AI Dictionary of Terms

Q: What is an AI dictionary?

An AI dictionary is a reference resource that defines artificial intelligence and machine learning terms in plain language, making complex concepts accessible to everyone.

Q: How many terms are in this AI terminology glossary?

This living dictionary contains 160+ AI and machine learning terms across 8 specialized categories, ranging from foundational Architecture to Legal AI and Healthcare AI.

Q: Who is this AI dictionary for?

It is built for everyone. Every term includes a Simple Version written in plain English for beginners, alongside detailed technical definitions and code examples.

A curated, living dictionary of artificial intelligence and machine learning terminology with 190+ definitions explained simply. Built for everyone — from complete beginners to seasoned practitioners.

Created and curated by Alex Nubla.

💡 Goal: To promote AI literacy by providing clear, accessible explanations of complex AI concepts.
Every term includes a “Simple Version” written in plain English — no jargon, no assumptions — so anyone can understand, regardless of their technical background.

🎯 Scope & Rigor: We strictly exclude “confabulations,” marketing fluff, or made-up buzzwords that lack real currency in standard AI terminology. This dictionary focuses exclusively on validated, widely recognized concepts with established technical definitions.

Browse by Category

🏗️ Architecture — 50 terms · Core model designs, neural network structures, and foundational paradigms

Activation Function · AGI (Artificial General Intelligence) · Algorithm · Artificial Intelligence (AI) · ASI (Artificial Superintelligence) · Attention Mechanism · Autoregressive · BERT · CNN · Computer Vision · Context Window · Deep Learning · Diffusion Model · Embedding · Encoder-Decoder · Expert Systems · Foundation Model · Frontier Model · GAN · Generative AI · GPT · GRU · Layer Normalization · LLM · LRM · LSTM · Machine Learning (ML) · Mixture of Experts (MoE) · Model · Multimodal · Narrow AI · Natural Language Processing (NLP) · NER (Named Entity Recognition) · Neural Network · Object Detection · Open-Weight Model · Parameter · Positional Encoding · RNN · Reasoning Model · Residual Connections (Skip Connections) · Robotics · RoPE (Rotary Position Embedding) · Spatial Intelligence · Softmax Function · Token · Transformer · Weights · XGBoost · YOLO

⚙️ Training — 40 terms · How models learn, adapt, and improve

Adam / AdamW (Adaptive Moment Estimation) · Backpropagation · Batch Size · Catastrophic Forgetting · Chain of Thought · Context Engineering · Convergence · Data Augmentation · Distillation · DPO · Federated Learning · Few-Shot Learning · Fine-tuning · Gradient Descent · Gradient / Gradient Flow · Hyperparameter · In-Context Learning · Learning Rate · LoRA · Loss Function · Optimizer · Overfitting / Underfitting · PEFT · Pre-training · Prompt · Prompt Engineering · Reinforcement Learning (RL) · Regularization · Reward Model · RLHF · Scaling Laws · Self-Supervised Learning · Supervised Learning · Synthetic Data · Training · Transfer Learning · Unsupervised Learning · Validation Set · Vanishing Gradient Problem · Zero-Shot Learning

🚀 Deployment — 26 terms · Putting models into production

Batch Processing · Beam Search · Caching · Edge Computing · Feature Store · Graphics Processing Unit (GPU) · Grounding · Inference · Inference-Time Compute · KV Cache · Latency · MCP · Model Serving · Observability · Orchestration · Prompt Injection · Quantization · RAG · Sampling · Semantic Search · Speculative Decoding · Streaming · Temperature · Throughput · Top-p Sampling (Nucleus Sampling) · Vector Database

📏 Evaluation — 8 terms · Measuring model performance and quality

Benchmarking · Cross-Validation · Data Contamination · Deterministic · Generalization Error · Hallucination · Non-Deterministic · Perplexity

️🛡️ Ethics & Safety — 18 terms · Responsible AI and risk mitigation

AI Safety · AI Slop · AI Sycophancy · AI Washing · Alignment · Bias · Data Privacy · Deepfake · Digital Labor · Ethical AI · Explainability / XAI · Guardrails · HITL · Human-Centered AI (HCAI) · Jailbreak · Parasitic AI · Responsible AI · Spiralism

🏢 Enterprise AI — 14 terms · AI infrastructure, governance, and organizational adoption

Agent · Agentic AI · AI Gateway · Chatbot · Compliance · Conversational AI · Copilot · MLOps / LLMOps · Model Monitoring / Drift Detection · Open Source · Predictive Analytics · Shadow AI · Tool Use / Function Calling · Vibe Coding

🏥 Healthcare AI — 24 terms · AI applications, standards, and regulations in healthcare delivery and clinical care

Clinical Decision Support (CDS) · Clinical NLP · Clinical Prediction Model · Clinical Trials AI · Computer-Aided Detection (CAD) · Diagnostic AI · DICOM · Digital Biomarker · Digital Therapeutics (DTx) · Digital Twin · Drug Discovery AI · EHR Integration · FDA Approval (SaMD) · FHIR · Health Informatics · In Silico Trials · Interoperability · Medical Imaging AI · PHI · Precision Medicine · Radiomics · Remote Patient Monitoring (RPM) · Regulatory AI · Wearable AI

⚖️ Legal AI — 18 terms · AI terminology, regulatory frameworks, and legal principles governing AI systems

AI Act · AI Copyright & IP · AI Forensics (Media Forensics) · AI-Generated Evidence · AI Liability · Algorithmic Accountability · Algorithmic Audit · Algorithmic Risk Assessment · Automated Decision-Making (ADM) · eDiscovery · Generative AI Disclosure · High-Risk AI System · Impact Assessment · Regulatory Sandbox · Right to Explanation · Systemic Risk · TAR (Technology-Assisted Review) · Trustworthy AI

🔤 Browse All Terms (A-Z)

Letter	Terms
A	Activation Function, Adam / AdamW (Adaptive Moment Estimation), Agent, Agentic AI, AGI (Artificial General Intelligence), AI Act, AI Copyright & IP, AI Forensics (Media Forensics), AI Gateway, AI-Generated Evidence, AI Liability, AI Safety, AI Slop, AI Sycophancy, AI Washing, Algorithm, Algorithmic Accountability, Algorithmic Audit, Algorithmic Risk Assessment, Alignment, Artificial Intelligence (AI), ASI (Artificial Superintelligence), Attention Mechanism, Automated Decision-Making (ADM), Autoregressive
B	Backpropagation, Batch Processing, Batch Size, Beam Search, Benchmarking, BERT, Bias
C	Caching, Catastrophic Forgetting, Chain of Thought, Chatbot, Clinical Decision Support (CDS), Clinical NLP, Clinical Prediction Model, Clinical Trials AI, CNN, Compliance, Computer-Aided Detection (CAD), Computer Vision, Context Engineering, Context Window, Convergence, Conversational AI, Copilot, Cross-Validation
D	Data Augmentation, Data Contamination, Data Privacy, Deep Learning, Deepfake, Deterministic, Diagnostic AI, DICOM, Diffusion Model, Digital Biomarker, Digital Forensics, Digital Labor, Digital Therapeutics (DTx), Digital Twin, Distillation, DPO, Drug Discovery AI
E	Edge Computing, Embedding, Encoder-Decoder, eDiscovery (Electronic Discovery), EHR Integration, Ethical AI, Explainability / XAI, Expert Systems
F	FDA Approval (SaMD), Federated Learning, Feature Store, Few-Shot Learning, FHIR (Fast Healthcare Interoperability Resources), Fine-tuning, Foundation Model, Frontier Model
G	GAN, Generalization Error, Generative AI, Generative AI Disclosure, GPT, Gradient Descent, Gradient / Gradient Flow, Graphics Processing Unit (GPU), Grounding, GRU, Guardrails
H	Hallucination, Health Informatics, High-Risk AI System, HITL, Human-Centered AI (HCAI), Hyperparameter
I	Impact Assessment (Algorithmic), In Silico Trials, In-Context Learning, Inference, Inference-Time Compute, Interoperability
J	Jailbreak
K	KV Cache
L	Latency, Layer Normalization, Learning Rate, LLM, LoRA, Loss Function, LRM, LSTM
M	Machine Learning (ML), MCP, Medical Imaging AI, Mixture of Experts (MoE), MLOps / LLMOps, Model, Model Monitoring / Drift Detection, Model Serving, Multimodal
N	Narrow AI, Natural Language Processing (NLP), NER (Named Entity Recognition), Neural Network, Non-Deterministic
O	Object Detection, Observability, Open Source, Open-Weight Model, Optimizer, Orchestration, Overfitting / Underfitting
P	Parameter, Parasitic AI, PEFT, Perplexity, PHI (Protected Health Information), Positional Encoding, Precision Medicine, Predictive Analytics, Pre-training, Prompt, Prompt Engineering, Prompt Injection
Q	Quantization
R	RAG, Radiomics, Reasoning Model, Regularization, Regulatory AI, Regulatory Sandbox, Reinforcement Learning (RL), Remote Patient Monitoring (RPM), Residual Connections (Skip Connections), Responsible AI, Reward Model, Right to Explanation, RLHF, RNN, Robotics, RoPE (Rotary Position Embedding)
S	Sampling, Scaling Laws, Semantic Search, Self-Supervised Learning, Shadow AI, Softmax Function, Spatial Intelligence, Speculative Decoding, Spiralism, Streaming, Supervised Learning, Synthetic Data, Systemic Risk
T	TAR (Technology-Assisted Review), Temperature, Throughput, Token, Tool Use / Function Calling, Top-p Sampling (Nucleus Sampling), Training, Transfer Learning, Transformer, Trustworthy AI
U	Unsupervised Learning
V	Validation Set, Vanishing Gradient Problem, Vector Database, Vibe Coding
W	Wearable AI, Weights
X	XGBoost
Y	YOLO
Z	Zero-Shot Learning

🔍 Search the Dictionary

📚 View All Terms

❓ Frequently Asked Questions

What is an AI dictionary?
An AI dictionary is a curated reference resource that defines artificial intelligence and machine learning terminology in plain language, making complex concepts accessible to everyone.

How many terms are in this AI terminology glossary?
This living dictionary contains 190+ AI and machine learning terms across 8 specialized categories, ranging from foundational Architecture to Legal AI and Healthcare AI.

Who is this AI dictionary for?
It is built for everyone. Every term includes a “Simple Version” written in plain English for beginners, alongside detailed technical definitions and code examples for developers and practitioners.

Is this AI dictionary free to use?
Yes! This is an open-source project licensed under the MIT License. You are free to use, share, and contribute to this machine learning glossary.

🤝 Contributing & License

Want to add a term? Check out our Contributing Guidelines. This project is open source under the MIT License.