Towards Abductive Reasoning-Based ComputationalDesign Tools: Using Machine Learning as a way to explore the combined design spaces of multiple parametric models

Abductive Reasoning - reasoning based on experience - is important for design.This research tries to lay the groundwork for using Variational Autoencoders(VAE) - currently one of the most established deep learning architectures forgenerative modelling, a subfield of Machine Learning (ML) - as a way to supportabductive reasoning in early design stages. While our research is still in its earlystages, the first results look promising. In this paper we explain the current stateof our research, its premises and methods, and discuss the results achieved thusfar. We also explain the motivation behind our work and the potential we see inusing VAE in this way and why we believe our approach could represent aparadigmatic shift in the way parametric models can be used in design

(PDF) Towards Abductive Reasoning-Based Computational Design Tools Using Machine Learning as a way to explore the combined design spaces of multiple parametric models. Available from: https://www.researchgate.net/publication/354948638_Towards_Abductive_Reasoning-Based_Computational_Design_Tools_Using_Machine_Learning_as_a_way_to_explore_the_combined_design_spaces_of_multiple_parametric_models [accessed May 02 2022].