MolQAE: Quantum Autoencoder for Molecular Representation Learning

TL;DR

MolQAE introduces a quantum autoencoder that directly encodes complete molecular structures into quantum states, enabling efficient molecular representation learning with high fidelity, promising advancements in cheminformatics and drug discovery.

Contribution

It is the first quantum autoencoder to process full molecular structures using a quantum architecture tailored for NISQ-era devices, advancing quantum cheminformatics.

Findings

Effective capture of molecular features with high fidelity

Significant dimensionality reduction with minimal information loss

Potential for improved drug and materials discovery

Abstract

We introduce Quantum Molecular Autoencoder (MolQAE), the first quantum autoencoder to leverage the complete molecular structures. MolQAE uniquely maps SMILES strings directly to quantum states using parameterized rotation gates, preserving vital structural information. Its quantum encoder-decoder framework enables latent space compression and reconstruction. A dual-objective strategy optimizes fidelity and minimizes trash state deviation. Our evaluations demonstrate effective capture of molecular characteristics and a remarkable preservation of fidelity, approaching robust molecular reconstruction even with substantial dimensionality reduction. Our model establishes a quantum pathway in cheminformatics by being the first to process complete molecular structural information with a dedicated quantum architecture considering the Noisy Intermediate-Scale Quantum (NISQ)-era development and promising significant advances in drug and materials discovery.