jqi41/QuantumDot

Classical-to-Quantum Transfer Learning for Quantum Dots

Pre-X+VQC

An Implementation of Classical-to-Quantum Transfer Learning Algorithm with an Illustration of Quantum Dots

Our codes include the experiments of PCA+VQC, Pre-ResNet18+VQC, and Pre-ResNet50+VQC for charge stability diagrams for Quantum Dots

Installation

The main dependencies include pytorch and torchquantum

Torch Quantum

pip3 install torchquantum

0. Downloading the dataset

git clone https://gitlab.com/QMAI/mlqe_2023_edx.git

1. Simulating Pre-ResNet18+VQC experiments

1.1 Assessing the representation power of Pre-ResNet18+VQC

python Pre-ResNet+VQC.py --num_qubits=8 --test_kind='rep' --model_kind='ResNet18'

1.2 Assessing the generalization power of Pre-ResNet18+VQC

python Pre-ResNet+VQC.py --num_qubits=8 --test_kind='gen' --model_kind='ResNet18'

2. Simulating Pre-ResNet50+VQC experiments

2.1 Assessing the representation power of Pre-ResNet18+VQC

python Pre-ResNet+VQC.py --num_qubits=8 --test_kind='rep' --model_kind='ResNet50'

3. Simulating PCA+VQC experiments

3.1 Assessing the representation power of PCA+VQC

python PCA+VQC.py --num_qubits=8 --test_kind='rep' 

3.2 Assessing the representation power of PCA+VQC

python PCA+VQC.py --num_qubits=8 --test_kind='gen' 

3.2 Assessing the generalization power of PCA+VQC

python PCA+VQC.py --num_qubits=8 --test_kind='gen' 

4. Comparing Pre-ResNet+VQC with Pre-ResNet+NN regarding representation and generalization powers

4.1 Comparing their representation power

python Pre-ResNet_nn.py --num_qubits=8 --test_kind='rep'

4.2 Comparing their generalization power

python Pre-ResNet_nn.py --num_qubits=8 --test_kind='gen'