CTFR 18/889,486 CTFR 82222 DETAILED ACTION Notice of Pre-AIA or AIA Status 07-03-aia AIA 15-10-aia The present application, filed on or after March 16, 2013, is being examined under the first inventor to file provisions of the AIA. Response to Arguments Applicant’s arguments with respect to claim(s) 1-20 have been considered but are moot in view of the new ground of rejection. Claim Rejections - 35 USC § 103 07-06 AIA 15-10-15 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. 07-20-aia AIA 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. 07-23-aia AIA 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. 07-21-aia AIA Claim (s) 1-20 are rejected under 35 U.S.C. 103 as being unpatentable over Park et al ., “Circuit-Based Quantum Random Access Memory for Classical Data,” and further in view of Havlicek et al. , “Supervised learning with quantum-enhanced feature spaces” (hereinafter “Havlicek”) . Claim 1: Park teaches a system comprising: a memory configured to store a set of data packets, and a processor, operably coupled to the memory (e.g. Park teaches a quantum random access memory (QRAM) architecture, which is a system comprising a memory for storing classical data entries ("a set of data packets") and a quantum circuit-based processor (the "flip-flop" QRAM circuit) that is operably coupled to and acts upon that memory to create quantum states (see Page 1: "a circuit-based flip-flop quantum random access memory to construct a quantum database of classical information")), and configured to: receive the set of data packets (e.g. Park's method involves registering or updating a set of M classical data entries. This prerequisite step of receiving/inputting the data is explicitly described: "For registering or updating classical data consisting of M entries..." (Page 1). The process operates on a given input dataset); convert each of the set of data packets into a respective quantum state array, wherein the respective quantum state array indicates a value of each quantum bit associated with a respective binary bit from among the set of data packets (e.g. The FF-QRAM circuit converts classical bit strings (\vec{d}^{(l)}, e.g., 1101) into quantum computational basis states (|\vec{d}^{(l)}\rangle, e.g., |1101 ⟩ ). The method "superpose[s] them in the computational basis states" (Page 1), and the resulting quantum state directly encodes the value of each classical bit into a specific qubit's basis state (|0 ⟩ or |1 ⟩ ). (See Page 2, Eq. 1: |\psi \rangle_{\mathrm{QDB}} = \sum_{l = 0}^{M - 1}b_l|\vec{d}^{(l)}\rangle and the description of \vec{d}^{(l)} as a string of quantum bits)). generate a unified buffer array associated with the set of data packets (e.g. Park's process results in a single, unified quantum state known as a quantum database (QDB). This QDB is a superposition state that aggregates all the individually converted data packets. This is explicitly stated as the output of the QRAM operation: \mathrm{QRAM}(D)\sum_{j}\psi_{j}|j\rangle_{B}|0\rangle_{R}\equiv\sum_{l}\psi_{l}|\vec{d}^{(l)}\rangle_{B}|b_{l}\rangle_{R} (Page 2, Eq. 2). This summed state is the "unified buffer array."); validate that the unified buffer array corresponds to a quantum representation of the set of data packets, wherein validating that the unified buffer array corresponds to the quantum representation of the set of data packets comprises: comparing the unified buffer array with a vector that comprises the set of data packets (e.g., page 3, post-selection validation: “Finally, the queried QDB derived from Eq. (1) can be obtained by selecting an appropriate angle \theta^{(l)} to match the desired probability amplitude b_l, and **post-selecting the measurement outcome); and determining that the unified buffer array corresponds to the vector based at least in part upon the comparison (e.g., page 3, post-selection validation: “Finally, the queried QDB derived from Eq. (1) can be obtained by selecting an appropriate angle \theta^{(l)} to match the desired probability amplitude b_l, and **post-selecting the measurement outcome). Not explicitly taught by Park is that converting each of the set of data packets into the respective quantum state array comprises implementing superposition and entanglement properties of quantum bits in each quantum state array. However, Havlíček teaches mapping classical data into quantum states through a quantum feature-map circuit. Specifically, Havlíček teaches that: “A core element in the computational speed-ups enabled by quantum algorithms is the exploitation of an exponentially large quantum state space through controllable entanglement and interference” (p. 209, left column). Havlíček further teaches that classical data is mapped into quantum states by a feature-map circuit: “The data is mapped non-linearly to a quantum state Φ(x)” (p. 209, right column; Fig. 1a). Havlíček teaches a feature-map circuit employing Hadamard operations to generate superposition and multi-qubit interaction terms that generate entanglement between qubits during encoding of the classical data into quantum states (Fig. 1b; p. 209; Fig. 2c, p. 210). Havlíček further explains that circuits producing only product states are insufficient and that quantum advantage is obtained through feature maps exploiting entanglement and interference (p. 209). Therefore, Havlíček expressly teaches implementing superposition and entanglement properties during the conversion of classical data into quantum-state representations. It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention, to modify Park ’s quantum-data encoding system to employ Havlíček ’s feature-map encoding techniques because Havlíček teaches that encoding classical information into quantum states using controllable entanglement and superposition provides richer and more expressive quantum representations of the underlying data and improves subsequent quantum information processing tasks. A person of ordinary skill in the art would have recognized that incorporating Havlíček ’s known entangling quantum-state preparation techniques into Park ’s QRAM-based data-encoding architecture would have predictably improved the quality and expressiveness of the generated quantum-state representations while preserving Park ’s ability to store and process classical data in quantum form. Accordingly, it would have been obvious to modify Park in view of Havlíček to arrive at the claimed invention. As per claims 8 and 15, the claimed features are rejected similarly to claim 1 above. Claim 2: Park and Havlíček teach the system of Claim 1, wherein: the set of data packets is associated with an application programming interface (API) request to perform a set of tasks; and the set of data packets comprises a first data packet associated with a first task and a second data packet associated with a second task. For instance, Park 's entire disclosure is based on encoding classical data entries consisting of n bits (e.g., n-bit strings). The quantum database (QDB) is constructed from classical bit strings \vec{d}^{(l)} = d_0^{(l)}d_1^{(l)}\dots d_{n-1}^{(l)} \in \{0,1\}^{n} (Page 2, Eq. 1). This expressly teaches that the input data packets are bit strings. As per claims 9 and 16, the claimed features are rejected similarly to claim 2 above. Claim 3: Park and Havlíček teach the system of Claim 2, wherein converting each of the set of data packets into the respective quantum state array comprises: converting the first data packet into a first quantum state array; converting the second data packet into a second quantum state array; and initializing a qubit for each binary bit from among the set of data packets, wherein each binary bit 0 is converted into a qubit |0> and each binary bit 1 is converted into a qubit |1>. For instance, Park teaches that the FF-QRAM circuit (Fig. 1) explicitly starts with the register qubit in the |0> state (|0 ⟩ _R). Furthermore, the process of rearranging computational basis states using classically-controlled X gates (ΞX) inherently involves initializing and flipping qubits to specific basis states (|0> or |1>) as part of the encoding process (Page 3, Eqs. 3-4). This is a fundamental and necessary step in any quantum circuit-based state preparation. As per claims 10 and 17, the claimed features are rejected similarly to claim 3 above. Claim 4: Park and Havlíček teach the system of Claim 3, wherein generating a unified buffer array associated with a quantum representation of the set of data packets comprises: determining which quantum state arrays have a corresponding length and position in a vector space; determining that the first quantum state array has the corresponding length and position in the vector space as the second quantum state array, indicating that the first quantum state array carries corresponding qubits as the second quantum state array; pairing the first quantum state array with the second quantum state array in response to determining that the first quantum state array has the corresponding length and position in the vector space as the second quantum state array; and populating the unified buffer array with the paired first quantum state array and the second quantum state array. For instance, Park 's method uses controlled rotation gates, C^n R_p(\theta), which perform a rotation by angle θ around the p-axis of the Bloch sphere (Page 2). The axis p is a design choice (e.g., the y-axis rotation R_y is used in the quantum support vector machine example, Page 4). Specifying alignment along a particular axis is therefore an inherent and disclosed aspect of the state preparation circuit. As per claims 11 and 18, the claimed features are rejected similarly to claim 4 above. Claim 5: Park and Havlíček teach the system of Claim 4, but fail to teach that validating that the unified buffer array corresponds to the quantum representation of the set of data packets is in response to: performing a convolution operation between the unified buffer array and the vector; and determining that the convolution operation results in zero or less than a threshold percentage difference between the unified buffer and the vector. Havlíček , however, teaches comparing a quantum state with a classical vector using a kernel method that is mathematically equivalent to a convolution-like operation in the feature space. For instance, Havlicek teaches, in page 1: “The data is mapped non-linearly to a quantum state … the quantum state space as feature space … inner products K(\mathbf{x},\mathbf{z}) = |\langle\Phi(\mathbf{x})|\Phi(\mathbf{z})\rangle|^2.” In machine learning, convolution is a specific type of kernel operation. Therefore, a person of ordinary skill in the art (PHOSITA) would recognize that comparing a quantum representation to a classical vector can be done by evaluating a similarity measure (e.g., fidelity, inner product) and checking whether the difference is below a threshold. Havlicek explicitly uses a decision rule based on a threshold: “assign the label \hat{m}(x)=y whenever \hat{p}_y(x) > \hat{p}_{-y}(x) - yb” (page 2). That is a threshold-based comparison. Using a convolution operation (a linear kernel) to compute similarity and then comparing the result to zero or a threshold is a well-known, obvious alternative to the post-selection method of Park. Therefore, It would have been obvious to a PHOSITA to validate the quantum representation using a convolution (or any inner-product-based similarity measure) followed by threshold comparison, as taught by Havlicek. As per claims 12 and 19, the claimed features are rejected similarly to claim 5 above. Claim 6: Park and Havlíček teach the system of Claim 4, but fail to teach that validating that the unified buffer array corresponds to the quantum representation of the set of data packets is in response to: performing an inverse matrix multiplication between the unified buffer array and the vector; and determining that the inverse matrix multiplication results in an identity unit matrix. However, Havlicek teaches the kernel matrix approach: estimating the kernel matrix K for training data and using it to find support vectors. A PHOSITA would understand that if the quantum state representation is correct, then the kernel matrix (inner products between quantum states) should be consistent with the classical data. One way to check consistency is to compute K^{-1}K and verify that the result approximates the identity matrix. This is a standard linear algebra validation technique. The step of performing inverse matrix multiplication to obtain an identity matrix is a routine mathematical check for correctness of a representation. Therefore, a PHOSITA would have found it obvious to validate the unified buffer array by computing the matrix of inner products between the quantum state and the classical vector, inverting that matrix, and checking for identity, which is a standard consistency check in linear algebra and quantum information (e.g., verifying that the Gram matrix is well-behaved). As per claims 13 and 20, the claimed features are rejected similarly to claim 6 above. Claim 7: Park and Havlíček teach the system of Claim 4, but fail to teach that the processor is further configured to determine that the first data packet and the second data packet are associated with a corresponding task in response to determining that the first quantum state array has the corresponding length and position in the vector space as the second quantum state array. However, Park's express purpose for creating the quantum database (QDB) is to serve as input for a specific quantum algorithm or computational task (e.g., a quantum support vector machine, Page 4). And Havlicek explicitly teaches classifying data based on the position of quantum states in feature space – i.e., if two quantum states are close (have similar length and position), they receive the same label (task). Havlicek, page 2: “The mapped data can be separated by the hyperplane … States with a positive expectation value … receive a [+1] label.” That is, identical or nearby positions in the vector space correspond to the same classification task. A PHOSITA would immediately recognize that if two quantum state arrays have the same length and position (i.e., are identical or nearly identical), then the underlying data packets are associated with the same task. Therefore, it would have been obvious to determine task correspondence by comparing the length and position of quantum state arrays in the vector space, as taught by Havlicek’s classification decision rule. As per claim 14, the claimed features are rejected similarly to claim 7 above. Conclusion 07-40 AIA 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 GUERRIER MERANT whose telephone number is (571)270-1066. The examiner can normally be reached Monday-Friday 8:00 Am - 5:00 PM. 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, Mark Featherstone can be reached at 571-270-3750. 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. /GUERRIER MERANT/Primary Examiner, Art Unit 2111 6/10/2026 Application/Control Number: 18/889,486 Page 2 Art Unit: 2111 Application/Control Number: 18/889,486 Page 3 Art Unit: 2111 Application/Control Number: 18/889,486 Page 4 Art Unit: 2111 Application/Control Number: 18/889,486 Page 5 Art Unit: 2111 Application/Control Number: 18/889,486 Page 6 Art Unit: 2111 Application/Control Number: 18/889,486 Page 7 Art Unit: 2111 Application/Control Number: 18/889,486 Page 8 Art Unit: 2111 Application/Control Number: 18/889,486 Page 9 Art Unit: 2111 Application/Control Number: 18/889,486 Page 10 Art Unit: 2111 Application/Control Number: 18/889,486 Page 11 Art Unit: 2111 Application/Control Number: 18/889,486 Page 12 Art Unit: 2111
Prosecution Insights