The past year 2017 has been great. But as we move forward in the technology world, the word we hear more and more often in our everyday lives is AI or Artificial Intelligence. Surely giving intellect to an inanimate object seems to be a bit bizarre but it is exciting nonetheless. AI has given us a simpler way to carry out our jobs and will hopefully continue to do so in at least the next five years. It has given us our personal assistants built right into our smartphones and cars that can drive themselves at a price point that people can actually buy. But the uses are obviously much much greater than smartphones and building a car in which you can both text and drive, the goal is to make all kinds of efficient systems. For example an Astronomical centre where complex operations are carried out in the fastest way possible or an AI guided electrical power grid that can provide maximum power at minimum wastage. The possibilities here, are endless.
So how does AI actually work?
Here are the basic steps in which an AI system is built…
STEP 1. Neural Framework
If this topic has interested you, you’ve probably heard about the term “Neural Network”. A neural network refers to the architectural blueprint in which the the “brain cells” or “neurons” which will be placed in the body of the system. We have public projects like Google’s TensorFlow which is a proven architectural neural network model which can be optimized to do a variety of different tasks. It’s more like the DNA structure in an intelligent system.
STEP 2. Modification
So of course, in order to carry out different tasks, we need to optimize how the system works to get the maximum benefit. For example, we will never put powerful GPUs in a server computer, and will not put powerful processors in a computer for mining cryptocurrencies.
STEP 3. Training
Now comes the most interesting and the most difficult part of AI programming. AI utilizes machine learning. Just like a baby child, at this point, the computer has a brain but the computer doesn’t know what to do with it. It now needs training and guidance to successfully carry out the tasks it was meant to do. For example, if you ask the system what is 2+3 at this point, it won’t be able to answer. But, if you say manipulate 2 and 3 in such a way using mathematical operations such as plus, minus, multiplication and division so that the result is 5, it might be able to figure it out that addition is the answer after several reruns and optimizations for the first time. At this time, the system has gained IQ because it knows the result as well as the procedure from which it came. With more tasks and optimizations for hundreds of days, the AI is finally ready for practical use. And like living beings, it is based on learning experiences. It learns from its previous mistakes and improves upon it each time, acquiring greater and greater efficiency after each use.
Share how you feel about the post in the comment box below or drop a message on Facebook.