Welcome to the Docs for QCompress!

QCompress is a Python framework for the quantum autoencoder (QAE) algorithm. Using the code, the user can execute instances of the algorithm on either a quantum simulator or a quantum processor provided by Rigetti Computing’s Quantum Cloud Services. For a more in-depth description of QCompress (including the naming convention for the types of qubits involved in the QAE circuit), please go to section Introduction to QAE and QCompress.

For more information about the algorithm, see Romero et al. Note that we deviate from the training technique used in the original paper and instead introduce two alternative autoencoder training schemes that require lower-depth circuits (see Sim et al).

Features

This code is based on an older version written during Rigetti Computing’s hackathon in April 2018. Since then, we’ve updated and enhanced the code, supporting the following features:

• Executability on Rigetti’s quantum processor(s)
• Several training schemes for the autoencoder
• Use of the RESET operation for the encoding qubits (lowers qubit requirement)
• User-definable training circuit and/or classical optimization routine

Contents

• Installing QCompress
  • Installing QCompress on QMI
• Introduction to QAE and QCompress
  • What is a quantum autoencoder (QAE)?
  • QAE model in QCompress
  • Overview of QCompress
  • Examples
  • How to cite QCompress
• Demo: Compressing ground states of molecular hydrogen
  • Background: Full cost function training
  • Background: Halfway cost function training
  • Demo Outline
  • QAE Settings
  • Qubit labeling
  • Generating input data
    • Plotting the energies of the input data set
  • Training circuit for QAE
  • Initialize the QAE engine
  • Training
    • Printing the optimized parameters
    • Plot training losses
  • Testing
• Demo: Two-qubit QAE instance
  • QAE Settings
    • Aside: Running on the QPU
  • Data preparation circuits
  • Qubit labeling
  • Data generation
  • Training circuit preparation
  • Define the QAE instance
  • Training
    • Printing the optimized parameters
    • Plot training loss
  • Testing
• Demo: Running parameter scans
  • QAE Settings
  • Data preparation circuits
  • Qubit labeling
  • Data generation
  • Training circuit preparation
  • Define the QAE instance
  • Loss landscape generation and visualization
  • A slightly larger example

API Reference

• qcompress

Indices and tables

• Index
• Module Index
• Search Page