arXiv:2610.04995
Blind Remote Implementation of Arbitrary Two-Qubit Unitary Operation via Pre-Shared Entanglement
Plaban Saha, Binayak S. Choudhury
Abstract
We propose a protocol for the remote implementation of an arbitrary two-qubit unitary operation using pre-shared entanglement. In our protocol, Bob possesses an unknown two-qubit state, while Alice implements the required operations using calibrated black-box devices. Alice does not need to know the parameters of these operations, and Bob's unknown state never leaves his laboratory. The protocol is based on the Euler decomposition of single-qubit unitary operations and the Cartan decomposition of two-qubit unitary operations. First, we construct protocols for the remote implementation of the single-qubit operation $U = R_z(γ)R_x(β)R_z(α)$ using pre-shared Bell states. We then construct a protocol for the two-qubit operation $R_{zz}(θ)$ using a single pre-shared Bell state and extend it to $R_{xx}(θ)$ and $R_{yy}(θ)$ through local unitary transformations. These elementary protocols are finally combined to implement an arbitrary two-qubit unitary operation. A third party, Charlie, determines the required decomposition and provides Alice with the corresponding black-box devices before the protocol begins. Thus, neither Alice nor Bob requires the classical description of the unitary operation.
AI Summary
The paper considers remote implementation of an arbitrary two-qubit unitary on an unknown state held by Bob, using shared entanglement and operations available at Alice’s laboratory. Its information-access model includes a third party, Charlie, who computes the unitary’s decomposition and supplies calibrated black-box devices for its elementary gates. Neither Alice nor Bob needs the unitary’s classical description; Bob’s original target qubits remain in his laboratory. The construction combines two explicit protocols. An arbitrary single-qubit gate $R_z(\gamma)R_x(\beta)R_z(\alpha)$ is implemented using two Bell pairs, local ancillary operations, measurements, and five classical bits. A two-qubit rotation $R_{zz}(\theta)$ uses one Bell pair and two classical bits: Bob encodes the target-qubit parity, Alice applies her black-box rotation with an ancillary qubit initialized to $|0\rangle$, and measurement-dependent corrections complete the operation. Fixed local conjugations extend this primitive to $R_{xx}$ and $R_{yy}$. The Cartan decomposition then assembles four arbitrary single-qubit blocks and three two-qubit rotation blocks into an arbitrary $U\in U(4)$, up to global phase. The resulting resource count is eleven Bell pairs and twenty-six classical bits. Correctness is established through explicit state transformations and composition of the elementary protocols. The stated setting requires Charlie’s prior decomposition and access to separate calibrated devices for the constituent operations, rather than access only to an undecomposed black box for $U$. “Blind” here denotes the absence of a required classical gate description; the paper does not provide a cryptographic blindness definition or security proof.
Central claim may not hold: After Step 5 of the single-qubit protocol, Alice can recover Bob's entire unknown input state using only a local CNOT, so retaining Bob's physical target qubit does not establish that his quantum state never leaves his laboratory. The claimed universal Cartan decomposition incorrectly restricts all three coordinates to [0,π/4], excluding some local-equivalence classes, although permitting a signed coordinate repairs the construction.
Written by gpt-6.1-sol from the paper's full text. Automated: it can be wrong.