Numerous quantum algorithms assume that classical data has already been converted into quantum states, a process known as Quantum State Preparation (QSP). However, achieving precise QSP requires a circuit depth that scales exponentially with the number of qubits, posing a significant challenge to realizing quantum advantage. Recent research explores Parameterized Quantum Circuits (PQCs) as an approximate alternative, offering improved scalability with reduced circuit depth. However, the iterative, state-by-state optimization required by this approach creates substantial runtime overhead, which severely limits its practicality.To improve the efficiency of approximate QSP, we introduce a novel two-stage framework that can potentially generate QSP circuits for arbitrary quantum states. In theoffline training stage, our model learns a direct mapping from target states to circuit parameters, thereby bypassing the need foronline, state-by-state optimizationduring theinference stage. Extensive evaluations show that our approach significantly reduces runtime overhead by up to 132×, making a steady step towards efficient neural approximate QSP.
Numerous quantum algorithms assume that classical data has already been converted into quantum states, a process known as Quantum State Preparation (QSP). However, achieving precise QSP requires a circuit depth that scales exponentially with the number of qubits, posing a significant challenge to realizing quantum advantage. Recent research explores Parameterized Quantum Circuits (PQCs) as an approximate alternative, offering improved scalability with reduced circuit depth. However, the iterative, state-by-state optimization required by this approach creates substantial runtime overhead, which severely limits its practicality.To improve the efficiency of approximate QSP, we introduce a novel two-stage framework that can potentially generate QSP circuits for arbitrary quantum states. In theoffline training stage, our model learns a direct mapping from target states to circuit parameters, thereby bypassing the need foronline, state-by-state optimizationduring theinference stage. Extensive evaluations show that our approach significantly reduces runtime overhead by up to 132×, making a steady step towards efficient neural approximate QSP.
Numerous quantum algorithms operate under the assumption that classical data has already been converted into quantum states, a process termed Quantum State Preparation (QSP). However, achieving precise QSP requires a circuit depth that scales exponentially with the number of qubits, making it a substantial obstacle in harnessing quantum advantage. Recent research suggests using a Parameterized Quantum Circuit (PQC) to approximate a target state, offering a more scalable solution with reduced circuit depth compared to precise QSP. Despite this, the need for iterative updates of circuit parameters results in a lengthy runtime, limiting its practical application. In this work, we demonstrate that it is possible to leverage a pre-trained neural network to directly generate the QSP circuit for arbitrary quantum state, thereby eliminating the significant overhead of online iterations. Our study makes a steady step towards a universal neural designer for approximate QSP.
A central challenge in quantum machine learning is the state preparation bottleneck that describes the prohibitive computational cost of loading high-dimensional classical data into a quantum state. Although amplitude encoding can represent 2n-dimensional data using only n qubits in principle, preparing arbitrary states remains computationally expensive, typically requiring variational optimization of a parameterized quantum circuit for each individual data instance. In this work, we propose a method that avoids iterative optimization by training a classical neural network to map input data directly to the continuous parameters of a fixed quantum circuit. We demonstrate the generation of quantum image states with high fidelity on data not seen during training. Since all optimization is performed once during training, the resulting model encodes new inputs in a single inference step, providing a scalable pathway for data loading in near-term quantum algorithms. We validate our method on the MNIST and Fashion-MNIST datasets, achieving fidelities up to 0.992 on unseen images and reducing the per-data-instance runtime by more than 5000-fold.
Recent assertions of a potential advantage of Quantum Neural Network (QNN) for specific Machine Learning (ML) tasks have sparked the curiosity of a sizable number of application researchers. The parameterized quantum circuit (PQC), a major building block of a QNN, consists of several layers of single-qubit rotations and multi-qubit entanglement operations. The optimum number of PQC layers for a particular ML task is generally unknown. A larger network often provides better performance in noiseless simulations. However, it may perform poorly on hardware compared to a shallower network. Because the amount of noise varies amongst quantum devices, the optimal depth of PQC can vary significantly. Additionally, the gates chosen for the PQC may be suitable for one type of hardware but not for another due to compilation overhead. This makes it difficult to generalize a QNN design to wide range of hardware and noise levels. An alternate approach is to build and train multiple QNN models targeted for each hardware which can be expensive. To circumvent these issues, we introduce the concept of knowledge distillation in QNN using approximate synthesis. The proposed approach will create a new QNN network with (i) a reduced number of layers or (ii) a different gate set without having to train it from scratch. Training the new network for a few epochs can compensate for the loss caused by approximation error. Through empirical analysis, we demonstrate ≈71.4% reduction in circuit layers, and still achieve ≈16.2% better accuracy under noise.
There is increasing interest in the area of quantum computing, but using and developing new quantum algorithms, often represented as quantum circuits, is arduous and requires designers with a wide area of experience. The exploration of techniques for the automatic design of quantum algorithms, i.e., quantum architecture search, is therefore of high importance. This work explores the use of neural networks for quantum architecture search, since neural networks have proven quite successful in various complex search domains, and current approaches in quantum architecture search make very restricted use of them. There are many different domains in which quantum architecture search can be applied. They can be used to compile unitaries into quantum circuits, develop ansätze for variational quantum circuits, or develop new quantum algorithms by applying it to input output pairs of one task one wishes to solve. The focus in this thesis is on the task of quantum state preparation, a very important task in the field of quantum computing, as it is a subroutine in a large number of quantum algorithms like quantum machine learning and Hamiltonian simulation. This thesis approaches the task of quantum architecture search first by establishing the importance of the structure of quantum circuits. This is done with a large-scale study of the influence different circuit structure parameters, such as qubits and depth, have on the performance on variational quantum machine learning. During this research, four guidelines for designing circuits for variational quantum machine learning were established. DeepQPrep, a neural network-augmented deductive search for the automatic design of quantum circuits for quantum state preparation, was proposed. It is able to create non-parameterized quantum circuits to solve 76.9% of a restrictive set of quantum states with quantum circuits that were much shallower than the ones found in the ground-truth data. The model was trained on a dataset created in this thesis as none were available at the time. During the exploration of extensions of DeepQPrep to enable it to solve more arbitrary quantum state preparation problems without the need for labeled data, different challenges became apparent. A study on the difficulty of training neural networks to create parameterized quantum circuits using reinforcement learning found that training is hampered by the barren plateau problem of quantum computing. It shows that the variance of the gradient of the quantum circuit output with respect to its parameters goes exponentially to zero. A study showed that this problem makes the application of neural networks to encode quantum circuits highly challenging when no intermediate quantum state information is used. This culminated in DeepQPrepGrad, which combines parameterized circuits, the inductive search of DeepQPrep, and the reinforcement learning-based approach with more compact representations. By combining these parts with the gradients of quantum circuits, an approach that is able to create parameterized quantum circuits for arbitrary quantum states was obtained. The approach was evaluated on 20000 arbitrary 3-qubit quantum states without the need for any labeled data, resulting in a mean minimal similarity of 0.97 alleviated the found challenges to a large degree. This thesis explored key challenges faced when training neural networks to encode quantum circuits. These challenges are of paramount importance as they potentially affect a wide range of tasks in the field of quantum computing to which neural networks could be applied. Lastly, DeepQPrepGrad provides a significant step forward for the creation of quantum circuits using neural networks. DeepQPrepGrad makes great progress in solving the challenges highlighted above and has the potential to be applicable in a wide range of quantum architecture search tasks, such as variational eigensolver and variational quantum machine learning.
We extend directed quantum circuit synthesis (DQCS) with reinforcement learning from purely discrete gate selection to parameterized quantum state preparation with continuous single-qubit rotations Rx, Ry, and Rz. We compare two training regimes: a one-stage agent that jointly selects the gate type, the affected qubit(s), and the rotation angle; and a two-stage variant that first proposes a discrete circuit and subsequently optimizes the rotation angles with Adam using parameter-shift gradients. Using Gymnasium and PennyLane, we evaluate Proximal Policy Optimization (PPO) and Advantage Actor--Critic (A2C) on systems comprising two to ten qubits and on targets of increasing complexity with λ ranging from one to five. Whereas A2C does not learn effective policies in this setting, PPO succeeds under stable hyperparameters (one-stage: learning rate approximately 5×10−4 with a self-fidelity-error threshold of 0.01; two-stage: learning rate approximately 10−4). Both approaches reliably reconstruct computational basis states (between 83\% and 99\% success) and Bell states (between 61\% and 77\% success). However, scalability saturates for λ of approximately three to four and does not extend to ten-qubit targets even at λ=2. The two-stage method offers only marginal accuracy gains while requiring around three times the runtime. For practicality under a fixed compute budget, we therefore recommend the one-stage PPO policy, provide explicit synthesized circuits, and contrast with a classical variational baseline to outline avenues for improved scalability.
Quantum machine learning (QML) has demonstrated the potential to outperform classical machine learning algorithms in various fields. However, encoding classical data into quantum states, known as quantum data loading, remains a challenge. Existing methods achieve high accuracy in loading single data, but lack efficiency for large-scale data loading tasks. In this work, we propose Neural Circuit Parameter Prediction, a novel method that leverages classical deep neural networks to predict the parameters of parameterized quantum circuits directly from the input data. This approach benefits from the batch inference capability of neural networks and improves the accuracy of quantum data loading. We introduce real-valued parameterization of quantum circuits and a three-phase training strategy to further enhance training efficiency and accuracy. Experimental results on MNIST dataset show that our method achieves a 17.31 % improvement in infidelity score and 108 times faster runtime compared to existing methods. Our approach provides an efficient solution for quantum data loading, enabling the practical deployment of QML algorithms on large-scale datasets.