AI glossary
Plain-English definitions of the artificial intelligence terms you meet in tutorials, tool reviews and release notes. No jargon for the sake of it.
40 terms.
A
- AI Agent Business & Product
- A system that uses a model to plan and carry out multi-step tasks, often calling tools along the way.
- API Business & Product
- An Application Programming Interface that lets software talk to a service programmatically.
- Artificial Intelligence Fundamentals
- The field of building computer systems that perform tasks normally requiring human intelligence, such as understanding language, recognising images or making decisions.
B
C
- Computer Vision Models & Architectures
- The field of teaching computers to interpret images and video.
- Context Window Prompting & Interfaces
- The maximum amount of text, measured in tokens, that a model can consider at once.
D
- Deep Learning Fundamentals
- Machine learning that uses multi-layered neural networks to learn complex patterns directly from raw data.
- Diffusion Model Models & Architectures
- A generative model that creates images by starting from noise and progressively removing it.
E
- Embedding Data & Evaluation
- A numeric vector that represents the meaning of text, an image or another object so that similar items sit close together.
F
- Few-shot Learning Machine Learning
- Guiding a model with a small number of examples in the prompt before asking it to complete a task.
- Fine-tuning Machine Learning
- Further training a pre-trained model on a smaller, task-specific dataset so it specialises.
- Foundation Model Models & Architectures
- A large model pre-trained on broad data that can be adapted to many downstream tasks.
- Function Calling Business & Product
- Letting a model request a specific external tool or API with structured arguments.
G
- Generative AI Fundamentals
- AI that creates new content such as text, images, audio or video.
- Guardrails Safety & Ethics
- Rules and checks that keep an AI system within safe, intended behaviour.
H
- Hallucination Safety & Ethics
- When a model confidently states something that is false or unsupported by its sources.
I
- Inference Fundamentals
- Running a trained model to produce an output, as opposed to training it.
L
- Large Language Model Models & Architectures
- A neural network trained on massive text corpora to predict the next token, enabling it to generate and understand language.
- Latency Fundamentals
- The time between sending a request and receiving a model's response.
M
- Machine Learning Fundamentals
- A branch of AI in which models learn patterns from data instead of being explicitly programmed with rules.
- Multimodal AI Models & Architectures
- AI that understands and generates more than one type of data, such as text, images and audio together.
N
- Natural Language Processing Fundamentals
- The branch of AI concerned with understanding and generating human language.
- Neural Network Models & Architectures
- A computing model made of connected nodes organised in layers that learn by adjusting the strength of their connections.
O
- Overfitting Data & Evaluation
- When a model memorises training examples instead of learning patterns that generalise.
P
- Parameters Models & Architectures
- The internal numbers a model learns during training; their count is a rough measure of model size.
- Prompt Engineering Prompting & Interfaces
- The practice of designing instructions and context so a model produces accurate, useful output.
R
- Reinforcement Learning Machine Learning
- Training an agent to take actions that maximise a reward through trial and error.
- Retrieval-Augmented Generation Models & Architectures
- A technique that lets a model look up relevant documents and include them in its answer, improving accuracy and freshness.
- RLHF Machine Learning
- Reinforcement Learning from Human Feedback, a method that aligns models with human preferences using ranked examples.
S
- Supervised Learning Machine Learning
- Training with labelled examples, where each input comes with the correct answer.
- System Prompt Prompting & Interfaces
- High-level instructions that set a model's role, tone and rules before any user message.
T
- Temperature Prompting & Interfaces
- A setting that controls how random or creative a model's output is.
- Token Fundamentals
- A chunk of text, roughly a word or part of a word, that a language model reads and generates.
- Tokenization Fundamentals
- The process of splitting text into tokens before it is fed to a language model.
- Training Data Data & Evaluation
- The examples a model learns from, whose quality and coverage shape its behaviour.
- Transformer Models & Architectures
- A neural network architecture that uses self-attention to weigh the importance of every word in a sequence relative to every other word.
U
- Unsupervised Learning Machine Learning
- Finding structure in data that has no labels attached.
V
- Vector Database Data & Evaluation
- A database optimised for storing embeddings and finding the nearest vectors quickly.
Z
- Zero-shot Learning Machine Learning
- Asking a model to perform a task it was not explicitly trained on, using only a description of the task.