Evolved chair design from Autodesk in upper right

Recent advances in artificial intelligence and the simulation of complex systems have allowed software to play an involved participatory role. Project Dreamcatcher from Autodesk Research is an experimental groundbreaking design platform with an emphasis on the study of generative design. Designers are in a position to explore trade-offs between several different automatically generated designs and to pick solutions which meet their design needs.

The Dreamcatcher framework searches through an automatically synthesized design space to evaluate a vast number of designs produced to meet design requirements. Designers may analyse the solutions created in real time, returning to the problem description at any point to modify goals and constraints. The problem description of Dreamcatcher is a framework for designers to define their design tasks. Solutions are modular and incremental thanks to pattern-based classification. The knowledgebase of the Dreamcatcher’s architecture, generated by machine learning techniques, is a categorized database of predefined objects.

The synthesis goals include physical specifications of structural, thermal, and fluid features. The model synthesis approaches used by Dreamcatcher compete with each other to solve design tasks most efficiently.

Design Explorer by Dreamcatcher presents the user with a set of possible solutions and their associated strategies for the solution. The Design Explorer user interface helps users create a conceptual model that provides high performance alternatives compared to all others by exploring the interactions between variables.

Since 2015, Autodesk offers Autodesk Within, an additive-manufacturing software tool. Within models lattices which are usually difficult or impossible to model. Lattice structures use less material because they are porous. This means lighter parts for models. In the medical field this means a more membrane-like way of imitating anatomical structures.

In a world of human and AI collaboration, the design approach will be increasingly using software in the future to run analytical models and create design concepts. Human intelligence is unsurpassed in essential decision taking especially if we take into account that the human brain might be able to tap into the quantum world of computing where everything is instantaneously connected to everything else which deep learning is incapable of doing at this very moment in time. All we humans need to do is learn to use more of our brain potential since our brain is the most complex system that we currently know of in the Universe. We as human designers simply need to learn how to use all available up and coming technology and with it create long-lasting complex and beautiful objects and structures.

The future is now.

Categories:

Tags: