Jet engines, lifelike movie monsters, cancer-fighting nanorobots and bespoke products. We live in a world where the objects all around us are designed by someone, and at the same time we are experiencing the biggest change in how we make things since the industrial revolution.<\/p>\n
But this change isn\u2019t just about new technology; designers and engineers are experiencing profound process changes and totally new and different ways of working. How do designers and engineers rise to meet the challenges that accompany these changes?<\/p>\n
At Autodesk, we asked ourselves this question: What if there was a better way to organize all design information and allow people\u2019s ideas and creations to emerge more organically?<\/p>\n
Imagine having Google search-like functionality for the world of 3D models. Think of the simplicity and the speed at which you could work. This is what we have been working on at Autodesk and it is called Design Graph.<\/p>\n
Design Graph is a powerful new machine learning system that uses algorithms to extract large amounts of rich 3D design data.\u00a0It then categorizes every single component and design your design team has ever created, by classification and relationship, to create a living catalog that is able to react to a constantly-evolving world and guide your designs of the future.<\/p>\n
Designers and engineers simply search across all of their files for a part type, such as a bolt or a bike seat, and Design Graph returns dozens or hundreds of options.<\/p>\n
\u201cWe created Design Graph to enable designers to focus more on solving design problems rather than the mechanics of representing their design,\u201d said Mike Haley, Sr. Director of Machine Intelligence at Autodesk. \u201cDesign Graph can save valuable time, eliminate redundant work and reduce costly errors.\u201d<\/p>\n
In the case of Design Graph, the goal was to teach computers to identify and understand designs based on their inherent characteristics\u2014their shape and structure\u2014rather than by any labeling (tags) or metadata. After all, whoever designed the part originally could label it any of dozens of ways, using full words or abbreviations. Metadata created by people, unless carefully managed, tends to be unreliable. With Design Graph, the computer uses its own observations about the 3D geometry contained in every 3D model.<\/p>\n
\nMachine Learning
\n<\/strong>In recent years, programming breakthroughs and the limitless compute power of the cloud have given a computer the ability not just to think, but to learn independently, identify patterns and make predictions based on what it learns \u2013 rather than relying solely on the limited set of logic that can be programmed into it. Machine learning acts as an accelerant that advances what our machines and tools are capable of doing.<\/p>\n
\nHow Does It Work?
\n<\/strong>Say that you\u2019re designing a motorcycle, and you need a certain gear box. While working inside A360 Drive (soon to be available across A360), you can use Design Graph to search based on name, shape, category, properties or a combination of these to find the right pre-existing gear box design from your own firm\u2019s design files.<\/p>\n