Computational Design Laboratory

DeepCloud in ACADIA 2018

Ardavan Bidgoli and Pedro Veloso, two Ph.D. students from CodeLab, recently released their paper “DeepCloud: The Application of a Data-driven, Generative Model in Design” on ACADIA 2018 conference in Machine Learning Data session. This year ACADIA conference took place at Universidad Iberoamericana in Mexico City from 18-20 October 2018.

DeepCloud

Abstract by the authors: Generative systems have a significant potential to synthesize innovative design alternatives. Still, most of the common systems that have been adopted in design require the designer to explicitly define the specifications of the procedures and in some cases the design space. In contrast, a generative system could potentially learn both aspects through processing a database of existing solutions without the supervision of the designer. To explore this possibility, Bidgoli and Veloso review recent advancements of generative models in machine learning and current applications of learning techniques in design. Then, they describe the development of a data-driven generative system titled DeepCloud. It combines an autoencoder architecture for point clouds with a web-based interface and analog input devices to provide an intuitive experience for data-driven generation of design alternatives. They delineate the implementation of two prototypes of DeepCloud, their contributions, and potentials for generative design.

About ACADIA

ACADIA was formed for the purpose of facilitating communication and critical thinking regarding the use of computers in architecture, planning and building science. The organization is committed to the research and development of computational methods that enhance design creativity, rather than simply production, and that aim at contributing to the construction of humane physical environments. A particular focus is education and the software, hardware and pedagogy involved in education.

This year the topic is “Recalibration” which refers to the imprecision and infidelity as an attempt to build upon the already vibrant cross-border dialog between the artistically, computationally and technologically driven communities with the potential not only to re-examine boundaries, but also to provide for a moment of self-reflection. In the face of uncertainty, the idea of a recalibration could come to mean achieving a new level of understanding of the inherent infidelities of the systems in which ACADIA have been working and then embracing that imprecision rather than trying to impose a means to control it.

Please visit the ACADIA2018 website for more information http://2018.acadia.org/index.html

Author: Yixiao Fu
Category: News