention.
Claims 1, 3-5, 9, 11, 13-15, and 19 are rejected under 35 U.S.C. 102(a)(1) as being anticipated by Jeon et al. (hereinafter Jeon) (Pat. App. Pub. No. US 2022/0114448 A1).
Regarding Claim 1, Jeon teaches a method of a first communication node, comprising (Para. [0040], “a UE . . . may include at least one processor and at least one memory . . . [which] stores instructions that, when executed, cause the at least one processor to perform operations according to some implementations of the present disclosure” and Para. [0036] – [0037], “The term UE may be referred to as terminal equipment, mobile station (MS), mobile terminal (MT), user terminal (UT), subscriber station (SS), wireless device, personal digital assistant (PDA), wireless modem, handheld device, etc. In the present disclosure, a BS refers to a fixed station that communicates with a UE and/or another BS and exchanges data and control information with a UE and another BS . . . a node refers to a fixed point capable of transmitting/receiving a radio signal to/from a UE or BS”, where a method comprising operations, “to perform operations according to some implementations of the present disclosure”, of a “UE”, which is a first communication node, “a BS . . . communicates with a UE . . . exchanges data and control information with a UE . . . a node refers to a fixed point capable of transmitting/receiving a radio signal to/from a UE”, is disclosed; see also Para. [0007], “In one aspect, provided is a method for updating a neural network (NN) model in a neural network (NN) based wireless communication system” and Fig. 1-6, where additional aspects of the “method” are further discussed):
configuring at least one intelligent functionality among a plurality of intelligent functionalities based on functionality configuration information received from a second communication node (Para. [0054], “In each round of PARTEL, a server partitions a global model under training into blocks of parameters, called parametric blocks, and allocate each of them to a single device for updating. In other words, PARTEL supports the distributed training of a large-scale AI model by dynamically partitioning the model and allocating the resultant parametric blocks to different devices for updating”; Para. [0060], “Each worker (e.g. UE) performs one task assigned by the server”; Para. [0083], “The training process in FIG. 2 iterates when all parametric blocks are updated in the round, i.e., the tasks of all workers are synchronized in each round”; and Fig. 6, “S602 Allocate partial model parameter w1 to UE 1”, where the first communication node, such as “UE 1”, configures, “all parametric blocks are updated in the round, i.e., the tasks of all workers are synchronized in each round”, at least one intelligent functionality, “Each worker (e.g. UE) performs one task ”, among a plurality of intelligent functionalities, “The training process . . . [of] tasks”, based on functionality configuration, “task assigned” and “Allocate partial model parameter w1”, information received from a “server”, “a server partitions a global model under training into blocks of parameters, called parametric blocks, and allocate each of them to a single device for updating” and “one task assigned by the server”, which is a second communication node, see Para. [0036] – [0037], “a BS refers to a fixed station that communicates with a UE and/or another BS and exchanges data and control information with a UE and another BS. The term BS may be referred to as [a] . . . server . . . a node refers to a fixed point capable of transmitting/receiving a radio signal to/from a UE or BS”);
obtaining a training dataset corresponding to the at least one intelligent functionality based on an associated identifier (ID) received from the second communication node (Fig. 6, “S606 Provide data sample indexes M1 for UE1”; Para. [0083] - [0084], “all parametric blocks are updated in the round, i.e., the tasks of all workers are synchronized in each round . . . W-stage: The parameter matrix W is divided into K blocks, with each being updated by one worker”; and Para. [0072], “W-stage: Fixing the values of Z, solve the problem of in which the problem of each neuron is independent and can be written as
PNG
media_image1.png
88
382
media_image1.png
Greyscale
”, where the first communication node, such as “UE1”, obtains a training dataset, “Provide data sample indexes M1 for UE1” of dataset “m”, see Para. [0045], Table 1, “data sample m”, corresponding to the at least one intelligent functionality, “W-stage: The parameter matrix W is divided into K blocks, with each being updated by one worker” which, as shown in Eq. (2), uses “m”, based on an associated identifier (ID), “indexes M1”, received, “Provide data sample indexes M1”, from the second communication node “Server”, as indicated by the arrow from “Server” to “UE 1” in step “S606”; alternatively, see Para. [0054], “In each round of PARTEL, a server partitions a global model under training into blocks of parameters, called parametric blocks, and allocate each of them to a single device for updating. In other words, PARTEL supports the distributed training of a large-scale AI model by dynamically partitioning the model and allocating the resultant parametric blocks to different devices for updating”; Para. [0083], “The training process in FIG. 2 iterates when all parametric blocks are updated in the round, i.e., the tasks of all workers are synchronized in each round”; and Fig. 6, “S602 Allocate partial model parameter w1 to UE 1”, where the first communication node, such as “UE 1”, obtains a training dataset, “Allocate partial model parameter w1 to UE 1”, corresponding to the at least one intelligent functionality, “all parametric blocks are updated in the round, i.e., the tasks of all workers are synchronized in each round”, based on an associated identifier (ID), such as the identifier of “w1”, received from the second communication node, “a server partitions a global model under training into blocks of parameters, called parametric blocks, and allocate each of them to a single device”);
training an intelligent model to perform the at least one intelligent functionality using the training dataset (Para. [0081], “Computation Phase: Each worker computes the update (e.g., gradients or the proximal gradients) of the its allocated parametric block”; Para. [0060], “Each worker (e.g. UE) performs one task assigned by the server”; and Para. [0083], “The training process in FIG. 2 iterates when all parametric blocks are updated in the round, i.e., the tasks of all workers are synchronized in each round”, where the first communication node, “Each worker (e.g. UE)” and “all workers”, trains an intelligent model to perform the at least one intelligent functionality, “computes the update (e.g., gradients or the proximal gradients) of the its allocated parametric block” “all parametric blocks are updated in the round, i.e., the tasks of all workers are synchronized in each round”, using the training dataset, see Para. [0072], “W-stage: Fixing the values of Z, solve the problem of in which the problem of each neuron is independent and can be written as
PNG
media_image1.png
88
382
media_image1.png
Greyscale
” and Para. [0084], “W-stage: The parameter matrix W is divided into K blocks, with each being updated by one worker”, where the training dataset, data sample “m”, is used for the “W-stage”, which “update[s]” the “parameter matrix”, or alternatively, the training dataset “parameter[s]” themselves are used in their own updating, “all parametric blocks are updated in the round”); and
performing intelligent operations according to the at least one intelligent functionality using the trained intelligent model (Para. [0054], “In each round of PARTEL, a server partitions a global model under training into blocks of parameters, called parametric blocks, and allocate each of them to a single device for updating . . . Then devices upload the updates to a server where they are assembled and then applied to updating the model. The two steps are iterated till the model converges”; Para. [0060], “Each worker (e.g. UE) performs one task assigned by the server”; Para. [0083], “The training process in FIG. 2 iterates when all parametric blocks are updated in the round, i.e., the tasks of all workers are synchronized in each round”; and Para. [0175], “FIG. 6 . . . S602 to S609 may be iterated periodically or upon request of the server for updating the CNN model”, where “training . . . steps are iterated till the model converges”, such that, for a subsequent “iterat[ion]”, the first communication node, performs, “worker (e.g. UE) performs”, intelligent operations, “updating”, see Para. [0081], “Each worker computes the update (e.g., gradients or the proximal gradients) of the its allocated parametric block”, according to the at least one intelligent functionality, “when all parametric blocks are updated in the round, i.e., the tasks of all workers are synchronized in each round”, using the trained intelligent model, “the CNN model”, as trained in the previous “iterat[ion]”; see also Para. [0068], “The auxiliary variables are used to partition a CNN model to make workers (e.g. UEs) train the respective partitions of the CNN model, and are not used in inference using the trained CNN model”, where “inference using the trained CNN model” is discussed; see also Para. [0082], “Pull Phase: All workers upload the updates of their corresponding parametric blocks to the server. The server updates the whole parameter vector”, where the “Pull Phase” could also be described performing intelligent operations, “All workers upload the updates”, according to the at least one intelligent functionality, the task “corresponding [to their] parametric blocks”, using the trained intelligent model, “corresponding parametric blocks” as trained in the current and previous “iterat[ion]”).
Regarding Claim 3, Jeon teaches the method according to claim 1, further comprising: after configuring the at least one intelligent functionality, transmitting additional information to the second communication node (Para. [0175, “FIG. 6 illustrates an example of the update procedure of a neural network (NN) model . . . two workers (UE 1 and UE 2) participate in the partitioned edge learning. The CNN model may be updated by partitioning the model parameters W and the auxiliary matrices Z to the workers for edge-learning (S601 to S 609). S602 to S609 may be iterated periodically or upon request of the server for updating the CNN model”; Para. [0060], “Each worker (e.g. UE) performs one task assigned by the server”; Para. [0083], “The training process in FIG. 2 iterates when all parametric blocks are updated in the round, i.e., the tasks of all workers are synchronized in each round”; and Fig. 6, wherein, after configuring the at least one intelligent functionality, “performs one task assigned” to “Update w1 given data set {xm}”, the first communication node, such as “UE 1”, transmits additional information, “Report updated partial model parameter w1” and “Report updated auxiliary matrices {zm}”, to the second communication node, “Server”; see also Fig. 1; Fig. 3; and Para. [0078] – [0082], “Considering the PARTEL system shown in FIG. 1, its operations shown in FIG. 3 may be elaborated as follows . . . Pull Phase: All workers upload the updates of their corresponding parametric blocks to the server. The server updates the whole parameter vector”; see also Para. [0054], “In each round of PARTEL, a server partitions a global model under training into blocks of parameters, called parametric blocks, and allocate each of them to a single device for updating . . . Then devices upload the updates to a server where they are assembled and then applied to updating the model”, wherein after configuring the at least one intelligent functionality, “a single device for updating”, additional information is transmitted by the first communication node, “devices upload the updates”, to the second communication node, “to a server”);
receiving functionality reconfiguration information from the second communication node based on the additional information; and reconfiguring the at least one intelligent functionality by incorporating the additional information based on the functionality reconfiguration information (Para. [0054], “In each round of PARTEL, a server partitions a global model under training into blocks of parameters, called parametric blocks, and allocate each of them to a single device for updating. In other words, PARTEL supports the distributed training of a large-scale AI model by dynamically partitioning the model and allocating the resultant parametric blocks to different devices for updating”; Para. [0060], “Each worker (e.g. UE) performs one task assigned by the server”; Para. [0083], “The training process in FIG. 2 iterates when all parametric blocks are updated in the round, i.e., the tasks of all workers are synchronized in each round”; and Fig. 6, “S602 Allocate partial model parameter w1 to UE 1”, where the first communication node, such as “UE 1”, configures, “all parametric blocks are updated in the round, i.e., the tasks of all workers are synchronized in each round”, at least one intelligent functionality, “Each worker (e.g. UE) performs one task ”, based on the functionality configuration, “task assigned” and “Allocate partial model parameter w1”, information received from the second communication node “server”, “a server partitions a global model under training into blocks of parameters, called parametric blocks, and allocate each of them to a single device for updating” and “one task assigned by the server”, and where for subsequent iterations, see Para. [0054], “In each round of PARTEL, a server partitions a global model under training into blocks of parameters, called parametric blocks, and allocate each of them to a single device for updating . . . Then devices upload the updates to a server where they are assembled and then applied to updating the model. The two steps are iterated till the model converges” and Para. [0175], “FIG. 6 . . . S602 to S609 may be iterated periodically or upon request of the server for updating the CNN model”, the configuring is reconfiguration based on the additional information, the “updates” to the “parametric blocks” from the previous iteration, that incorporates the additional information, the “updates” to the “parametric blocks” from the previous iteration, based on the functionality reconfiguration information, “task assigned” and “Allocate partial model parameter w1”).
Regarding Claim 4, Jeon teaches the method according to claim 1, further comprising: after configuring the at least one intelligent functionality, receiving additional information and functionality reconfiguration information from the second communication node; and reconfiguring the at least one intelligent functionality by incorporating the additional information based on the functionality reconfiguration information (Para. [0054], “In each round of PARTEL, a server partitions a global model under training into blocks of parameters, called parametric blocks, and allocate each of them to a single device for updating. In other words, PARTEL supports the distributed training of a large-scale AI model by dynamically partitioning the model and allocating the resultant parametric blocks to different devices for updating”; Para. [0060], “Each worker (e.g. UE) performs one task assigned by the server”; Para. [0083], “The training process in FIG. 2 iterates when all parametric blocks are updated in the round, i.e., the tasks of all workers are synchronized in each round”; and Fig. 6, “S602 Allocate partial model parameter w1 to UE 1”, where the first communication node, such as “UE 1”, configures, “all parametric blocks are updated in the round, i.e., the tasks of all workers are synchronized in each round”, at least one intelligent functionality, “Each worker (e.g. UE) performs one task ”, among a plurality of intelligent functionalities, “The training process . . . [of] tasks”, based on functionality configuration, “task assigned”, and by incorporating the additional information “Allocate partial model parameter w1”, both received from the second communication node, “server”, “a server partitions a global model under training into blocks of parameters, called parametric blocks, and allocate each of them to a single device for updating” and “one task assigned by the server”, and where for subsequent iterations, see Para. [0054], “In each round of PARTEL, a server partitions a global model under training into blocks of parameters, called parametric blocks, and allocate each of them to a single device for updating . . . Then devices upload the updates to a server where they are assembled and then applied to updating the model. The two steps are iterated till the model converges” and Para. [0175], “FIG. 6 . . . S602 to S609 may be iterated periodically or upon request of the server for updating the CNN model”, the configuring is reconfiguration that occurs after configuring the at least one intelligent functionality in the previous iteration).
Regarding Claim 5, Jeon teaches the method according to claim 1, wherein the obtaining of the training dataset comprises: receiving dataset information corresponding to the at least one intelligent functionality from the second communication node (Fig. 6, “S606 Provide data sample indexes M1 for UE1”; Para. [0083] - [0084], “all parametric blocks are updated in the round, i.e., the tasks of all workers are synchronized in each round . . . W-stage: The parameter matrix W is divided into K blocks, with each being updated by one worker”; and Para. [0072], “W-stage: Fixing the values of Z, solve the problem of in which the problem of each neuron is independent and can be written as
PNG
media_image1.png
88
382
media_image1.png
Greyscale
”, where the first communication node, such as “UE1”, receives dataset information, “Provide data sample indexes M1 for UE1” of dataset “m”, see Para. [0045], Table 1, “data sample m”, corresponding to the at least one intelligent functionality, “W-stage: The parameter matrix W is divided into K blocks, with each being updated by one worker” which, as shown in Eq. (2), uses “m”, based on an associated identifier (ID), “indexes M1”, received, “Provide data sample indexes M1”, from the second communication node “Server”, as indicated by the arrow from “Server” to “UE 1” in step “S606”; alternatively, see Para. [0054], “In each round of PARTEL, a server partitions a global model under training into blocks of parameters, called parametric blocks, and allocate each of them to a single device for updating. In other words, PARTEL supports the distributed training of a large-scale AI model by dynamically partitioning the model and allocating the resultant parametric blocks to different devices for updating”; Para. [0083], “The training process in FIG. 2 iterates when all parametric blocks are updated in the round, i.e., the tasks of all workers are synchronized in each round”; and Fig. 6, “S602 Allocate partial model parameter w1 to UE 1”, where the first communication node, such as “UE 1”, obtains a training dataset, “Allocate partial model parameter w1 to UE 1”, corresponding to the at least one intelligent functionality, “all parametric blocks are updated in the round, i.e., the tasks of all workers are synchronized in each round”, based on an associated identifier (ID), such as the identifier of “w1”, received from the second communication node, “a server partitions a global model under training into blocks of parameters, called parametric blocks, and allocate each of them to a single device”); and
receiving a dataset corresponding to the dataset information from the second communication node and acquiring the dataset as the training dataset (Para. [0081], “Computation Phase: Each worker computes the update (e.g., gradients or the proximal gradients) of the its allocated parametric block”; Para. [0060], “Each worker (e.g. UE) performs one task assigned by the server”; Para. [0083] - [0084], “all parametric blocks are updated in the round, i.e., the tasks of all workers are synchronized in each round . . . W-stage: The parameter matrix W is divided into K blocks, with each being updated by one worker”; and Para. [0072], “W-stage: Fixing the values of Z, solve the problem of in which the problem of each neuron is independent and can be written as
PNG
media_image1.png
88
382
media_image1.png
Greyscale
”, where the first communication node, “worker (e.g. UE)”, must receive a dataset “m”, see Para. [0045], Table 1, “data sample m”, acquiring the dataset as the training dataset in order to use it to “compute[] the update (e.g., gradients or the proximal gradients) of the its allocated parametric block”, as shown in Eq. 2, and where the dataset “m” corresponds with the dataset corresponding to the dataset information from the second communication node, see Fig. 6, “S606 Provide data sample indexes M1 for UE1”, where “data sample indexes M1” corresponds with m as its “indexes”, or alternatively, see Fig. 6, “S602 Allocate partial model parameter w1 to UE 1”, where the dataset information, “Allocate partial model parameter” corresponds with the data used to train it, “update (e.g., gradients or the proximal gradients) of the its allocated parametric block”).
Regarding Claim 9, Jeon teaches the method according to claim 1, further comprising: after the training of the intelligent model, transmitting model information for the intelligent model to the second communication node (Para. [0175, “FIG. 6 illustrates an example of the update procedure of a neural network (NN) model . . . two workers (UE 1 and UE 2) participate in the partitioned edge learning. The CNN model may be updated by partitioning the model parameters W and the auxiliary matrices Z to the workers for edge-learning (S601 to S 609). S602 to S609 may be iterated periodically or upon request of the server for updating the CNN model” and Fig. 6, wherein, after the training of the intelligent model, “Update w1 given data set {xm}”, the first communication node, such as “UE 1”, transmits model information for the intelligent model, “Report updated partial model parameter w1” and “Report updated auxiliary matrices {zm}”, to the second communication node, “Server”; see also Fig. 1; Fig. 3; and Para. [0078] – [0082], “Considering the PARTEL system shown in FIG. 1, its operations shown in FIG. 3 may be elaborated as follows . . . Pull Phase: All workers upload the updates of their corresponding parametric blocks to the server. The server updates the whole parameter vector”; see also Para. [0054], “In each round of PARTEL, a server partitions a global model under training into blocks of parameters, called parametric blocks, and allocate each of them to a single device for updating . . . Then devices upload the updates to a server where they are assembled and then applied to updating the model”, wherein after the training of the intelligent model, “a single device for updating”, model information for the intelligent model is transmitted by the first communication node, “devices upload the updates”, to the second communication node, “to a server”).
Regarding Claim 11, Jeon teaches a first communication node comprising at least one processor, wherein the at least one processor causes the first communication node to perform . . . (Para. [0040], “a UE . . . may include at least one processor and at least one memory and additionally further include at least one transceiver. The at least one memory stores instructions that, when executed, cause the at least one processor to perform operations according to some implementations of the present disclosure which are described hereinafter”, where, as discussed above, the “UE” is the first communication node, which comprises at least one processor, “may include at least one processor”, wherein the at least one processor causes the first communication node to perform, “The at least one memory stores instructions that, when executed, cause the at least one processor to perform operations”; see also Para. [0041], “The functions, procedures, proposals, and/or methods disclosed in this document may be implemented using firmware or software, and the firmware or software may be configured to include the modules, procedures, or functions. Firmware or software configured to perform the functions, procedures, proposals, and/or methods disclosed in this document may be included in the at least one processor or stored in the at least one memory so as to be driven by the at least one processor. The functions, procedures, proposals, and/or methods disclosed in this document may be implemented using firmware or software in the form of code, commands, and/or a set of command”).
The remaining limitations are substantially the same as limitations of claim 1, therefore it is rejected under the same rationale.
Regarding Claim 13, the additional elements of the dependent claim are substantially the same as limitations of Claim 3, therefore it is rejected under the same rationale.
Regarding Claim 14, the additional elements of the dependent claim are substantially the same as limitations of Claim 4, therefore it is rejected under the same rationale.
Regarding Claim 15, the additional elements of the dependent claim are substantially the same as limitations of Claim 5, therefore it is rejected under the same rationale.
Regarding Claim 19, the additional elements of the dependent claim are substantially the same as limitations of Claim 9, therefore it is rejected under the same rationale.
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 2, 10, 12, and 20 are rejected under 35 U.S.C. 103 as being unpatentable over Jeon in view of Meyer et al. (hereinafter Meyer) (Pat. No. US 12,265,888 B1).
Regarding Claim 2, Jeon teaches the method according to claim 1, wherein the configuring of the at least one intelligent functionality comprises: . . . [allocation of] each of the plurality of intelligent functionalities . . . [and] in response to the functionality configuration information received from the second communication node . . . , configuring the at least one intelligent functionality corresponding to the functionality configuration information among the plurality of intelligent functionalities (Para. [0054], “In each round of PARTEL, a server partitions a global model under training into blocks of parameters, called parametric blocks, and allocate each of them to a single device for updating. In other words, PARTEL supports the distributed training of a large-scale AI model by dynamically partitioning the model and allocating the resultant parametric blocks to different devices for updating”; Para. [0060], “Each worker (e.g. UE) performs one task assigned by the server”; and Fig. 6, “S602 Allocate partial model parameter w1 to UE 1”, where the first communication node, such as “UE 1”, configures, “all parametric blocks are updated in the round, i.e., the tasks of all workers are synchronized in each round”, at least one intelligent functionality, “Each worker (e.g. UE) performs one task ”, among a plurality of intelligent functionalities, “The training process . . . [of] tasks”, in response to and corresponding to the functionality configuration, “task assigned” and “Allocate partial model parameter w1”, information received from a “server”, “a server partitions a global model under training into blocks of parameters, called parametric blocks, and allocate each of them to a single device for updating” and “one task assigned by the server”, which is the second communication node, see Para. [0036] – [0037], “a BS refers to a fixed station that communicates with a UE and/or another BS and exchanges data and control information with a UE and another BS. The term BS may be referred to as [a] . . . server . . . a node refers to a fixed point capable of transmitting/receiving a radio signal to/from a UE or BS”; see also Para. [0083], “The training process in FIG. 2 iterates when all parametric blocks are updated in the round, i.e., the tasks of all workers are synchronized in each round”, where each of the plurality of intelligent functionalities, “all parametric blocks are updated in the round, i.e., the tasks”, are allocated to the first communication nodes, “all workers”); and
transmitting a completion report to the second communication node upon completion of the configuring of the at least one intelligent functionality (Para. [0175, “FIG. 6 illustrates an example of the update procedure of a neural network (NN) model . . . two workers (UE 1 and UE 2) participate in the partitioned edge learning. The CNN model may be updated by partitioning the model parameters W and the auxiliary matrices Z to the workers for edge-learning (S601 to S 609). S602 to S609 may be iterated periodically or upon request of the server for updating the CNN model”; Para. [0060], “Each worker (e.g. UE) performs one task assigned by the server”; Para. [0083], “The training process in FIG. 2 iterates when all parametric blocks are updated in the round, i.e., the tasks of all workers are synchronized in each round”; and Fig. 6, wherein, upon completion of the configuring of the at least one intelligent functionality, “Each worker (e.g. UE) performs one task ” to “Update w1 given data set {xm}”, the first communication node, such as “UE 1”, transmits a completion report, “Report updated partial model parameter w1” and “Report updated auxiliary matrices {zm}”, to the second communication node, “Server”; see also Fig. 1; Fig. 3; and Para. [0078] – [0082], “Considering the PARTEL system shown in FIG. 1, its operations shown in FIG. 3 may be elaborated as follows . . . Pull Phase: All workers upload the updates of their corresponding parametric blocks to the server. The server updates the whole parameter vector”; see also Para. [0054], “In each round of PARTEL, a server partitions a global model under training into blocks of parameters, called parametric blocks, and allocate each of them to a single device for updating . . . Then devices upload the updates to a server where they are assembled and then applied to updating the model”, wherein upon completion of the configuring of the at least one intelligent functionality, “a single device for updating”, a completion report is transmitted by the first communication node, “devices upload the updates”, to the second communication node, “to a server”).
Jeon does not explicitly disclose . . . in response to a capability information request received from the second communication node, transmitting, to the second communication node, a capability report including information on . . . supportable by the first communication node . . . based on the capability report . . . .
However, Meyer teaches . . . [a method for training a global model using a first and second communication node] (Abstract, “A system and method for performing machine learning in a mobile computing device which is configured to be coupled with a cloud computing system . . . and periodically transmitting changes to the local machine learning system from the mobile computing device to the global machine learning system hosted by the cloud computing system”, where the “method” is used for training a global model, “transmitting changes . . . to the global machine learning system”, using a first communication node, “a mobile computing device”, and a second communication node, “a cloud computing system”)
in response to a capability information request received from the second communication node, transmitting, to the second communication node, a capability report including information on [each of the plurality of intelligent functionalities] supportable by the first communication node . . . (Col. 5-6, ln. 61-33, “FIGS. 4 and 5 are flowcharts illustrating a process for providing a local machine learning system according to an embodiment . . . The mobile computing device 100 may receive a request from the global machine learning system to provide the capacity of the mobile computing device 100 (operation 410). For example, the capacity may include the available memory in memory 250, which can be used to store a local machine learning system. The capacity may also include additional information regarding the configuration of the mobile computing device 100 and other applications such as the system application 252. The mobile computing device 100 may transmit a mobile computing device capacity from the mobile computing device 100 to the global machine learning system stored in the memory 220 of the cloud computing system 210 through network 210 by using one or more of the communication components 270 of the mobile computing device 100 (operation 415) . . . The local machine model includes training parameters such as labels and weights. The local machine learning system may include one or more than one local machine learning mode”; Fig. 4; and Fig. 5, where in response to a capability information request received by the first communication node and from the second communication node, “The mobile computing device 100 may receive a request from the global machine learning system to provide the capacity of the mobile computing device 100 (operation 410)”, the first communication node transmits a capability report to the second communication node, “The mobile computing device 100 may transmit a mobile computing device capacity from the mobile computing device 100 to the global machine learning system stored in the memory 220 of the cloud computing system 210 through network 210 by using one or more of the communication components 270 of the mobile computing device 100 (operation 415)”, which includes including information, on each of the plurality of intelligent functionalities supportable by the first communication node, “the capacity may include the available memory in memory 250, which can be used to store a local machine learning system. The capacity may also include additional information regarding the configuration of the mobile computing device 100 and other applications such as the system application 252” which demonstrates supportability of a plurality of intelligent functionalities, “The local machine learning system may include one or more than one local machine learning mode”)
[receiving functionality configuration information from the second communication node] based on the capability report . . . (Col. 6, ln. 24-32, “Based on the mobile computing device capacity, the mobile computing device 100 may receive from the cloud computing system 210 a local machine learning system including a local machine learning model at the direction of the global machine learning system (operation 420). The local machine model includes training parameters such as labels and weights. The local machine learning system may include one or more than one local machine learning model”; Fig. 4; and Fig. 5, where the first communication node receives functionality configuration information from the second communication node, “the mobile computing device 100 may receive from the cloud computing system 210 a local machine learning system including a local machine learning model at the direction of the global machine learning system (operation 420) . . . The local machine learning system may include one or more than one local machine learning model”, based on the capability report, “Based on the mobile computing device capacity”).
Before the effective filing date of the invention, it would have been obvious to one of ordinary skill in the art to combine the configuring of the at least one intelligent functionality, which comprises allocation of each of the plurality of intelligent functionalities and in response to the functionality configuration information received from the second communication node, configuring by the first communication node of the at least one intelligent functionality corresponding to the functionality configuration information among the plurality of intelligent functionalities of Jeon with the method for training a global model using a first and second communication node, wherein in response to a capability information request received from the second communication node, the first communication node transmits, to the second communication node, a capability report including information on each of the plurality of intelligent functionalities supportable by the first communication node and receives functionality configuration information from the second communication node based on the capability report of Meyer in order to allocate intelligent functionalities across the distributed learning system (Jeon, Para. [0083], “all parametric blocks are updated in the round, i.e., the tasks of all workers are synchronized in each round”) based on the capacity of the worker devices to perform machine learning tasks and store machine learning data (Meyer, Col. 6, ln. 9-32, “The mobile computing device 100 may receive a request from the global machine learning system to provide the capacity of the mobile computing device 100 (operation 410). For example, the capacity may include the available memory in memory 250, which can be used to store a local machine learning system . . . Based on the mobile computing device capacity, the mobile computing device 100 may receive from the cloud computing system 210 a local machine learning system . . . The local machine learning system may include one or more than one local machine learning model”), which may need dynamic adjustments across training iterations based on evolving computer processes (Meyer, Col. 7-8, ln. 64-2, “the process or system may periodically check for any changes to the local machine learning system (e.g. changes in one or more training parameters in a local machine learning model) and forward any changes in order to reduce traffic (decrease latency), which greatly increases processing efficiency and reduces the amount of computing resource”; Meyer, Col. 6, ln. 14-17, “The capacity may also include additional information regarding the configuration of the mobile computing device 100 and other applications such as the system application 252”; see generally Jeon, Para.[0004], “Deploying machine learning algorithms at the network edge enables low-latency access to distributed data and computation resources, resulting in an active research area called edge learning”).
Regarding Claim 10, Jeon in view of Meyer teaches the method according to claim 1, wherein the performing of the intelligent operations comprises (Jeon, Para. [0054], “In each round of PARTEL, a server partitions a global model under training into blocks of parameters, called parametric blocks, and allocate each of them to a single device for updating . . . Then devices upload the updates to a server where they are assembled and then applied to updating the model. The two steps are iterated till the model converges”; Jeon, Para. [0060], “Each worker (e.g. UE) performs one task assigned by the server”; Jeon, Para. [0083], “The training process in FIG. 2 iterates when all parametric blocks are updated in the round, i.e., the tasks of all workers are synchronized in each round”; and Jeon, Para. [0175], “FIG. 6 . . . S602 to S609 may be iterated periodically or upon request of the server for updating the CNN model”, where “training . . . steps are iterated till the model converges”, such that, for a subsequent “iterat[ion]”, the first communication node, performs, “worker (e.g. UE) performs”, intelligent operations, “updating”, see Jeon, Para. [0081], “Each worker computes the update (e.g., gradients or the proximal gradients) of the its allocated parametric block”, according to the at least one intelligent functionality, “when all parametric blocks are updated in the round, i.e., the tasks of all workers are synchronized in each round”, using the trained intelligent model, “the CNN model”, as trained in the previous “iterat[ion]”; see also Jeon, Para. [0068], “The auxiliary variables are used to partition a CNN model to make workers (e.g. UEs) train the respective partitions of the CNN model, and are not used in inference using the trained CNN model”, where “inference using the trained CNN model” is discussed):
operating the intelligent model to perform the at least one intelligent functionality based on a function activation signal received from the second communication node; and deactivating operations of the intelligent model based on a function deactivation signal received from the second communication node (Jeon, Para. [0054], “In each round of PARTEL, a server partitions a global model under training into blocks of parameters, called parametric blocks, and allocate each of them to a single device for updating . . . Then devices upload the updates to a server where they are assembled and then applied to updating the model. The two steps are iterated till the model converges”; Jeon, Para. [0060], “Each worker (e.g. UE) performs one task assigned by the server”; Jeon, Para. [0083], “The training process in FIG. 2 iterates when all parametric blocks are updated in the round, i.e., the tasks of all workers are synchronized in each round”; and Jeon, Para. [0175], “FIG. 6 . . . S602 to S609 may be iterated periodically or upon request of the server for updating the CNN model”, where “training . . . steps are iterated till the model converges”, such that, the first communication nose operates the intelligent model to perform the at least one intelligent functionality, “updating”, based on signals received from the second communication node, “a server partitions a global model under training into blocks of parameters, called parametric blocks, and allocate each of them to a single device for updating”, which, in view of Meyer, include function activation signals and function deactivation signals, see Meyer, Col. 6, ln. 24-32, “Based on the mobile computing device capacity, the mobile computing device 100 may receive from the cloud computing system 210 a local machine learning system . . .The local machine learning system may include one or more than one local machine learning model”, where the data containing the “one or more than one local machine learning model” is within the broadest reasonable interpretation of a function activation signal because it signals the “model[s]” to be actively used in a particular training iteration, “The training process in FIG. 2 iterates”, and is also within the broadest reasonable interpretation of a function deactivation signal when a model transmitted in a previous iteration, “The training process in FIG. 2 iterates when all parametric blocks are updated in the round” is absent from the current iteration’s “receive[d] . . . one or more than one local machine learning model”, which allows the first communication node to activate or deactivate operations, “In each round of PARTEL, a server partitions a global model under training into blocks of parameters, called parametric blocks, and allocate each of them to a single device for updating”).
The reasons for obviousness were discussed in regard to the rejection of claim 2 above and remain applicable here.
Regarding Claim 12, the additional elements of the dependent claim are substantially the same as limitations of Claim 2, therefore it is rejected under the same rationale.
Regarding Claim 20, the additional elements of the dependent claim are substantially the same as limitations of Claim 10, therefore it is rejected under the same rationale.
Claims 6-7 and 16-17 are rejected under 35 U.S.C. 103 as being unpatentable over Jeon in view of Hao et al. (hereinafter Hao) (Pat. App. Pub. No. US 2025/0088232 A1) and Cashmore et al. (hereinafter Cashmore) (Pat. App. Pub. No. US 2023/0263420 A1).
Regarding Claim 6, Jeon teaches the method according to claim 1, wherein the obtaining of the training dataset comprises: collecting a plurality of data units . . . (Para. [0081], “Computation Phase: Each worker computes the update (e.g., gradients or the proximal gradients) of the its allocated parametric block”; Para. [0060], “Each worker (e.g. UE) performs one task assigned by the server”; Para. [0083] - [0084], “all parametric blocks are updated in the round, i.e., the tasks of all workers are synchronized in each round . . . W-stage: The parameter matrix W is divided into K blocks, with each being updated by one worker”; and Para. [0072], “W-stage: Fixing the values of Z, solve the problem of in which the problem of each neuron is independent and can be written as
PNG
media_image1.png
88
382
media_image1.png
Greyscale
”, where the first communication node, “worker (e.g. UE)”, must collect a dataset “m”, see Para. [0045], Table 1, “data sample m”, in order to use it to “computes the update (e.g., gradients or the proximal gradients) of the its allocated parametric block”, as shown in Eq. 2)
information received from the second communication node (Fig. 6, “S606 Provide data sample indexes M1 for UE1”; Para. [0083] - [0084], “all parametric blocks are updated in the round, i.e., the tasks of all workers are synchronized in each round . . . W-stage: The parameter matrix W is divided into K blocks, with each being updated by one worker”; and Para. [0072], “W-stage: Fixing the values of Z, solve the problem of in which the problem of each neuron is independent and can be written as
PNG
media_image1.png
88
382
media_image1.png
Greyscale
”, where the first communication node, such as “UE1”, receives information, “Provide data sample indexes M1 for UE1” of dataset “m”, see Para. [0045], Table 1, “data sample m”, from the second communication node “Server”);
in response to receipt of a data request signal from the second communication node, transmitting . . . [data] to the second communication node (Para. [0175], “FIG. 6 illustrates an example of the update procedure of a neural network (NN) model according to some implementations of the present disclosure. For convenience description, in FIG. 6, it is assumed that two workers (UE 1 and UE 2) participate in the partitioned edge learning. The CNN model may be updated by partitioning the model parameters W and the auxiliary matrices Z to the workers for edge-learning (S601 to S 609). S602 to S609 may be iterated periodically or upon request of the server for updating the CNN model” and Fig. 6, where in response to receipt of a data request signal from the second communication node, “Provide data sample indexes M1 for EU 1”, transmitting data to the second communication node from the first communication node, “Report updated auxiliary matrices {zm}”); and
receiving a dataset including one or more data units corresponding to the at least one intelligent functionality among the plurality of data units from the second communication node and acquiring the dataset as the training dataset (Para. [0081], “Computation Phase: Each worker computes the update (e.g., gradients or the proximal gradients) of the its allocated parametric block”; Para. [0060], “Each worker (e.g. UE) performs one task assigned by the server”; Para. [0083] - [0084], “all parametric blocks are updated in the round, i.e., the tasks of all workers are synchronized in each round . . . W-stage: The parameter matrix W is divided into K blocks, with each being updated by one worker”; and Para. [0072], “W-stage: Fixing the values of Z, solve the problem of in which the problem of each neuron is independent and can be written as
PNG
media_image1.png
88
382
media_image1.png
Greyscale
”, where the first communication node, “worker (e.g. UE)”, must receive a dataset “m”, see Para. [0045], Table 1, “data sample m”, , which includes one or more data units, “m=1” to “i”, which are acquired as the training dataset in order to use it to “computes the update (e.g., gradients or the proximal gradients) of the its allocated parametric block”, as shown in Eq. 2, and where the dataset “m” corresponds with the dataset information from the second communication node, see Fig. 6, “S606 Provide data sample indexes M1 for UE1”, where “data sample indexes M1” corresponds with m as its “indexes”, and corresponds, as training data, to the at least one intelligent functionality, “Each worker (e.g. UE) performs one task ”, among a plurality of intelligent functionalities, “The training process . . . [of] tasks”, among the plurality of data units, “a server partitions a global model under training into blocks of parameters, called parametric blocks”, from the second communication node, “Server”, see Para. [0054], “In each round of PARTEL, a server partitions a global model under training into blocks of parameters, called parametric blocks, and allocate each of them to a single device for updating. In other words, PARTEL supports the distributed training of a large-scale AI model by dynamically partitioning the model and allocating the resultant parametric blocks to different devices for updating”; Para. [0060], “Each worker (e.g. UE) performs one task assigned by the server”; Para. [0083], “The training process in FIG. 2 iterates when all parametric blocks are updated in the round, i.e., the tasks of all workers are synchronized in each round”; and Fig. 6, “S602 Allocate partial model parameter w1 to UE 1”).
Jeon does not explicitly disclose . . . based on data collection configuration . . . the plurality of data units . . . .
However, Hao teaches . . . [a first communication node that collects training data] based on data collection configuration [information received from a second communication node] . . . (Para. [0094], “the UE 104 may receive, from the first BS 102a, an indication . . . such as a . . . index or identifier”; Para. [0122], “the UE 104 may receive, from the BS 102 . . . the QCL reference signals may be indicated via reference signal resource identifiers (IDs)”; Para. [0129], “the UE 104 may receive, from the BS 102 . . . an indication that the CSI-RS is in-distribution with the training data associated with the first antenna architecture”; and Para. [0133], “the UE 104 may receive, from the BS 102, one or more reference signals associated with the transmission spatial information”, where a first communication node, “UE”, receives data collection configuration information, “an indication . . . such as a . . . index or identifier”, “QCL reference signals may be indicated via reference signal resource identifiers (IDs)”, “an indication that the CSI-RS is in-distribution” or, “reference signals associated with the transmission spatial information”, from a second communication node, “BS”, and where the first communication node collects training data based on data collection configuration, see Para. [0169], “the indication of the transmission spatial information may include an indication of an in-distribution configuration and/or training data . . . The training data may be indicative of the transmission spatial information. The UE may receive the training data for the machine learning module. In some cases, the in-distribution configuration may indicate the training data to use for the machine learning module”; Para. [0122], “The UE 104 may also receive, from the BS 102, the QCL reference signal (e.g., the TRS), where the QCL reference signal may be indicative of training data for a machine learning module”; and Para. [160], “The reference signal may include . . . training data”)
[and the first communication node receiving a dataset from the second communication node as the training dataset] (Para. [077], “the network entity (e.g., the BS 102) may provide the UE 104 with training data to train the machine learning module 510”; see also Para. [0133], “the UE 104 may receive, from the BS 102, one or more reference signals associated with the transmission spatial information” and Para. [160], “The reference signal may include . . . training data”, where the first communication node, “UE” receives “reference signals”, which comprise a dataset as the training dataset, “The reference signal may include . . . training data”, from the second communication node, “BS”).
Before the effective filing date of the invention, it would have been obvious to one of ordinary skill in the art to combine the collecting a plurality of data units by the first communication node, receiving information received from the second communication node, and in response to receipt of a data request signal from the second communication node, transmitting data to the second communication node of Jeon with the first communication node that collects training data based on data collection configuration information received from the second communication node and the first communication node receiving a dataset from the second communication node as the training dataset of Hao in order to allow centralized control of training data (Hao, Para. [077], “the network entity (e.g., the BS 102) may provide the UE 104 with training data to train the machine learning module 510”) through data collection configuration information (Hao, Para. [0169], “the indication of the transmission spatial information may include an indication of an in-distribution configuration and/or training data . . . The training data may be indicative of the transmission spatial information. The UE may receive the training data for the machine learning module. In some cases, the in-distribution configuration may indicate the training data to use for the machine learning module”; Hao, Para. [0122], “The UE 104 may also receive, from the BS 102, the QCL reference signal (e.g., the TRS), where the QCL reference signal may be indicative of training data for a machine learning module”; and Hao, Para. [160], “The reference signal may include . . . training data”), which will allow for adaptive adjustments (Hao, Para. [0173], “the network entity may obtain (e.g., receive), from the UE, CSI based on the reference signal and the transmission spatial information. For example, the CSI may indicate a CQI, a PMI, a RSRP, and/or SINR associated with the reference signal(s). In some cases, the network entity may adjust a configuration at the UE based on the CSI. For example, the network entity may adjust the modulation and coding scheme configured at the UE based on the CSI”) and specialized model training (Hao, Para. [0128], “At activity 1404, the UE 104 may receive, from the BS 102, synthesized training data associated with the first antenna architecture. At activity 1406, the BS 102 may determine a spatial correlation associated with a second antenna architecture and/or TxRU mapping, for example, as described herein with respect to FIGS. 8-10. At activity 1408, the UE 104 may receive, from the BS 102, synthesized training data associated with the second antenna architecture. At activity 1410, the UE 104 may train a first machine learning module based on the training data associated with the first antenna architecture of the BS 102. At activity 1412, the UE 104 may train a second machine learning module based on the training data associated with the second antenna architecture of the BS 102”) designed for a device’s specific constraints (Hao, Para. [0040], “The channel estimation processing complexity at a UE may be proportional to the frequency bandwidth associated with the ports. As the number of ports increase, the UE may be tasked to process an increasing frequency bandwidth associated with the ports for channel estimation and/or CSI reporting. In certain cases, the UE may not have the capability to process the corresponding frequency bandwidth associated with a large number of ports and corresponding antenna elements associated with a network entity”) and implemented to reduce computational overhead (Hao, Para. [0042], “The transmission spatial information described herein may facilitate CSI-RS overhead reduction with usage of a machine learning module to perform channel estimation and/or determine CSI and reduce the CSI-RS overhead. The transmission spatial information may be used to train, select, and/or configure the machine learning module”).
Additionally, Cashmore teaches [a first communication node transmitting] the plurality of data units [to a second communication node] . . . (Fig. 1 and Para. [0130], “if the processor 201 determines to update the machine-learned model 207 then the wearable article 20 transmits the data and the inference to the base station 30 (FIG. 1) . . . Transmitting the training data to the base station 30”, where a first communication node, “the wearable”, transmits a plurality of data units, “transmits the data and the inference . . . the training data”, to a second communication node, “the base station 30”).
Before the effective filing date of the invention, it would have been obvious to one of ordinary skill in the art to combine the collecting the plurality of data units based on data collection configuration information received from the second communication node and, in response to receipt of a data request signal from the second communication node, transmitting data to the second communication node of Jeon in view of Hao with the first communication node transmitting the plurality of data units to the second communication node of Cashmore in order to dynamically adjust the training location, based on computational constraints, in order to take advantage of the computational resources of the second communication device or the efficiency of the first communication device (Cashmore, Para. [0129] – [0130], “if the processor 201 determines to update the machine-learned model 207, then the processor 201 may update the machine-learned model 207. In these implementations, it will be appreciated that the base station 30 (FIG. 1) and the server 40 (FIG. 1) may not be required . . . if the processor 201 determines to update the machine-learned model 207 then the wearable article 20 transmits the data and the inference to the base station 30 (FIG. 1) so as to be used as training data to update a machine-learned model stored on the base station 30 . . . This approach is beneficial as the computationally intensive task of model updating is not performed on the wearable article 20 but rather a base station 30. Due to size, battery and portability constraints, wearable articles 20 typically have limited computational capability to perform machine-learning operations. Transmitting the training data to the base station 30 therefore allows for local and secure model updating to be performed for the wearable device 20 even if the wearable device 20 does not have the computational or power requirements to perform on-device model re-training. Performing inference operations on the wearable device 20 may be beneficial in allowing for inferences to be generated without requiring the wearable article 20 to be in communication with the base station 30”), which may change based on the computational intensity of the assigned task (Jeon, Para. [0083] - [0084], “all parametric blocks are updated in the round, i.e., the tasks of all workers are synchronized in each round . . . W-stage: The parameter matrix W is divided into K blocks, with each being updated by one worker”; Meyer, Para. [0015], “The base station is able to perform the relatively computationally intense task of performing the model update operation which may, in some circumstances, be undesirable to perform on the wearable article due to considerations such as size, power, and computational resources”) or over the course of the training iterations based on evolving device constraints (Jeon, Para. [0175], “FIG. 6 . . . S602 to S609 may be iterated periodically or upon request of the server for updating the CNN model”; Hao, Para. [0140], “over the time, the channel estimation and CSI feedback algorithm changes (e.g., adapts or updates) based on new learning. The channel estimation and CSI feedback procedures may be used for adapting various characteristics of the communication link between a UE and a network entity, such as transmit power control, modulation and coding scheme(s), code rate, subcarrier spacing, etc”).
Regarding Claim 7, Jeon in view of Hao and Cashmore teaches the method according to claim 6, wherein the obtaining of the training dataset comprises: transmitting configuration change information of the first communication node to the second communication node (Jeon, Para. [0081], “Computation Phase: Each worker computes the update (e.g., gradients or the proximal gradients) of the its allocated parametric block”; Jeon, Para. [0060], “Each worker (e.g. UE) performs one task assigned by the server”; Jeon, Para. [0083] - [0084], “all parametric blocks are updated in the round, i.e., the tasks of all workers are synchronized in each round . . . W-stage: The parameter matrix W is divided into K blocks, with each being updated by one worker”; and Jeon, Para. [0072], “W-stage: Fixing the values of Z, solve the problem of in which the problem of each neuron is independent and can be written as
PNG
media_image1.png
88
382
media_image1.png
Greyscale
”, where the first communication node, “worker (e.g. UE)”, must collect a dataset “m”, see Jeon, Para. [0045], Table 1, “data sample m”, in order to use it to “computes the update (e.g., gradients or the proximal gradients) of the its allocated parametric block”, as shown in Eq. 2, which, in view of Hao, the obtaining of the training dataset, “the UE 104 may receive, from the BS 102, an indication of transmission spatial information” “the transmission spatial information may include an indication of . . . training data” and “the UE 104 may receive, from the BS 102, one or more reference signals” “The reference signal may include . . . training data”, occurs iteratively, “over the time, the channel estimation and CSI feedback algorithm changes (e.g., adapts or updates) based on new learning”, such as to comprise transmitting configuration change information of the first communication node to the second communication node, “the UE 104 may communicate with the BS 102 based on the CSI. For example, the BS 102 may adjust the modulation and coding scheme used for communications with the UE 104 based on the CSI”, see Hao, Fig. 16; Hao, Para. [0132] – [0140], “At activity 1602, the UE 104 may receive, from the BS 102, an indication of transmission spatial information . . . At activity 1604, the UE 104 may receive, from the BS 102, one or more reference signals associated with the transmission spatial information . . . At activity 1614, the UE 104 may communicate with the BS 102 based on the CSI. For example, the BS 102 may adjust the modulation and coding scheme used for communications with the UE 104 based on the CSI. . . over the time, the channel estimation and CSI feedback algorithm changes (e.g., adapts or updates) based on new learning. The channel estimation and CSI feedback procedures may be used for adapting various characteristics of the communication link between a UE and a network entity, such as transmit power control, modulation and coding scheme(s), code rate, subcarrier spacing, etc”; Hao, Para. [0169], “the indication of the transmission spatial information may include an indication of . . . training data . . . The training data may be indicative of the transmission spatial information. The UE may receive the training data for the machine learning module”; and Hao, Para. [160], “The reference signal may include . . . training data”; see also Hao, Para. [0173], “the network entity may obtain (e.g., receive), from the UE, CSI based on the reference signal and the transmission spatial information. For example, the CSI may indicate a CQI, a PMI, a RSRP, and/or SINR associated with the reference signal(s). In some cases, the network entity may adjust a configuration at the UE based on the CSI. For example, the network entity may adjust the modulation and coding scheme configured at the UE based on the CSI”),
and the transmitting of the plurality of data units to the second communication node comprises: transmitting the plurality of data units (Jeon, Para. [0175], “FIG. 6 illustrates an example of the update procedure of a neural network (NN) model according to some implementations of the present disclosure. For convenience description, in FIG. 6, it is assumed that two workers (UE 1 and UE 2) participate in the partitioned edge learning. The CNN model may be updated by partitioning the model parameters W and the auxiliary matrices Z to the workers for edge-learning (S601 to S 609). S602 to S609 may be iterated periodically or upon request of the server for updating the CNN model” and Jeon, Fig. 6, where in response to receipt of a data request signal from the second communication node, “Provide data sample indexes M1 for EU 1”, transmitting data to the second communication node from the first communication node, “Report updated auxiliary matrices {zm}”, which, in view of Cashmore, the data includes the plurality of data units, see Cashmore, Fig. 1 and Cashmore, Para. [0130], “if the processor 201 determines to update the machine-learned model 207 then the wearable article 20 transmits the data and the inference to the base station 30 (FIG. 1) . . . Transmitting the training data to the base station 30”)
collected up to a time of changing configuration of the first communication node to the second communication node upon receipt of a data request signal transmitted from the second communication node based on the configuration change information (Cashmore, Fig. 1 and Cashmore, Para. [0126] - [0130], “The wearable article 20 performs on-device machine-learning using . . . data sensed by the sensor 211 . . . if the processor 201 determines to update the machine-learned model 207 then the wearable article 20 transmits the data and the inference to the base station 30 (FIG. 1) . . . Transmitting the training data to the base station 30”, where the plurality of data units collected up to a time, “data sensed by the sensor 211”, is transmitted upon receipt of a data request signal transmitted from the second communication node, “if the processor 201 determines to update the machine-learned model 207 then the wearable article 20 transmits the data and the inference to the base station 30 (FIG. 1)”, which, in view of Hao, is up until a time of changing configuration of the first communication node to the second communication node and based on the configuration change information, see Hao, Fig. 16 and Hao, Para. [0138] – [0140], “At activity 1614, the UE 104 may communicate with the BS 102 based on the CSI. For example, the BS 102 may adjust the modulation and coding scheme used for communications with the UE 104 based on the CSI. . . over the time, the channel estimation and CSI feedback algorithm changes (e.g., adapts or updates) based on new learning. The channel estimation and CSI feedback procedures may be used for adapting various characteristics of the communication link between a UE and a network entity, such as transmit power control, modulation and coding scheme(s), code rate, subcarrier spacing, etc”).
The reasons for obviousness were discussed in regard to the rejection of claim 6 above and remain applicable here.
Regarding Claim 16, the additional elements of the dependent claim are substantially the same as limitations of Claim 6, therefore it is rejected under the same rationale.
Regarding Claim 17, the additional elements of the dependent claim are substantially the same as limitations of Claim 7, therefore it is rejected under the same rationale.
Claims 8 and 18 are rejected under 35 U.S.C. 103 as being unpatentable over Jeon in view of Hao.
Regarding Claim 8, Jeon in view of Hao teaches the method according to claim 1, wherein the obtaining of the training dataset comprises: receiving the associated ID and reference signals (RSs) from the second communication node; and collecting a plurality of data units corresponding to the at least one intelligent functionality based on the RSs and the associated ID and acquiring the plurality of data units as the training dataset (Jeon, Fig. 6, “S606 Provide data sample indexes M1 for UE1”; Jeon, Para. [0083] - [0084], “all parametric blocks are updated in the round, i.e., the tasks of all workers are synchronized in each round . . . W-stage: The parameter matrix W is divided into K blocks, with each being updated by one worker”; and Jeon, Para. [0072], “W-stage: Fixing the values of Z, solve the problem of in which the problem of each neuron is independent and can be written as
PNG
media_image1.png
88
382
media_image1.png
Greyscale
”, where the first communication node, such as “UE1”, obtains a training dataset, “Provide data sample indexes M1 for UE1” of dataset “m”, see Jeon, Para. [0045], Table 1, “data sample m”, which includes a plurality of data units, such as “m=1” through “I”, that are acquired as the training dataset “data sample m”, and are corresponding to the at least one intelligent functionality, “W-stage: The parameter matrix W is divided into K blocks, with each being updated by one worker” which, as shown in Eq. (2), uses “m”, based on an associated identifier (ID), “indexes M1”, received, “Provide data sample indexes M1”, from the second communication node “Server”, as indicated by the arrow from “Server” to “UE 1” in step “S606”; alternatively, see Jeon, Para. [0054], “In each round of PARTEL, a server partitions a global model under training into blocks of parameters, called parametric blocks, and allocate each of them to a single device for updating. In other words, PARTEL supports the distributed training of a large-scale AI model by dynamically partitioning the model and allocating the resultant parametric blocks to different devices for updating”; Jeon, Para. [0083], “The training process in FIG. 2 iterates when all parametric blocks are updated in the round, i.e., the tasks of all workers are synchronized in each round”; and Jeon, Fig. 6, “S602 Allocate partial model parameter w1 to UE 1”, where the first communication node, such as “UE 1”, obtains a training dataset, “Allocate partial model parameter w1 to UE 1”, corresponding to the at least one intelligent functionality, “all parametric blocks are updated in the round, i.e., the tasks of all workers are synchronized in each round”, based on an associated identifier (ID), such as the identifier of “w1”, received from the second communication node, “a server partitions a global model under training into blocks of parameters, called parametric blocks, and allocate each of them to a single device”, and where, in view of Hao the receiving is of IDs, see Hao, Para. [0094], “the UE 104 may receive, from the first BS 102a, an indication . . . such as a . . . index or identifier”; Hao, Para. [0122], “the UE 104 may receive, from the BS 102 . . . the QCL reference signals may be indicated via reference signal resource identifiers (IDs)”; and Hao, Para. [0129], “the UE 104 may receive, from the BS 102 . . . an indication that the CSI-RS is in-distribution with the training data associated with the first antenna architecture”, and reference signals, Hao, Para. [0133], “the UE 104 may receive, from the BS 102, one or more reference signals associated with the transmission spatial information”, which are a basis for the collection of data units as the training data, see Hao, Para. [0169], “the indication of the transmission spatial information may include an indication of an in-distribution configuration and/or training data . . . The training data may be indicative of the transmission spatial information. The UE may receive the training data for the machine learning module. In some cases, the in-distribution configuration may indicate the training data to use for the machine learning module”; Hao, Para. [0122], “The UE 104 may also receive, from the BS 102, the QCL reference signal (e.g., the TRS), where the QCL reference signal may be indicative of training data for a machine learning module”; and Hao, Para. [160], “The reference signal may include . . . training data”).
The reasons for obviousness were discussed in regard to the rejection of claim 6 above and remain applicable here.
Regarding Claim 18, the additional elements of the dependent claim are substantially the same as limitations of Claim 8, therefore it is rejected under the same rationale.
Conclusion
The prior art made of record and not relied upon is considered pertinent to applicant's disclosure.
Yi et al. (“RHFedMTL: Resource-Aware Hierarchical Federated Multi-Task Learning”) discloses additional methods for resource constrained machine learning implemented using communication nodes (see Pg. 1, Abstract).
Any inquiry concerning this communication or earlier communications from the examiner should be directed to MATTHEW BRYCE GOLAN whose telephone number is (571)272-5159. The examiner can normally be reached Monday through Friday, 8:00 AM to 5:00 PM ET.
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, Alexey Shmatov can be reached at (571) 270-3428. 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.
/MATTHEW BRYCE GOLAN/Examiner, Art Unit 2123
/ALEXEY SHMATOV/Supervisory Patent Examiner, Art Unit 2123
