Autodesk have recently revealed Design Graph, a Machine Learning system that will be available to 3D engineers operating across the A360 cloud. Design Graph will record every component designed on a cloud platform, and will be able to categorise them by learning about structures, instead of relying on unreliable manual tagging. This software will save time for designers wishing to make use of previous ideas, as well as opening up design collaboration.
Machine Learning, which is a crucial part of the progress of Artificial Intelligence (AI), refers to the drive to enable computers to learn without instruction. There are many pioneering movements going on within Machine Learning (ML), and amongst these is the manufacturing industry, which is being transformed by the increasing power of computers to develop and expand object databases.
ML first emerged out of the larger field of AI in the middle of the 20th century, inspired by Alan Turing’s question ‘can machines do what we (as thinking entities) can do?’ Despite early exploration, it was not until the 1990s when ML started becoming the influential force we see today, when the movement’s key figures moved to solve achievable, short-term problems.
Despite recent progress, ML is still a somewhat unknowable process. As Wired report, not even the best programmer can understand exactly how a computer is managing to learn by itself. This makes mastery of ML, in order to distribute it as part of software, an immense challenge, which explains why Google for instance appointed an ML expert as their head of search, and started training its engineers in the field.
Google are not the only major corporation to have invested heavily in the potential of ML; Amazon, Microsoft, Facebook and many more are all putting in a serious effort. Autodesk are at the forefront; CEO Jeff Kowalski envisions ML being a major part of the company’s work in the exciting fields of generative design and the Internet of Things.