A reproducibility study of "Augmenting Genetic Algorithms with Deep Neural Networks for Exploring the Chemical Space"

02/01/2021
by   Kevin Maik Jablonka, et al.
20

Nigam et al. reported a genetic algorithm (GA) utilizing the SELFIES representation and also propose an adaptive, neural network-based penalty that is supposed to improve the diversity of the generated molecules. The main claims of the paper are that this GA outperforms other generative techniques (as measured by the penalized logP) and that a neural network-based adaptive penalty increases the diversity of the generated molecules. In this work, we investigated the reproducibility of their claims. Overall, we were able to reproduce comparable results using the SELFIES-based GA, but mostly by exploiting deficiencies of the (easily optimizable) fitness function (i.e., generating long, sulfur containing chains). In addition, we also reproduce results showing that the discriminator can be used to bias the generation of molecules to ones that are similar to the reference set. Moreover, we also attempted to quantify the evolution of the diversity, understand the influence of some hyperparameters, and propose improvements to the adaptive penalty.

READ FULL TEXT

page 10

page 11

page 12

research
09/25/2019

Augmenting Genetic Algorithms with Deep Neural Networks for Exploring the Chemical Space

Challenges in natural sciences can often be phrased as optimization prob...
research
08/28/2017

ChemGAN challenge for drug discovery: can AI reproduce natural chemical diversity?

Generating molecules with desired chemical properties is important for d...
research
09/08/2022

Adaptive Combination of a Genetic Algorithm and Novelty Search for Deep Neuroevolution

Evolutionary Computation (EC) has been shown to be able to quickly train...
research
07/30/2020

Research on Fitness Function of Tow Evolution Algorithms Using for Neutron Spectrum Unfolding

Using evolution algorithms to unfold the neutron energy spectrum, fitnes...
research
07/26/2021

A Partial Reproduction of A Guided Genetic Algorithm for Automated Crash Reproduction

This paper is a partial reproduction of work by Soltani et al. which pre...
research
11/21/2017

Genetic Algorithms for Evolving Deep Neural Networks

In recent years, deep learning methods applying unsupervised learning to...
research
01/18/2019

Infeasibility and structural bias in Differential Evolution

This paper thoroughly investigates a range of popular DE configurations ...

Please sign up or login with your details

Forgot password? Click here to reset