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 .
Amendments of 11/20/2025 are entered.
Claims 1-19, 21-31 are pending. Claims 16 and 28 are withdrawn.
Response to Arguments
The arguments in Remarks of 10/20/2025 have been considered. They are directed to the new clarification in the claims, specifically the claims now clarifying that the UE to determine the training time itself. Thus the arguments are moot in view of a new ground of rejection.
Claim Rejections - 35 USC § 103
In the event the determination of the status of the application as subject to AIA 35 U.S.C. 102 and 103 (or as subject to pre-AIA 35 U.S.C. 102 and 103) is incorrect, any correction of the statutory basis (i.e., changing from AIA to pre-AIA ) for the rejection will not be considered a new ground of rejection if the prior art relied upon, and the rationale supporting the rejection, would be the same under either status.
The following is a quotation of 35 U.S.C. 103 which forms the basis for all obviousness rejections set forth in this Office action:
A patent for a claimed invention may not be obtained, notwithstanding that the claimed invention is not identically disclosed as set forth in section 102, if the differences between the claimed invention and the prior art are such that the claimed invention as a whole would have been obvious before the effective filing date of the claimed invention to a person having ordinary skill in the art to which the claimed invention pertains. Patentability shall not be negated by the manner in which the invention was made.
The factual inquiries for establishing a background for determining obviousness under 35 U.S.C. 103 are summarized as follows:
1. Determining the scope and contents of the prior art.
2. Ascertaining the differences between the prior art and the claims at issue.
3. Resolving the level of ordinary skill in the pertinent art.
4. Considering objective evidence present in the application indicating obviousness or nonobviousness.
Claim(s) 1-6, 14, 19, 21-24, 31 is/are rejected under 35 U.S.C. 103 as being unpatentable over Bai et al. (US 2021/0091838) in view of Ykovlev et al. (US 2020/0327448).
As to claim 1:
BAI discloses:
An apparatus for wireless communication at a user equipment (UE),comprising:
memory; and at least one processor coupled to the memory (¶0090, UE with processor/memory) and configured to:
determining, at the UE, data related to a machine learning (ML) or a neural network model at the UE for channel state information (CSI) derivation; and transmit a message indicating the processing time to a first network entity. (See ¶ 0065, 0068, 0069, UE generates and sends ML information associated with a machine learning model for estimating CSI to a base station for evaluation of the model. ¶0079-0081, the UE determine one or more parameters associated with the predictive model and communicate those to the base station)
Bai, however, does explicitly discloses the determined ML model information generated at the UE being, namely, a processing time of training such machine learning model.
Yakovlev, in a related field of machine federated learning, discloses in at least Abstract, ¶0073-0077, discloses a computer device 600 that trains and use ML models, such as ML model 610. The computer device uses a regressor 650 to determine a training time for the ML model 610 and output the training time for evaluation (¶0018-0019, 0079).
It would have been obvious to one of ordinary skill in the art before the effective filing time of the invention that UE of Bai to incorporate the training time estimation feature as in Yakovlev and include them into the ML learning related information to be sent to the base station. Per ¶0019 of Yakovlev, having this information determined allows for effective task management and cost-aware scheduling. Given both the base station and UE to employ the service the ML models, sharing the ML model between UE and the base station allows the base station or network operator to obtain this information readily (without having to figuring it out themselves)
This implementation advantageously further improves system transparency and efficiency via information sharing.
As to claim 19:
Bai discloses a method of wireless communication at a user equipment (UE), comprising: determining, at the UE, data related to a machine learning (ML) or a neural network model at the UE for channel state information (CSI) derivation; and transmit a message indicating the processing time to a first network entity. (See ¶ 0065, 0068, 0069, UE generates and sends ML information associated with a machine learning model for estimating CSI to a base station for evaluation of the model. ¶0079-0081, the UE determine one or more parameters associated with the predictive model and communicate those to the base station)
Bai, however, does explicitly discloses the determined ML model information generated at the UE being, namely, a processing time of training such machine learning model.
Yakovlev, in a related field of machine federated learning, discloses in at least Abstract, ¶0073-0077, discloses a computer device 600 that trains and use ML models, such as ML model 610. The computer device uses a regressor 650 to determine a training time for the ML model 610 and output the training time for evaluation (¶0018-0019, 0079).
It would have been obvious to one of ordinary skill in the art before the effective filing time of the invention that UE of Bai to incorporate the training time estimation feature as in Yakovlev and include them into the ML learning related information to be sent to the base station. Per ¶0019 of Yakovlev, having this information determined allows for effective task management and cost-aware scheduling. Given both the base station and UE to employ the service the ML models, sharing the ML model between UE and the base station allows the base station or network operator to obtain this information readily (without having to figuring it out themselves)
This implementation advantageously further improves system transparency and efficiency via information sharing.
As to claim 21 and 31:
Bai discloses:
A method and an apparatus for wireless communication at a, comprising: memory; and at least one processor coupled to the memory (¶0051, base station with processor and memory) and configured to performs said method:
receive a message from a user equipment (UE) that includes an indication related to a ML model or neural network model, wherein the indication indicates one or more parameters associated with said ML model or neural network model for channel state information (CSI) derivation (See ¶ 0065, 0068, 0069, UE generates and sends ML information associated with a machine learning model for estimating CSI to a base station for evaluation of the model. ¶0079-0081, the UE determine one or more parameters associated with the predictive model and communicate those to the base station)
transmit a configuration to the UE based on the processing time; and receive the CSI from the UE based on the configuration. (See ¶0068, 0069, 0051, 0070, 0082, at the base station, receives a ML information, wherein UE generates and sends ML information associated with a machine learning model for estimating CSI to base station. The UE then receives configuration information from the base station that included negotiated ML model and how to conduct CSI specifically. The UE reports the CSI to the base station based on the configuration.)
Bai, however, does explicitly discloses the determined ML model information generated at the UE being, namely, a processing time of training such machine learning model.
Yakovlev, in a related field of machine federated learning, discloses in at least Abstract, ¶0073-0077, discloses a computer device 600 that trains and use ML models, such as ML model 610. The computer device uses a regressor 650 to determine a training time for the ML model 610 and output the training time for evaluation (¶0018-0019, 0079).
It would have been obvious to one of ordinary skill in the art before the effective filing time of the invention that UE of Bai to incorporate the training time estimation feature as in Yakovlev and include them into the ML learning related information to be sent to the base station. Per ¶0019 of Yakovlev, having this information determined allows for effective task management and cost-aware scheduling. Given both the base station and UE to employ the service the ML models, sharing the ML model between UE and the base station allows the base station or network operator to obtain this information readily (without having to figuring it out themselves)
This implementation advantageously further improves system transparency and efficiency via information sharing.
As to claim 2:
Bai in view of Yakovlev discloses all limitations of claim 1, wherein the first network entity is a base station, a transmission reception point (TRP), or another UE and the apparatus further includes: at least one antenna; and a transceiver coupled to the at least one antenna and the at least one processor. (Bai, ¶0069, 0051, 0053, base station/TRP and the UE has antenna with transceiver controlled by the processor)
As to claim 3:
Bai in view of Yakovlev discloses all limitations of claim 1, wherein the processing time corresponds to an amount of time between a first time associated with a reference signal used for training in the ML or NN model a second time at which the UE has successfully trained the ML or NN model. (Bai, ¶0071-0072, the ML measuring the received the reference signal, which indicates the start of the processing operation. In context of Yakovlev, the client reports the training time that ends when the training is finished. Therefore, naturally, the processing time is when the model receives the input (CSI) to when it finishes the training)
As to claim 4:
Bai in view of Yakovlev discloses all limitations of claim 3, wherein the ML or NN model is successfully trained when the UE is able to report back the CSI trained by ML or NN model, or is able to use trained weights of ML or NN model to achieve an accuracy or a quality of service (QoS). (See Bai, ¶0076, 0077, 0069, using machine learning model, the UE successful report back the CSI, thus the machine learning model has been successfully trained)
As to claim 5:
Bai in view of Yakovlev discloses all limitations of claim 1, wherein the processing time corresponds to a time between reception of a command to train the ML or NN model and completion of the training of the ML or NN model. (See Yakovlev, 4.2, “the aggregator asks every client to train on the local data and waits for their acknowledgement for 𝑇𝑚𝑎𝑥 seconds. All clients that respond within 𝑇𝑚𝑎𝑥 have their response latency value 𝑅𝑇𝑖 incremented with the actual training time, while the ones that have timed out are incremented by 𝑇𝑚𝑎x”).
As to claim 6:
Bai in view of Yakovlev discloses all limitations of claim 1, wherein the memory and the at least one processor are further configured to: receive a configuration from a second network entity at least based on the processing time; and transmit channel state information (CSI) to the second network entity based on the configuration. (See BAI, ¶0068, 0069, 0051, 0070, 0082, at the base station, receives a ML information, wherein UE generates and sends ML information associated with a machine learning model for estimating CSI to base station. The UE then receives configuration information from the base station that included negotiated ML model and how to conduct CSI specifically. The UE reports the CSI to the base station based on the configuration.)
As to claim 14:
Bai in view of Yakovlev discloses all limitations of claim 1, wherein the processing time for training the ML or NN model is based on a type of wireless signal procedure performed by the ML or NN model , the type of the wireless signal procedure including at least one of: channel state information (CSI) determination, demodulation, positioning determination, or waveform determination. (See Bai, ¶0063, “issues (e.g., delay) experienced due to CSI reporting by a UE may be exacerbated when the UE is in a high-mobility state (e.g., when the UE is traveling at relatively high rates of speeds), as the high-mobility state may introduce the Doppler effect to communication between the UE and the base station thereby making CSI reporting by the UE time-variant. In other words, the faster the UE is moving, the faster channel conditions change over time”. The CSI processing is uniquely affected by mobility state of the UE due to the nature of computation. Thus training of a learning model for CSI computation is similarly affected by UE’s mobility state over other type of signal processing)
As to claim 22:
Bai in view of Yakovlev discloses all limitations of claim 21, wherein the wireless communication is at a base station, a transmission reception point (TRP), or another UE, the apparatus further comprising: at least one antenna; and a transceiver coupled to the at least one antenna and the at least one processor. (Bai, ¶0069, 0051, 0053, base station/TRP having antenna coupled to processor with transceiver, and the UE has antenna with transceiver controlled by the processor)
As to claim 23:
Bai in view of Yakovlev discloses all limitations of claim 21, wherein the processing time corresponds to an amount of time between a first time associated with a reference signal used for training in the ML or NN model and a second time at which the UE has successfully trained the ML or NN model , (Bai, ¶0071-0072, the ML measuring the received the reference signal, which indicates the start of the processing operation. In context of Yakovlev, the client reports the training time that ends when the training is finished. Therefore, naturally, the processing time is when the model receives the input (CSI) to when it finishes the training) and wherein the ML or NN model is successfully trained when the UE is able to report back the CSI trained by the ML or NN model , or is able to use trained weights of the ML or NN model to achieve an accuracy or a quality of service (QoS). (See Bai, ¶0076, 0077, 0069, using machine learning model, the UE successful report back the CSI, thus the machine learning model has been successfully trained)
As to claim 24:
Bai in view of Yakovlev discloses all limitations of claim 21, wherein the processing time corresponds to a time between reception of a command to train the ML or NN model and completion of the training of the ML or NN model . (See Yakovlev, 0077, 0018, training time is associated a period of time where a new configuration or new dataset are being used to train the ML model, thus occurring after the computer receives a trigger signal to start the training and when it complete training).
Claim(s) 7, 9-13, 17, 23-27, 29 is/are rejected under 35 U.S.C. 103 as being unpatentable over Bai et al. (US 2021/0091838) in view of in view of Ykovlev et al. (US 2020/0327448 and in further view of Jung et al. (Restructuring Batch normalization to Accelerate CNN training) - 2019.
As to claim 7:
Bai in view of Yakovlev discloses all limitations of claim 1, while they do not explicitly mention the processing time for training the ML or NN model is based on at least one of: a number of layers in the ML or NN model , a number of weights in the ML or NN model , or a type of one or more layers of the ML or NN model . However, claim 7 is directed to natural causation fact, i.e. the more load to be processed (layers, weights), the long the process takes.
Indeed, Jung in at least page 2, right column, Fig. 1, Section 2.1 show that types and number of layers significantly impact processing time (i.e. training time measured) for a given learning model.
It would have been obvious to one of ordinary skill in the art before the effective filing time of the invention that processing time for training the ML or NN model is based on at least one of: a number of layers in the ML or NN model , a number of weights in the ML or NN model , or a type of one or more layers of the ML or NN model . This limitation is a narration of a natural cause-effect fact than a point of novelty. Simply put, more layers, more weights for the learning model will demand more processing load, thus time for a training system, which is made clear in page 2 of Jung.
As to claim 9:
Bai in view of Yakovlev discloses all limitations of claim 1, while they do not explicitly mention the processing time for training the ML or NN model is based on an amount of models or layers to be trained. However, claim 9 is directed to natural fact, i.e. the more processing loads to be processed (layers, weights), the long the process takes.
Indeed, Jung in at least page 2, right column, Fig. 1, Section 2.1 show that types and number of layers significantly impact processing time (i.e. training time measured) for a given learning model.
It would have been obvious to one of ordinary skill in the art before the effective filing time of the invention that processing time for training the ML or NN model is based on at least one of: a number of layers in the ML or NN model , a number of weights in the ML or NN model , or a type of one or more layers of the ML or NN model . This limitation is a narration of a natural cause-effect fact than a point of novelty. Simply put, more layers, more weights for the learning model will demand more processing load, thus time for a training system, which is made clear in page 2 of Jung.
As to claim 10:
Bai in view of Yakovlev and Jung discloses all limitations of claim 9, wherein the processing time is further based on whether a single ML or NN model or multiple models are to be trained simultaneously. (Jung, Fig. 1, section 2.1, each network has a number of layers, thus more networks to be trained, the more layers to be processed, thus increasing processing loads, thus processing time)
As to claim 11:
Bai in view of Yakovlev and Jung discloses all limitations of claim 10, wherein the processing time is based on training the multiple models are trained simultaneously based on concurrent training in at least one of: a same component carrier, a same band, a same bandwidth part, a same band combination, a same frequency range, a same slot, a same subframe, or a same frame, wherein the multiple models are trained simultaneously until the UE responds with a complete training message. (Jung, Fig. 1, section 2.1, each network has a number of layers, thus more networks to be trained, the more layers to be processed, thus increasing processing loads, thus processing time. For example, as shown in Fig. 1, REsNet, DenseNet require their own individual need of time dedicated for training execution. When they are trained at the same time, the resource for training will naturally be split, thus causing training time for each model to be significantly increased. As the networks are trained by the same client, they are trained using the same resource range available for the UE)
As to claim 12:
Bai in view of Yakovlev and Jung discloses all limitations of claim 10, wherein the processing time is further based on whether a single layer or multiple layers of the ML or NN model are to be trained simultaneously. (Jung, 2.1, Fig. 1, “For example, AlexNet (Krizhevsky et al., 2012) consists of only 5 CONV layers and 3 FC layers, whereas VGGNet (Simonyan & Zisserman, 2014) has 13-16 CONV layers and 3 FC layers. On these early and shallow CNN models, the portion of execution time on CONV and FC layers is dominant, accounting for up to 95% of total execution time as shown in Figure 1”, i.e. number of layers trained simultaneously will split training resources, i.e. less training power per layer, thus increasing training time )
As to claim 13:
Bai in view of Yakovlev discloses all limitations of claim 1, however is silent on the processing time for training the ML or NN model is based on a sequence order of multiple layers of the ML or NN model .
Jung, in a related field of training learning model, discloses in Fig. 7 and page 9 text that processing time can be reduced by altering sequence order of multiple layers, for example merging different specific layers’ processing together, thereby speeding up learning.
It would have been obvious to one of ordinary skill in the art before the effective filing time of the invention that in the system of Bai/Yakovlev, the processing time for training the ML or NN model is based on a sequence order of multiple layers of the ML or NN model . This implementation allows for more efficient learning as well as reduce memory usage (left column, page 9 of Jung).
As to claim 17:
Bai in view of Yakovlev discloses all limitations of claim 1, however is silent on the processing time is for at least one of: a bandwidth part, a numerology, a component carrier, a band, a band combination, a frequency range, or one or more timeline factor, and wherein the one or more timeline factor includes at least one of: a layer, a layer type, a combination of layers, an input vector length, an output vector length, an intermediate vector length, a number of layers, or a sequence of layers.
Jung, in a related field of training learning model, discloses in Fig. 7 and page 9 text that processing time is depended on which layer type (Fig. 1) and/or can be reduced by altering sequence order of multiple layers, for example merging different specific layers’ processing together, thereby speeding up learning.
It would have been obvious to one of ordinary skill in the art before the effective filing time of the invention that in the system of Bai/Yakovlev, the processing time for training the ML or NN model is based on a sequence order of multiple layers of the ML or NN model . This implementation allows for more efficient learning as well as reduce memory usage (left column, page 9 of Jung).
As to claim 25:
Bai in view of Yakovlev discloses all limitations of claim 21, while they do not explicitly mention the processing time for training the ML or NN model is based on at least one of: a number of layers in the ML or NN model , a number of weights in the ML or NN model , a type of one or more layers of the ML or NN model , quasi co-location (QCL) information from a previously trained ML or NN model or a ML or NN model state indication from the previously trained neural network.
Indeed, Jung in at least page 2, right column, Fig. 1, Section 2.1 show that types and number of layers significantly impact processing time (i.e. training time measured) for a given learning model.
It would have been obvious to one of ordinary skill in the art before the effective filing time of the invention that processing time for training the ML or NN model is based on at least one of: a number of layers in the ML or NN model , a number of weights in the ML or NN model , or a type of one or more layers of the ML or NN model . This limitation is a narration of a natural cause-effect fact than a point of novelty. Simply put, more layers, more weights for the learning model will demand more processing load, thus time for a training system, which is made clear in page 2 of Jung.
As to claim 26:
Bai in view of Yakovlev discloses all limitations of claim 21, however is silent on wherein the processing time for training the ML or NN model is based on at least one of an amount of neural networks or layers to be trained and whether a single neural model or multiple models are to be trained simultaneously, wherein the multiple models are trained simultaneously based on concurrent training in at least one of: a same component carrier, a same band, a same bandwidth part, a same band combination, a same frequency range, a same slot, a same subframe, or a same frame.
Jung discloses the processing time is further based on whether a single layer or multiple layers of the ML or NN model are to be trained simultaneously(Jung, 2.1, Fig. 1, “For example, AlexNet (Krizhevsky et al., 2012) consists of only 5 CONV layers and 3 FC layers, whereas VGGNet (Simonyan & Zisserman, 2014) has 13-16 CONV layers and 3 FC layers. On these early and shallow CNN models, the portion of execution time on CONV and FC layers is dominant, accounting for up to 95% of total execution time as shown in Figure 1”, i.e. number of layers trained simultaneously will split training resources, i.e. less training power per layer, thus increasing training time ) wherein the processing time is based on training the multiple models are trained simultaneously based on concurrent training in at least one of: a same component carrier, a same band, a same bandwidth part, a same band combination, a same frequency range, a same slot, a same subframe, or a same frame, wherein the multiple models are trained simultaneously until the UE responds with a complete training message.
(Jung, Fig. 1, section 2.1, each network has a number of layers, thus more networks to be trained, the more layers to be processed, thus increasing processing loads, thus processing time. For example, as shown in Fig. 1, REsNet, DenseNet require their own individual need of time dedicated for training execution. When they are trained at the same time, the resource for training will naturally be split, thus causing training time for each model to be significantly increased. As the networks are trained by the same client, they are trained using the same resource range available for the UE)
It would have been obvious to one of ordinary skill in the art before the effective filing time of the invention that processing time for training the ML or NN model is based on at least one of: a number of layers in the ML or NN model , a number of weights in the ML or NN model , or a type of one or more layers of the ML or NN model . This limitation is a narration of a natural cause-effect fact than a point of novelty. Simply put, more layers, more weights for the learning model will demand more processing load, thus time for a training system, which is made clear in page 2 of Jung.
As to claim 27:
Bai in view of Yakovlev discloses all limitations of claim 21, however is silent on the processing time is based on at least one of whether a single layer or multiple layers of the ML or NN model are to be trained simultaneously, a sequence order of the multiple layers of the ML or NN model , a type of wireless signal procedure performed by the ML or NN model , or an accuracy level.
Jung, in a related field of training learning model, discloses in Fig. 7 and page 9 text that processing time can be reduced by altering sequence order of multiple layers, for example merging different specific layers’ processing together, thereby speeding up learning.
It would have been obvious to one of ordinary skill in the art before the effective filing time of the invention that in the system of Bai/Yakovlev, the processing time for training the ML or NN model is based on a sequence order of multiple layers of the ML or NN model . This implementation allows for more efficient learning as well as reduce memory usage (left column, page 9 of Jung).
As to claim 29:
Bai in view of Yakovlev discloses all limitations of claim 21, however is silent on the processing time is for at least one of: a bandwidth part, a numerology, a component carrier, a band, a band combination, a frequency range, or one or more timeline factor, and wherein the one or more timeline factor includes at least one of: a layer, a layer type, a combination of layers, an input vector length, an output vector length, an intermediate vector length, a number of layers, or a sequence of layers.
Jung, in a related field of training learning model, discloses in Fig. 7 and page 9 text that processing time is depended on which layer type (Fig. 1) and/or can be reduced by altering sequence order of multiple layers, for example merging different specific layers’ processing together, thereby speeding up learning.
It would have been obvious to one of ordinary skill in the art before the effective filing time of the invention that in the system of Bai/Yakovlev, the processing time for training the ML or NN model is based on a sequence order of multiple layers of the ML or NN model . This implementation allows for more efficient learning as well as reduce memory usage (left column, page 9 of Jung).
Claim(s) 8, 15 is/are rejected under 35 U.S.C. 103 as being unpatentable over Bai et al. (US 2021/0091838) in view of Ykovlev et al. (US 2020/0327448) and in further view of Black (US 6,269,351).
As to claim 8:
Bai in view of Yakovlev and Jung discloses all limitations of claim 1, however is silent on the processing time for training the ML or NN model is based on at least one of quasi co-location (QCL) information or a neural network state indication from a previously trained model.
Black, in a related field of training learning model, discloses training a neural network wherein the NN is initially trained with initial set of parameters. Upon the NN outputs an error ratio that satisfies a threshold condition, the trained ANN is further trained at a modified learning rate (which affects learning time). (See at least Abstract, Fig. 4 and associated text).
It would have been obvious to one of ordinary skill in the art before the effective filing time of the invention that processing time for training the ML or NN model is based on at least one of quasi co-location (QCL) information or a neural network state indication from a previously trained neural network in the system of Bai/Yakovlev. This implementation advantageously address the need for training an artificial neural network having the capability to quickly and efficiently minimize prediction error without the susceptibility to local minima in the error surface experienced by current neural network training systems and methods (Black, col. 4, L10-60).
As to claim 15:
Bai in view of Yakovlev and Jung discloses all limitations of claim 1, however is silent on wherein the processing time for training the ML or NN model is based on an accuracy level.
Black, in a related field of training learning model, discloses training a neural network wherein the NN is initially trained with initial set of parameters. Upon the NN outputs an accuracy indicia (error ratio) satisfying a threshold condition, the trained ANN is further trained at a modified learning rate (which affects learning time). (See at least Abstract, Fig. 4 and associated text).
It would have been obvious to one of ordinary skill in the art before the effective filing time of the invention that processing time for training the ML or NN model is based on at least one of quasi co-location (QCL) information or a neural network state indication from a previously trained neural network in the system of Bai/Yakovlev. This implementation advantageously address the need for training an artificial neural network having the capability to quickly and efficiently minimize prediction error without the susceptibility to local minima in the error surface experienced by current neural network training systems and methods (Black, col. 4, L10-60).
Allowable Subject Matter
Claim 18 and 30 is/are objected to as being dependent upon a rejected base claim, but would be allowable if rewritten in independent form including all of the limitations of the base claim and any intervening claims and overcoming all outstanding formality issues. The references of record discloses all limitations of claim 1 and 21, however do not disclose: “report a first processing time and a second processing time, wherein the memory and the at least one processor are further configured to: receive a configuration from a second network entity to use the first processing time or the second processing time; and apply the first processing time or the second processing time based on a power saving feature of the UE”.
Conclusion
The prior art made of record and not relied upon is considered pertinent to applicant's disclosure.
US 2021/0351885 - Methods and systems for reporting CSI and selecting optimal beams using ML. The CSI report is sent to a gNB, which includes feedback parameters, computed and predicted using ML. The feedback parameters are computed using measurements performed using CSI-RS. Values of the feedback parameters likely at future, based on channel variation and the measurements, are pre-dieted using ML. The computed and predicted feedback parameters are included in the CSI report. Optimal CSI-RS resource allocation and optimal CSI reporting periodicity are determined using ML and sent to the gNB. The CSI report is encoded using the ML based model. The RSRP of the beams are predicted using ML for beam selection.
Applicant's amendment necessitated the new ground(s) of rejection presented in this Office action. Accordingly, THIS ACTION IS MADE FINAL. See MPEP § 706.07(a). Applicant is reminded of the extension of time policy as set forth in 37 CFR 1.136(a).
A shortened statutory period for reply to this final action is set to expire THREE MONTHS from the mailing date of this action. In the event a first reply is filed within TWO MONTHS of the mailing date of this final action and the advisory action is not mailed until after the end of the THREE-MONTH shortened statutory period, then the shortened statutory period will expire on the date the advisory action is mailed, and any nonprovisional extension fee (37 CFR 1.17(a)) pursuant to 37 CFR 1.136(a) will be calculated from the mailing date of the advisory action. In no event, however, will the statutory period for reply expire later than SIX MONTHS from the mailing date of this final action.
Any inquiry concerning this communication or earlier communications from the examiner should be directed to QUAN M HUA whose telephone number is (571)270-7232. The examiner can normally be reached 10:30-6:30.
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, Anthony Addy can be reached at 571-272-7795. 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.
/QUAN M HUA/Primary Examiner, Art Unit 2645