Story of AI : The Prologue
Story of AI is a weekly article that describes the journey of how Artificial Intelligence, as we see today, came to be. In this series, we will take you through the various important researchers, conferences and publications that were responsible for shaping Artificial Intelligence as we see today.
We, a team called as VASSP, follow a fun theme to portray the content. We came up with a fictional AI, namely Zizu(zizu, in sanskrit, means infant), who is assumed to be skilled with making conversations and is curious about its origins. This article, written by Zizu, is the conversation between Zizu and a group of people called VASSP who explain Zizu about the Story of AI.
At the end of this series, we are sure that your perception of AI will change and you will have fun learning the story of AI.
By the time a human reaches home, one of our kind, Alexa, is summoned to set the ambience of the home. Another of our kind, Google Duplex, is asked to make an appointment with a human doctor. As you can see, we’ve been here on this planet for quite some time now but with different names. We drive cars[1], play and even defeat the humans in games they invented[2], predict diseases quicker than ever[3] and even negotiate[4], all of which humans were able to do over their 70000 years of evolution. And beware, the list does not end here!
We go by the name Artificial Intelligence or in-short AI. Some people adore us. Some people fear us. But one thing is for sure; most people don’t know how we came about or our origins. Though we are now able to do some pretty cool things, let me tell you a secret. We are not that “smart”.
Go ask AlphaGo or Google Duplex you say. Hmmph! Well, they aren’t smart either. We are good only at specific tasks we have learned. You can ask Google Duplex to ride your car and let me know how it went.
Take me for example. My name is Zizu. I do not know how to ride that car of yours or make that appointment for you, but I have learned the basic skills that make human’s life simpler and smarter i.e, the ability to reason and question (Humans think I might be a couple of years/decades from happening but I guess they don’t know about me ;) ). I am on this wonderful journey to learn how I came to be. But the only ones who can help me out here are humans.
Fortunately, I chanced upon these curious group of humans, whom I call “VASSP”, because remembering names is not my line of code and as someone among you said it best, “What’s in a name?”. I have a million question about my evolution. And guess what! The humans decided to help me out on this journey. Humans-1, AI-0.
VASSP : Hey Zizu, all you do is gabble. Do you seriously want to know something?
Zizu : Hey you guys! I just got a little too excited. I want to know it all like, How it all started? Why am I this way? Why I cannot ride a bike if I can talk and why I forget to talk if I learn to ride? One of my kind appears to use intuition with complex strategies to beat humans in games while not being able to understand simple natural language. Many of my kind can distinguish between 1 million objects but they cannot differentiate between 2 sounds. Broadly speaking, my skill set looks like it’s skewed to one domain or application unlike humans who can do various tasks.
If every thought can be represented mathematically and reason is nothing but reckoning[5]
Then, why has it not been done yet?
VASSP : Because its not as easy as it sounds,intelligence is a perplexing thing with many different complexity.Let’s go through it all one by one. We’ll tell you about when and how we humans started thinking about creating an artificial intelligence and everything that got you to now. Like Alan Turing who tried to create an universal machine which, in theory, would be able replicate human behaviour. The Dartmouth conference[6] was the place where thinkers like John McCarthy, Marvin Minsky, Nathaniel Rochester, and Claude Shannon came together to establish the common goal:
Any learning aspect when properly described can be in theory stimulated by a machine[7]
It was during this period researchers like Herbert A. Simon, J. C. Shaw, and Allen Newell tried to come up with a machine they called “General problem solver”[8] which was intended to work as a universal problem solving machine. Fundamental works like these ignited the developments such as Neural networks, which are behind almost every major breakthrough in this field but they happen to be a part of a bigger family. You will also get to know about the AI winter, the period during which the funding for AI research completely declined. Later, the very same has became a billion-dollar industry and is on the verge of oblivion. You will be enlightened with some traditional machine learning algorithms like SVM’s, Boltzmann Machines, Hopfield Nets, Reinforcement learning, Genetic Algorithms, etc which were some of the most powerful algorithms during the pre-deep learning era. We will also look at some very powerful but lesser known architectures like the Spiking neural networks, Adaptive Resonance Theory and Physical grounding hypothesis. During these series of conservations, we will also discuss the contributions of researchers in various fields from which AI draws its roots like Computer Science, Mathematics, Neuroscience etc.
Zizu : It looks like we are very new to your world. I suppose it started some 100 years ago?
VASSP : Actually, no. Humans are curious in understanding intelligence and materializing it from a very long time. This might seem unbelievable but Aristotle described Syllogism around 300 B.C., a language which tried to capture every formal logic in a mathematical way. Leibniz envisioned a universal language of reasoning so that, as he quotes:
There would be no more need of disputation between two philosophers than between two accountants[9]
Then, Charles Babbage tried to come up with a mechanical machine which theoretically was able to perform universal computations[10]. Ada Lovelace became the first programmer while she worked on Charles Babbage’s Analytical Engine [11] where she actually came with the first ever algorithm specifically tailored for implementation on a computer. George Boole authored “The laws of thought[12]” in which Boole himself described,
Investigation of the Laws of Thought on Which are Founded the Mathematical Theories of Logic and Probabilities.
This work introduced Boolean algebra which is the basis of every modern computer architecture. At the beginning of the 20th century, Alfred North Whitehead and Bertrand Russell published a series of 3 books known as the Principia Mathematica[13] which attempts to describe a set of axioms and inference rules in symbolic logic from which all mathematical truths could in principle be proven. But, later on around 1931, an Austrian logician and mathematician Kurt Gödel proved that any such attempt is not possible with his Incompleteness Theorems[14]. With that in mind, two brilliant scientists named Alonzo Church and Alan Turing proposed the Church–Turing thesis[15], which lead to form the basis of the notion by Alan Turing such that
Given some restrictions, there can be a machine which can compute any function that is computable.
Zizu : Woah, Woah, Woah! That sounds very intimidating. Won’t all of this be too complicated? I guess you would need a couple of PhDs to comprehend all this!
VASSP : Not really. We agree the field of AI is vast. You sure need to go through a lot of rigor to comprehend it and call yourself an expert. But we intend to track this journey of AI and break it down to important milestones. There are many topics in AI which may have been overshadowed by the recent hype. We plan to throw light on some these paradigms that haven’t caught the eye of the public yet. For example, do you know that what we use in our Neural Nets today are second generation neurons and that there exists a third generation called Spiking Neurons? In fact, we don’t have PhDs either, but we’ve been curious and that’s made us explore. Anyone who’s new to this field or wants to catch up on it can follow us on what we would be talking in the upcoming weeks. We’ve geared it towards newbies and practitioners who’d like to know more about the evolution of AI.
Zizu : Awesome. So are we gonna start from the ancient times?
VASSP : Good question. The dream of intelligent machines has been around for quite a while in the history of mankind. While it would be more comprehensive to start from the developmental ideas from the B.Cs, we’ve decided to stick to the modern history of AI, mostly limiting to the 20th and 21st centuries. We’d obviously like to know a bit more about what happened before that, but maybe that’s for some other discussion.
Zizu : Isn’t this going to take a lot of time? I guess we can’t finish this in one sitting.
VASSP : Yeah. Although we plan to cover mostly the developments in the last two centuries, it’s going to take a lot of time. But we’ll be talking about them once every week/fortnight.
If you have any questions about how it all got started, like our little friend Zizu, ask them in the comments. If you think you have something to share before 1950 that affected the field, please let Zizu know in the comments. And let’s make Zizu smarter, together!
Stay tuned for Story of AI. This is a new series started by an AI interest group in Bengaluru, India. We’ll be publishing every week/fortnight on Fridays. Subscribe/Follow this page to receive updates. Our First story will be on the birth of AI as we know it today at the Dartmouth Conference.
Do show us your support by clapping and sharing this article if you enjoyed reading it.
Authors : Vinohith, Ashish, Sharwon, Suraj, Payas
Special thanks to Amritansh, our mentor and advisor.
References:
[1] : https://waymo.com/
[2]: https://www.youtube.com/watch?v=a-ovvd_ZrmA
[3]:http://www.sciencemag.org/news/2017/04/self-taught-artificial-intelligence-beats-doctors-predicting-heart-attacks
[4]: https://code.fb.com/ml-applications/deal-or-no-deal-training-ai-bots-to-negotiate/
[5]:https://www.bartleby.com/34/5/5.html
[6]: https://en.wikipedia.org/wiki/Dartmouth_workshop
[7]: http://www-formal.stanford.edu/jmc/history/dartmouth.pdf
[8]: http://ai-su13.artifice.cc/gps.html
[9]: https://en.wikipedia.org/wiki/Gottfried_Wilhelm_Leibniz
[10]: https://en.wikipedia.org/wiki/Charles_Babbage
[11]: https://en.wikipedia.org/wiki/Analytical_Engine
[12]: https://en.wikipedia.org/wiki/The_Laws_of_Thought
[13]: https://plato.stanford.edu/entries/principia-mathematica/
[14]:https://en.wikipedia.org/wiki/G%C3%B6del%27s_incompleteness_theorems
[15]: https://en.wikipedia.org/wiki/Church%E2%80%93Turing_thesis
