Skip to main content

Important AI concepts

AI is the catch all phrase for many important innovations in Computer Science. These are trickling into other parts of the world and are important to understand

Generative AI​

Machine Learning​

Machine Learning is a fundamentally different approach to software. People used to take knowledge and processes and code them directly to machines. That means that a machine could only do things it was explicitly designed to do. This is great, but doesn't address all of the intangibles in our knowledge, which we cannot measure.

In the age of Machine Learning, we can teach the computer to learn from examples and use feedback mechanisms

Reinforcement Learning​

Reinforcement Learning is a way to help AI systems optimize their result. It creates a reward mechanism, which allows the system to learn from the outcome.

Refinforcement learning loop

Transformers​

Transformers are the current state of the art in Machine Learning. The major change it makes is adopting the practice of self-attention, which allows the model to prioritize which words are most important.

This is important, because it allows the model to produce words in smaller chunks and create more precise models. This architecture is the impetus of the pretrained algorithms we saw in 2022.