AI Overview

What is AI?

"Algorithms enabled by constraints, exposed by representations that support models targeted at loops that tie thinking, perception and action together." - Patrick Winston, professor of AI and computer science at MIT.

Two Types

The two types of AI are Narrow AI and General AI.

  1. Narrow AI has the ability to complete a specific objective well because it is very constrained. This makes Narrow AI helpful in autonomous cars.

  2. General AI, or Strong AI, has a wider range of intelligence that results in a large variety of capabilities.

Machine Learning

AI operates using the three types of machine learning: supervised intelligence, unsupervised intelligence, and reinforcement learning.

  1. Supervised intelligence learns as a human would by using a past instance of machine learning to understand a labeled dataset.

  2. Unsupervised intelligence interprets an unlabeled dataset when humans can't. It can find patterns in a scrambled mess of data.

  3. Reinforcement learning uses a system of rewards and punishments to improve the intelligence. It consists of an agent that is being trained, an environment where the agent is trained, the state of the agent in the environment, the action performed by the agent, and the rewards which are given to the agent. The agent performs the task, gets a reward, and changes itself based on the reward.