When I took an online course on Big Data, the instructor instructed me that “Big Data is commonly characterized by a number of Vs.” She did not, however, specify the number, and the Vs were introduced with great suspense:

The first three are Volume, Velocity, and Variety.

Later on, she introduced two more Vs: Veracity and Valence

I was pretty excited at this point, especially since I hadn’t heard from Valence since high school chemistry class. Not a minute later, she dropped a bombshell:

Moreover, we must be sure to never forget our sixth V, Value.

With no more Vs mentioned for several lectures, I settled down with this model of Big Data, and indeed I passed the course with high marks. It wasn’t until months later, my world was turned upside-down with a shocking revelation: Big Data has no fewer than TEN Vs!

The Incognito Four are: Variability, Venue, Vocabulary, and Vagueness

Would you believe that the Incognito Four remained hidden for more than 13 years before revealing themselves to the world?! Before anyone starts thinking, “This must be a joke”, I warn you: It is not. I know this because the one who helped reveal the Incognito Four, Dr. Kirk Borne, says it is not.

As I began to think more and more about Big Data, however, I realized that ten is just not enough. Big Data is too big and important to be contained by a mere ten Vs.

Therefore, I submit to you, dear Reader, that there are no fewer than 23 Vs necessary to characterize Big Data. Here are the more recent discoveries, the Lucky Thirteen, along with the requisite explanations:

Vastness: Big Data is not just big, it is vast.

Voluminousness: Because the volume of Big Data is so big, it is voluminous, and therefore it has voluminousness.

Voluptuousness: Big Data is not just voluminous, it has an innate beauty and therefore has a certain voluptuousness. Don’t take my word for it. 93% of Data Scientists agree on this.

Voodoo-magic: If you are not employing voodoo-magic in your Big Data Analytics, you should at least tell people you are.

Victory: If Big Data doesn’t give you victory against your opponents, you’re doing it wrong.

Vegetarian: You probably never realized this, but Big Data doesn’t eat meat.

Viscosity: Big Data is like a huge tanker truck full of molasses. It has crashed, and the molasses is spilling everywhere. It is going to crush you and everything you love. But it will take time.

Venom: When Big Data is really working for you, it is a bit like venom, isn’t it? See Victory above.

Vernacular: Big Data is often expressed in the vernacular. Vernacular is, by definition, how people actually express things. Therefore, Big Data cannot be anything but vernacular, right?

Vagility: Big Data will give your company or scientific career lots of vagility. Vagility is like agility, but it starts with a V; therefore, it is better.

Vectors: Big Data has lots of vectors. It has even vectors of vectors!

Virality: Big Data has obviously gone viral. Therefore, it has the essential quality of virality.

Vorticality: Vortexes are things that spin, often out of control. They are vortical. Big Data is like this, and therefore, it has vorticality. In fact, this word was invented specifically to describe how vortical Big Data is.

Venality: Big Data will really do whatever you want it to do, no matter how good or bad your personal intentions are. This is what ontologists call venality.

(Note: If you aren’t familiar with ontologists or ontologies, see my other article “The 64 Os of Ontologies”)

There you have it. The 23Vs of Big Data. For future students of Big Data, I sure hope it is enough. But if history is any lesson, it is only a matter of time before more Vs are discovered. While 23 was enough to satisfy Dr. Pepper, it wasn’t enough for Messrs. Baskins, Robbins, or Heinz. We may even be in a period of hockey stick growth for the Vs of Big Data. For the sake of humanity, let’s hope Vs don’t contribute to global warming. I repeat, this is not a joke.

Ok……yes, it is. ;-)

The 23Vs of Big Data was originally published in HackerNoon.com on Medium, where people are continuing the conversation by highlighting and responding to this story.

While you are here, enjoy this lovely quote:

In recent years, artificial intelligence, machine learning, and deep learning have become buzzwords. For those who have been working in these fields for awhile, this is somewhat amusing and occasionally even gratifying. I certainly don’t view it as a bad thing that these areas of technology are receiving more widespread attention. With this increased attention also comes the need to explain these concepts to the masses, which is a noble challenge to tackle.

Nonetheless, there is one recent trend that has a similar effect on me as fingernails scratching a blackboard:

Come and learn the difference between machine learning and deep learning!

Aaaaaah!!!!!!! Shivers down my spine!!!

Take this tweet for example:

Wen👩🏻‍💻 on Twitter

Machine learning vs deep learning: what's the difference? https://t.co/7oGXzmlY06

Or this one:

Imaginovation on Twitter

I'm sure you've heard of Machine Learning and Deep Learning. But do you know the difference? https://t.co/J057mSAqo3

Or one of my all-time “favorites”:

Allay Technologies on Twitter

@SumitGup illustrates the difference between ML and DL at #think2018 Read his thoughts on accelerating #MachineLearning algorithms: https://t.co/NVTGVdNJN5

One would think that the VP of AI and deep learning at IBM, Sumit Gupta, would be a reliable source on the matter, but sadly, he is leading his audience astray.

Deep learning is not something inherently different from machine learning; it is, in fact, one of the latest innovations in the field of machine learning that actually brings us closer to the original goals for which machine learning pioneers had strived. Let us briefly review those goals.

Machine Learning Roots

As I recounted in an earlier article, machine learning is not a particularly new idea. The fundamental idea was already articulated by Alan Turing as early as 1947:

What we want is a machine that can learn from experience.

This basic desire has been re-articulated over the years by several generations of machine learning researchers, such as Arthur Samuel, Claude Shannon, Nils Nilsson, and Tom Mitchell. Consider this passage, published in 1959, by Arthur Samuel:

At the outset it might be well to distinguish sharply between two general approaches to the problem of machine learning. One method, which might be called the Neural-Net Approach, deals with the possibility of inducing learned behavior into a randomly connected switching net (or its simulation on a digital computer) as a result of a reward-and-punishment routine.

For those familiar with deep learning, this should sound strikingly familiar. In fact, the basic ideas in artificial neural networks were already established in the 1940s. These ideas were discussed and studied widely in the machine learning community over many decades, but at some point they fell out of fashion.

Feature Engineering

A curious thing happened in the field of machine learning sometime around the 1980s. I haven’t been able to pinpoint exactly when it occurred, but over time, more and more effort was exerted by researchers in the task of coming up with clever “features” to serve as inputs to machine learning algorithms. For example, instead of feeding the raw pixel values of images into a image recognition algorithm, fancy “pre-processing” algorithms were devised to detect edges, corners, textures, and other higher-level features from the images. These would then serve as the input features into the recognition algorithms. This somewhat laborious process is essentially what is depicted in Mr. Gupta’s slide with the heading “Machine Learning”, shown above.

This process, which is generally known as feature engineering, also tends to be application specific. That is, features that work well in, say, computer vision tasks are usually completely useless in other tasks, such as activity recognition from motion sensors.

In many cases, the number of available features might even become large and unwieldy. In comes the idea of feature selection. New algorithms had to be devised just to figure out which are the best features to use for a given machine learning task.

This emphasis on feature engineering and feature selection, however, is not what defines machine learning. If anything, an over-sized focus on these areas is a distraction from the ultimate aim of machine learning, which is to develop computer systems that automatically learn from experience. This is because in feature engineering the researcher or engineer inserts himself or herself into the equation, trying to optimize the learning process with clever new ideas for features (or deciding on sets of features).

Deep Learning to the Rescue

I won’t go into the nitty-gritty of what deep learning is and how it works in detail; there are plenty of good resources on that topic, such as Michael Nielsen’s excellent online book Neural Networks and Deep Learning. Instead I just want to point out how deep learning, in many ways, rectifies the above issues and restores machine learning closer to its original goals.

Instead of coming up with elaborate feature sets, deep learning largely works with high-dimensional “raw data” (e.g. pixels in an image). For sure, deep learning still employs application-specific tricks, such as the idea of convolutional neural networks used mostly in image recognition. In fact, one could argue that restricting a neural network in such fashion is a form of feature engineering. But even so, these techniques are considerably more general than ad hoc features, such as edge detectors from images or “variance of the dynamic acceleration in the horizontal plane” from accelerometer data. The basic idea is to suck whatever data you have into a large and deep neural network and let the learning algorithm do the hard work. Lack of expressiveness in features is made up for with an abundance of data and multiple hidden layers in the network that automatically learn the best features.

By all measures, however, this is still what we call machine learning. Setting aside “deep learning” as something different from “machine learning” is just bad pedagogy. It reveals a lack of understanding for what the goal of machine learning has been ever since the concept was first formulated. The relationship between the two is best expressed in the diagram below. Deep learning falls under the general umbrella of machine learning, which is itself part of the wider picture of artificial intelligence.

In summary, when you are impressing your hot date by describing what deep learning is all about, try not to make the mistake of saying it’s something completely new and way cooler than that old school machine learning stuff. And by all means, stop retweeting those nonsensical tweets pitting machine learning against deep learning. For in reality, deep learning is one of machine learning’s long lost friends.

Please stop saying Deep Learning is something different from Machine Learning was originally published in HackerNoon.com on Medium, where people are continuing the conversation by highlighting and responding to this story.

In this article, which is Part 2 of a series tracing the concept of artificial intelligence from its inception, we pick up the story with Alan Turing, who is considered by many to be the “father” of computer science. As we shall see, Alan Turing has a credible right to be called not only the father of computer science but one of the earliest pioneers in artificial intelligence (or machine intelligence, as he would have called it). For this reason, I devote considerable attention to his brief but remarkable career, including some biographical details.

I repeat my disclaimer that I am not a professional historian. Instead, I hope this series of articles inspires others to further study this fascinating history, as well as providing insight into what “artificial intelligence” actually means.

On Computable Numbers (1936)

In 1936, Alan Turing published one of the landmark papers in the history of science, “On Computable Numbers, with an Application to the Entscheidungproblem” [1]. In it, he described what he called “the universal computing machine”. In his honor, this theoretical construct is now called the universal Turing machine, and it is not only of theoretical interest: Many consider it to be the foundational idea of modern computers.

Turing’s early work in computational science is relevant to AI because it brought forward the idea of computation represented by Babbage’s Analytical Engine. Turing’s paper proved from a mathematical perspective the possibility of general computation performed by a single machine. That is, Turing proved that any number or sequence that can be computed, can be computed by a single type of machine, the so-called universal Turing machine. While this may not prove that machines can think, it shows that machines for general computational tasks can be devised (at least in theory).

Enigma and the Bombe (1939–1942)

On 4 September 1939, the day after France and the UK declared war on Germany, Turing moved from Cambridge to Bletchley Park, headquarters of the UK’s Government Code and Cypher School (GC & CS) [3]. There he became the principal designer of the code-breaking machine Bombe, which was used to decipher codes from the German Enigma cipher machine.

The story of the Enigma and the Bombe is a long and interesting one, but it is already well documented. I will recount only the essential details. I refer the interested reader to Jack Copeland’s excellent book The Essential Turing and references cited therein [3].

What I would like to explore briefly in this article is the connection between Turing’s work and the development of AI, which I believe sheds light on the concept of what AI actually is.

In many ways, the Bombe can be thought of as a primitive implementation of machine learning. A Bombe was a machine that houses elements replicating several copies of an Enigma, built for the purpose of discovering keys that would unlock coded German messages. It used heuristic search to find a candidate key that would solve a certain set of constraints (i.e. a heuristic, represented by the specific set-up of the Bombe), and then a human operator would check whether this key, applied to the coded message, would produce coherent German. Turing would later conjecture that “intellectual activity consists mainly of various kinds of search.” [4].

I argue that the above procedure was a primitive form of machine learning because this is, at a high level, what many machine learning algorithms do today: (1) select a configuration from a large set of possible configurations, representing parameters to a function that satisfie certain constraints, (2) check/measure the candidate solution (i.e. parameter set) according to some measure of correctness, (3) repeat as necessary, and (4) once satisfied with the performance, use the chosen configuration as a model for computing some output on other yet unseen data.

There are three major differences between the Bombe and modern machine learning, not so much in procedure but in implementation. First, the Bombe required that a human operator set up the machine using cables connecting different components (see figure above), in order to impose certain constraints/heuristics. A human was also required to transfer the candidate “solution” to a separate machine (a replica Enigma), in order to try to decipher a full German message. In principle, these manual steps could have been also automated, but in this time period, it was probably easier to have a human operator perform them than build/extend a machine to perform the same tasks.

Secondly, whereas the Bombe could select and check 17,576 configurations in about two hours [3], modern machine learning performed on modern computers can perform this type of procedure millions of times per second in many cases (the details depend, of course, on the number of parameters, the function between computed, size of computer, etc.).

Thirdly, the Bombe was a machine designed to perform a particular task (i.e. learning keys to Enigma codes), and it could be used only for that task. Today, we usually execute machine learning tasks on general purpose computers, which can be “reconfigured” for other tasks simply by loading other programs into memory.

Still, despite the manual nature and “slowness” of the Bombe in modern terms, it represented a major advancement in automation and computational speed for its particular task, compared to anything that existed prior to that point. Note that there were other computational machines in existence at this time, such as the IBM 601. I’m not sure it is possible though to make an “apples to apples” comparison of computational speed between different machines of this era like we do today. Lastly, I’d like to point out that the Bombe was based closely on the design of an early Enigma-cracking machine, the Bomba, designed by Polish mathematicians.

Before moving on to other topics, I want to point out that in a span of only three or four years, Turing shifted from the very theoretical (On Computable Numbers) to the highly practical (cracking German military communications). In the words of the official historian of the British Secret Service, Sir Harry Hinsley:

I won’t say that what Turing did made us win the war, but I daresay we might have lost it without him.

We also know that in early 1941 the UK was at risk of running out of food and other basic supplies, due to attacks on ships by German U-boats. At that time, the Naval version of Enigma had not been cracked and many thought it to be unbreakable, but finally when Turing’s team began regularly decoding Naval Enigma messages in June 1941, British ships were able to be successfully routed away from the U-boats [3]. Disaster was averted.

If these outcomes— helping the British population avert starvation and the Allies to win WWII — don’t epitomize practical research, I don’t know what does. In my view, the highly practical nature of his work near the beginning of Turing’s tragically short career was highly impactful for the development of AI. Had Turing continued in a more theoretical vein, the field may not have developed as rapidly as it did during the period of 1941–1953, which we will now explore.

Turing’s shift to machine intelligence (c. 1941–1953)

According to interviews with Donald Michie, who worked with Turing at Bletchley Park, Turing began pondering “machine intelligence” as early as 1941. He circulated a paper on machine intelligence among his colleagues at GC & CS, but this is now lost. We also know that towards the end of 1941, there were not many theoretical problems concerning Enigma left for Turing to work on, so he began to work on other problems, joining the Enigma team only for brief periods [3].

In November 1942, Turing went to the US, where (among other assignments) he worked at Bell Labs on speech encryption. There he met Claude Shannon, whom we will cover in Part 3. Turing returned to the UK in March 1943 and continued to work mostly on automatic speech encryption until the end of the war.

ACE: The first general-purpose computer design (1945–1947)

Although Turing’s 1936 paper On Computable Numbers described a universal computer machine, it was not intended as a practical design for a computer. In October 1945, Turing joined the Mathematics Division of the National Physical Laboratory (NPL) to work on such a design. In late 1945, Turing produced a technical report titled “Proposed Electronic Calculator,” which specified the design of such a machine in great detail (see image below). In this report, he also raises the topic of machine intelligence, describing what would become a “classic” AI problem:

Given a position in chess the machine could be made to list all the ‘winning combinations’ to a depth of about three moves on either side. This…raises the question ‘Can the machine play chess?’ It could fairly easily be made to play a rather bad game. It would be bad because chess requires intelligence. We stated at the beginning of this section that the machine should be treated as entirely without intelligence. There are indications however that it is possible to make the machine display intelligence at the risk of its making occasional serious mistakes. By following up this aspect the machine could probably be made to play very good chess.

When a pilot version of this machine, which was eventually named the Automatic Computing Engine (ACE; in homage to Babbage), was built in May 1950, it was the world’s fastest computer, running at 1 MHz [3, p. 367]. Another computer, built using the same basic principles as Turing’s ACE design, is the Bendix G15, considered by some to be the first personal computer.

When Turing was designing the ACE, machine intelligence was at the forefront of his mind, as seen in a letter to W. Ross Ashby (see image above), circa 1946:

In working on the ACE I am more interested in the possibility of producing models of the action of the brain than in the practical applications to computing…

…It would be quite possible for the machine to try out variations of behaviour and accept or reject them in the manner you describe and I have been hoping to make the machine do this…The ACE is in fact analogous to the ‘universal machine’ described in my paper on computable numbers. This theoretical possibility is attainable in practice…Thus, although the brain may in fact operate by changing its neuron circuits…. we could nevertheless make a model, within the ACE, in which this possibility was allowed for…

On 20 February 1947, Turing gave a lecture on the ACE to the London Mathematical Society. We know something about this lecture through a draft copy, which is available from [5] and has been reprinted in several volumes (e.g. [7] and [3]). Not only is this the first known public lecture to describe machine intelligence, Turing also clearly articulated the concept of machine learning:

It has been said that computing machines can only carry out the processes that they are instructed to do…It is also true that the intention in constructing these machines in the first instance is to treat them as slaves, giving them only jobs which have been thought out in detail…Up till the present machines have only been used in this way. But is it necessary that they should always be used in such a manner?

Let us suppose we have set up a machine with certain initial instruction tables [i.e. programs], so constructed that these tables might on occasion, if good reason arose, modify those tables…Possibly it might still be getting results of the type desired when the machine was first set up, but in a much more efficient manner…It would be like a pupil who had learnt much from his master, but had added much by his own work. When this happens I feel that one is obliged to regard the machine as showing intelligence. As soon as one can provide a reasonable large memory capacity it should be possible to begin to experiment on these lines…What we want is a machine that can learn from experience.

(Boldface added by me for emphasis.)

Mostly due to internal politics and mismanagement of the ACE project by NPL, it took until May 1950 to complete a ‘pilot model’ of the ACE.

The Cambridge Sabbatical (1947–1948)

In July 1947, while still employed by the NPL, Turing left for a twelve-month sabbatical at University of Cambridge. The purpose of this leave is best expressed by the Director of NPL, Sir Charles Darwin [5]:

[Turing] wants to extend his work on the machine still further towards the biological side. I can best describe it by saying that hitherto the machine has been planned for work equivalent to that of the lower parts of the brain, and he wants to see how much a machine can do for the higher ones; for example, could a machine be made that could learn by experience?

After returning from this sabbatical, Turing produced a report titled, “Intelligent Machinery.” This is a highly original work, introducing ideas such as genetic algorithms and neural networks (what he called “unorganised machines”) with learning capabilities, and reinforcement learning. Rather than giving a detailed summary, I will just quote Turing’s own summary that appears at the end of the report:

The possible ways in which machinery might be made to show intelligent behaviour are discussed. The analogy with the human brain is used as a guiding principle. It is pointed out that the potentialities of the human intelligence can only be realised if suitable education is provided. The investigation mainly centres round an analogous teaching process applied to machines. The idea of an unorganised machine is defined, and it is suggested that the infant human cortex is of this nature. Simple examples of such machines are given, and their education by means of rewards and punishments is discussed. In one case the education process is carried through until the organisation is similar to that of an ACE.

Turing never published the report, but again it is available via [5] and several anthologies (i.e. [8] and [3]).

The Manchester Years (1948-1953)

Frustrated with the slow progress, Turing left his position at the NPL in 1948 to join the Computing Machine Laboratory at the Victoria University of Manchester.

The Computing Machine Laboratory was set up by Max Newman (another veteran of Bletchley Park) in 1946. When Turing joined, they were deep into the development of the “Manchester Baby,” which became the world’s first stored-program computer when it ran its first program on 21 June 1948. Although I won’t go into the details of Turing’s contributions to the early computers developed in Manchester, it should suffice to say that his influence was substantial.

Turing’s most famous achievements during the period of 1948–1953 were a series of articles and public lectures on the topic of machine intelligence, including the article “Computing Machinery and Intelligence,” published in 1950 in the leading philosophy journal Mind. In this article, the famous “imitation game” was proposed, which is now known as the Turing Test:

I propose to consider the question, ‘Can machines think?' This should begin with definitions of the meaning of the terms ‘machine’ and ‘think’…Instead of attempting such a definition I shall replace the question by another, which is closely related to it and is expressed in relatively unambiguous words.

The new form of the problem can be described in terms of a game which we call the ‘imitation game’. It is played with three people, a man (A), a woman (B), and an interrogator (C) who may be of either sex. The interrogator stays in a room apart from the other two. The object of the game for the interrogator is to determine which of the other two is the man and which is the woman…The interrogator is allowed to put questions to A and B…

After giving some examples of the type of questions the interrogator might ask, Turing continues:

We now ask the question, ‘What will happen when a machine takes the part of A in this game?’ Will the interrogator decide wrongly as often when the game is played like this as he does when the game is played between a man and a woman? These questions replace our original, ‘Can machines think?’

The remainder of the article describes various possible objections to the idea that machines can think and Turing’s rebuttals to these objections. There is also a chapter on “Learning Machines,” which is largely an expansion on the ideas he outlined in his 1948 article. He closes the article with an insightful observation:

We can only see a short distance ahead, but we can see plenty there that needs to be done.

In 1951, gave several radio presentations broadcast by the BBC, including one titled “Intelligent Machinery, A Heretical Theory” and a second one titled “Can Digital Computers Think?” In 1952, he participated in a panel discussion broadcast on BBC with the topic “Can Automatic Calculating Machines Be Said To Think?”. Other panelists included Richard Braithwaite (a philosopher), Geoffrey Jefferson (a neurosurgeon), and Max Newman.

During this time period, Turing also became interested in the topic of Artificial Life. This is a bit beyond the scope of the current article, so I won’t attempt to summarize his work in this area. In 1953, he published a short essay titled “Chess”, culminating several years of effort in programming a computer to play chess. Finally, in 1954, Turing published his last article, titled “Solvable and Unsolvable Problems,” published in Science News, a journal that popularized science. The purpose of this article was to present his earlier and perhaps most lasting contribution On Computable Numbers to a general audience.

Turing’s Trial and Death

In 1952 Turing was brought to trial and convicted of homosexual acts, which were a criminal offense in the UK at that time. On 8 June 1954, Turing was found dead in his home. Although the death was ruled a suicide, this finding has been the subject of controversy [9].

Despite his short life and career, Turing’s impact on human and computer ingenuity is lasting. I’d like to close with a quote of Turing from an article in The Times, dated 11 June 1949:

This is only a foretaste of what is to come, and only the shadow of what is going to be. We have to have some experience with the machine before we really know its capabilities. It may take years before we settle down to the new possibilities, but I do not see why it should not enter any of the fields normally covered by the human intellect and eventually compete on equal terms.

References

[1] Turing, A. M. (1937). On computable numbers, with an application to the Entscheidungsproblem. Proceedings of the London mathematical society, 2(1), 230–265. (Note: Technically, the paper was published in 1937, but during this era papers were often literally read aloud at meetings and would later be published in a proceedings of paper form. This paper was read on 12 November 1936.)

[2] Hodges, A. (n.d.). Computable Numbers and the Turing Machine, 1936. The Alan Turing Internet Scrapbook, Retrieved from http://www.turing.org.uk/scrapbook/machine.html.

[3] Copeland, B. J. (Ed.). (2004). The Essential Turing: Seminal Writings in Computing, Logic, Philosophy, Artificial Intelligence, and Artificial Life plus The Secrets of Enigma. Oxford: Clarendon Press.

[4] Turing, A. (1948). “Intelligent Machinery”. In B. J. Copeland (Ed.). (2004). The Essential Turing: Seminal Writings in Computing, Logic, Philosophy, Artificial Intelligence, and Artificial Life plus The Secrets of Enigma. Oxford: Clarendon Press.

[5] Copeland, J. (n.d.). AlanTuring.net: The Turing Archive for the History of Computing, Retrieved from http://www.alanturing.net/.

[6] King’s College (Cambridge). (n.d.). The Turing Digital Archive, Retrieved from http://www.turingarchive.org.

[7] Carpenter, B. E. and Doran R. W. (1977). The other Turing machine. The Computer Journal. Vol. 20(3).

[8] Evans, C. R. and Robertson, A. D. J. (Eds.). (1968) Key Papers: Cybernetics. London: Butterworths.

[9] Wikipedia contributors. (2018, March 15). Alan Turing. In Wikipedia, The Free Encyclopedia. Retrieved 20:08, March 20, 2018, from https://en.wikipedia.org/w/index.php?title=Alan_Turing&oldid=830576161

Other references not cited:

Copeland, B. J. (2005). Alan Turing’s Automatic Computing Engine: The Master Codebreaker’s Struggle to Build the Modern Computer. New York: Oxford University Press.

Copeland, B. J. (Ed.). (2012). Alan Turing’s Electronic Brain: The Struggle to Build the ACE, the World’s Fastest Computer. New York: Oxford University Press.

Copeland, J., Bowen, J., Sprevak, M., & Wilson, R (Ed.). (2017). The Turing Guide. New York: Oxford University Press.

Katsuhiko, S. and M. Sugimoto (2017). “From Computing Machines to Learning Intelligent Machines: Chronological Development of Alan Turing’s Thought on Machines”. In Understanding Information: From the Big Bang to Big Data. (A. J. Schuster) Cham: Springer Nature.

What is Artificial Intelligence? Part 2 was originally published in TDS Archive on Medium, where people are continuing the conversation by highlighting and responding to this story.

Ok, perhaps that is a bit too harsh. Of course, people using the term AI have some idea of its meaning, but I have noticed in many cases that the usage by non-experts (and even some supposed experts) strays quite far from how the term was originally intended in academic circles.

Nobody really knows who coined the term “artificial intelligence”. Many people attribute it to the belated computer scientist John McCarthy, but according to interviews conducted by Daniel Crevier, McCarthy denies having come up with the term [1, p. 50].

Nonetheless, the term AI became prevalent in use in the mid-1950s, albeit within a small circle of primarily American scientists. A synonymous term, machine intelligence, seems to have originated in the UK, and it is still in use today, thought not as widely as AI. In this article, I mainly use the term AI to refer to the concept of “thinking machines,” except where I’m discussing authors who consistently used “machine intelligence” instead.

Overall, this article aims to shed light on the concept of AI from a historical perspective and briefly trace its development up to the modern era. As a disclaimer, I need to point out that I’m not a professional historian. There are likely many missing episodes in this brief history, but I will try to make up for this by citing numerous more authoritative references.

I intend to publish the article in several parts. In Part 1, I focus on ideas leading up to the concept of machines that can think. Part 2 will focus on developments in the UK between 1936 and roughly 1954. Part 3 will continue the story from around 1954 when focus started to shift to the US.

A Brief History of Early Artificial Intelligence

In order to understand what artificial intelligence is, we look to the past to see what great thinkers have thought was possible for machines to do. I have divided this early history of AI is divided into three periods: “From Automata to the Analytical Engine,” “From Turing Machines to Checkers,” and “From the Logic Theorist to Self-Writing Programs”. This takes us up to roughly the end of the 1950s, a point at which we can say artificial intelligence is a firmly-rooted scientific discipline.

From Automata to the Analytical Engine

Some authors trace the idea of thinking machines back to ancient Egypt or ancient Greece (e.g. [2]), but personally I don’t find these examples, such as the conversation between Socrates and Euthypro about a standard of piety, as having much to do with thinking machines or artificial intelligence (see [3]). Another example is Talos, the figure made of bronze from Greek mythology [4]. As he was supposedly made by the Greek God Hephaestus, he hardly qualifies as an artificial creature.

Based on my limited research, I think the idea of thinking machines can be traced back to the 17th century [5]. In particular, René Descartes published Discourse on the Method in 1637 (where the famous phrase, “I think, therefore I am” comes from), in which he conjectured that it would be impossible to create a machine that could reason like humans [6]:

“…it must be morally impossible that there should exist in any machine a diversity of organs sufficient to enable it to act in all the occurrences of life, in the way in which our reason enables us to act.”

In fact, in this same passage he outlined what amounts to the Turing Test, a topic we will turn to later:

Nor will this appear at all strange to those who are acquainted with the variety of movements performed by the different automata, or moving machines fabricated by human industry, and that with help of but few pieces compared with the great multitude of bones, muscles, nerves, arteries, veins, and other parts that are found in the body of each animal…

…but if there were machines bearing the image of our bodies, and capable of imitating our actions as far as it is morally possible, there would still remain two most certain tests whereby to know that they were not therefore really men. Of these the first is that they could never use words or other signs arranged in such a manner as is competent to us in order to declare our thoughts to others: for we may easily conceive a machine to be so constructed that it emits vocables, and even that it emits some correspondent to the action upon it of external objects which cause a change in its organs; for example, if touched in a particular place it may demand what we wish to say to it; if in another it may cry out that it is hurt, and such like; but not that it should arrange them variously so as appositely to reply to what is said in its presence, as men of the lowest grade of intellect can do.

The second test is, that although such machines might execute many things with equal or perhaps greater perfection than any of us, they would, without doubt, fail in certain others from which it could be discovered that they did not act from knowledge, but solely from the disposition of their organs…

The automata that he refers to are essentially mechanized imitations of various living things. Such automata are reported to have been built throughout antiquity, but they hardly qualify as machines mimicking human thought.

One exception is Wolfgang von Kempelen’s Great Chess Automaton, which was built in 1769. It was designed to play chess, and play it well. It even beat Benjamin Franklin! Excelling at chess is certainly a capability most would agree requires intelligence. The only problem is that von Kempelen’s chess automaton turned out to be a fake. There was actually a person inside the contraption! This wasn’t revealed until 1837, nearly seventy years after its creation [7].

This example demonstrates, however, that the idea of thinking machines was certainly alive and well by the 18th century, and some were willing to try to prove Descartes wrong (if they were even aware of his assertion). In reality, nobody seemed to have any idea how to make thinking machines, but at least the dream was there.

A giant leap forward was taken in the first half of the 19th century, when inventor Charles Babbage proposed the Analytical Engine. It is widely regarded as the first design of a general-purpose computer. Although the Analytical Engine has never been fully built, many consider it to be influential in pushing forward the dream of mechanized general-purpose computation.

In 1843, Ada Lovelace translated and annotated a paper by Luigi Menabrea, which was a description of Babbage’s Analytical Engine, based on lectures Babbage gave in Turin in 1840 [8]. Nobody seems to have thought that the Analytical Engine was indeed a thinking machine:

for the machine is not a thinking being, but simply an automaton which acts according to the laws imposed upon it.

Lovelace adds in her notes:

The Analytical Engine has no pretensions whatever to originate anything. It can do whatever we know how to order it to perform. It can follow analysis; but it has no power of anticipating any analytical relations or truths. Its province is to assist us in making available what we are already acquainted with.

Despite the Analytical Engine not being regarded as a thinking machine, it was a major step forward in the design of a computer that could carry out complex mathematical computations. This would prove to be inspirational for other computing pioneers, such as Alan Turing, who we will turn to next.

Continue to Part 2.

References

[1] Crevier, D. (1993). AI: the tumultuous history of the search for artificial intelligence. New York: Basic Books.

[2] Haack, S. (n.d.). A BRIEF HISTORY OF ARTIFICIAL INTELLIGENCE. Retrieved from https://www.atariarchives.org/deli/artificial_intelligence.php.

[3] Plato. (ca. 399 BC) Euthyphro. Retrieved from http://www.gutenberg.org/ebooks/1642.

[4] Talos. Retrieved from https://www.greekmythology.com/Myths/Creatures/Talos/talos.html.

[5] Artificial Intelligence. (2010) Did you know?. Retrieved from https://didyouknow.org/ai/.

[6] Descartes, R. (1637). Discourse on the Method of Rightly Conducting One’s Reason and of Seeking Truth. Retrieved from http://www.gutenberg.org/ebooks/59.

[7] The Great Chess Automaton. (n.d.). The Museum of Hoaxes. Retrieved from http://hoaxes.org/archive/permalink/the_great_chess_automaton.

[8] Menabrea, L. F. (1843). Sketch of the Analytical Engine invented by Charles Babbage , Esq. (Trans. A. Lovelace). Scientific Memoirs. Vol. 3. Retrieved from https://www.fourmilab.ch/babbage/sketch.html (Original work published in 1842).

What is Artificial Intelligence? Part 1 was originally published in TDS Archive on Medium, where people are continuing the conversation by highlighting and responding to this story.

On September 5, 2016, I made a drastic life decision: I would make one last post to Facebook, and then I would never make another post again. If you happen to be on Facebook yourself, you can see what that ultimate post was, but basically it was an announcement that I wouldn’t be posting to Facebook ever again, along with a link to my blog, where I explained the reasons for this decision.

I won’t bore you with the details of my reasons (that’s what my blog is for!), but in short, Facebook had become an addiction for me, and I decided it was best to quit cold turkey (well, almost. The true story is a bit more complicated). There were other non-personal reasons — societal reasons actually — why I felt compelled to make this move, but the most driving reason was a personal one (to make my life better).

I should perhaps add that I was an active member of the Facebook community since its primordial phase in Spring 2004, and I even had some contact with Facebook’s now famous founder, Mark Zuckerburg, back in those early days. For many years, I was posting photos and cheesy jokes on a daily basis or more.

In recent years, however, I became more and more weary about what Facebook was doing to my life and those around me. Not to mention, the world in general. The problem was, I was hooked! I felt at least that it would be incredibly difficult to pull away. I would lose contact with the world! (So I thought…)

My experience — seven months down the road — is that nothing could be further from the truth. In fact, I feel more connected to the world I live in, perhaps more so than I have felt in my entire life. Most importantly, I feel more connected to the people around me, who happen to be the people I care about most (e.g. my own family). I feel I’m even in better contact with my close family members (mother, father, brother, sister, etc.) who happen to live on another continent. Instead of assuming they know what’s going on in my life via Facebook, I engage in actual conversation with them to let them know what’s going on in my life. Amazing concept!

I also spend more time with friends. Granted, this is limited mainly to friends who happen to live in the same country as me, but I’ve noticed that actually, I have more friends nearby than I would have time to spend with them. And for those friends far away, email and phone calls work great, too!

Perhaps most importantly, I’ve managed to cure my addiction. Although I still have an account on Facebook, I no longer do I feel the urge in every idle moment to tap on the blue box and scroll through endless feed of so-called *news*. I certainly don’t feel the need to check and re-check how many people have liked my recent posts. There are no recent posts.

The only time recently I felt the urge to check Facebook was on my birthday. I was sure that many friends had sent their birthday wishes there, so I went to Facebook to make sure the mail was delivered. That experience, which I also wrote about on my blog, left me feeling more sure than ever that I had made the right decision to call it quits on Facebook.

One side effect of not being addicted to Facebook is I feel less tired, less stressed, and I generally have more time and energy to think. Sometimes I just spend time alone (preferably outdoors), and I think about stuff. I no longer think something and then immediately react, “I’ve gotta share this idea on Facebook!” I let ideas simmer. I talk about my ideas with friends, one-on-one, before blurting them out across the whole world. Sometimes I even let my friends convince me my idea is a bad one.

I believe the world is better off for this. I also believe the world would be a whole lot better off if more people took a similar course of action, whether it be about Facebook, Twitter, or whatever new-shiny-thing that comes along. Does the world really need a mega-megaphone? Could we try just turning it off for a little while? Maybe cooler heads would prevail a bit more often. It seems that even Zuckerburg himself has pondered these questions.

Facebook stated mission is to “give people the power to share” and “make the world more open and connected.” But in the 21st century, in my opinion, we have more than enough powers to share, especially in the developed world. And in 12+ years of having discussions on Facebook, many of them heated debates, I highly doubt any of them made the world more open or connected.

At least in my own case, I’m convinced life is much better off without Facebook. Critics might point out that I’m still on Twitter, but I use that mainly for professional purposes nowadays, and rarely outside of business hours. My main social media are media that have been around for ages: hugs, conversations, and smiles. I’m certainly not anti-tech though, so when direct contact isn’t possible, I get by with email, FaceTime, or my new favorite communication solution Wire (yeah, open source!).

Moral of the story: If you have yourself felt that Facebook has taken over your life, but you’re too afraid to pull back, then I hope my story will give you some courage. I have noticed nothing but positive effects in my life. It wouldn’t be an exaggeration to say that I feel like a new man. A more free one. There is nothing wrong with not trying to keep up with everything in everyone’s life, and there is no shame in being more private about your own.

I stopped posting to Facebook seven months ago —  Here’s how my life has been transformed was originally published in Privacy Revolution on Medium, where people are continuing the conversation by highlighting and responding to this story.