A prominent feature in science fiction over the years has been artificial intelligence. Since the earliest days of computing, scientists and other thinkers have been fascinated by the notion of creating a machine capable of replicating the human brain. It used to be thought that the analogy of the human brain is like a computer ran deep. However, we now know that the picture is much more complicated, the way that the brain works goes beyond a simple computer.
We still do not fully understand how consciousness arises in the human brain, and there is still much debate surrounding whether consciousness can be separated from advanced intelligence. But artificial intelligence need not be this complex; we see far simpler examples of what we might describe as artificial intelligence on a regular basis. The voice assistants pre-installed on every modern smartphone are just one example and now these same AIs are being integrated into alarm clocks and speakers so that they can be used to control a variety of smart devices around the home.
Artificial intelligence is increasingly finding its way into industrial and manufacturing contexts. There are even AIs being used to conduct high-frequency trading on the stock market. AIs are now everywhere, meaning that it is becoming easy to forget just how amazingly complex they are. AIs have a great deal to offer the world of engineering. Some of the most exciting current and prospective uses of artificial intelligence are within the field of engineering.
What is Artificial Intelligence?
The term was first used at a conference at Dartmouth College in 1956. However, while artificial intelligence has long been considered and discussed in an abstract, theoretical sense, it is only in the last decade that we have begun to see it being used in consumer technology. It has now become so ubiquitous in our everyday lives that it is easy to forget what a complex demonstration of technological prowess and understanding artificial intelligence represents.
In answering the question of what artificial intelligence is, and what the term means today, we need to consider what constitutes intelligence. This is not as simple as many people assume it should be. For example, would you consider all animals to be intelligent? Or rather, to have intelligence?
Some animals, such as cats, octopuses, and even dolphins, among others, demonstrate high levels of intelligence. When comparing two different animals, such as a mouse and a gorilla, there are a number of ways that scientists can measure their relative intelligence. But objectively defining and measuring intelligence is difficult.
The AIs that are used in the engineering sector combine both software and hardware components. Think of the robots on a car assembly line and the software that controls them. They are in themselves quite impressive feats of engineering, but are they intelligent?
You might be surprised to learn just how smart and sophisticate our uses of artificial intelligence in engineering are becoming. Smart production lines are definitely the future. But how exactly does artificial intelligence make such a big difference to the engineering sector?
The rise of artificial intelligence promises to allow us to develop machines capable of performing ever more complicated manufacturing, and even design, tasks. Machines that are capable of learning and improving without human intervention are the ultimate goal, and this would have significant, and far-reaching implications. Furthermore, in our pursuit of creating ever more powerful AIs, we are discovering information about how our own brains work and how we approach the learning process, both consciously and unconsciously.
Many engineers fear that their jobs could soon be taken over by sufficiently advanced robots. As our manufacturing and design capabilities have continued to expand, we have been able to build machinery that is capable of replicating just about everything that a human can do on an assembly line. These fears are not unfounded then, as automation is continuing to take jobs away from people in a number of different areas.
Things aren’t entirely bleak, however, a Stanford University study entitled ‘One Hundred Year Study of Artificial Intelligence’, reported that there was no imminent threat to workers. The study argued that even if or when artificial intelligence does have a significant impact on jobs, this will be balanced by numerous other positive effects on society.
Perhaps the most prominent example of artificial intelligence being used in engineering is in the field of automobile manufacturing.
The combination of software and hardware that has made its way on to the manufacturing line has grown progressively more sophisticated over the years. Initially, these robots were performing simple engineering tasks that involved relatively large components and movements. Today, they are capable of precision movements and of emulating the most intricate parts of the process.
It would not be unreasonable to say that we are now living in an age of data. Data is a commodity unlike any other that the world has ever known. It is extremely valuable financially, but it can also be used directly in order to give a business a massive edge over the competition.
Artificial intelligence, especially in its most sophisticated implementations, relies heavily on large data sets and algorithmic learning.
One of the most exciting applications of artificial intelligence within the field of engineering is machine learning. Machine learning is dependent upon the constant generation and analysis of data. It is via this process, of extensively collecting data about performance and subsequently analyzing it, that an artificial intelligence is able to learn. If the program is equipped with the right algorithms to identify mistakes, and formulate solutions, then it can perform a process and continually refine it.
For engineers who are working on large scale public projects, big data will be a staple of their work. Big data analysis can tell researchers, in unprecedented detail, where the flow of people in urban environments is at its densest. This, in turn, means that public infrastructure decisions can be based on objective scientific analyses.
Also, within the context of engineering for public works, big data can be used to analyze how well certain solutions have performed when implemented elsewhere. Big data can also allow for an objective and detailed comparison of how similar the current environment is to ones where the solution has been used before. This is relatively simple when using big data analytics techniques, but would be a long and expensive process to complete otherwise.
One of the most significant technological concepts for the future of artificial intelligence-led engineering is machine learning. Machine learning is the study of exactly how machines learn. The ultimate goal of artificial intelligence isn't just to have machines that can learn, but to have machines that are capable of self-analysis. Such a machine could assess the efficiency of its learning methods and so refine its processes to a much greater degree.
But what would the practical applications of machine learning look like? Well, imagine if every one of those robotic arms you see putting cars together contained a tiny camera. Each arm could then look over the work of the previous robots along the assembly line. If they identify an issue then they could formulate a solution.
We already have the technology to accomplish the first part. We can take a high-resolution video of a half-assembled car and develop algorithms to identify whether there are any clear faults. We could then have the robots respond to the fault based on what they 'see'.
Machine learning takes this process to the next level. With machine learning, the data collected by all of the robots involved in production can be pooled together. With a central artificial intelligence to control each one, it can learn which problems are most likely to appear. With machine learning, that central artificial intelligence would also be able to formulate solutions to problems, rather than simply following pre-defined routines.
Natural Language Processing
Natural language processing is a field of study dedicated to improving the ability of humans and machines to communicate. In particular, natural language processing aims to improve the sophistication with which machines can respond to the human voice. Like with machine learning, natural language processing makes heavy use of large data sets and algorithm-based learning.
Think of the voice assistant in your smartphone. If you have owned a number of smartphones over the last decade or so, then you may well have noticed how much the accuracy with which they hear and transcribe our voices has improved. While your phone might be able to identify the words that you've said, this isn't the same as understanding.
Right now, your phone looks for certain keywords that it understands and works out what you are asking it to do based on context. It then responds or performs an action, and sometimes vocalizes a response. Natural language processing aims to refine this process by allowing the machine to develop a deeper understanding of language. If this understanding is refined enough, then it will reach a point where the machine can deduce what someone wants when presented with an entirely new command or request.
In the Iron Man films, Tony Stark is able to have long conversations with his home assistant, an artificial intelligence called Jarvis. When Tony is designing his Iron Man suits he holds conversations with Jarvis, Jarvis is able to produce schematics according to specifications which Tony expresses, in usual conversational language. This seems like pure science fiction, yet this is exactly the direction that researchers hope to one day take the field in.
For example, if an engineer is trying to work out how to reinforce a particular feature in their design, wouldn't it be great if they could just ask their computer? Or in the case of an assembly line, imagine if a human overseer could give the robots feedback. They could ask the robots to perform their roles in a slightly different way, to make adjustments, or even to try new things and analyze the result.
These applications are some way off, we still have much to learn about machine learning. However, in recent years we have already made some significant advances that few people could have predicted.
You might be wondering what image processing could have to do with engineering? The connection might not immediately seem obvious, but this is another technology which is vital to implementing artificial intelligence to its full potential in the field of engineering.
When humans see an object, it is because light is entering the eye and being converted into an electric signal. This signal is then carried to the brain via the optic nerve. The brain turns this electronic signal into an image, it is this image that we 'see'.
Machines work in a very similar way. We can set up a camera in order to record an image, and we can display this image to a user. However, this is not the same as the machine understanding the image. With image processing algorithms, we can have machines analyze what they see and react accordingly. From an engineering perspective, this means we could have machines which are able to identify structural abnormalities and other issues that have identifiable visible signs.
This kind of image processing technology could also make a significant difference to the workplace safety of engineers. There may often be visual clues indicating structural deficiencies and weaknesses that are not immediately obvious until the structure fails. By combining image processing with data input from other sensors, artificial intelligence can be used in a variety of contexts. For example, on both construction sites and the scenes of fires, structural integrity can become a concern. Having a more reliable way for engineers to assess integrity could save lives.
Internet of Things
Many of you can probably still remember a time when being connected to other people meant being at home. Once you ventured outside, there was no 3G or 4G network in place for internet browsing. Eventually, very slow and expensive mobile internet came in the form of WAP.
Today, we are used to having vast amounts of data flying through the airwaves all around us. As smart devices become more common in our homes, we are also beginning to see the practical potential of being able to link devices together.
The Internet of Things refers to a hypothetical network, which would connect everyday devices and things together, in the same way that the internet connects computers from around the world. Allowing the various devices in our lives to collect and share data would open up some exciting new possibilities.
As the Internet of Things gradually becomes a reality, it will increasingly become something that engineers consider during the design process. With the Internet of Things as a reality, the virtually endless number of ways that we can connect devices and have them work together will allow new and innovative solutions to many problems.
No discussion of the impact that artificial intelligence is having on engineering would be complete without mentioning the impact of automation on jobs. In many places, there are widespread fears and anxieties surrounding automation. As machines begin to replace humans in certain jobs, there are worries that we will eventually have no need to hire people at all.
It should be acknowledged that the threat to jobs is very real, and in some areas, it has a significant effect on communities. However, most researchers agree that the long-term benefits of automation outweigh the potential drawbacks.
In the case of engineers specifically, artificial intelligence is opening some exciting new horizons for the field. These new opportunities should be embraced. It is important to realize that many of these advances will make a big difference to our ability to tackle the largest issues facing our civilization.
Al Affects Blockchain and Cryptocurrency Tech
A fantastic illustration of how innovative use of AI can affect cryptocurrencies and blockchain technology is the Magnus Collective. They comprise a decentralized network of AIs, including sensors, hardware, computers, robots, and human. It is a hybrid token, might be an evolution of the ICO concept.
Artificial intelligence has had an impact on just about every conceivable industry and sector, engineering is no exception. There are a number of different applications of artificial intelligence that are of considerable use to engineers. From allowing more intuitive and innovative interactions with software and machinery, to keeping a watchful eye over the work of engineers as well as other machines, artificial intelligence has many roles to play.
As our methods of collecting an analyzing large data sets become more refined, we are able to unlock the full potential of big data and algorithmic learning. We have always understood that both of these concepts can yield impressive results, but the transformative nature that they both have on engineering demonstrate that they are even more powerful than we once thought.