The balance of power in the tech landscape is shifting towards artificial intelligence with IT bellwethers baking AI into existing products. But what's more crucial to the future of businesses is an AI workforce. And this starts with training the brains behind the business and setting up online program to help employees explore new roles.
While many programmers can code, they are not yet versed in machine learning. Despite the hype in the IT industry, and startups working in this space, developers or freshers aren't sure how to get started in the field of AI. Entrepreneur and product enthusiast Shival Gupta gave an interesting perspective on this - the relevance of a full stack developer will not be enough in the changing scenario and in the next two years, full stack will not be full stack without AI skills. Analytics India Magazine curates the best way to get started with learning AI.
Getting Started AI - The Most Important Skill Of The 21st Century
1.Prime Yourself For AI With These Free Books
Shival Gupta shared his experience of getting started with learning AI and emphasized how it is important to familiarize oneself with basic AI terms and approaches. This is true and a good way to prime yourself for AI would be starting with some free books. Peter Norvig and Stuart J. Russell's Artificial Intelligence: A Modern Approach.
The book deals with not just the basic AI concepts and algorithms (expert systems, depth-first and breadth-first search, knowledge representation, etc.) but also the fundamental of mathematics such as Bayesian Reasoning, First Order Logic, NL n-grams et al. For those interested in Deep Learning, here's a book
by Ian Goodfellow, Yoshua Bengio and Aaron Courville. Also, check out this free book on Logic For Computer Science
which explains the mathematical logic for understanding computer science and emphasizes algorithmic methods for solving proofs. Here's a books specifically aimed at senior undergraduate students explains the fundamentals of AI with a framework to study the design of intelligent computational agents.
2. Spruce Up The Required Math
Since both Calculus and Linear Algebra have a wide application in AI/ML techniques, it will be a good idea to learn it. AI enthusiasts argue that most of machine learning techniques can be reduced to Linear Algebra, and Calculus, such as the backpropagation algorithm for training neural networks. Also, take a deep dive in Discrete Math, Calculus (differential, integral, and multivariate), Probability and Statistics, Linear Algebra, Regression Analysis, and Stochastic Processes (Poisson processes, Markov chains, and Brownian Motion) for a career in AI/ML. You can check out these free learning resources on the web for Probabilistic Theory
, Introduction to Statistical Learning With R
& on Inference & Learning Algorithms
3. Get acquainted with Python, (C/C++) & data structures
AI practitioners believe that one can do AI/ML in any mainstream language and also non-mainstream languages. The biggest difference lies in the performance and availability of libraries / tools. For example, C++,is all set to outperform Java or Python and allows the developer to maximize the capabilities of the hardware. On the other hand, Python has a really good FFI, and is often used in conjunction with C or C++. Meanwhile, Octave/MATLAB, R, Python, C++, Java, R and a few other languages features high quality libraries, which is equally important to you depending on what you want to do. The general consensus is that one must stick to something popular, like Python that has a great toolkit/libraries.
4.There Are Many Open-source Frameworks To Start Experimenting With
Post this, one can pick a framework and implement it for basic classifications. According to developer Akash Paul, picking a framework can be a challenging task because they are all built with different purposes in mind. He cites an example: Caffe uses a declarative approach to define models in it whereas with TensorFlow, one can programmatically create and use models, even visualize and deploy them with ease across platforms. Some of the recommendations for the hardware are, buying a mighty Pascal series GPU (1060 6GB), i3, 8GB RAM and a SSD to get a minimal rig for AI workloads. Check out Nvidia's CUDA toolkit, it is a good place to start for experimentation.
5. Open A GitHub Account & Search For Popular Projects
GitHub has the world's largest collection of open source data Goal and it has a lot of resources for machine learning enthusiasts. You can also check out the most popular projects on GitHub here. Plan to do a project every month and try to finish it. For the latest AI resources and tutorials, check out this link.
6. Create Your First Chatbot
Try building your own chatbot as an AI project. Now, before you start programming a bot, know the three parts that go into the making of the chatbot - input text, sending button and output text. According to an AI practitioner, web crawlers used by search engines giant Google are the best example of advanced bot. Before you start programming bots, check out these open source platforms:
xpath: Developers use XPath expressions to select XML nodes or node-sets based on a variety of criteria.
Regex: Regular expression (Regex) is a special text string for describing a search pattern and is used for building basic chatbots
7. Free Resources
You should also open up accounts on learning sites and check out projects that put you on the steepest learning slope. There are also free AI academies such as Intel AI Academy
that provide essential learning materials, tools and technology to beginners. NVIDIA runs free self-paced labs
that provide training on latest techniques and training on how to deploy neural networks across a wide spectrum of applications.
San Francisco based Youtube AI educator Siraj Raval is on a mission to teach AI to developers. Known as more of a Youtubeur, he has a very nonchalant way of teaching and his curated videos have been thumbed down for not being too educational on popular forums. We recommend skipping this free resource.