Artificial Intelligence — The Skynet of Tomorrow?
Are the machines ready to take over?
Artificial Intelligence
Artificial Intelligence, or A. I. as we like to call it, has been a major buzzword since the advent of Machine Learning. There have been many different views on the rise of A. I. — both positive as well as negative. Some take it as the revolution that has and will completely change the way we live yet some take it as an attack on our way of living. Most of these views are influenced by science-fiction literature, movies or TV shows that we’ve seen — Terminator, Battlestar Galactica, The Rise of The Machines, Westworld and so on.
But, is the A. I. of today really on the path to become the Skynet of tomorrow?
Well, not really — at least not yet.
The most amazing things that the A. I. of today can do is recognize faces or objects, classify data, or predict some quantities — after extremely intense training that is. So, it’s safe to say that we’re far from such an Artificial General Intelligence.
Trending AI Articles:
1. How to train a neural network to code by itself ?
But, these tasks for which the systems have been trained on a good amount of data can be done very effectively.
Image Recognition
Source: www.clarifai.com
One such example can be seen in the image above. The image classification engine at clarifai.com can accurately classify an image and present it’s results as probabilities. The above image is classified as “sunset” with a probability of 0.997, “water” with a probability of 0.995 and so on.
This system is able to give such good results because it’s been trained on an enormous amount of data. That is the fascinating thing about machines. Once they learn something, they can do it amazingly fast with an unmatched accuracy. Such systems can be trained to do handwriting recognition with an accuracy of more than 99%.
But even though these systems are so good at what they do, they can’t do much else. They’ll do what they’re trained to do — classify images — and nothing else.
Autonomous Driving
Autonomous Driving
Next, we have another application of A. I. that we have seen in the real world — self-driving cars. There are a lot of different systems in this field. One of them involves just the images of the road and the actions taken by the driver to navigate that road. It’s trained like so:
A neural network is set up to receive images of the road from a camera mounted on top of the vehicle as input. And the degree of the turn of the steering wheel by the driver to navigate that road is also fed in to the network as the “required” output. After going through the process of training, the network learns how much the steering wheel has to be turned and in what direction when such roads are encountered. Thus, the network learns how to “drive” a car.
If you’d like to learn more about neural networks, you can read my previous article:
[Artificial Neural Networks and Deep Learning
Ever since the advent of computers in particular and technology in general, the idea of creating intelligent systems…becominghuman.ai](https://becominghuman.ai/artificial-neural-networks-and-deep-learning-a3c9136f2137 "https://becominghuman.ai/artificial-neural-networks-and-deep-learning-a3c9136f2137")
But, this is just an oversimplified scenario. We are all aware that driving a car is more than just knowing how to turn the steering wheel. Advanced systems have been built that use multiple inputs like radars and sonars in addition to the visual camera systems and accurately classify the environment. This includes pedestrian detection, distance and speed of surrounding vehicles, road markings and so on. All of these when used together make up a complex system that can safely and efficiently navigate the packed roadways of today.
Natural Language Processing
NLP
NLP is the field of A. I. that helps us to communicate with the machines in a way that is natural to us. It is the system behind virtual assistants like Siri, Google Assistant as well as a part of fictional systems like J.A.R.V.I.S.
Whenever we have to do any task on our computer systems, we use some form of communication: “typed commands” in a CUI environment, Point-and-Click in a normal GUI environment and Touch-and-Type in a touch-enabled GUI environment.
But, now when we have to set an alarm, instead of tapping on the alarm icon on our phones, then tapping on the “add” button, and then setting the time, we just say — “Ok Google, set an alarm at 5:30 am tomorrow”.
And it’s done. Just as if we had a personal assistant waiting on our every beck and call!
That is the power of NLP systems of today.
This is just the tip of the iceberg. A. I. in general and Machine Learning in particular have a lot up their sleeves which makes our lives more comfortable and getting things done a little easier.
But, these machines can perform the above functions only after learning how to do them. So, despite all this, A. I. is nowhere near the level of A. G. I that can effectively do the thinking that a human being can do.
So, we’re safe from Skynet — for now.
Thank you for reading.
Disclaimer: Although every care has been taken to make sure the information presented above is correct, no guarantees can be given.
