Computer science has been able to make progress in the field of technology time and again with new inventions and creations. In every decade or so, we come across something that is well and truly unique and something which defines the next few decades to come.
Back in the day, a portable computer (what we call a laptop) was a deal changer all across the world.
A few generations ago, the internet was the biggest thing of its time. People could not have even imagined talking to each other and sharing different forms of media at the ease of a button, sitting in front of a pile of metal contraption.
But here we are in the 21st century, doing the unthinkable. If you would have told someone in the 90’s that our phones would be as powerful as some of the PC’s back then, people would call you crazy.
So what is the equivalent to all this in today’s time and age?
Over the past few years, the hot topic of discussion in the field of computer science has happened to be about Artificial Intelligence. You have probably inadvertently heard of it in movies and other form of media by the name of AI which is basically its short form.
But before we can get to the more fun part about Artificial Intelligence, you should first be introduced to its basic ideas. What is AI anyway? Is it just as simple as making robots and having them act like humans, eventually leading them to take over the world? Well sci-fi fantasies aside, there is more to it than that. Although by theory, that is just as much a possibility as anything else.
What is AI anyway?
Other than the fact that AI is an acronym for Artificial Intelligence, what are the other basic things that one should know about it? The core idea of AI is something like this.
People do a ton of different tasks on a daily basis, ranging from the simple to the overly complicated. To us humans, these tasks can be really easy to complete or very difficult. It depends on the kind of work that has been given to us.
The task of an Artificially Intelligent program or machine would be to make all of these tasks easy and trivial to complete. The way this is achieved is by teaching the machine to ‘think’ in the same way a human being would think. This may sound confusing regarding how something like this can be achieved.
After all, a computer’s thinking procedure is nothing but a bunch of 0’s and 1’s right? If so, then how can something as complex as a human being’s thinking process be emulated in a machine. The key lies in the word ‘emulation’ itself. We merely try to recreate this thinking process but there is obviously no way (theoretically) that it can be perfected to a 100%.
The General Method of Learning
The procedure used for this ‘teaching’ mechanism goes something like follows:
- A huge bunch of data is fed into the machine as a learning set.
- These data sets act as examples, something which the machine can use as reference.
- Based off of these already existing data sets, a proper prediction can be made regarding the nature of a certain entity that is not yet registered in the database.
This is basically the general idea with which AI programs are initiated and trained. As you can clearly see, greater the number of training data sets to work with, higher is the accuracy of the output given by said program.
Keep in mind however that just randomly feeding the program with huge data sets is not what we are trying to achieve. What we really want to do is feed as varied information to the machine as possible. Once we achieve the highest possible variation, we should then try to increase the frequency of each type. This way we can expand on the knowledge in both a lengthwise as well as breadthwise fashion.
Now, several different methods and structures can be used to get this into motion. After all, how does the machine process all of the information being fed into it?
It needs to be able to differentiate between all of them, and the only way to do that is by segregating the various features within the given entity.
The Clash in Concept
As mentioned right above, the main idea behind working with AI to even begin with is to try and teach machines to do what we humans can do simply. But therein lies a certain contradiction here that needs to be addressed before you can try to move on to the greater concepts.
- There are certain tasks that human beings find very difficult or resource consuming to do. For instance, the multiplication of two 7-digit numbers is something even the most talented mathematicians will struggle to perform. However, for even the most basic computers, this can be done in milliseconds.
- On the other hand, something as simple as recognizing someone’s face or voice is a trivial task amongst humans. It is the sort of operation that we do subconsciously on a daily basis. But try coding a simple program that can do this same task and you will find yourself in a world of complications. Pretty amusing isn’t it?
It is this very contradiction in the nature of problem solving that needs to be observed in order to try and really understand the problems that you are trying to solve. Most of the times, it can be found that the things that we are spending a lot of time teaching to these machines which should be better than us by theory, are the ones that are really easy for a human being to normally do.
It is this way by nature and we can draw a lot of help from it as well when trying to solve our own set of problems. There have also been plenty of research papers regarding this same topic and you should probably check out a few of them to gain more experience regarding this subject matter.
So when people are going through all of this trouble, just to create something similar to human beings but not exactly, what do they have in mind? What are the different applications that we can enjoy when we have finally perfected the art of AI? Hell, surely it does not need to be at a 100% to enjoy to the fullest does it?
Here are some of the major ones:
- Image Processing/Recognition:
Using AI, it is possible to train the machine into learning from millions and millions of images. With the help of these images, the machine will become capable of recognizing a new image and classify it into its corresponding category. This is very useful in camera systems where knowing the identity of the person being recorded is the main objective.
- Speech Recognition:
The same also applies for recognizing speech too. There are some smart systems already that are capable of detecting language, voice and even accents. It can also deal with background noise, different dialects, and other disturbances and so on. This mechanism can be used for identifying an individual as well, which is very useful in security systems and the like.
- Visions Systems:
These are intelligent systems that are capable of understanding and interpreting spatial input in and around them. Given a certain frame of reference, such a system would be able to detect the relative positions of each and every point within said reference. These vision systems in turn are useful in a variety of different places like:
- Map making uses a variety of different vision systems attached to airplanes that can scan the area in and around it. This helps interpret spatial information in a more visual way that everyone can read and understand.
- It is also used by doctors to scan the insides of patients and observe the area within. This helps them conduct their operations in a more precise way, such that treatment can be meted out appropriately.
- Forensic artists also find an application for such systems in the recognition of faces of criminals from their sketched portraits. There are plenty of software available that can do this.
- Intelligent Robots:
This obviously refers to the type of machines that you usually see on television in sci-fi movies which tell you every bit of information starting from the temperature of the room to the status of the outside pressure. Obviously, real life technology has a long way to go before it can hope to match up to things like that, but the idea is there.
- Game AI:
Quite possibly one of the most entertaining applications of AI is in the field of games. Teaching a computer how to play a game and then to play against it, is one of the most satisfying things to do. Chess is possibly the game where the most resources have been thrown to make an AI that is well and truly worthy of being on the same level as a human.
The Benefits & Risks Involved
It is obvious that the concept of AI has a huge amount of potential to be explored. There are plenty of benefits that we will be able to experience for ourselves once we have perfected this technology. This is the driving force that is making our research go in directions that we never thought would be possible. However, there are some risks involved as well which we cannot ignore. Let’s discuss some of them.
- The obvious advantage of using AI in our daily lives (which we already do to an extent) is to make things easier for us. If machines are capable of doing some of the things that we already can, then using them as an alternative will make things much faster and simpler for us.
- Because of the methodical nature of machines in general, they tend to be more accurate in their calculations and working methods. As such, the efficiency for machines is something that will always be higher than that of an average human being. Thus, we save a lot of time.
- There may even be problems that humans are not capable of solving altogether. If trained and programmed right, some machines will become capable of doing these problems, that too in the most time efficient and resource efficient manner possible.
- With the trivial tasks handed over to AI, we can concentrate our research on other important fields as well. This will help us increase our quality of life and overall lifespans in general. The entire human race will profit from it.
- If an AI oriented machine is capable of doing all the work that a human being already can, there would be no need for us to work anymore. As such, jobs would presumably become a bit scarcer. Or at the very least, people would need some time to adjust themselves.
- In time, AI may become so advanced that their method of problem solving becomes misaligned from our line of thought completely. This would lead to less than optimal solutions for us humans, and that is something that needs to be avoided.
- AI can also be misused by those who mean harm to do a lot of malicious things. Terrorism is already a concern in the modern world. Giving them these ‘toys’ to play around with will make things even worse.
No matter how many risks involved, by nature we humans will continue to explore this field time and again. As such, AI is a platform that is being researched a lot in today’s times. There is not a single piece of technology out there that does not use AI in some form or another. So the best you can do is to arm yourself with knowledge regarding how it all works.
Nancie L Beckett happens to be well informed and well trained personnel in the field of computer technologies. Although self-trained for the most part, she keeps up with all of the latest trends in the current market. With an MBA degree from the University of Massachusetts, her experience of 6 years is sure to provide you with some really insightful pieces of information.