Artificial Intelligence: The Technology That Will Change The World

(A Quick Introduction)

When most people hear the term Artificial Intelligence, they immediately think of futuristic, Terminator-style robots who are ready to take over the world. This is not the case! Artificial Intelligence is already here, and it has been for decades.

How Long Has AI Really Been Around?

Contrary to most people’s beliefs, the concept of AI has been around for centuries. AI-first appeared in Greek Mythology with the giant, animated, bronze warrior protector named Talos, whose purpose was to protect the island of Crete. Obviously, nobody knows if this protector actually existed, but the concept of AI was present centuries ago.

Greek Warrior Talos

Fast-forwarding to the 19th century, Alan Turing published a paper in the 1950s which speculated that machines might be able to think. In 1956 the term “Artificial Intelligence” was coined by John McCarthy.

Alan Turing

Since then, AI has improved exponentially. In 1997 IBM created Deep Blue, an AI system that beat Gary Kasparov, the world chess champion, in chess. In 2016, Google DeepMind created Alpha Go, which beat the world champion Go player. Now I know what you are thinking. Next, there are going to be Robots that are taking over the world. You are right… Just kidding! This is very unlikely to occur in the near future. Scientists can barely get a robot to pour itself a glass of water, let alone conquer the world. However, these are potentially realistic outcomes, so we need to make sure that we take the right precautions and measures to avoid this from ever materializing.

What Exactly Is Artificial Intelligence?

AI might seem very complicated, but the concept is quite straight forward. AI can be summed up as the following: The process of developing a machine that can learn and behave like a human. Simple right?

The way AI works is that it analyzes large amounts of data to learn how to complete a task. This is called Machine Learning. Machine learning is the process where a machine can learn how to perform a certain task, without being specifically programmed to do so. There are generally three types of Machine Learning: Supervised Learning, Unsupervised Learning, and Reinforcement Learning.

To explain these different types of Machine Learning, allow me to present a scenario. You are at the grocery store, and your job is to sort different fruits and vegetables. Now there are a couple of ways that this could happen.

Supervised Learning

What if you were given labels? Then, all you would have to do is sort the apples with the fruits, lettuce with the vegetables, bananas with the fruits, and so on. Fairly simple, right. Data being sorted with labels. Supervised Learning is a very effective technique. However, if the machine gets trained too much, it can become brittle. If a new fruit (Piece of data) is introduced that the machine does not recognize, it will not know what to do. This is what you call overfitting.

In this example, the green line would be considered Overfitting

Unsupervised Learning

Now, what if, instead, you were not given any labels? Well, what you could do is group things together that look alike. You could sort the data based on similar features. What color is the fruit? What shape is it? What ethnicity? Machines can learn how to sort different data into different groups based on similarities. This is called clustering.

Unsupervised Learning can be preferred over Supervised because not all data comes with labels, and it can be very time consuming if humans have to label every single piece of data. Clustering also does not have the problem of overfitting.

Despite its advantages, Unsupervised Learning still has its downfalls. If the data that you are asking it to sort is too complicated, the machine will have less accuracy than the Supervised method. Unsupervised learning can also be less accurate if it is dealing with problem-solving and decision making related tasks.

Reinforcement Learning

The third way that you could complete this task would be through trial and error. You could put one object in either the fruit or vegetable group, and if you chose the correct group, then you would get a reward. If it is incorrect, then you would get negative feedback. The machine learns by repeating the process over and over, constantly learning what is correct vs. what is wrong. At the moment, Reinforcement Learning is used more for games and Robotics. It is not nearly as good for tasks such as sorting data.

Now, each of these different types of Machine Learning has advantages and disadvantages. The real magic and the things that top-level AI engineers are working on is creating machines that combine these methods. This combination of methods has resulted in the creation of Self-driving cars as well as amazing machines like Alpha Go!

Why Is AI Becoming Increasingly Important Now?

There are numerous reasons for this:

  1. Computational power. AI takes a lot of machine power if it wants to be done successfully. The recent advancements in technology, with the higher power of CPUs and GPUs, have really allowed AI to grow exponentially.
  2. There is so much more data. The whole way that Machine Learning works is that you input data to the system, and then it can process all the data and give you an output.
  3. The algorithms. Scientists have figured out how to build different types of Neural Networks that allow computers to process data much quicker.
  4. Another reason why AI is becoming so popular is that so many Universities, Governments, Startups, and tech giants like Google, Amazon, and Microsoft, etc. have all started investing in AI.

Why AI Can Revolutionize Society

AI is already used in a lot of different ways. The google search engine, Amazon Alexa, and the Netflix recommendation system are all examples of AI. AI is used to help doctors locate tumors in patients, to help astrologers find new stars and planets in the galaxy, and to help environmentalists find new ways to address Climate Change. The possibilities for using are endless.

But why is it so helpful?

The reason AI can be so useful is its ability to reduce the time it takes to perform tasks from years to minutes. An example of this could be house pricing. The process of listing prices for homes requires many steps. You must go to every single home and look at the dimensions, the square footage, the number of bathrooms, bedrooms, etc. These are all different variables that you must consider when pricing a house. This could take you a very long time if you have to price a thousand homes. Now, with AI, you could perform this in a couple of minutes. The amount of data computers can process at such high speeds is just so much quicker than humans.

Let Us Examine What The Future Might Bring

The possible applications of AI in the future are amazing. Imagine a world of self-driving cars, taxis, planes. A world where the human race will be merged with technology. A place where every single thing that we do now will be exponentially faster. AI can greatly improve the speed and efficiency of healthcare. Robots and drones will be doing tasks that are too dangerous for humans (Mining, Inspecting Faulty Structures, Pilots, etc.). We will also start incorporating AI with other technologies such as Quantum physics, Virtual Reality, and Advanced Transportation.

Moreover, this is only a small portion of the opportunities that will be created from AI. I think we can all agree that in the future, AI will be greatly beneficial to Society.

Why Learn About AI

Like I mentioned before, AI will be used in almost every single field possible. If you start learning it now, you will be ahead of the curve. When AI and Robots do start replacing jobs, you will not be the one sitting at home without one. Instead, you could be designing the algorithms that are impacting the world! Even if you do not exactly want to design AI, it will become incorporated into every aspect of Society, so even just knowing the basics will be of benefit to you.

To Recap

History

  • AI was first introduced in 1950 by Alan Turing
  • 1997 Deep Blue — 2016 Alpha Go

What Is AI

  • AI is the process of developing a machine to learn like a human
  • Machine Learning is the process where a machine can learn how to perform tasks without being programmed to do so
  • Three types: Supervised Learning, Unsupervised Learning, Reinforcement Learning
  • Supervised is with labels.
  • Unsupervised is without labels. Sorts based on Similarities
  • Reinforcement is trial and error

Why Now

  • Computational power
  • Data
  • Algorithms
  • Investment

How Can AI Help

  • Wide variety of fields
  • Medicine, Astrology, Climate Change
  • Processes Data much quicker

The Future Of AI

  • Self-driving vehicles
  • Humans will become integrated with technology
  • Robots/Drones
  • AI will be combined with other technologies
  • Quantum Physics, Virtual Reality, Space Technologies, etc…

Why You Should Learn About AI

  • It will be in every field
  • Eventually, it will replace jobs
  • Good to know

What You Can Do To Learn More

I just recently started looking into AI. However, some good sources that I recommend for you are this Udacity training course. It teaches you about Neural Networks and working with Pytorch in Python.

These are some really good Youtube Videos teach you the basics of AI:

Citations