Artificial intelligence is the art of making computers that behave like the ones in movies.
What is Artificial Intelligence?
Artificial intelligence is software programs that mimic the way humans learn and solve a complex problem. These systems are different from other applications which mainly process transactions and takes decisions which are explicitly programmed. Such applications cannot learn on their own.
Artificial Intelligence Learning Paths
Even before you chalk out the learning path for AI and ML you need to have a clear understanding as to why you want to explore this domain and what are the preliminaries.
You can check some fascinating stuff and innovation taking place because of AI.
- To start with AI, you can read up on some popular ML algorithms without spending too much time on the underlying math behind these algorithms.
- You would need to build your programming skills in languages like Python to building AI and ML models using libraries like sci-kit-learn, tensorflow, keras etc.
- There are some good videos on ML on YouTube – A complete series on ML types like supervised learning, unsupervised learning etc. Its recommended that you watch an entire series rather than watching short videos.
- Try to understand the various elements involved in ML and the most important ones like feature engineering and variable reduction.
- Start building your ML models once you’re comfortable and try different problem statements and different Machine Learning
So hopefully these initial guidelines will help you to set up good foundation for AI.
AI and Deep Learning Salary Comparison
Check out an article on the salary comparison for AI and ML skills globally (ref: “Huge Salaries for ‘scarce’ A.I. Talent: And why it matters”).
Multinational tech companies with R&D centers including Google, Microsoft, and others are paying large salaries and joining bonuses – Numbers floating around range from 1.5 – 2 million rupees/annum. However, here is the caveat:
- Candidates with some experience and excellent educational background are valued
- A strong grounding in the fundamentals is essential (Just a few courses on MOOC won’t cut it)
Startups may offer higher salaries and interesting work but may not have long-term prospects (if you are looking for job-security)
What is the right path to Learn AI?
AI is not a subject. It’s a very wide field. All of the popular stuff mentioned below can be loosely interchanged with the term AI.
- Machine Learning (ML)
- Natural Language Processing (NLP)
- Autonomous Robots
- Perception and etc.
Well, the list above is not limited to what I have mentioned. It should be noted that quite a few underlying concepts appear across the domain of AI. The correct way to head into AI is to first get the basic concepts right as they appear quite frequently.
- Know some object-oriented programming language
- Know data structures and algorithms and how they are implemented – graphs are very important
- Be good at math.
What math concepts should I know to self-study AI?
- Linear algebra
- Differential calculus
- Probability and statistics
- Numerical optimization
- Differential calculus
Artificial Intelligence Learning Path
There are two ways to enter AI, the most popular one that I think is obviously the hyped up ML route. Prof. Andrew Ng’s Coursera course on ML is famous and gets you a good head-start into ML https://www.coursera.org/learn/m….. This course provides an introduction to various Machine Learning algorithms are out there and, more importantly, the general procedures/methods for machine learning, including data preprocessing, hyper-parameter tuning, etc.
Later you can switch to Prof. Geoffery Hinton’s course on Neural Networks/Deep Learning on Coursera.- Neural Networks for Machine Learning – University of Toronto | Coursera
The second way is probably to explore search problems in AI. These are some algorithms that will quickly allow you to build some AI systems that can solve a very specific problem.
Projects are a great way to learn the implementation of the concepts in AI. It’ll help you develop a profile that will allow others to realize that you are good at it and at the same time act as a motivation for you to learn further.
Bottom line – Multiple ways to enter the field. All working towards making “stuff” more optimized and (or) autonomous.
I feel lectures from reputed Institute will help you build a strong base. Try and avoid directly jumping in MOOC courses because of their own pace with stipulated time frame.
Keep in mind doing multiple courses and getting certificates will add weight to your resume but not on your knowledge.
Courses, lectures, and learning materials
My learning preference is to watch lecture videos and thankfully there are several excellent courses online. Here are few classes I liked
- The lectures from the 2012 IPAM Summer School on Deep Learning:
Graduate Summer School: Deep Learning, Feature Learning (Schedule) – IPAM
- The 2015 deep learning summer school: Deep Learning Summer School, Montreal 2015
- It’s well explained, and all of the material you need to do the course is available here: Artificial Intelligence: Principles and Techniques
- Deep learning at Oxford 2015Taught by Nando de Freitas who expertly explains the basics, without overcomplicating it. Start with Lectures 9 if you are already familiar with Neural Networks and want to go deep. He uses Torch framework in his examples. (Videos on Youtube)
- Neural Networks Classby Hugo Larochelle: Another excellent course
- Yaser Abu-Mostafa’s machine learning course: More theory if you are interested
- Read the NIPS 2015 Deep Learning Tutorialby Geoff Hinton, Yoshua Bengio, and Yann LeCun for an introduction at a slightly lower level.
If you are more into books, here are some excellent resources. Go ahead and check them out, I won’t judge
- [Deep Learning Book] (http://www.deeplearningbook.org/), MIT
- Deep Learning Book by Ian Goodfellow, Yoshua Bengio and Aaron Courville: Bit denser but never the less a great resource
- Machine Learning, Andrew Ng https://www.coursera.org/learn/machine-learning#
- Neural Networks for Machine Learning, University of Toronto, Geoffrey Hinton https://www.coursera.org/learn/neural-networks/home/info
Neural Networks for Machine Learning: Geoffrey Hinton’s class on Coursera. Hinton is an excellent researcher who demonstrated the use of generalized backpropagation algorithm and was crucial to the development of deep learning.
Other channels and forum for knowledge sharing on AI
Projects on AI
Intro to AI
TED Talks on AI
Other Good Resources
If you come across more such wonderful resources, do comment below.