---
title: Dressed Quantum Network (DQN)
url: https://www.emergentmind.com/topics/dressed-quantum-network-dqn
type: topic
---

# Dressed Quantum Network (DQN)

The Dressed Quantum Network (DQN) is a hybrid classical–quantum variational classifier designed to address the limitations of variational Quantum Machine Learning (QML) on Noisy Intermediate-Scale Quantum (NISQ) hardware, specifically by reducing qubit requirements and minimizing noise sources. The DQN framework comprises a classical "super compressed encoding" layer followed by a compact quantum circuit composed exclusively of single-qubit gates and projective measurements. This architecture enables resource-efficient embedding of high-dimensional data into a minimal qubit register and leverages robust state preparation and variational transformations for effective supervised classification [2007.10242].

## 1. Architecture and Data Flow

The DQN consists of two core components:

- **Classical encoding layer:** Implements a fully connected transformation \( W\in\mathbb{R}^{d\times N}\), mapping the input vector \(\mathbf{x}\in\mathbb{R}^d\) to an \(N\)-dimensional vector \(\widetilde{\mathbf{x}}\in\mathbb{R}^N\), where \(N\) is the number of target classes. The encoding is given by
  \[
  \widetilde x_j = \sum_{i=1}^d w_i^j x_i\,,\quad j=1,\ldots,N.
  \]

- **Quantum variational circuit:** For each of the \(N\) qubits,
  1. State preparation is performed by applying a \(R_z(\widetilde x_j)\) rotation followed by a Hadamard gate:
     \[
     |\psi(\mathbf{x})\rangle = \bigotimes_{j=1}^N e^{i\sigma_3\widetilde x_j} H |0\rangle^j
     \]
  2. Variational SU(2) transformation with parameters \(\boldsymbol\alpha^j = (\alpha_1^j, \alpha_2^j, \alpha_3^j)\):
     \[
     U^j = e^{i\sigma_3\alpha_1^j}e^{i\sigma_2\alpha_2^j}e^{i\sigma_3\alpha_3^j}
     \]
  3. Projective measurement onto the \(\sigma_3=+1\) eigenstate yields the classwise probabilities \(P^j\).

This structure decouples input dimension \(d\) from qubit number \(N\) and circumvents the need for multi-qubit gates, a significant source of decoherence and operational infidelity in NISQ devices.

## 2. Super Compressed Encoding

A distinguishing feature of DQN is the "super compressed encoding," which maps high-dimensional input \(\mathbf{x}\) to an \(N\)-tuple \(\widetilde{\mathbf{x}}\), with \(N\ll d\) in typical applications. This transformation both reduces the dimensionality and encodes data for direct quantum state preparation:
- Each component \(\widetilde x_j\) is utilized as the rotation angle in the \(R_z\) operation for qubit \(j\).
- The resulting compression allows implementations where the number of qubits required equals only the number of output classes, independently of the sample's original feature dimension.
- This aggressive reduction mitigates the hardware constraints of current quantum processors and enables scalable QML solutions on NISQ devices [2007.10242].

## 3. Quantum Circuit Construction and Parameterization

The quantum portion of DQN comprises only single-qubit gates for each of the \(N\) qubits, structured as follows:
- **State preparation:** Each qubit is initialized by an \(R_z(\widetilde x_j)\) followed by a Hadamard.
- **Variational layer:** SU(2) gate sequence per qubit, decomposed as \(R_z(\alpha_1^j) R_y(\alpha_2^j) R_z(\alpha_3^j)\).
- **Circuit depth:** Per qubit, the sequence involves a single axial rotation, a Hadamard, and three more rotations.
- **Parameters:** The total quantum parameter count is \(3N\), with all parameters being independent and trainable during circuit optimization.
- **Measurement:** Each qubit is measured in the computational basis, with class probabilities inferred from the resulting statistics [2007.10242].

No two-qubit operations are present, which drastically reduces susceptibility to correlated errors and crosstalk.

## 4. Training Protocol and Optimization

DQN is trained using a standard supervised learning paradigm:
- Loss function: Multiclass tasks utilize cross-entropy loss:
  \[
  \mathcal{E}_{\rm CE} = -\sum_{s=1}^N P_{\rm target}^{(k)s} \ln[\mathrm{SoftMax}(\mathbf{P})^s]
  \]
  For binary classification, a linear loss function is applied:
  \[
  \mathcal{E} = \sum_{p=1}^{m_1}(1-P_+^p) + \sum_{q=1}^{m_2}[1-(1-P_+^q)]
  \]
- Optimization: Classical optimizers such as Adam or RMS-prop are employed. Gradients with respect to all weights (\(w_i^j\)) and quantum angles (\(\alpha_r^j\)) are computed via finite differences or automatic differentiation using PennyLane.
- No explicit regularization is used, but the restriction to single-qubit operations acts as a natural regularizer against circuit noise.
- The optimization loop iterates over the training set, accumulating loss, computing parameter gradients, and updating both classical and quantum parameters until convergence [2007.10242].

## 5. Experimental Results and Hardware Implementation

The DQN was benchmarked on multiple datasets using diverse platforms:

| Dataset            | Classes | Python Sim. | Qiskit Sim. | IBM-Q       |
|--------------------|---------|-------------|-------------|-------------|
| Fisher's Iris (d=4)| 3       | 90%         | 94%         | 82%         |
| WBC (d=30)         | 2       | 92.4%       | 96.5%       | 91.7%       |
| Abalone (d=8)      | 6       | 67.7%       | 67.4%       | 67.2%       |

- Hardware platforms: Python (noise-free), Qiskit with PennyLane, and real IBM-Q devices (Rome, Armonk, Melbourne).
- Only single-qubit gates are needed; no CNOTs are required.
- Typical number of shots per circuit: 8,000–20,000.
- Reported single-qubit gate errors are between approximately 0.05% and 0.4% on IBM-Q, with measurement error around 1%.
- No active error-mitigation was used; however, simulations incorporating noise models were validated by comparison to experiment [2007.10242].

## 6. Bloch-Sphere Clustering and Interpretability

For binary classification (single qubit case), DQN's operation can be interpreted via clustering on the Bloch sphere:
- Post-training, the classical encoding clusters the compressed inputs \(\widetilde x\) such that the corresponding states \(|\psi(\mathbf{x})\rangle\) are mapped to two distinct arcs at \(Q_3 = 0\) on the Bloch sphere.
- The variational SU(2) parameters then rotate these arcs toward the north (\(Q_3=+1\)) or south (\(Q_3=-1\)) pole, corresponding to output classes.
- The measurement probability for outcome \(+1\) is given by \(P_+ = (1+Q_3)/2\); the role of the quantum variational layer is thus both separable and explainable via geometric visualization.
- The Bloch-sphere clustering provides direct intuition for the interplay between classical and quantum parameters and the mechanism of class separation [2007.10242].

## 7. Significance and NISQ Suitability

DQN offers a variational QML solution tailored to NISQ hardware limitations:
- Requires only as many qubits as output classes, regardless of input dimensionality.
- Uses single-qubit gates exclusively, mitigating two-qubit gate-induced noise.
- Demonstrates competitive classification accuracy on standard benchmarks.
- Supports interpretability through geometric visualization on the Bloch sphere.

A plausible implication is that DQN's architectural simplicity and noise resilience may inform future designs for resource-constrained QML algorithms and foster further integration with NISQ-era hardware [2007.10242].

Source: https://www.emergentmind.com/topics/dressed-quantum-network-dqn