Machine Learning

• A way to learn from data
• The goal is a set of parameters that can be used to predict another thing
• Makes predictions based on data used to train the model

Inferencing

• Inferencing is when a model is making preciction on new data

• Real Time

• Favors speed over accuracy

• E.g., Chatbots

• Batch

• Speed of the result is usually not a concern

Edge computing

• Edge devices are usually device with less computing power that are close to where the data is generated
• Small Language Models (SML) are used on edge devices
• Large Language Models (LLM) are used on remote powerful servers

Approaches

Regression

• Linear regression parameters
• slope (a) and y-intercept (b)
• $y = ax + b$

Classification

Deep Learning