Evaluating Parameter-Based Training Performance of Neural Networks and Variational Quantum Circuits

This repository contains the source code and results for the experiments presented in Evaluating Parameter-Based Training Performance of Neural Networks and Variational Quantum Circuits.

Running Supervised Learning Experiments

Classical Neural Network

Supervised learning experiment using a classical neural network:

python src/sl_experiment.py \
--approach=NN \
--nn_num_hidden_layers=1 \
--nn_hidden_size=9 \
--dataset="iris" \
--lr=0.01 \
--num_epochs=50 \
--batch_size=8 \
--train_test_split=0.75 \
--seed=0

Variational Quantum Circuit

Supervised learning experiment using a variational quantum circuit:

python src/sl_experiment.py \
--approach=VQC \
--vqc_encoding="angle_embedding" \
--vqc_num_layers=2 \
--vqc_data_reuploading=True \
--vqc_output_scaling=True \
--dataset="iris" \
--lr=0.01 \
--num_epochs=50 \
--batch_size=8 \
--train_test_split=0.75 \
--seed=0

Running Reinforcement Learning Experiments

Classical Neural Network

Reinforcement learning experiment using a classical neural network:

python src/rl_experiment.py \
--approach=NN \
--nn_num_hidden_layers=1 \
--nn_hidden_size=12 \
--lr=0.01 \
--gamma=0.95 \
--epsilon=1.0 \
--epsilon_min=0.01 \
--epsilon_decay=0.99 \
--target_update_every=20 \
--replay_memory_capacity=1000 \
--batch_size=16 \
--num_episodes=500 \
--max_steps_per_episode=100 \
--seed=0

Variational Quantum Circuit

Reinforcement learning experiment using a variational quantum circuit:

python src/rl_experiment.py \
--approach=VQC \
--vqc_encoding="angle_embedding" \
--vqc_num_layers=3 \
--vqc_data_reuploading=True \
--vqc_output_scaling=True \
--lr=0.01 \
--gamma=0.95 \
--epsilon=1.0 \
--epsilon_min=0.01 \
--epsilon_decay=0.99 \
--target_update_every=20 \
--replay_memory_capacity=1000 \
--batch_size=16 \
--num_episodes=500 \
--max_steps_per_episode=100 \
--seed=0