A sincere note to the reader of this post: This post is not from a very expert ML enthusiast. It’s the research of a beginner who understands what difficulty his fellow beings will face while beginning with something new.
Technologies like Artificial Intelligence, Machine Learning, Blockchain seems fascinating to every newcomer to the world of programming. He or She is looking master basics of any programming language and quickly jump to these technologies. But as the name sounds fascinating, It’s more painful to learn especially when You have only one Master that is GOOGLE. Because It provides You everything but what is adequate it doesn’t guide. Or it depends to the skills of searching person who can get the most cream from the master GOOGLE. It took many years to master that skill too (just kidding not more than a year if you spend lots of time on the Internet). So today, I am writing about Artificial Intelligence you could have a better understanding of each thing(not all) in the world of Artificial Intelligence.
Note: First of all let me make clear Artificial Intelligence is not Python, or it doesn’t mean that you have to go for Python if you want to master the skills of Artificial Intelligence. You can go with your own Programming Language to the world of Artificial Intelligence. Most of the AI enthusiast prefer Python for AI, because of It’s simplicity, and lot’s of libraries are available to work with AI, If You are comfortable with your language then You need to just reach out the master GOOGLE to find out right tools that work with AI in your Programming Language.
What’s the Difference between Artificial Intelligence, Machine Learning, Deep Learning or What should i Go with, Is it Artificial Intelligence, Machine Learning, or Deep Learning?
I am not going to bore you with the theoretical definition of Artificial Intelligence, Machine Learning, and Deep Learning. If You have any interest in Knowing their definition you may google it. You can see the picture, it depicts the relation that deep learning is a subfield of Representational Learning, and it is a further subfield of ML and AI. But this picture makes the idea more vague for the learner that whether, to begin with, AI, ML or DL. At a very high level You can differentiate these terms as follows:
Artificial Intelligence: The study of how human performs a task, and how the machine can mimic that behavior of Human
Machine Learning: The study of human learns to perform the task, and how the machine can mimic the behavior of Human
Deep Learning: The study of how the human brain or neurons work and mimicking that behavior within the machine.
I really like the explanation of Leo Andrade written in his Blog
Artificial intelligence is like teaching a student directly the information that you want them to learn.
Machine intelligence is like giving a student a book and allowing them to learn and process the information on their own.
The process of deep learning is the same as machine learning, except in this case the student is capable of learning from mistakes made and constantly improving.
So what to begins with and what you will be learning on the pathway?
At first, You need to define your needs?:
Whether the system you are developing has a certain and fixed set of output(may range from 100 to thousand), and certain rules can manipulate the whole system – Then go With AI. In Artificial Intelligence, you define certain rules. Here You are playing with algorithms. Basics one(Min Max, Alpha beta Pruning)
If you have to deals with data(tons of data) which is manipulated with the help of statistical rules instead and need to predict the output after building and training a model then go with ML. Here You will be learning Regression(Linear and Logistic), Neural Networks. You will be to classify the data based on the given features.
If you need to build a cognitive system(Like Chatbots) that enhance itself with a period of time and you have high computing systems incorporated with GPUs and CPUs then go with deep learning(But you must have basic Knowledge of Neural Networks and Regression)
Conclusion: If you are a beginner with programming first grasp some skills of Programming(and remember AI is not Python, so you can begin with any one). Secondly after having enough confidence that you can learn and build something after reading a piece of code in that language, then begin with Machine Learning (because AI according to my point of view is an advanced version of Algorithm Design)(where? , go with Andrew Ng course at Coursera as it is mostly preferred by Beginner in ML, And at some moments You feel like quitting it, But You must not. Take a break(it’s not two three hours, and neither week) and then again begin with it . As you complete this have some practice over machine learning problems participate in challenges on Kaggle. Then begin with deep learning. Happy Learning and One more tip begin with these technologies as soon as you can(earlier bachelors or even if you can begin with in school) and don’t dwell too deep in any technology. You need to just understand the core and as you implement one then you will be master with time.