Conditional-Variational-Autoencoder-for-Concrete-Design

This repository contains the code for the paper 'Accelerated Discovery Sustainable Building Materials' submitted to the AAAI 2019 Spring Symposium. The code builds a Conditional Variational Autoencoder (CVAE) using the Concrete Compressive Strength Data Set from the UC Irvine Machine Learning Repository together with environmental impact data collected from a web-based tool (https://concrete-epd-tool.org/). The network demonstrates that a trained generative model can be used to design environment-friendly concrete that may help in meeting sustainable development targets.