Original Source Here
Get to know AI and ML in plain language.
What is AI?
Basically, you can call a computer or machine that can imitate our (human) intellect and behavior an AI.
AI is older than you might think.
AI is not new. It has been recognized as an academic field since 1956. What’s new is our experience with the improving abilities of AI and how we think about it.
What AI can do?
AI can learn and improve specific skills over time. This ability makes AI awesome and can be intimidating for some people. In fact, I like how we define ‘Intelligence’ for AI as the ability to continuously ‘learn’ (not how much it already knows).
What makes AI intelligent?
It’s what we call “algorithms”; it’s like the brain of AI. And we call the technology that enables AI’s algorithms to learn and improve their own skills “Machine Learning” (ML). So, you will hear AI and ML in a conversation about AI like they are twins. (But now you know that they are not.)
Specifically, AI sets up the initial rules to learn and improve the performance of a task. ML continuously learns and adjusts its own actions to improve the specified task.
Okay, how does ML learn?
We ‘train’ ML to learn. ML systems are very good at finding patterns from enormous batches of existing data (Big Data) in a short time. So, we can train ML by giving it a set of Big Data and letting it learn from the patterns it can find in the given data set. (Yes, that’s why the data set we use to train an ML is very important. And that’s why some people argue about the bias patterns in the training data set!) So, basically, ML systems learn from past events, then help us determine ‘similar’ future events based on the past they know. But they can also continue learning from our interaction with the machine interface as well.
ML-powered AI uses statistics to develop self-learning algorithms through enormous trials and errors. Just like us, but way faster.
For example, AlphaZero mastered the game of chess in four hours, while human typically takes about 10 years.
With this capability, we use ML-powered algorithms for marketing, manufacturing, medical research, speech recognition, and many other fields.
The next article will discuss different types of AI and ML.
Trending AI/ML Article Identified & Digested via Granola by Ramsey Elbasheer; a Machine-Driven RSS Bot