Reference
AI Glossary
Every important term, defined two ways, a simple intuition and a precise technical definition, with an example, related terms and a link to where it appears in the curriculum.
All terms
Agent
A model that plans and acts through tools in a loop.
Artificial Intelligence
Software that performs tasks associated with human intelligence.
Attention
A weighting of how much each token attends to others.
Chunking
Splitting documents into retrievable pieces.
Context Window
The most text a model can consider at once.
Deep Learning
Machine learning with many-layered neural networks.
Embedding
A vector representing meaning in a geometric space.
Evaluation Metrics
Numbers that measure how well a system performs.
KV Cache
Cached attention keys and values for fast generation.
Large Language Model
A large model trained to predict the next token in text.
Machine Learning
Learning patterns from data instead of hand-coded rules.
Model Context Protocol (MCP)
An open protocol connecting AI apps to tools and data.
Neural Network
Layers of simple units with learned connection weights.
Prompt Injection
Untrusted input that hijacks a model instruction.
Prompting
Steering a model by how you phrase the input.
Quantization
Storing model weights in fewer bits.
Reinforcement Learning
Learning to act from rewards through trial and error.
Retrieval
Finding the most relevant items for a query.
Retrieval-Augmented Generation
Answering with retrieved documents in the context.
Temperature
A knob for randomness in token selection.
Token
The unit of text a model actually reads.
Tool Use
Letting a model call functions to act or fetch data.
Transformer
The attention-based architecture behind modern language models.
Vector Search
Nearest-neighbour search over embeddings.