Augmentation of Artificial Intelligence in Design

By: Adam Sykes

 

For the past 2.5 million years we have been using passive tools. These, along with our hands led us to achieve what we are as a species today. However for the first time, we’re building generative tools and working within set goals in the forms of Artificial Intelligence (A.I.).

In the thousands of years of advanced civilizations, our current existence stands out as an anomaly, with the limits of nature being overcome using technology. This technology has allowed us to live connected, fast paced, industrialized lives. Artificial Intelligence is now being introduced in various industries, but no more so than for design and manufacture.

 

Alan Turing’s code breaking computer

 

Computer-based Artificial Intelligence has existed since the 40s, with Alan Turing demonstrating that any form of mathematical reasoning could be processed by a machine. Recent innovations have led to new milestones in the field. In 2016, Google DeepMind’s ‘AlphaGo A.I.’ beat the South Korean Go master, Lee Sedol. This historic event marked the first time a machine has beaten the very best player at this complex game.

The game of Go has 10 to the power of 170 possible board formations (1.0000000000000004e). Previous A.I incarnations used brute force computing power, and a ‘search tree’ (data structure) to try and calculate every possible move. It proved impossible for an A.I. to calculate every conceivable option and formulate a response.

Instead, the team at ‘AlphaGo’ mimicked the learning process of the brain – using deep neural networking and reinforcement leaning. ‘AlphaGo’ utilizes two neural networks together, leading to continuous improvement without the need for human intervention. However, this revolutionary leap in A.I. has, as of October 2017, been superseded by ‘AlphaGo Zero’.

A simpler yet more powerful architecture comprising the two networks of the previous version. The development of a form of computer-based intuition as opposed to purely running on logic is how this A.I. has been able to supersede humans. Often, DeepMind’s programmers did not understanding why AlphaGo was doing what it was. Intuition and super intelligence pose potential solutions to many of mankind’s inherent problems, and those we’re yet to face. A.I. is augmenting our mental facilities, leading to designs and concepts that we, as a species, would never be able to comprehend.

 

Go Master vs AlphaGo Zero

 

Maurice Conti, the director of strategic innovation at Autodesk, gave an uplifting and thought-provoking TED talk on intuitive A.I., stating that “we are at the dawn of a new age in human history”. The ‘Augmented Age’ will initiate a massive shift in procedures including dynamic decision making, pattern recognition, and advisory services. This process will involve natural human aptitudes becoming augmented by advanced computational systems. Helping the way in which we think, mechanical robotic systems, and digital nervous systems that connect humanity to a world.

Scary as it might seem, most agree that the only prudent future that doesn’t see humanity becoming obsolete is to create devices which can be implanted in the human brain. These would be made with the purpose of helping human beings merge with A.I software and its progress.

So, is this all Sci-fi nonsense? We carry mobile phones consisting of 150 million times more powerful computational power than the Saturn V rockets (which took men to the moon). Have we already made the leap into augmented intelligence?

 

Generative design is uses algorithms to create new designs using the constraints of the designer

 

Generative design is an A.I software that uses algorithms to create new designs, synthesizing geometry using the constraints and goals of the designer. Then optimising the design and manufacturing method to create forms with calculations for precise amounts of material, achieving maximum performance with zero wastage. Yet, this example is only the beginning, with Alexa, Google Assistant and Siri currently remaining as passive A.I tools. Currently they are being used for maps and appointments, not for higher thinking or learning. Their potential is far and beyond what we use them for and it will not be long before they begin to show their true potential.

These designs often mimic that of evolution, creating organic forms and structures: racing car endoskeletons take the form of a Cheetahs ribcage and the chassis of a quadcopter looking not unlike that of the pelvis of a flying squirrel. With advancements in SLS, and other forms of metal 3D printing, generative design A.I. tools have already been utilised in real world applications. Notably the Airbus A320, where they reduced its weight by half, whilst retaining its strength. This, in turn, resulted in large scale reductions in fuel costs and carbon dioxide emissions.

Prototypes can now be equipped with sensors providing performance data which is fed back into the generative design process, incrementally improving the outcome. To settle upon the perfect result, generative design creates millions of complex forms, which meet the specified constraints of the designer. These forms and solutions would have been impossible to make using traditional manufacturing methods and design processes.

While the human in this scenario is not required to put pen to paper, with the A.I. tool used to ideate; the initial parameters are still input by the creative and logical mind of the designer. The designer that is attempting to achieve the best possible solution to a design challenge. Technology is already augmenting human aptitudes in the form of robotic systems, with entire assembly lines running 24/7, controlled autonomously by machines. While the introduction of A.I into design and manufacture industries will inevitably lead to job loss, it is important to remember that the pre-cursing stage will be the human intervention in the implementation of these tools. It is also high likely that there will be a supervising human factor in all technologies of this type for the foreseeable future.

It’s hard to predict when A.I. will progress past the stage of needing the initial human input to solve problems. When (not if) it does, the most important constraint that we must control is ensuing unequivocally that both the A.I’s and humanities goals line up. We have already witnessed that when working together, the power of A.I in our hands can be a power unlike any other and that.

When applied correctly, we can explore realms that were never at our fingertips. Under the correct set of circumstances, we will be able to revolutionize every aspect of our society.

 

Asimo

 

– – –

As always, we’d love to hear what you have to say. So let us know what you think or if you have any questions or comments! You can email me or a member of the team via the website contact page or at make@christian-watson.com

We look forward to hearing from you. Follow us on Instagram and receive all of our new articles and updates by signing up for our newsletter.

 

Disclaimer: This blog is a non-profit educational facility for a worldwide community. All images are taken under the Education Fair Use title. All images used are sourced by us for use in educating and informing our audience to the best of our ability. If any persons or agencies find us in fault and would like us to stop using their image/s please contact us directly and we will remove the image indefinitely without issue.