The best way to start AI
Disruptions and changes are inevitable and once in a while, some new technology always disrupts industries. Same is the case with AI, there was so much noise in earlier 90s followed by disappointment and the loss of funding which we called AI winter and now again AI has picked up the pace and it has been changing the face of many industries since last 10-12 years. There were some doubts in the between but thanks to Advance Research in Algorithms, acceleration in Computing power, Industry support in terms of funding has led to AI now becoming a mainstream technology. At the same time, there is no doubt that AI will lead the future.
Source" alt="Source" />
Everyone in the Indian and global industry is talking about AI. And since we are optimistic, we are only talking good about AI But there is one more talk happening on AI and that is "Skill for AI". I have come across many articles which have been claiming Shortage of AI workforce, Leaving behind in AI race which is kind of right. Few of article has been also questioning about Math skills of Indian professionals and student to which I think, we are not lacking in math skills. Anyway, a significant percentage of Indians are AI engineer in all these global AI-powered companies such as NVIDIA, Intel, Google etc. Now the real question is are we really lacking in skills which can support the AI? Because as much I know with basics knowledge of Function, Probability, Statistics and programming anyone can START the AI (Please remember I am only talking about the START). And once you are able to start then it's just a matter of time to become master. Now what do you think, are we lacking in Math skills for a while let's say we don't know Function, Probability, and Statistics? How hard is it for us to learn these skills in a digital era where most of the knowledge is available totally free.
If someone is telling you AI is an Anaconda and it will eat you if you will try to play with it, that person is totally wrong. Anyone can start AI with some basics knowledge of Math and programming. And even if you don't know the math basics you can learn from any platform and that too totally free with an average spending 10 to 15 hours per week. Anyone can start the AI whether you are a student at a college or you are a working professional. But starting AI at your colleges has its own advantage. (I will try to write a post especially for college students). For now, suppose you don't know anything and still want to start AI you can follow these steps.
1. Start with Math (Functions, Probability, and Statistics ): You can learn all these from Khan Academy. Statistics is important for AI. Understanding of graphs and their various use cases are extremely important. But it is easy to learn.
2. Start Programming language ( Python or Julia): You can learn Python absolutely free. You only need to search "How to learn Python" or one of the books which I am totally in love with "Python Data Science Handbook" by Jake VanderPlas. This book is available free of cost. If you are starting to practice Python, installing various libraries is bit tedious work. You might face some common errors while installing, so to not encounter such errors, Anaconda is the best environment you can use because most of the libraries will be pre-installed. If you are completely new then you can also use Julia because it is extremely easy to use and the speed is much faster then Python. The only reason everyone prefers Python for the data science is that Python has a strong community and you can find almost everything in Python.
3. Start any Machine Learning course: There are millions of free courses available on the internet. But I always suggest starting with Coursera "Machine Learning Course" because of
a. It is free,
b. The course is totally flexible with the time frame,
c. The course is designed in such a way that anyone can understand and learn very easily.
d. Andrew has explained the math behind the ML in an excellent way which means you will start understanding the fundamentals. The only one problem with this course is that they have used Octave language but by doing this course you will learn basics of the math behind AI/ML. You can start with many other online courses such as Nanodegree from Udacity, Introduction to Machine Learning from DataCamp etc. But you can learn everything without spending any single amount, of course, you have to spend some time.
Doing all these courses will take a maximum of 4 months with an average spending of 10 to 15 hours per week. And once you will finish this you will start doing further AI work by yourself.
Initially, I thought of writing about Google Cloab but I will write about Google Cloab in some another post.
If I summarize this in one line, I’d say "You can Start AI without any prior knowledge and you can start working on actual AI after 6 months”
Starting AI doesn't mean you will start building the big models of CNN, RNN etc. but you will start understanding the fundamentals. And further AI knowledge depends on your further interest.
It is a bad idea to start AI with some drop and down tools, API etc. I will also write why it's not a good idea to use such tools for starting the AI soon.