AI Inference vs Training with Machine Learning: How AI Learns [Video]

AI inference and AI training are processes that work together. Let’s look at AI inference first. This is simply asking the AI a question and getting an answer.

You ask a question, and the AI gives a prediction. For AI training, we take that prediction, check for mistakes, and tell the AI where it went wrong. The AI then tweaks its internal setup to be more accurate next time.

During AI inference, the input and output are done through matrix multiplications. You ask a question, it gets turned into numbers, those numbers form a matrix, and then matrix multiplication gives you the answer. Now, in AI training, we look at the AI’s answer and see how wrong it is.

First, AI inference gives a predicted answer. Then in phase two, the training part, we compare it to the correct answer, highlight the mistake, and adjust the internal matrices.