DETAILED ACTION
Notice of Pre-AIA or AIA Status
The present application, filed on or after March 16, 2013, is being examined under the first inventor to file provisions of the AIA .
This action is made non-final.
Continued Examination Under 37 CFR 1.114
A request for continued examination under 37 CFR 1.114, including the fee set forth in 37 CFR 1.17(e), was filed in this application after final rejection. Since this application is eligible for continued examination under 37 CFR 1.114, and the fee set forth in 37 CFR 1.17(e) has been timely paid, the finality of the previous Office action has been withdrawn pursuant to 37 CFR 1.114. Applicant's submission filed on July 14, 2026, and August 13, 2026, has been entered. Claims 1-20 are pending. Claims 1, 9, and 17 are independent claims.
Response to Arguments
Applicant’s arguments, dated July 14, 2026, regarding the 35 U.S.C. 112(b) rejection of claim 8 in the previous office action have been fully considered. In light of the amendment, the 112(b) rejection for claim 8 has been withdrawn.
Applicant’s arguments, dated July 14, 2026, regarding the 35 U.S.C. 103 rejections of the previous office action have been fully considered but are moot as they do not apply to the reference Shi being used in the current rejections of claims 1, 9, and 17 and their associated dependent claims 2-8, 10-16, and 18-20, respectively, to teach the amended claim limitation directed to “wherein the pre-trained is already trained via supervised training, unsupervised training or reinforcement learning to perform a defined inferencing task and is also employed to perform reprogrammable federated learning of the pre-trained and frozen neural network using local training datasets respectively stored by the set of client devices.”
Specifically, Shi teaches “wherein the pre-trained [and frozen neural network] is already trained via… unsupervised training” (Page 3, Section III and Page 6, Section IV, A. “Experimental Setup”), “to perform a defined inferencing task” (Page 3, Fig. 2, Caption and Page 4, Section A), and “employed to perform reprogrammable federated learning of the pre-trained and frozen neural network” (Page 4, Col. 1, Section A, ¶1, Page 6, Col. 1, Section A, ¶1 as explained in the rejection for claim 1 below). Additionally, Shi replaces Han in teaching “a newly-insertable trainable output layer” (Page 3, Fig. 2, Caption, Page 4, Section A and Page 6, Section IV, A. “Experimental Setup”).
With the addition of the reference Shi teaching the subject matter introduced in the amendments, the rejections under 103 U.S.C. 103 stand.
Claim Objections
Claim 1 is objected to because of the following informalities: “wherein the pre-trained is already trained…” should read “wherein the pre-trained and frozen neural network is already trained…” Appropriate correction is required.
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.
Claims 1-3, 5-6, and 8 are rejected under 35 U.S.C. 103 as being unpatentable over McMahan et al. (US 20190227980 A1) in view of Shi et al. (“FedCoCo: A Memory Efficient Federated Self-supervised Framework for On-Device Visual Representation Learning,” 2022, hereinafter Shi) and further in view of An et al. (“Input-Tuning: Adapting Unfamiliar Inputs to Frozen Pretrained Models,” hereinafter An).
Regarding claim 1:
McMahan teaches a server device (Fig. 2 – 200, 210, ¶134 “The system 200 includes a server 210”), comprising: a processor that executes computer-executable components stored in a non-transitory computer-readable memory to perform operations comprising… (Fig. 2 – 212-216, ¶136 “The one or more memory devices 214 can store information accessible by the one or more processors 212, including computer-readable instructions 216 that can be executed by the one or more processors 212”).
Regarding the limitation sharing a pre-trained frozen neural network with a set of client devices, wherein the pre-trained is already trained via supervised training, unsupervised training, or reinforcement learning to perform a defined inferencing task and is also employed to perform reprogrammable federated learning of the pre-trained and frozen neural network using local training datasets respectively stored by the set of client devices, McMahan teaches sharing a… neural network with a set of client devices (Fig. 1 – 102-106, ¶127 “The one or more server computing device(s) 104 can be configured to access machine-learned model 106, and to provide model 106 to a plurality of client computing devices 102. Model 106 can be… a neural network”) and to perform… federated learning of the… neural network using local training datasets respectively stored by the set of client devices (¶125 “private training, particularly in a federated learning framework, shows the possibility of training models with significant privacy guarantees,” Fig. 1 – 102-106, ¶127 “an example system 100 for training one or more machine learning models 106 using respective training data 108 stored locally on a plurality of client computing devices 102 (i.e., a local dataset”). However, McMahan fails to teach a pre-trained frozen neural network and wherein the pre-trained is already trained via supervised training, unsupervised training, or reinforcement learning to perform a defined inferencing task and is also employed to perform reprogrammable federated learning of the pre-trained and frozen neural network…
Shi, in the same field of endeavor, teaches a pre-trained frozen neural network (Page 6, Col. 1, Section A, ¶1 “To evaluate the performance of the representations generated by the encoder in the trained model… We freeze the parameters in the encoder,” Col. 2, ¶2 “We use Resnet-18 as the architecture for encoders”) and wherein the pre-trained is already trained via supervised training, unsupervised training, or reinforcement learning to perform a defined inferencing task (Page 3, Col. 2, Section III, ¶1 “This paper proposes a framework FedCoCo aiming at training a global model to extract good visual representations from unlabeled streaming data in a network of edge devices with limited storage,” Page 6, Col. 1, Section IV, ¶1 “we evaluate the performance of the visual representation learned by FedCoCo,” please note that supervised training, unsupervised training, or reinforcement learning is being interpreted to mean that the items in the list are disjunctive, hence training a model through “unlabeled streaming data” encompasses unsupervised training; furthermore, “to extract good visual representations” encompass a defined inferencing task) and is also employed to perform reprogrammable federated learning of the pre-trained and frozen neural network (Page 4, Col. 1, Section A, ¶1 “there are two stages in the proposed FCL framework FedCoCo. In the first stage, the model is collaboratively trained on streaming unlabeled data by distributed devices to generate good visual representations. An additional classifier is then trained on top of the online encoder from the learned model with few (e.g. 1%) labeled data, in the second stage,” Page 6, Col. 1, Section A, ¶1 “To evaluate the performance of the representations generated by the encoder in the trained model, we use linear evaluation… We freeze the parameters in the encoder and train a classifier on top of it,” wherein adding a trainable “classifier” on top of “the online encoder of the learned model” which was pre-trained using federated learning encompasses reprogrammable federated learning of the pre-trained and frozen neural network when given its broadest reasonable interpretation of further modifying and fine-tuning a model architecture that has already been trained using federated learning).
McMahan and Shi are analogous art to the claimed invention as both are in the same field of endeavor of federated learning. Therefore, it would have been obvious to one of ordinary skill in the art, before the effective filing date of the claimed invention, to combine the pre-trained and frozen neural network of Shi with the system of McMahan. The motivation to do so is “to handle… huge amounts of decentralized unlabeled data with limited storage resources on edge [devices]” (Shi, Abstract).
Regarding the limitation and performing training via orchestrating reprogrammable federated learning of the pre-trained and frozen neural network among the set of client devices, wherein the training via orchestrating the reprogrammable federated learning comprises foregoing altering already-trained internal parameters of the pre-trained and frozen neural network and sandwiching the pre-trained and frozen neural network in series between a newly-inserted trainable input layer and a newly-insertable trainable output layer, McMahan teaches and performing training via orchestrating… federated learning of the… neural network among the set of client devices (¶54 “Example algorithms provided by the present disclosure are based on… federated learning. For example, in federated learning, a shared model can be trained while leaving the training data on each user's client computing device”). However, McMahan fails to teach and performing training via orchestrating reprogrammable federated learning of the pre-trained and frozen neural network among the set of client devices, wherein the training via orchestrating the reprogrammable federated learning comprises foregoing altering already-trained internal parameters of the pre-trained and frozen neural network and sandwiching the pre-trained and frozen neural network in series between a newly-inserted trainable input layer and a newly-insertable trainable output layer.
Shi teaches reprogrammable federated learning of the pre-trained and frozen neural network (Page 4, Col. 1, Section A, ¶1, Page 6, Col. 1, Section A, ¶1 as explained above) … wherein the training via orchestrating the reprogrammable federated learning comprises foregoing altering already-trained internal parameters of the pre-trained and frozen neural network and sandwiching the pre-trained and frozen neural network in series between… and a newly-insertable trainable output layer (Page 3, Fig. 2, Caption “The online encoder is first collaboratively trained on clients with data selected from streaming unlabeled data to generate good representations. Then a classifier is trained with few labeled data on top of the online encoder,” Page 6, Col. 1, Section A, ¶1 “We freeze the parameters in the encoder and train a classifier on top of it with 1%, 10%, or 100% labeled data,” Figure 2 depicts inserting a trainable “Classifier,” which functions in substantially the same way as a newly-insertable trainable output layer at the end of a trained “Backbone” or the pre-trained and frozen neural network). However, Shi fails to teach sandwiching the pre-trained and frozen neural network in series between a newly-inserted trainable input layer…
An, in the same field of endeavor, teaches sandwiching the pre-trained and frozen neural network in series between a newly-inserted trainable input layer (Abstract: “By only tuning continuous prompts with a frozen pre trained language model (PLM), prompt-tuning takes a step towards deploying a shared frozen PLM to serve numerous downstream tasks,” Page 2, Fig. 2 depicts an “Input-Adapter” inserted below a “Frozen PLM,” Page 2, Column 2, ¶1 “we add a lightweight trainable module between word embeddings and the bottom layer of the PLM, to adjust the encoding of unfamiliar inputs directly (i.e., the “input-adapter” module in Figure 2),” Page 5, Fig. 4 and Equation (9) depict the “Input-Adapter,” Page 5, Col. 1, ¶1 “Here σ is an element-wise nonlinear activation function (ReLU). W1 and W2 are learnable matrices. We denote T (·) as an input-adapter, since it plays the role that adapts the surface representations of x to better utilize the frozen PLM. Figure 4 illustrates this module,” Page 7, Column 1, Section 6: “Section 6.1 analyzes the performance with different PLM backbones,” Section 6.1: “We evaluate the effectiveness on different back bones from two aspects: backbone architectures (GPT2-Large and T5-Large)… with word embeddings as input,” an “input-adapter” that has “learnable matrices,” uses a “nonlinear activation function (ReLU),” and adapts input representations before forwarding the results to the “Frozen PLM” functions in substantially the same way as a newly-inserted trainable input layer) …
McMahan, Shi, and An are analogous art to the claimed invention as all are from the same field of endeavor of machine learning. Therefore, it would have been obvious to one of ordinary skill in the art, before the effective filing date of the claimed invention, to combine the newly-inserted trainable input layer and pre-trained and frozen neural network of An and the newly-insertable trainable output layer and pre-trained and frozen neural network using unsupervised learning for a defined inferencing task of Shi with the neural network, federated learning, client devices, and server device of McMahan. The motivation to do so is “to handle… huge amounts of decentralized unlabeled data with limited storage resources on edge [devices]” (Shi, Abstract) and to “achieve a comparable or even better performance than fine-tuning” (An, Abstract).
Regarding claim 2, McMahan in view of Shi and further in view of An teaches the server device of claim 1 (and thus the rejection of claim 1 is incorporated).
Regarding the limitation and wherein the reprogrammable federated learning involves the at least one trainable input layer and the at least one trainable output layer, but not the pre-trained and frozen neural network, being locally adjusted by the set of client devices, McMahan teaches and wherein the… federated learning involves the at least one trainable input layer and the at least one trainable output layer… being locally adjusted by the set of client devices (Fig. 1 – 106, “Model 106 can be… a neural network,” Fig. 3 – 306-308, ¶24 “after receiving the machine-learned model, each selected client computing device can then determine a local update,” although not explicitly stated, one of ordinary skill in the art would recognize that it is implicit that federated learning involves the at least one trainable input layer and the at least one trainable output layer of a neural network when training a neural network with federated learning). However, McMahan fails to teach reprogrammable federated learning and but not the pre-trained and frozen neural network, being locally adjusted…
Shi teaches reprogrammable federated learning (Page 4, Col. 1, Section A, ¶1, Page 6, Col. 1, Section A, ¶1 as explained above with respect to claim 1) and but not the pre-trained and frozen neural network being adjusted (Page 6, Col. 1, Section A, ¶1 “We freeze the parameters in the encoder and train a classifier on top of it”).
McMahan and Shi are analogous art to the claimed invention as both are in the same field of endeavor of federated learning. Therefore, it would have been obvious to one of ordinary skill in the art, before the effective filing date of the claimed invention, to combine the pre-trained and frozen neural network of Shi with the trainable input and output layers of McMahan. The motivation to do so is “to handle… huge amounts of decentralized unlabeled data with limited storage resources on edge [devices]” (Shi, Abstract).
Regarding claim 3, McMahan in view of Shi and further in view of An teaches the server device of claim 2 (and thus the rejection of claim 2 is incorporated).
Regarding the limitation wherein, during an iteration of the reprogrammable federated learning, the processor also: shares a global internal parameter value array of the at least one trainable input layer and of the at least one trainable output layer with the set of client devices, McMahan teaches wherein, during an iteration of the… federated learning, the processor also: shares a global internal parameter value array of the at least one trainable input layer and of the at least one trainable output layer with the set of client devices (Fig. 1 – 106, “Model 106 can be… a neural network,” Fig. 3 – 304-306, ¶149 “At (304), the method (300) can include providing, by the one or more server computing devices, the machine-learned model to the selected client computing devices, and at (306), the method (300) can include receiving, by the selected client computing devices, the machine-learned model. In some implementations, the machine-learned model can be, for example, a global set of parameters,” a person having ordinary skill in the art would recognize that sharing a global internal parameter value array of the at least one trainable input layer and of the at least one trainable output layer with the set of client devices is implicit when training a neural network with federated learning). However, McMahan fails to teach the reprogrammable federated learning.
Shi teaches the reprogrammable federated learning (Page 4, Col. 1, Section A, ¶1, Page 6, Col. 1, Section A, ¶1 as explained above with respect to claim 1).
McMahan further teaches and instructs the set of client devices to locally update the global internal parameter value array of the at least one trainable input layer and of the at least one trainable output layer using local training datasets, thereby causing the set of client devices to respectively generate a set of locally-updated internal parameter value arrays of the at least one trainable input layer and of the at least one trainable output layer (Fig. 1 – 106, “Model 106 can be… a neural network,” Fig. 2 – 238, Fig. 3 – 306-310, ¶143 “a client device 230 can receive a machine-learned model (such as a set of global parameters) from the server 210, train the machine-learned model based at least in part on the local dataset to generate a locally-trained model (such as updated local values for the global set of parameters for the machine-learned model), determine a difference between the machine-learned model and the locally-trained model (such as a difference between the global parameters and the updated local values), and clip the difference to generate the local update. In some implementations, the local update can be expressed in a vector, a matrix, or other suitable format,” one of ordinary skill in the art would recognize that locally updating the global internal parameter value array of the at least one trainable input layer and of the at least one trainable output layer using local training datasets is implicit when training a neural network with federated learning).
McMahan and Shi are analogous art to the claimed invention as both are in the same field of endeavor of federated learning. Therefore, it would have been obvious to one of ordinary skill in the art, before the effective filing date of the claimed invention, to combine the fine-tuning method of Shi with the federated learning system of McMahan. The motivation to do so is “to handle… huge amounts of decentralized unlabeled data with limited storage resources on edge [devices]” (Shi, Abstract).
Regarding claim 5, McMahan in view of Shi and further in view of An teaches the server device of claim 3 (and thus the rejection of claim 3 is incorporated).
Regarding the limitation wherein, during the iteration of the reprogrammable federated learning, the processor also: accesses the set of locally-updated internal parameter value arrays from the set of client devices, McMahan teaches wherein, during the iteration of the… federated learning, the processor also: accesses the set of locally-updated internal parameter value arrays from the set of client devices (Fig. 3 – 310-312, ¶151 “At (310), the method (300) can include providing, by each selected client computing device, the local update to the one or more server computing devices, and at (312), the method (300) can include receiving, by the one or more server computing devices, the local updates. In some implementations, the local updates can be, for example, expressed in one or more matrices, vectors, or other suitable format”). However, McMahan fails to teach the reprogrammable federated learning.
Shi teaches the reprogrammable federated learning (Page 4, Col. 1, Section A, ¶1, Page 6, Col. 1, Section A, ¶1 as explained above with respect to claim 1).
McMahan further teaches and aggregates the set of locally-updated internal parameter value arrays into a new global internal parameter value array (Fig. 3 – 314-316, ¶152 “At (314), the method (300) can include determining a differentially private aggregate of the local updates,” ¶156 “At (316), the method (300) can include determining an updated machine-learned model based at least in part on the bounded-sensitivity data-weighted average of the local updates”).
McMahan and Shi are analogous art to the claimed invention as both are in the same field of endeavor of federated learning. Therefore, it would have been obvious to one of ordinary skill in the art, before the effective filing date of the claimed invention, to combine the fine-tuning method of Shi with the federated learning system of McMahan. The motivation to do so is “to handle… huge amounts of decentralized unlabeled data with limited storage resources on edge [devices]” (Shi, Abstract).
Regarding claim 6, McMahan in view of Shi and further in view of An teaches the server device of claim 5 (and thus the rejection of claim 5 is incorporated).
Regarding the limitation wherein, during a next iteration of the reprogrammable federated learning, the processor also: shares the new global internal parameter value array with the set of client devices, McMahan teaches wherein, during a next iteration of the… federated learning, the processor also: shares the new global internal parameter value array with the set of client devices (Fig. 3 – 318-320, ¶157 “At (318), the method (300) can include providing, by the one or more server computing devices, the updated machine-learned model to one or more client computing devices, and at (320), the method (300) can include receiving, by the one or more client computing devices, the updated machine-learned model… the updated machine-learned model can be a global model provided to the pool of available client computing devices”). However, McMahan fails to teach the reprogrammable federated learning.
Shi teaches the reprogrammable federated learning (Page 4, Col. 1, Section A, ¶1, Page 6, Col. 1, Section A, ¶1 as explained above with respect to claim 1).
McMahan further teaches and instructs the set of client devices to locally update the new global internal parameter value array using the local training datasets (Fig. 3, ¶158 “Any number of iterations of local and global updates can be performed. That is, method (300) can be performed iteratively to update the machine-learned model based on locally stored training data over time,” although not explicitly stated, one of ordinary skill in the art would recognize that instructs the set of client devices to locally update the new global internal parameter value array using the local training datasets is implied to occur when “Any number of iterations of local and global updates can be performed”).
McMahan and Shi are analogous art to the claimed invention as both are in the same field of endeavor of federated learning. Therefore, it would have been obvious to one of ordinary skill in the art, before the effective filing date of the claimed invention, to combine the fine-tuning method of Shi with the federated learning system of McMahan. The motivation to do so is “to handle… huge amounts of decentralized unlabeled data with limited storage resources on edge [devices]” (Shi, Abstract).
Regarding claim 8, McMahan in view of Shi and further in view of An teaches the server device of claim 5 (and thus the rejection of claim 5 is incorporated).
Regarding the limitation wherein, during the iteration of the reprogrammable federated learning, the operations also comprise: determining, via a moment accounts technique, how much of a privacy budget associated with the pre-trained and frozen neural network has been consumed by the iteration, McMahan teaches wherein, during the iteration of the… federated learning, the operations also comprise: determining, via a moment accounts technique, how much of a privacy budget associated with the pre-trained and frozen neural network has been consumed by the iteration (¶113 “a moments accountant M can be used to achieve privacy bounds. For example, the moments accountant… can upper bound the total privacy cost of T steps”). However, McMahan fails to teach the reprogrammable federated learning.
Shi teaches the reprogrammable federated learning (Page 4, Col. 1, Section A, ¶1, Page 6, Col. 1, Section A, ¶1 as explained above with respect to claim 1).
McMahan and Shi are analogous art to the claimed invention as both are in the same field of endeavor of federated learning. Therefore, it would have been obvious to one of ordinary skill in the art, before the effective filing date of the claimed invention, to combine the fine-tuning method of Shi with the federated learning system of McMahan. The motivation to do so is “to handle… huge amounts of decentralized unlabeled data with limited storage resources on edge [devices]” (Shi, Abstract).
Claim 4 is rejected under 35 U.S.C. 103 as being unpatentable over McMahan in view of Shi and further in view of An, and further in view of Cheng et al. (US 20230039182, hereinafter Cheng).
Regarding claim 4, McMahan in view of Shi and further in view of An teaches the server device of claim 3 (and thus the rejection of claim 3 is incorporated).
Regarding the limitation wherein the set of client devices perform such local updates via differentially private stochastic gradient descent, McMahan teaches wherein the set of client devices perform such local updates (Fig. 2 – 238, Fig. 3 – 306-310, ¶143 “a client device 230 can receive a machine-learned model (such as a set of global parameters) from the server 210, train the machine-learned model based at least in part on the local dataset to generate a locally-trained model”). However, McMahan fails to teach updates via differentially private stochastic gradient descent.
Cheng, in the same field of endeavor, teaches updates via differentially private stochastic gradient descent (Fig. 6 – 601, ¶120 “each edge node device may independently select a differential privacy mechanism… such as a differentially-private stochastic gradient descent (DP-SGD)”).
McMahan and Cheng are analogous art to the claimed invention as all are from the same field of endeavor of machine learning. Therefore, it would have been obvious to one of ordinary skill in the art, before the effective filing date of the claimed invention, to combine the differentially private stochastic gradient descent of Cheng with the local client devices of McMahan. The motivation to do so is to increase privacy during federated learning (Cheng, ¶120 “DP-SGD is a method that improves a stochastic gradient descent algorithm to realize differentially-private machine learning”).
Claim 7 is rejected under 35 U.S.C. 103 as being unpatentable over McMahan in view of Shi and further in view of An, and further in view of Baracaldo et al. (“Federated Learning: A Comprehensive Overview of Methods and Applications,” hereinafter Baracaldo).
Regarding claim 7, McMahan in view of Shi and further in view of An teaches the server device of claim 5 (and thus the rejection of claim 5 is incorporated).
Regarding the limitation wherein the aggregation of the set of locally-updated internal parameter value arrays is performed via federated averaging, McMahan teaches wherein the aggregation of the set of locally-updated internal parameter value arrays is performed (Fig. 3 – 314, ¶152 “At (314), the method (300) can include determining a differentially private aggregate of the local updates”). However, McMahan fails to teach performed via federated averaging.
Baracaldo, in the same field of endeavor, teaches performed via federated averaging (Page 2, ¶2 “Once the aggregator has received the model updates from the parties, they can then be merged into a common model… this can be as simple as averaging the weights, as proposed in the FedAvg algorithm,” Page 8, ¶2 “The aggregator’s fusion algorithm F averages the parameters of each party,” Page 9, ¶2 “Using, for example, FedAvg as the fusion function F, we can then compute locally the new local model weights… All parties send their model weights to the aggregator where the weights are averaged”).
McMahan and Baracaldo are analogous art to the claimed invention as all are from the same field of endeavor of machine learning. Therefore, it would have been obvious to one of ordinary skill in the art, before the effective filing date of the claimed invention, to combine the federated averaging of Baracaldo with the locally-updated internal parameter value arrays of McMahan. The motivation to do so is to increase the efficiency of federated learning (Baracaldo, Page 8, ¶2 “FedAvg… is more effective by taking advantage of independent processing at each party… Experiments show that this approach performs well for different model types”).
Claims 9-16 recite a computer-implemented method that parallels the apparatus claims of 1-8, respectively. Therefore, the analysis discussed above with respect to claims 1-8 also applies to claims 9-16, respectively. Accordingly, claims 9-16 are rejected based on substantially the same rational as set forth above with respect to claims 1-8, respectively.
Claims 17-20 recite a computer-implemented method that parallels the apparatus claims of 1-4, respectively. Therefore, the analysis discussed above with respect to claims 1-4 also applies to claims 17-20, respectively. Accordingly, claims 17-20 are rejected based on substantially the same rational as set forth above with respect to claims 1-4, respectively.
Conclusion
Any inquiry concerning this communication or earlier communications from the examiner should be directed to WILLIAM M LEE whose telephone number is (571)272-4761. The examiner can normally be reached Mon-Fri. 8am-5pm.
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, Cesar Paula can be reached at (571)272-4128. 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.
/WILLIAM M LEE/
Examiner, Art Unit 2145
/CESAR B PAULA/Supervisory Patent Examiner, Art Unit 2145
Prosecution Insights