I have considered a number of candidates for the most significant event that in some major way either contributed or changed, a computing trend.
The main event that started it all was the invention of the steam-powered machine that could do the calculations with simple numbers. The so-called ‘Difference Engine’ is considered to be the world’s first computer” (Difference engine, 2016). It was designed, albeit never constructed, by an English mechanical engineer Charles Babbage in 1822. In the history of computing, this would become one of the most celebrated moments.
The two centuries that came after the invention of difference engine have seen the number of events and discoveries that deserve a place in this introduction.
Initially, it was the development of a modern computer concept in 1936 by Alan Turing. Then in 1943, two of the University of Pennsylvania professors, J. Mauchly and J. Eckert built the world’s first digital computer ENIAC, which filled a 20-foot by 40-foot room. “More than 60 feet of panels, together with some 18,000 vacuum tubes and 1500 relays have been assembled into this machine” (Goldstine, H, & Goldstine, 1946). The next breakthrough happened on 23 December 1947, when Brattain and H. R. Moore demonstrated the first prototype of the transistor to managers at Bell Labs (History of the transistor, 2016). This was followed by the development of the first computer language in 1953 by Grace Hopper and less than a decade later, in 1961, by an introduction of the concept that later became known as the World Wide Web. The concepts of Internet were introduced by an American computer scientist Leonard Kleinrock. And it was the invention of the Internet and its rapid adoption in the mid-1990s, which changed our world and also how we consume and share information today.
It is largely thanks to these major discoveries that we live in such a vibrant, highly social and interconnected world. The past advancements in computing bring me to Artificial Intelligence, the field that in my opinion shows the biggest promise of all.
I became very interested in AI and the area of study myself and decided to exercise my Java programming skills to create an Android AI application called ANDY.
Creation of Andy was my very first attempt at building a voice-controlled software; a program that would act as an intelligent personal assistant. I have decided to approach the task in a very simplistic way. Allow me to provide an actual example. If the user of Andy app asked e.g. “What time it is?” – I would first convert the voice into text, then look for the main keywords (such as ‘time’) to find the appropriate algorithm that I’ve built to deal with time-related questions. Then I would use GPS location of the user’s phone to locate the user’s whereabouts and based on the data I gathered, use one of the open Time APIs to look for the current time in a given location.
Eventually, Andy got pretty fancy. It could translate any English word or sentence to over 60 languages, call anyone in user’s address book, send emails or text messages by voice, give directions, answer math questions, do unit conversions, control Android device features and support many other functions. And with over 1 million downloads, the application became very popular. You can see the demo of the program on Youtube at: https://www.youtube.com/watch?v=2ESpqt7iYes
So, as you can see, simply by implementing numerous very specific algorithms, I was able to create a program that felt very much like an actual artificial intelligence. However, this wasn’t a real artificial intelligence. It was only a pretend, an exercise in being able to parse the meaning of the voice input programmatically. All of the features inside the program had to be hard-coded. In other words, Andy was never smart; it was only cleverly put together to fake the intelligence.
Fortunately, the field of Artificial Intelligence is a lot more. It’s a broad discipline that “draws upon computer science, mathematics, psychology, linguistics, philosophy, neuroscience and artificial psychology” (Artificial Intelligence, 2016).
I want to focus on one the subfields of computer science called the Machine Learning, that evolved from the study of pattern recognition, computational intelligence, and computational learning theory in artificial intelligence.
What is the Machine Learning? Perhaps the best description was given in 1959 by Arthur Samuel, who “gave his famous definition of machine learning as the field of study that gives computers the ability to learn without being explicitly programmed” (Schuld, M., Sinayskiy, I. and Petruccione, F., 2015).
As I mentioned earlier, my Andy application never does anything more than just blindly following the pre-programmed actions. But try to imagine a brand new kind of machine, a robot that can acquire new knowledge and make unique decisions. This is a truly perplexing thought, but it’s exactly what the machine learning is all about.
Perhaps you’re asking, but how can we achieve the goal of creating an intelligent machine? How can machine learning help?
One of the areas which shows a lot of promise is the so-called “Deep Learning”. Deep Learning is a new area of Machine Learning, introduced with the goal of bringing machine learning closer to artificial intelligence. It’s the idea of creating a system that is only powered by the initial deep learning algorithms, but which is later capable of evolving new algorithms on its own. Something that can be done by predicting and adapting to changes in environment. The final goal of machine learning and deep learning, is to create a self-learning, self-evolving mechanism.
Tech titans like Amazon, Apple, Google, Microsoft and Intel are already investing heavily in artificial intelligence and deep learning. Just recently Intel bought the Artificial Intelligence start-up Nervana Systems, a two-year-old start-up considered among the leaders in developing machine learning technology. “Nervana has built an extensive machine learning system… The platform is used for everything from analyzing seismic data to find promising places to drill for oil to looking at plant genomes in search of new hybrids.” (Pressman, 2016)
As we can see, many of these platforms are built on analyzing datasets, but in my view, it’s the nature which provides the best examples of the processes AI field should adopt. Just look at the biodiversity that surrounds us. Today, taxonomists recognize some 350,000 different, very highly adapted species. It is this evolution over successive generations; that I believe can provide the true source of inspiration. We should look into ideas of natural selection, adaptation, cooperation, specialization. But also learn from the extinction of functions that do not serve the purpose. We need to explore how these mechanisms work and see if can reproduce them, perhaps that’s the way to build a self-evolving system.
Another area where I see a lot of potential, is the examinations of mechanisms behind neural networks. The life emerged almost 4 billion years ago and the brain is the most striking result of the evolution. As we know, the biological brain powers learning in all vertebrates. If we want to be successful in recreating the mechanisms of the learning process, perhaps that is another area to look for inspiration.
So, it is this notion of a ‘smart’ machine, a deep learning artificial intelligence, which in my view shows the most potential to propel the humankind into the great unknown. And I say ‘unknown’ for a reason, because as helpful as Artificial Intelligence could be in its initial stages, it’s hard to make any predictions about the impact of a full-blown super-intelligence. There are many unanswered questions surrounding the future of AI. What if such intelligence could outperform humans in every task and easily do everything better than we do?
Stephen W. Hawking, the current Director of Research at the Centre for Theoretical Cosmology at the University of Cambridge, is not very confident about the future when it comes to creating thinking machines. He told the BBC: “The development of full artificial intelligence could spell the end of the human race.” (Cellan-Jones, 2014). Prof Hawkins is probably right; there are many reasons to be alert, because it’s very apparent that we cannot overlook the dangers of Artificial Intelligence and its possible misuse. More specifically when it comes to military and policing.
However, I want to remain optimistic. There could be as many opportunities to use the power of the super-intelligence to our benefit. We will however need more than logic to find the best ways to employ AI in such a way that humanity and society can take the advantage. As Albert Einstein once said: “Logic will get you from A to B. Imagination will take you everywhere. “. So it’ll be up to us, to find the imaginative ways.
That said, I want to conclude by providing an example: Imagine a true Artificial Intelligence employed in the field of governance. A super intelligence could look at all existing protocols and procedures while taking in account all possible scenarios and outcomes. A positive consequence in this case, could perhaps be the construction of an enhanced set of regulations. Those that could help the society in general to further enhance the legislature, and at the end of the day, assist in building a better world.
Difference engine (2016) in Wikipedia. Available at: https://en.wikipedia.org/wiki/Difference_engine (Accessed: 12 August 2016).
History of the transistor (2016) in Wikipedia. Available at: https://en.wikipedia.org/wiki/History_of_the_transistor (Accessed: 13 August 2016).
Goldstine, H, & Goldstine, A 1946, ‘The electronic numerical integrator and computer (ENIAC)’, Mathematical Tables And Other Aids To Computation, 2, p. 97, MathSciNet via EBSCOhost, EBSCOhost, viewed 13 August 2016.
Artificial Intelligence (2016) Available at: https://en.wikipedia.org/wiki/Artificial_intelligence (Accessed: 13 August 2016).
Schuld, M., Sinayskiy, I. and Petruccione, F., 2015. An introduction to quantum machine learning. Contemporary Physics, 56(2), pp.172-185, viewed 13 August 2016.
Pressman, A. (2016) Why Intel bought artificial intelligence startup Nervana systems. Available at: http://fortune.com/2016/08/09/intel-machine-learning-nervana/ (Accessed: 13 August 2016).
Inc, E.S. (2016) ANDY (Siri like assistant). Available at: https://play.google.com/store/apps/details?id=andy.xml&hl=en (Accessed: 13 August 2016).
Bostrom, N., 2003. Ethical issues in advanced artificial intelligence. Science Fiction and Philosophy: From Time Travel to Superintelligence, pp.277-284, viewed 13 August 2016.