Examples
Latest Release
Getting Started

Quickstart
Start with the essentials and build your first photonic quantum layer.

Your first QuantumLayers (Iris)
Three practical ways to build and train QuantumLayer models on Iris.

From simple to the CircuitBuilder
How to use MerLin's circuit builder to easily create custom quantum layers.

Binary classification (Iris)
How to use MerLin for binary classification. Includes usage of a Variational Quantum Circuit (VQC) and a quantum kernel method.

Build a hybrid model
Build a hybrid PyTorch model with classical layers around a quantum layer.
Core Concepts

Data Encoding (Angle vs Amplitude)
Introduction to these two encoding strategies and their tradeoffs.

Measurement strategies
Exploration of the possible output types of the MerLin modules.

The computation spaces and the QuantumBridge
How to use MerLin's circuit builder to easily create custom quantum layers.

Running on hardware
A tutorial on how to run yor quantum layer on Quandela's photonic hardware.
Model Families

Kernel Methods: Fidelity kernel classification
Introduction to these two encoding strategies and their tradeoffs.

Kernel Methods: Quantum Kernel vs Clasical Kernel
Comparing classical kernels with quantum fidelity kernels on a simple dataset.

The Basics of Quantum Optical Reservoir Computing
Introduction to the quantum optical reservoir computing and classical readout.

Variational Quantum Models: Classification
Training a variational photonic circuit on a simple classification task.

Variational Quantum Models: Impact of the Photon Number
How photon number affects expressivity and Fourier components?

Computer Vision: Image Classification with QuantumLayers
Exploration of the hybrid classical-quantum pipeline for image classification.

Computer Vision: Image Classification with a Photonic QCNN
Building a photonic quantum convolutional neural network.

Generative Models: Photonic QGAN
How to generate image patches using a photonic generator and classical discriminator.
Application examples

Binary Classification
Comparing VQC and kernel approaches on a simple classification task.

Image Classification
Overview of MerLin approaches for computer vision tasks.

Self-Supervised Learning
Exploring self-supervised learning techniques for image classification

Quantum Transfer Learning
Hybrid transfer learning with frozen classical feature extractors and trainable quantum heads.

Time-series forecasting using QORC
Forecasting the Lorenz attractor time series with a quantum optical reservoir computing model

Time-series Prediction
Overview of sequential models (QORC, QRNN, QLSTM).

Adversarial Robustness
Studying vulnerability of quantum models to adversarial perturbations.
Hardware Examples

Remote processor usage
How to run yor quantum layer on Quandela's photonic hardware.

Comparing simulation and hardware reservoirs on Ascella QPU
A comparison of the performance of an optical reservoir on a simulated quantum layer and on Ascella QPU.

Reservoir Computing on Belenos
A tutorial on how to run yor quantum layer on Quandela's photonic hardware.

Hardware-Aware Model Design
Listing some practical guidelines for designing circuits compatible with hardware.
Research and Paper reproductions
QML research at Quandela

Efficient training of photonic quantum generative models
Paper on an efficient training procedure for photon-native quantum generative models.

Perceval Quest
A paper listing all of the submissions of the Perceval Quest: Classifying MNIST with a QML approach.
Highlights of the reproduced papers

Fock State-Enhanced Expressivity
Expressivity study of photonic variational models with increasing photon counts.

QORC
Quantum optical reservoir computing powered by boson sampling.

QSSL
Quantum self-supervised learning with photonic and gate-model quantum backends.

Explore Reproduced Papers
Browse our catalog of 21 papers reproduced with reusable QML baselines.
Benchmarks and Evaluations
Browse through our notebooks

Reproduced paper: Fock State-Enhanced Expressivity's Fourier Series Study
Experiment on the relation between the number of photons and expressivity of the model.

Reproduced paper: Fock State-Enhanced Expressivity's Kernel Study
Experiments to benchmarks the kernel used in the paper.

Reproduced paper: Fock State-Enhanced Expressivity's Quantum-Enhanced Random Kitchen Sinks Study
Experiments to benchmarks the quantum-enhanced random kitchen sinks used in the paper.
Discover more usages of MerLin
- First Quantum Layers: Classifying Iris with MerLin
- Bringing the Data Scientist Up to Speed With MerLin
- Hello World: Quantum Machine Learning with Merlin (Cloud)
- Conclusion
- Kernel Methods with MerLin
- Binary Classification Example with MerLin
- Classifying a non-linear dataset: A MerLin introduction from installation to classification
- Quantum-Classical Hybrid Neural Network Comparison
- Quantum Optical Reservoir Computing (QORC) with MerLin
- QORC on a Lorenz Time Series
- MerLin release 0.3 highlights
Based on reproduced papers
- Linear quantum photonic circuit as variational quantum classifiers
- Linear quantum photonic circuits as Gaussian kernel samplers
- Quantum-enhanced random kitchen sinks
- Theory validation experiment: Expressive power of the variational linear quantum photonic circuit (dependent of photon number)
- Limitations of Amplitude Encoding on Quantum Classification (AA_study)
- Photonic Quantum Convolutional Neural Network with Adaptive State Injection
- Hybrid Quantum-Classical Model Exploration
