A Systematic Assessment of Deep Learning Models for Molecule Generation

TL;DR

This paper provides a comprehensive evaluation of various deep learning models, specifically VAEs, for molecule generation in drug discovery, highlighting their strengths and limitations through systematic testing.

Contribution

It introduces an extensive testbed for comparing generative models for drug discovery, filling a gap in systematic evaluation of VAE methods.

Findings

Different VAE models show varying effectiveness in molecule generation.

The testbed enables fair comparison and benchmarking of generative models.

Insights into model performance can guide future research in drug discovery applications.

Abstract

In recent years the scientific community has devoted much effort in the development of deep learning models for the generation of new molecules with desirable properties (i.e. drugs). This has produced many proposals in literature. However, a systematic comparison among the different VAE methods is still missing. For this reason, we propose an extensive testbed for the evaluation of generative models for drug discovery, and we present the results obtained by many of the models proposed in literature.