Selectively Blind Quantum Computation

Abbas Poshtvan, Oleksandra Lapiha, Mina Doosti, Dominik Leichtle, Luka Music, Elham Kashefi · 2025

Read Selectively Blind Quantum Computation by Abbas Poshtvan, Oleksandra Lapiha, Mina Doosti, Dominik Leichtle, Luka Music, Elham Kashefi (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2504-17612.txt.

Introduction

Abbas Poshtvan and co-authors study how a client can hand a quantum computing task to a server without giving the whole task away. Known safe methods need the client to send quantum states, and the team ask whether that traffic can be cut down.

First they prove a hard limit: tricks done on the server side alone cannot cut the quantum traffic in a safe scheme. Then they give a new scheme that hides just one task among a known set of tasks. How alike the tasks in the set are sets the traffic needed. The count is the best of its kind, and it trades a bit of leaked data for far less traffic.

Catalog record

CREATORAbbas Poshtvan, Oleksandra Lapiha, Mina Doosti, Dominik Leichtle, Luka Music, Elham Kashefi
YEAR2025
COLLECTIONPAPERS
FORMATTXT
FILE SIZE117 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2504-17612

Browse the full library · About the Tower of Babel

Full text

The opening of the work, from the library's hosted text:

Selectively Blind Quantum Computation Abbas Poshtvan, Oleksandra Lapiha, Mina Doosti, Dominik Leichtle, Luka Music, Elham Kashefi Abstract Known protocols for the secure delegation of quantum computations from a client to a server in an information-theoretic setting require quantum communication. In this work, we investigate methods to reduce the communication overhead. First, we establish an impossibility result by proving that server-side local processes cannot decrease the quantum communication requirements of secure delegation protocols. We develop a series of no-go results that prohibit such a process within an information-theoretic framework. Second, we present a possibility result by introducing Selectively Blind Quantum Computing (SBQC), a novel functionality which allows the client to hide one among a known set of possible computations. We precisely characterise how the differences between the computations in the protected set influence the number of qubits sent during our SBQC implementation, therefore yielding a communication-optimal protocol. This approach can reduce qubit communication drastically and demonstrates the trade-off between information leaked to the server and the protocol’s communication cost. 1 Introduction 1.1 Context Currently, in order to delegate a quantum computation to a quantum service provider, a client sends over the input and the computation they want to perform, and the quantum server is supposed to do the computation and send back the outcome. For some important computations, the client may want the following privacy and security guarantees: the input, computation, and the output should stay unknown to the server (blindness); furthermore, the client should be able to check that the computation has been performed correctly (verifiability). Delegating quantum computations in this secure way to a powerful server is a crucial step toward making quantum computing widely accessible and establishing a secure quantum cloud. Hence, improving the efficiency and practicality of these functionalities together with the fundamental limitations on them has been a long-lasting topic of interest in quantum cryptography, complexity theory and physics [ 1 , 2 , 3 , 4 , 5 , 6 , 7 , 8 , 9 , 10 , 11 , 12 ] . In this manuscript, we focus on the blindness property only. The Universal Blind Quantum Computing (UBQC) protocol proposed by [ 1 ] achieves this functionality provided that the client can generate random single qubits and send them over a quantum channel to the server. However, the number of qubits that must be sent in this protocol (i.e. its communication complexity) is linear in the size of the circuit. From a practical point of view, the flying qubit from client has to be integrated into the remote quantum computer of the server. In the process, it might be affected by the noise of the channel established between the parties, which results in a reduction of the computation performance on the server’s side. The main question addressed in this paper is whether techniques can be used to reduce this communication complexity and also establish lower bounds on the number of encrypted qubits necessary to achieve blindness. There are three existing approaches for constructing the blind quantum computation functionality, each using a specific resource: 1. Blind quantum computing with Prepare-and-Send [ 1 ] . This approach has been formulated in the measurement based quantum computation (MBQC) model [ 13 ] . Here, the client is able to apply random single qubit rotations and sends multiple encrypted qubits to the server (where the keys are the rotation angles). The server uses them to construct an encrypted graph state. The computation is performed interactively with the client sending measurement instructions and the server returning the measurement outcomes for subsets of the graph’s qubits. At the end, the server sends back the output qubits to the client. The main resource for this variant of UBQC is the ability to prepare single qubits on the server’s machine without the server knowing the state of these qubits, also called secure remote state preparation (RSP). This is the main variant of UBQC we will be considering throughout this paper, and will discuss it in more details later. A closely related alternative approach are Receive-and-Measure protocols [ 14 , 3 ] , in which the server prepares the computation graph state and sends it qubit-by-qubit to the client, who has the ability of measuring these qubits in a certain number of bases. 2. Blind quantum computing with non-communicating entangled servers [ 6 ] . This device-independent approach requires two non-communicating servers to share maximally entangled states with each other prior to the protocol’s execution. Then the client instructs one of the servers to measure their part of this resource state and send back the outcome without the other server learning it. This information asymmetry is leveraged by the client to both perform the computation and test the servers’ honesty by utilizing the rigidity of non local games [ 15 ] . However, sharing high amounts of entanglement between distant servers and making sure that the severs are not communicating (e.g. by enforcing space-like separations so the parties can not have causal effects on each other) are the downsides of this approach. 3. Blind quantum computing with computational hardness assumptions and classical client [ 16 ] . The idea behind this approach is to perform secure remote state preparations using only classical public-key cryptography [ 5 , 17 ] , in particular Trapdoor Claw-Free Functions (TCF) based only the Learning with Errors problem. In this approach, quantum communication is removed entirely, at the cost of having to implement the TCFs complex and deep circuits on the server’s quantum hardware. Compared to the other approaches which are information-theoretically secure, these protocols provide only computational security. Among these approaches, the first one seems closer to implementation, as it has already been demonstrated in the lab [ 9 ] . From a practical perspective, the main obstacle here is the quantum communication cost which is equal to the size of the graph state implemented on the server (i.e. proportional to the number of gates in the circuit that the client wants to run). Reducing the communication cost and computing its fundamental lower bounds is an important step toward making this approach more feasible. Apart from the experimental significance of achieving UBQC with reduced qubit communication, the possibility of accomplishing this with minimal resources has important implications for complexity theory. The first question in this direction is whether it is possible to make the client fully classical while preserving information-theoretic security. As discussed, this can be done with computational security using TCFs. However, achieving blindness with a fully classical client in an information-theoretic context is unlikely, as it would lead to unlikely collapses in complexity classes. As shown in [ 18 ] , delegating Boson Sampling in such a way would imply the collapse of the polynomial hierarchy to the third level. This relates to the foundational question of verifying BQP computations with BPP verifiers, which remains an open question in the field. Therefore, studying the communication requirements of the [ 1 ] protocol can shed light into possible answers to this question. An important question in that direction is: Can we construct more efficient RSPs with less quantum communication? This question was positively addressed in the Random Oracle Model [ 19 ] . They demonstrate

Read the complete text