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 .
Claim Objections
Claim 4 objected to because of the following informalities:
“for each node according to a trainable transfonnation” should read “for each node according to a trainable transformation”
. Appropriate correction is required.
Claim Rejections - 35 USC § 103
The following is a quotation of 35 U.S.C. 103 which forms the basis for all obviousness rejections set forth in this Office action:
A patent for a claimed invention may not be obtained, notwithstanding that the claimed invention is not identically disclosed as set forth in section 102, if the differences between the claimed invention and the prior art are such that the claimed invention as a whole would have been obvious before the effective filing date of the claimed invention to a person having ordinary skill in the art to which the claimed invention pertains. Patentability shall not be negated by the manner in which the invention was made.
Claim(s) 1-2, 4, 6, 9, 11-13 are rejected under 35 U.S.C. 103 as being unpatentable over DUDASH et al. (U.S. Pub. No. US 20230325461 A1) in view of MARTIN et al. (“Combinatorial Optimization with Physics-Inspired Graph Neural Networks”).
Regarding claim 1, DUDASH teaches the invention substantially as claimed, including:
A computer-implemented method for selecting processing hardware based on a given quadratic unconstrained binary optimization, QUBO, problem, said method comprising the steps of:encoding the QUBO problem in a corresponding QUBO graph problem, wherein each binary variable of the QUBO problem corresponds to a node of the QUBO graph problem and edges between two nodes encode coefficients of terms containing both binary variables corresponding to the two nodes, or receiving a QUBO graph problem in a QUBO graph representation; ([0075] First, at step 106, the QUBO problem generated at step 104 may be represented as an undirected graph. For example, the system may generate data (e.g., a graph data structure) representing the problem as an undirected graph. An undirected graph may be defined by a plurality of vertices (or nodes) and a plurality of edges that link respective pairs of vertices to one another. An undirected graph may be represented visually as labeled points representing vertices and by lines, representing edges, connecting the points to one another. In some embodiments, the undirected graph generated to represent the QUBO problem may be a mathematical representation of a configuration of logical qubits used to solve the QUBO problem. Specifically, the vertices of the undirected graphs may represent the logical qubits while the edges may represent coupling between logical qubits.
[0076] FIG. 2 illustrates an exemplary undirected graph according to some embodiments of the present disclosure. Specifically, FIG. 2 illustrates an undirected graph 200 comprising vertices {A, B, C} and edges {{A, B}, {A, C}, {B, C}}. As shown, the vertices may be represented visually by labeled points 202, while the edges may be represented graphically by lines 204 connecting the vertices to represent which vertices are paired with one another.)
While DUDASH does teach encoding the QUBO problem onto a graph problem, it does not explicitly teach:
providing the QUBO graph problem as an input to a trained graph neural network on a processing system; retrieving a predicted performance metric for solving the QUBO problem with a variational quantum solver from an output of the trained graph neural network to the QUBO graph problem provided at the input;
However, in analogous art that similarly handles solving QUBO problems, MARTIN teaches:
providing the QUBO graph problem as an input to a trained graph neural network on a processing system; ((Page 5, Section 4, Paragraph 1)The class of Hamiltonians described above are not differentiable and cannot be used straightforwardly within the GNN training process. Therefore, for a given problem Hamiltonian H and graph G, we generate a differentiable loss function L(θ), as required for standard back propagation, by promoting the binary decision variables xi ∈ {0,1} to continuous (parametrized) probability parameters pi(θ) with the following (heuristic) relaxation approach xi −→pi(θ) ∈ [0,1] The soft assignments pi can be viewed as class probabilities. They are generated by our GNN Ansatz as final node embeddings pi = hK i ∈ [0,1] at layer K, after the application of a non-linear softmax activation function. Then, they are used as input for the loss function L(θ). In particular, for QUBO-type problems: HQUBO −→LQUBO(θ) = pi(θ)Qijpj(θ), i,j (6) which is differentiable with respect to the parameters of the GNN model θ, and similarly for PUBO problems on hyper-graphs with higher-order terms of the form pipjpk etc., thereby establishing a straightforward, general connection between combinatorial optimization problems, Ising Hamiltonians and GNNs. )retrieving a predicted performance metric for solving the QUBO problem with a variational quantum solver from an output of the trained graph neural network to the QUBO graph problem provided at the input; ((Page 7, Paragraph 2) We perform the benchmarks as follows. For graphs with up to a few hundred nodes, we compare our GNN-based solver to the (approximate) polynomial time Goemans-Williamson (GW) algorithm [81], which provides the current record for an approximate answer within some fixed multiplicative factor of the optimum (referred to as approximation ratio α), using semidefinite programming and randomized rounding. Specifically, the GW algorithm achieves a guaranteed approximation ratio of α ∼ 0.878 for generic graphs. This lower bound can be raised for specific graphs such as unweighted 3-regular graphs where α ∼ 0.9326 )
It would have been obvious to a person skilled in the art before the effective filing date of the invention to have combined with MARTIN’s teaching of inputting QUBO problem data into a GCN and, with DUDASH’s, teaching of encoding QUBO problem data onto a graph problem, to realize, with a reasonable expectation of success, a method that predicts a performance metric of QUBO graph problem data, as in MARTIN, using the QUBO problem data that was encoded onto the QUBO graph problem, as in DUDASH. A person of ordinary skill would have been motivated to improve the optimization of the solvers(MARTIN Page 1, introduction paragraph 1)
DUDASH further teaches:
based on the predicted performance metric, providing the QUBO problem to the variational quantum solver implemented on quantum hardware. ([0006] and updating the best local current graph, wherein the updating comprises: copying the best local current graph to form a best local current graph copy, modifying the best local current graph copy to form a candidate local graph, computing an evaluation rating for the candidate local graph, and determining, based on the evaluation rating for the candidate local graph and an evaluation rating for the best local current graph, whether one or more replacement criteria are met; in accordance with a determination that the one or more replacement criteria are met, replacing the best local current graph with the candidate local graph; storing updated best local current graphs associated respectively with each of the one or more GPU thread blocks in a local results array; identifying an updated best local current graph in the local results array as the best global graph; and configuring the quantum annealer based on the best global graph.)
MARTIN further teaches:
The method of claim 1, wherein the trained graph neural network has been trained to estimate the quantum approximation ratio, when solving the QUBO problem with the variational quantum solver, or a performance metric based thereon. ((Page 7, Paragraph 2) We perform the benchmarks as follows. For graphs with up to a few hundred nodes, we compare our GNN-based solver to the (approximate) polynomial time Goemans-Williamson (GW) algorithm [81], which provides the current record for an approximate answer within some fixed multiplicative factor of the optimum (referred to as approximation ratio α), using semidefinite programming and randomized rounding. Specifically, the GW algorithm achieves a guaranteed approximation ratio of α ∼ 0.878 for generic graphs. This lower bound can be raised for specific graphs such as unweighted 3-regular graphs where α ∼ 0.9326 “Combinatorial Optimization with Physics-Inspired Graph Neural Networks”)
Regarding claim 4, MARTIN further teaches:
The method of claim 1, wherein the trained graph neural network comprises a plurality of graph convolution layers, wherein each convolution layer comprises an aggregation step for aggregating, for each node of the QUBO graph problem, feature vectors of neighboring nodes into an aggregated feature vector,( (page 2, paragraph 2) The GNN follows a standard recursive neighborhood aggregation scheme [43, 44], where each node ν = 1,2,...,n collects information (encoded as feature vectors) of its neighbors to compute its new feature vector hk ν at layer k = 0, 1,...,K. After k iterations of aggregation, a node is represented by its transformed feature vector hk ν, which captures the structural information within the node’s k hop neighborhood [28]. ) and a transformation step, wherein an original feature vector of the node and the aggregated feature vector are transformed to an updated feature vector for each node according to a trainable transfonnation. ((page 2, paragraph 2) After k iterations of aggregation, a node is represented by its transformed feature vector hk ν, which captures the structural information within the node’s k hop neighborhood [28]. )
Regarding claim 6, MARTIN further teaches:
The method of claim 1, wherein the trained neural network has been trained by:receiving a training set, the training set comprising a plurality of QUBO graph problems and a performance metric corresponding to each of the QUBO graph problems, ((page 2, paragraph 2) based solver to (approximately) solve combinatorial optimization problems with up to millions of variables. The approach is schematically depicted in Fig. 1, and works as follows: First, we identify the Hamiltonian (cost function) H that encodes the optimization problem in terms of binary decision variables xν ∈ {0,1} and we associate this variable with a vertex ν ∈ V for an undirected graph G = (V,E) with vertex set V = {1,2,...,n} and the edge set E = {(i,j) : i,j ∈ V} capturing interactions between the decision variables. We then apply a relaxation strategy to the problem Hamiltonian to generate a differentiable loss function with which we perform unsupervised training on the node representations of the GNN. The GNN follows a standard recursive neighborhood aggregation scheme [43, 44], where each node ν = 1,2,...,n collects information (encoded as feature vectors) of its neighbors to compute its new feature vector hk ν at layer k = 0, 1,...,K. After k iterations of aggregation, a node is represented by its transformed feature vector hk ν, which captures the structural information within the node’s k hop neighborhood [28]. ) the performance metric indicating a ratio between a quality indicator for a solution candidate determined using the variational quantum solver for the QUBO graph problem and the same quality indicator for a classical solution to the QUBO graph problem, ((Page 7, paragraph 2) The complexity of MaxCut depends on the regularity and connectivity of the underlying graph. Following an existing trend in the community [76], we first consider the MaxCut problem on random (unweighted) d-regular graphs, where every vertex is connected to exactly d other vertices. We perform the benchmarks as follows. For graphs with up to a few hundred nodes, we compare our GNN-based solver to the (approximate) polynomial time Goemans-Williamson (GW) algorithm [81], which provides the current record for an approximate answer within some fixed multiplicative factor of the optimum (referred to as approximation ratio α), using semidefinite programming and randomized rounding. Specifically, the GW algorithm achieves a guaranteed approximation ratio of α ∼ 0.878 for generic graphs. This lower bound can be raised for specific graphs such as unweighted 3-regular graphs where α ∼ 0.9326 [82]. )) wherein the quality indicator is in particular a cost associated with the solution candidate; ((page 2, paragraph 2) First, we identify the Hamiltonian (cost function) H that encodes the optimization problem in terms of binary decision variables xν ∈ {0,1} and we associate this variable with a vertex ν ∈ V for an undirected graph G = (V,E) with vertex set V = {1,2,...,n} and the edge set E = {(i,j) : i,j ∈ V} capturing interactions between the decision variables. ) training a graph convolutional network on the training set, wherein the input of the graph convolution network receives the QUBO graph problems, and trainable parameters of the graph convolutional network are iteratively updated, such that the output of the graph convolutional network for a given QUBO graph problem approaches the performance metric. ((page 2, paragraph 2) Finally, once the unsupervised training process has completed, we apply a projection heuristic to map these soft assignments pν back to integer variables xν ∈ {0,1} using, for example, xν = int(pν). We numerically showcase our approach with results for canonical NP-hard optimization problems such as maximum cut (MaxCut) and maximum independent set (MIS), showing that our GNN-based approach can perform on par or even better than existing well established solvers, while being broadly applicable to a large class of optimization problems. )
Regarding claim 12, DUDASH further teaches:
The system of claim 11, wherein obtaining the corresponding QUBO graph problem comprises encoding, by the processing system, the QUBO problem in the corresponding QUBO graph problem. ([0075] First, at step 106, the QUBO problem generated at step 104 may be represented as an undirected graph. For example, the system may generate data (e.g., a graph data structure) representing the problem as an undirected graph. An undirected graph may be defined by a plurality of vertices (or nodes) and a plurality of edges that link respective pairs of vertices to one another. An undirected graph may be represented visually as labeled points representing vertices and by lines, representing edges, connecting the points to one another. In some embodiments, the undirected graph generated to represent the QUBO problem may be a mathematical representation of a configuration of logical qubits used to solve the QUBO problem. Specifically, the vertices of the undirected graphs may represent the logical qubits while the edges may represent coupling between logical qubits.
[0076] FIG. 2 illustrates an exemplary undirected graph according to some embodiments of the present disclosure. Specifically, FIG. 2 illustrates an undirected graph 200 comprising vertices {A, B, C} and edges {{A, B}, {A, C}, {B, C}}. As shown, the vertices may be represented visually by labeled points 202, while the edges may be represented graphically by lines 204 connecting the vertices to represent which vertices are paired with one another.)
Regarding claim 13, DUDASH further teaches:
The system of claim 11, wherein the system in particular comprises a machine readable model executable on the processing system and/or wherein the system further comprises Al processing hardware configured to implement the trained graph neural network, wherein the Al processing hardware in particular comprises a GPU, a neural processing unit, analog memory based hardware, or neuromorphic hardware. ([0006] In some embodiments, a method for configuring a quantum annealer to solve a quadratic unconstrained binary optimization problem comprises: receiving data representing an initial graph representing an embedding of a QUBO problem into a physical qubit architecture of the quantum annealer; initializing one or more GPU thread blocks; for each of the one or more GPU thread blocks:)
Regarding claims 9 and 11, they comprise of limitations similar to those of claim 1 and are therefore rejected for similar rationale.
Claim(s) 3 is rejected under 35 U.S.C. 103 as being unpatentable over DUDASH et al. (U.S. Pub. No. US 20230325461 A1), MARTIN et al. (“Combinatorial Optimization with Physics-Inspired Graph Neural Networks”) in further view of YAO et al. (U.S. Pub. No. US 20200394499 A1) .
While DUDASH, as modified by MARTIN, does teach claim 1, which claim 3 is dependent upon, it does not explicitly teach:
The method of claim 1, wherein the trained graph neural network is a convolutional graph neural network, in particular a spatial graph neural network.
However, in analogous art that similarly handles a GCN, YAO teaches:
The method of claim 1, wherein the trained graph neural network is a convolutional graph neural network, in particular a spatial graph neural network. ([0067] A conventional hourglass network includes a series of down-sampling and up-sampling operations with skip connections. These conventional networks follow the principles of the information bottleneck approach to deep learning models for improved performance. These conventional networks have also been shown to work well for tasks such as human pose estimation, facial landmark localization, etc. In accordance with the techniques of the disclosure, STGCNs 210 are designed to include an hourglass architecture that leverages an encoder-decoder structure for multi-modal action identification and segmentation with improved accuracy. Particularly, architecture 600 hourglass includes a series of STGCN blocks 210. Each STGCN block 210 includes an STGCN layer 510 (e.g., one or more spatial graph convolution layers, one or more temporal graph layers, and one or more non-linear layers) followed by a strided convolution layer 604 form the basic building block for an encoding process. One or more deconvolution layers 606 form the basic building block for the decoding process that brings the spatial and temporal dimensions to an original size.)
It would have been obvious to a person skilled in the art before the effective filing date of the invention to have combined with YAO’s teaching of a GCN that is a spatial GCN and, with DUDASH’s as modified by MARTIN, teaching of encoding QUBO problem data onto a graph problem, to realize, with a reasonable expectation of success, a method that utilizes a spatial GCN, as in YAO, for inputting the QUBO graph problem data, as in DUDASH, as modified by MARTIN. A person of ordinary skill would have been motivated to improve the optimization of the GCN (YAO [0005])
Claim(s) 5 is rejected under 35 U.S.C. 103 as being unpatentable over DUDASH et al. (U.S. Pub. No. US 20230325461 A1), MARTIN et al. (“Combinatorial Optimization with Physics-Inspired Graph Neural Networks”) in further view of NING et al. (U.S. Pub. No. US 20210090284 A1) .
While DUDASH, as modified by MARTIN, does teach claim 4, which claim 5 is dependent upon, it does not explicitly teach:
The method of claim 4, wherein the trained graph neural network comprises a pooling layer for pooling the output of at least one of the plurality of the graph convolution layers, and a fully-connected layer between the pooling layer and an output of the trained graph neural network.
However, in analogous art that similarly handles a GCN, NING teaches:
The method of claim 4, wherein the trained graph neural network comprises a pooling layer for pooling the output of at least one of the plurality of the graph convolution layers, and a fully-connected layer between the pooling layer and an output of the trained graph neural network. ( [0014] first graphic convolutional network (GCN) layer; a first Relu Unit connected to the first GCN layer; a second GCN layer connected to the first Relu Unit; a second Relu Unit connected to the second GCN layer; an average pooling layer connected to the second GCN layer; a fully connected network (FCN);)
It would have been obvious to a person skilled in the art before the effective filing date of the invention to have combined with NING’s teaching of a GCN fully connected with a pooling layer and, with DUDASH’s as modified by MARTIN, teaching of encoding QUBO problem data onto a graph problem, to realize, with a reasonable expectation of success, a method that utilizes a fully connected GCN, as in NING, for inputting the QUBO graph problem data, as in DUDASH, as modified by MARTIN. A person of ordinary skill would have been motivated to improve accuracy (NING [0004])
Claim(s) 7 and 10 are rejected under 35 U.S.C. 103 as being unpatentable over DUDASH et al. (U.S. Pub. No. US 20230325461 A1), MARTIN et al. (“Combinatorial Optimization with Physics-Inspired Graph Neural Networks”) in further view of SUGUIURA et al. (U.S. Pub. No. US 11748707 B2) .
While DUDASH, as modified by MARTIN, does teach claim 1, which claim 7 is dependent upon, it does not explicitly teach:
The method of claim 1, wherein the variational quantum solver comprises a variational quantum network for determining an output indicative for the solution to the QUBO problem, wherein a gate configuration of the variational quantum network is in particular selected based on the QUBO problem, wherein the variational quantum network is preferably based on a Quantum Approximate Optimization Algorithm, QAOA, and/or wherein the gate configuration preferably depends on a cost function attributing a cost to a solution for the QUBO problem and an optimal solution is associated with a global extremum of the cost function.
However, in analogous art that similarly handles a GCN, SUGUIURA teaches:
The method of claim 1, wherein the variational quantum solver comprises a variational quantum network for determining an output indicative for the solution to the QUBO problem, wherein a gate configuration of the variational quantum network is in particular selected based on the QUBO problem, wherein the variational quantum network is preferably based on a Quantum Approximate Optimization Algorithm, QAOA, and/or wherein the gate configuration preferably depends on a cost function attributing a cost to a solution for the QUBO problem and an optimal solution is associated with a global extremum of the cost function. ((Column 8, lines 22-45) In other implementations the quantum computing resources can be a quantum gate processor. A quantum gate processor includes one or more quantum circuits, i.e., models for quantum computation in which a computation is performed using a sequence of quantum logic gates, operating on a number of qubits (quantum bits). Quantum gate processors can be used to solve certain optimization problems, e.g., problems that can be formulated as a QUBO problem. For example, some quantum gate processors can solve QUBO problems by simulating a corresponding adiabatic quantum annealing process using a gate model. This can be advantageous, e.g., compared to directly performing the corresponding adiabatic quantum annealing process using a quantum annealer device, since not all quantum annealer devices can realize physical quantum systems that represent an optimization problem. For example, some quantum annealer devices may not provide the physical interactions necessary to solve an optimization problem. In these examples, a Hamiltonian describing the optimization problem can be decomposed into a sequence of single or multi-qubit quantum gates, and a solution to the optimization problem can be obtained through application of the sequence of single or multi-qubit gates on a register of qubits and subsequent measurement of the register of qubits.)
It would have been obvious to a person skilled in the art before the effective filing date of the invention to have combined with SUGUIURA’s teaching of a quantum network that employs quantum gating and, with DUDASH’s as modified by MARTIN, teaching of finding a QUBO solution, to realize, with a reasonable expectation of success, a method that utilizes quantum gating, as in SUGUIURA, for finding a solution for the QUBO problem, as in DUDASH, as modified by MARTIN. A person of ordinary skill would have been motivated to improve data management (SUGUIURA Background)
SUGUIURA further teaches:
The method of claim 9, wherein the variational quantum solver comprises a variational quantum network for determining an output indicative for the solution to the QUBO problem, wherein the quantum gate configuration of the variational quantum network is specific to the QUBO problem, and wherein the variational parameters define the output of the variational quantum network specific to the QUBO problem, wherein the variational quantum network is in particular based on a Quantum Approximate Optimization Algorithm, QAOA. ((Column 8, lines 22-45) In other implementations the quantum computing resources can be a quantum gate processor. A quantum gate processor includes one or more quantum circuits, i.e., models for quantum computation in which a computation is performed using a sequence of quantum logic gates, operating on a number of qubits (quantum bits). Quantum gate processors can be used to solve certain optimization problems, e.g., problems that can be formulated as a QUBO problem. For example, some quantum gate processors can solve QUBO problems by simulating a corresponding adiabatic quantum annealing process using a gate model. This can be advantageous, e.g., compared to directly performing the corresponding adiabatic quantum annealing process using a quantum annealer device, since not all quantum annealer devices can realize physical quantum systems that represent an optimization problem. For example, some quantum annealer devices may not provide the physical interactions necessary to solve an optimization problem. In these examples, a Hamiltonian describing the optimization problem can be decomposed into a sequence of single or multi-qubit quantum gates, and a solution to the optimization problem can be obtained through application of the sequence of single or multi-qubit gates on a register of qubits and subsequent measurement of the register of qubits.)
Claim(s) 8 and 14 are rejected under 35 U.S.C. 103 as being unpatentable over DUDASH et al. (U.S. Pub. No. US 20230325461 A1), MARTIN et al. (“Combinatorial Optimization with Physics-Inspired Graph Neural Networks”) in further view of ROSSI et al. (U.S. Pub. No. US 20220309334 A1) .
While DUDASH, as modified by MARTIN, does teach claim 1, which claim 8 is dependent upon, it does not explicitly teach:
The method of claim 1, wherein the method further comprises, as part of providing the QUBO graph problem to the variational quantum solver implemented on quantum hardware:providing the QUBO graph problem to a second trained graph neural network, wherein the second trained graph neural network has been trained, for a plurality of QUBO graph problems, to predict variational parameters associated with an optimal solution by the variational quantum solver,
However, in analogous art that similarly handles a GCN, ROSSI teaches:
The method of claim 1, wherein the method further comprises, as part of providing the QUBO graph problem to the variational quantum solver implemented on quantum hardware:providing the QUBO graph problem to a second trained graph neural network, wherein the second trained graph neural network has been trained, for a plurality of QUBO graph problems, to predict variational parameters associated with an optimal solution by the variational quantum solver, ([0018] The trained graph neural network predicts features, types, class labels, etc. of data in the dataset regardless of heterophily or homophily in the dataset, which thereby improves the performance of graph neural networks using real-world datasets)
It would have been obvious to a person skilled in the art before the effective filing date of the invention to have combined with ROSSI’s teaching of a GCN that predicts a feature, with DUDASH’s as modified by MARTIN, teaching of finding a QUBO solution, to realize, with a reasonable expectation of success, a method that predicts a feature, as in ROSSI, where that feature is the variational parameters, as in DUDASH, as modified by MARTIN. A person of ordinary skill would have been motivated to improve performance (ROSSI [0002])
MARTIN further teaches:
receiving from the second trained graph neural network predicted variational parameters at an output of the second trained graph neural network, ((Page 7, paragraph 2)We perform the benchmarks as follows. For graphs with up to a few hundred nodes, we compare our GNN-based solver to the (approximate) polynomial time Goemans-Williamson (GW) algorithm [81], which provides the current record for an approximate answer within some fixed multiplicative factor of the optimum (referred to as approximation ratio α), using semidefinite programming and randomized rounding. Specifically, the GW algorithm achieves a guaranteed approximation ratio of α ∼ 0.878 for generic graphs. This lower bound can be raised for specific graphs such as unweighted 3-regular graphs where α ∼ 0.9326 )
DUDASH further teaches:
and providing the predicted variational parameters and the QUBO problem to the variational quantum solver for determining a candidate solution for the problem with the variational quantum solver. ([0006] and updating the best local current graph, wherein the updating comprises: copying the best local current graph to form a best local current graph copy, modifying the best local current graph copy to form a candidate local graph, computing an evaluation rating for the candidate local graph, and determining, based on the evaluation rating for the candidate local graph and an evaluation rating for the best local current graph, whether one or more replacement criteria are met; in accordance with a determination that the one or more replacement criteria are met, replacing the best local current graph with the candidate local graph; storing updated best local current graphs associated respectively with each of the one or more GPU thread blocks in a local results array; identifying an updated best local current graph in the local results array as the best global graph; and configuring the quantum annealer based on the best global graph.)
Regarding claim 14, it comprises of limitations similar to those of claim 8 and is therefore rejected for similar rationale.
Conclusion
Any inquiry concerning this communication or earlier communications from the examiner should be directed to SKIELER A KOWALIK whose telephone number is (571)272-1850. The examiner can normally be reached 8-5.
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, Mariela D Reyes can be reached at (571)270-1006. 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.
/SKIELER ALEXANDER KOWALIK/Examiner, Art Unit 2142 /Mariela Reyes/Supervisory Patent Examiner, Art Unit 2142
Prosecution Insights