Common Terms

Want a quick and easy way to remember important AI and ML words? This glossary explains the most common terms in simple language.

Here are the key terms you’ll see when learning about ML stacks and how AI is trained. Each word is explained in plain English, just like the rest of this knowledge base.

Common Terms & Definitions

Model

The AI system that learns patterns from data and makes predictions or decisions.

Training

The process of showing data to a model so it can learn and improve. This is where the AI actually gets smarter.

Features

Useful pieces of information extracted from raw data that help the model learn (e.g., house size, word count, pixel values).

Dataset

A collection of data used for training, testing, or evaluating a model.

Overfitting

When a model learns the training data too well, including noise and mistakes, so it performs poorly on new data.

Underfitting

When a model is too simple and fails to learn important patterns from the data.

ML Stack

The complete set of tools, layers, and steps used to build, train, and run a machine learning system from start to finish.

Deployment

Putting a trained model into real use so it can make predictions for users or applications.

Monitoring

Checking if a deployed model is still performing well over time and detecting problems like data drift.

Hyperparameters

Settings you choose before training (such as learning rate or number of layers) that affect how the model learns.

Supervised Learning

Training a model using labeled data (data with correct answers provided).

Unsupervised Learning

Training a model on unlabeled data to find hidden patterns by itself.

LLM (Large Language Model)

A very large AI model trained on massive amounts of text, used for chatbots, writing, and understanding language.

Quick Tip

Bookmark this page! You can come back anytime you see a word you’re not sure about. As you practice more, these terms will start to feel natural.

Ready to test your knowledge? Try explaining a few of these terms in your own words or search for simple examples on Kaggle.