Artificial Intelligence and its subsets

Artificial Intelligence (AI):

As the name suggest an Intelligence which is Artificial, unlike the natural Intelligence of human beings. So it is basically a Computer Science and Technological advancement that human have made to enable computers/machines to think, act, behave, perform tasks and/or adapt reasonably and rationally like we humans do. And when we say AI the first think that comes in mind is Siri, Cortana, Google Assistant, Alexa, Robots, chatbots, rockets, self-driving cars and etc. AI is a very broad term, say like universe. So when we see a machine performing a task which a human may or may not be able to perform is all under AI. So when we think of our desktop and laptop we can consider it too as an AI.

Now let us understand more about this AI concept. As mentioned earlier there is a lot of advancement in Computer Science and Technology and this advancement has also lead to an advancement of generating lots and lots of data, so much data which we human cannot handle, but of course we need this data. This data is very important which can help us — to derive some useful information that can give some meaningful insights for an informed conclusion or decision-making or it is useful in adding some more information into already existing information. Also with lots of data has also lead to lots of information and with information came lots of questions — what, how, why, when, what-if, what-else and etc. Now the problem has arisen not only how to derive meaningful information from this lots and lots of data but also how to find the answers to our questions from this data. And as usual we took the help of Artificial Intelligence i.e. computers, so when there is a problem there is a computer for us to solve it and so came into picture Machine Learning.

And Machine Learning is nothing but as the name suggest the ability of machines / computers to learn from this lots of data and derive some meaningful insights for us and to answer our questions. And the machine learns and does this with the help of some pre-defined Algorithms without any need of more human intervention in this process.

Machine Learning

A subset of AI and in Machine Learning we have:

1. Supervised Machine Learning: Now in these lots of data we knew some answers but we needed more answers based on these answers we knew from the data that was continuously generated. And as mentioned earlier for our every problem there is a computer for us and so Supervised Machine Learning Algorithms is nothing but where the machine/computer learns from these data and the answers we knew on that data and then predicts for us the answers for more data that is continuously generated.
2. Unsupervised Machine Learning: Now in these lots of data we wanted to find some meaning, some depth and some insights which we actually cannot find ourselves or maybe which we know but we are unable to picture out or word out or segregate that thing from these data and so here Unsupervised Machine Learning Algorithms comes which is nothing but where the machine/computer learns from these data and predicts some meaning, information and insights in this data for us use it.
3. Semi-Supervised Machine Learning: Now unlike what is mentioned in Unsupervised Machine Learning, here may be we have some answers or may be the answers we know we are not sure whether they are the actual answers. And so, here Semi-Supervised Machine Learning Algorithms is the rescue factor which enables machine/computer to learn from these data and predicts that more rational answers for us to use.
4. Reinforcement learning: Now as we are prone to make errors so does the machine/computers too, as Algorithms are nothing but something that we humans have made for machine/computers to learn from data independently without our intervention to give desired predictable outputs for us to use. And cases may arise where the output the machine gives after learning from data may or may not be logically, reasonably or rationally right and so this are errors. Reinforcement learning Algorithms is nothing but where the machine/computer learns from the data makes prediction and if the prediction is right it is rewarded and if it is wrong then it is punished and further on bases of these rewards and punishments it learns more to give more and more right output/answers, which we are finding.

Artificial Neural Networks

Another subset of AI and comes within Machine Learning.

Artificial Neural Networks: Now since with the huge amount of data, there also arise complex computations in our quest to get fast more better and liable output, which are logically, reasonably and rationally sound. So Artificial Neural Networks came in which works like the way human brain processes information and its algorithm is based on same functionality as how the neurons function in a human brain. And this algorithm can be used in all the situations we mentioned earlier — that is Supervised, Unsupervised and Reinforcement.

Deep Learning

Another subset of AI and comes within Artificial Neural Networks which itself is within Machine Learning.

Deep Learning: But then with the increase in the complexity of computations and more depth required into the data Deep Learning Algorithms is the solution. And Deep Learning Algorithms is nothing but adding something more in the Artificial Neural Network, making it more enhance with more capabilities and thus making it Deep Neural Network.

And saying this all, yet our journey has just started with Artificial Intelligence and its subsets and there are a lots more to come and it is waiting for to give us an unbelievable and enriching life experience far beyond any expectations or anything we can think of.

This is what I have learnt during my journey of learning Machine Learning and Deep Learning. Please do leave your valuable feedback below and give me a chance to improve.