Skip to content

Back to All concepts

The Concept Library
AI & Engineering

AI & Engineering concepts

Machine learning, software, systems, and technical practice.

70 explained· 242 coming soon · See alsoResearch Product Psychology

Start with these five

Responsive Design
Design Tokens
Pareto Principle
Retrieval-Augmented Generation
Mixture of Experts

AI & Engineering

312concepts

Machine learning, software, systems, and technical practice.

70Explained · every one with checked sources

New
Responsive Design
A way of building one website whose layout, images and text adjust to fit any screen width, using fluid grids, flexible images and media queries.
Design Tokens
Named values that store a design system's colours, type, spacing and motion once, so every platform reads the same decision from one source.
Pareto Principle
The Pareto Principle says that in many systems a small share of causes, often about 20 percent, produces most of the results, often about 80 percent.
Retrieval-Augmented Generation
A method that answers a question by searching a document collection first, then having a language model write the answer from the passages it found.
Mixture of Experts
A model design that stores many sub-networks, called experts, and runs only a few of them per input, so capacity grows faster than compute cost.
Prompt Injection Attacks
Prompt injection is an attack that hides instructions inside text an AI system reads, causing it to follow the attacker's commands instead of its own.
Reinforcement Learning from Human Feedback
Reinforcement Learning from Human Feedback (RLHF) trains a language model to produce the responses human raters prefer, using a reward model built from their comparisons.
Pruning
Pruning is removing weights, neurons, or layers from a trained neural network to shrink it while keeping accuracy close to the original.
Quantization
Quantization stores a neural network's weights and activations using fewer bits, shrinking memory use and speeding up inference at some accuracy cost.
Distillation
A technique where a small model is trained to copy the behavior of a large one, keeping most of the accuracy while being cheaper and faster to run.
New
Foundation Model
A large AI model trained once on broad data so it can be adapted, with prompts or extra training, to many different tasks.
API
An agreed set of rules that lets one program ask another to do something, without knowing how the other works inside.
Deployment
Deployment is moving a built version of software onto the machines where it runs, then configuring it, starting it and checking it works.
Latency
The time between sending a request to a system and receiving its response, summed over network travel, waiting and processing.
New
Guardrails
Checks placed around an AI model that block, change or flag the inputs and outputs that break a set of rules.
Adversarial Attack
A deliberately crafted input, often changed imperceptibly, that makes a machine learning model give a wrong answer.
AI Safety
The field that works to prevent or reduce harm caused by AI systems, from misuse, from accidents and from loss of control.
Agent Loop
The cycle in which an AI agent decides, runs a tool, reads the result and decides again, until the task is done or a limit is reached.
Agentic Workflow
A process in which a language model completes a task in several steps, planning, using tools, checking and revising its work.
AI Agent
A system in which a language model chooses its own steps, calls tools, reads the results and repeats until a goal is met or a limit stops it.
Cross-Validation
A way to estimate how well a model predicts unseen data by training and testing it on rotating splits of one dataset.
Bias-Variance Tradeoff
The pattern that a model's prediction error splits into bias from being too simple and variance from being too sensitive to its training sample, which move in opposite directions.
Human Evaluation
Judging a system's output by asking people to rate, rank or mark errors in it, instead of computing a score automatically.
Evaluation Set
Labelled data a machine learning model has never trained on, used to score how well it performs on new examples.
Benchmark
A fixed task, dataset and scoring rule that many teams run, so their results can be compared on equal terms.
ROUGE
A set of automatic scores that rate a generated summary by counting the words and word sequences it shares with summaries written by people, mainly measuring how much of the reference it covers.
BLEU Score
An automatic machine translation score that compares output with human reference translations by counting matching word sequences of up to four words, with a penalty for output that is too short.
Perplexity
The metric that measures how well a language model predicts a text: the exponential of its average negative log-probability per token.
AUC
The area under a classifier's ROC curve: the probability that the model scores a random positive example higher than a random negative one, across every possible cutoff.
ROC Curve
A graph of a classifier's true positive rate against its false positive rate at every decision threshold, showing how well it separates two classes.
Confusion Matrix
A table that counts a classifier's true positives, false positives, false negatives and true negatives, from which metrics like precision and recall are calculated.
F1 Score
The harmonic mean of precision and recall, a single score from 0 to 1 that is high only when a classifier finds most positives and few of its flags are wrong.
Recall
The share of actual positive cases that a classifier or search system finds: true positives divided by all real positives, so a miss lowers it and a false alarm does not.
Precision
How closely repeated measurements of the same thing agree with each other; in classification, the share of positive predictions that are correct.
Accuracy
How close a measurement or prediction is to the true value; in classification, the share of predictions that are correct.
Data Leakage
In machine learning, information reaching a model during training that it would not have at prediction time, producing test scores that do not survive deployment.
CNN (Convolutional Neural Network)
A neural network that learns image features by sliding small reusable filters across the input, making it efficient at recognizing patterns like edges, shapes, and objects.
Activation Function
An activation function is the rule a single neuron in a neural network applies to its combined input before passing a signal onward, turning a plain weighted sum into a nonlinear value so the network can learn curves rather than only straight lines.
Active Learning
A machine learning approach where the model selects which unlabeled examples to have labeled next, usually the ones it is least certain about, to reach higher accuracy with fewer labels.
Fine-Tuning
Continuing to train a model that already learned general patterns during pretraining on a smaller, task-specific dataset, so it adapts to a narrower job without starting from scratch.
AI Alignment
The work of making AI systems pursue the goals people actually intend, not only the literal objective they were trained to maximise.
Self-Supervised Learning
Training a model on unlabeled data by hiding part of each example and having the model predict the hidden part from the rest.
Hallucination
When an AI model generates fluent, confident text that is factually wrong or unsupported by its source material, such as a fabricated citation or invented statistic.
Transformer
A neural network architecture that processes a whole sequence at once, letting every token draw information directly from every other token through attention.
Deep Learning
Deep learning trains neural networks with many layers to learn useful features directly from raw data such as images, audio and text.
Large Language Model (LLM)
A neural network trained on very large amounts of text to predict what comes next, then adjusted with human feedback so it follows instructions.
Chain of Thought
A prompting technique that asks a large language model to write out its intermediate reasoning steps before giving a final answer, instead of jumping straight to the answer.
Backpropagation
The algorithm that trains a neural network by computing how much each weight contributed to a wrong output, then adjusting every weight to reduce that error.
Beam Search
Beam search is a decoding method that keeps the k highest-scoring partial outputs at each step, so a model can compare several candidate sequences without scoring every possible one.
Self-Attention
The mechanism inside a Transformer that lets every element of a sequence weigh and combine information from every other element in that sequence.
Zero-Shot Learning
Zero-shot learning lets a model recognise classes or do tasks it has no labelled examples of, using a description such as a label name instead.
Few-Shot Learning
Few-shot learning is giving a model a handful of example input-output pairs inside the prompt itself, so it infers the task and applies it to a new input without any retraining.
Online Learning
In machine learning, updating a model's parameters incrementally on each new example as it streams in, rather than training once on a fixed batch, so the model keeps adapting to new data without being retrained from scratch.
Context Window
The total amount of text a language model can consider in a single request, measured in tokens and shared between the input and the response.
Loss Function
A loss function is the single number a training algorithm tries to minimize, measuring exactly how wrong a model's prediction is for one example.
Tesler's Law
Tesler's Law states that every application has an irreducible amount of complexity, so design can only choose whether the user, the application developer, or the platform developer deals with it.
Planning Fallacy
A systematic underestimate of the time, cost and risk of your own plans, made even by people who know that comparable projects overran.
Hindsight Bias
The tendency to see a past outcome as having been predictable, once you already know that it happened.
Triage
Triage is the recurring practice of sorting incoming bugs, tickets, or incidents by urgency and impact under time pressure, deciding what gets attention now, later, or never.
Accessibility
Accessibility is how well people with disabilities can perceive, understand, navigate, and use a product, service, or place.
Design System
The maintained set of standards, components and code a team uses to design and build a product consistently at scale.
Modularity
Modularity is dividing a system into modules that each hide one design decision behind a stable interface, so that decision can change without touching other modules.
Throughput
The amount of work a system finishes in a given period of time, such as requests per second, counting only completed work.
Doherty Threshold
Walter Doherty and Ahrvind Thadhani's 1982 finding that cutting a computer's response time below about a second raises how much work a person finishes per hour more than proportionally.
Postel's Law
Jon Postel's 1980 design rule for network software: be strict and conformant in what your own software sends, but tolerant of variations in what it accepts, as long as the sender's meaning is still clear.
Anthropomorphism
The tendency to attribute human traits, such as intentions, feelings and thoughts, to things that are not human, including animals, machines and storms.
Poka-Yoke
Shigeo Shingo's mistake-proofing from the Toyota Production System: a device or process step that prevents a human error, or flags it at once, before it becomes a defect.
Uncanny Valley
The dip in comfort people feel toward a robot, animated character, or synthetic face as its human likeness approaches, but does not quite reach, fully realistic.
Feedback Loops
A chain of cause and effect in which a system's output returns as an input and changes the next output.
Inattentional Blindness
Failing to see a fully visible object because attention was committed to something else, even while looking directly at it.

242Coming soon · not written yet

Coming soon
The Butterfly Effect
Coming soon
Swiss Cheese Model
Coming soon
Hebbian Learning
Coming soon
Bayesian Network
Coming soon
Complex Systems
Coming soon
CrI2SP Framework
Coming soon
Case-based Reasoning
Coming soon
Artificial Intelligence Models
Coming soon
System and Real World Alignment
Coming soon
Natural Language Generation
Coming soon
Speech Recognition
Coming soon
Virtual Agents
Coming soon
Systems Theory
Coming soon
Blockchain
Coming soon
Smishing
Coming soon
Markov Decision Processes
Coming soon
Root Cause Analysis
Coming soon
Ethical AI
Coming soon
Discrete-event Simulation
Coming soon
Fault Tree Analysis
Coming soon
Linear Discriminant Analysis (LDA)
Coming soon
Sentiment Analysis
Coming soon
CAR Model
Coming soon
Parallel Processing
Coming soon
Network Topology
Coming soon
Spoofing
Coming soon
Phishing
Coming soon
Decision Support Systems
Coming soon
Decision Intelligence
Coming soon
Vishing
Coming soon
Predictive Modelling
Coming soon
Recommender System
Coming soon
Algorithmic Bias
Coming soon
Smart Cities
Coming soon
Counterfactual Reasoning in AI
Coming soon
Surveillance and Artificial Intelligence (AI)
Coming soon
AI Maturity Models
Coming soon
AI in Healthcare Equity
Coming soon
Ethics of Automated Decision-Making
Coming soon
Automation Bias
Coming soon
Parasocial Trust in AI
Coming soon
Delegation Creep
Coming soon
Accountability Diffusion in AI
Coming soon
AI Literacy Gap
Coming soon
Temporal Consistency
Coming soon
Causal Discovery
Coming soon
Robustness Evaluation Metrics
Coming soon
Reinforcement Learning from Demonstrations
Coming soon
Neural-Symbolic Integration Risks
Coming soon
Large Language Model Cascading Failures
Coming soon
Human-AI Collaborative Filtering
Coming soon
Graph-Based Explainability
Coming soon
Generative Sparring
Coming soon
Explainable AI Toolkit
Coming soon
Adversarial Robustness Evaluation
Coming soon
Value Alignment Frameworks
Coming soon
Explainable Multimodal Fusion
Coming soon
Diffusion Probabilities
Coming soon
Robust Out-of-Distribution Generalization
Coming soon
Prompt Engineering Frameworks
Coming soon
Memory-Efficient Transformers
Coming soon
Knowledge Graph Refinement
Coming soon
Explainable Multimodal Learning
Coming soon
Data Quality Metrics
Coming soon
Token-Free Models
Coming soon
Neural Rendering
Coming soon
Human-in-the-Loop Feedback
Coming soon
Frozen Pre-trained Transformers
Coming soon
Prompt-Based Explainability Methods
Coming soon
Multimodal Transfer Learning with Vision-Language Models
Coming soon
Diffusion-Based Video Generation
Coming soon
Value-Alignment Risks
Coming soon
Prompt Verification
Coming soon
Neural Provenance
Coming soon
Multimodal Embeddings
Coming soon
Long-Tail Learning
Coming soon
Diffusion Denoising
Coming soon
Vision Transformers
Coming soon
Stable Diffusion 2.0
Coming soon
Meta Llama
Coming soon
Llama Model
Coming soon
Homomorphic Encryption as a Service
Coming soon
Diffusion-Based Image Editing
Coming soon
Physical Adversarial Attack
Coming soon
Human-AI Collaboration Framework
Coming soon
Explainable Transfer Learning
Coming soon
Embodied AI
Coming soon
Cognitive Graph
Coming soon
Robust Outlier Detection
Coming soon
Out-of-Distribution Detection
Coming soon
Lifelong Learning
Coming soon
Causal Graph
Coming soon
Prompt Chaining
Coming soon
Efficient Transformers
Coming soon
Dream Fusion
Coming soon
Multimodal Transfer Learning
Coming soon
Explainable Graph Neural Networks
Coming soon
Disentanglement
Coming soon
Cognitive Architecture
Coming soon
Zero-Knowledge Proofs
Coming soon
Vision-and-Language Pretraining
Coming soon
Value Alignment
Coming soon
Uncertainty-Aware Reinforcement Learning
Coming soon
Transfer Learning for Robotics
Coming soon
Safe Reinforcement Learning
Coming soon
Robustness Evaluation Framework
Coming soon
Prompt-Based Few-Shot Learning
Coming soon
Multimodal Sentiment Analysis
Coming soon
Large Language Model Cascades
Coming soon
Human-Centered AI
Coming soon
Foundation Model Fine-Tuning
Coming soon
Explainable AI Framework
Coming soon
Diffusion-Based Inpainting
Coming soon
Causal Attention
Coming soon
Sparse Attention
Coming soon
Multimodal Fusion
Coming soon
Influence Maximization
Coming soon
Explainable Reinforcement Learning
Coming soon
Diffusion-Based Generation
Coming soon
Causal Learning
Coming soon
ASI (Artificial Superintelligence)
Coming soon
AGI (Artificial General Intelligence)
Coming soon
Proprietary Model
Coming soon
Open Source Model
Coming soon
Model Serving
Coming soon
MLOps
Coming soon
AI Accelerator
Coming soon
GPU
Coming soon
Model Drift
Coming soon
Responsible AI
Coming soon
Red Teaming
Coming soon
Model Poisoning
Coming soon
Fairness
Coming soon
Explainability
Coming soon
Interpretability
Coming soon
Alignment
Coming soon
Reward Signal
Coming soon
Goal Conditioning
Coming soon
Task Decomposition
Coming soon
Copilot
Coming soon
RPA (Robotic Process Automation)
Coming soon
Workflow Automation
Coming soon
Function Calling
Coming soon
Retrieval
Coming soon
Multi-Agent System
Coming soon
Autonomous System
Coming soon
Underfitting
Coming soon
Generalization
Coming soon
Annotation
Coming soon
Class Imbalance
Coming soon
Regression
Coming soon
Classification
Coming soon
Clustering
Coming soon
Feature Engineering
Coming soon
Data Augmentation
Coming soon
Test Set
Coming soon
Sampler
Coming soon
Inpainting
Coming soon
Upscaling
Coming soon
Synthetic Data
Coming soon
Style Transfer
Coming soon
Speech to Text STT
Coming soon
Text to Speech TTS
Coming soon
Knowledge Graphs
Coming soon
RNN Recurrent Neural Network
Coming soon
Hidden Layer
Coming soon
Objective Function
Coming soon
Orchestration
Coming soon
Model
Coming soon
Neural Network
Coming soon
Algorithm
Coming soon
Data Point
Coming soon
System Prompt
Coming soon
Parameter
Coming soon
Embedding
Coming soon
Ground Truth
Coming soon
Audio Generation
Coming soon
Code Generation
Coming soon
Text Generation
Coming soon
Image Generation
Coming soon
Video Generation
Coming soon
Text-to-Image
Coming soon
Latent Diffusion
Coming soon
Artificial Intelligence
Coming soon
Imitation Learning
Coming soon
Supervised Learning
Coming soon
Reinforcement Learning
Coming soon
Unsupervised Learning
Coming soon
Diffusion Model
Coming soon
GAN (Generative Adversarial Network)
Coming soon
Artificial Neural Network (ANN)
Coming soon
Temperature
Coming soon
Generative AI
Coming soon
Machine Learning
Coming soon
Multimodal AI
Coming soon
Natural Language Processing (NLP)
Coming soon
Virtual Assistant
Coming soon
Data Preprocessing
Coming soon
Vision-Language Model (VLM)
Coming soon
Prompt Engineering
Coming soon
Training data
Coming soon
Training Set
Coming soon
Outpainting
Coming soon
Greedy Decoding
Coming soon
Federated Learning
Coming soon
Anomaly Detection
Coming soon
Top-p Sampling
Coming soon
Prediction
Coming soon
Latent Space
Coming soon
Hyperparameter
Coming soon
Weight
Coming soon
Inference
Coming soon
Seed
Coming soon
Prompt
Coming soon
Instruction Tuning
Coming soon
Semi-Supervised Learning
Coming soon
LSTM
Coming soon
Encoder
Coming soon
Encoder-Decoder Model
Coming soon
Turing test
Coming soon
Decoder
Coming soon
Normalization
Coming soon
Top-k Sampling
Coming soon
Dimensionality Reduction
Coming soon
Vector Database
Coming soon
Training
Coming soon
Validation Set
Coming soon
Tokenization
Coming soon
Machine Translation
Coming soon
Transfer Learning
Coming soon
Garbage In Garbage Out
Coming soon
Stress Testing
Coming soon
Fail-safes
Coming soon
The Irony of Automation
Coming soon
The Paradox of Automation
Coming soon
Entropy
Coming soon
Communication Overhead
Coming soon
Conservation of Energy
Coming soon
Sniper Links
Coming soon
Band-Aid Fixes
Coming soon
Syntactic Complexity
Coming soon
Serial Search

Other subjects

Psychology
Product
Design
Marketing
Research
Business
Finance
Leadership
Communication
Sales
Economics
Statistics
Neuroscience

The Concept Library.A structured concept library for tech professionals.

Missing a concept?

Tell us what to write next. We’ll email you when it’s live.

Which concept?

Several? Separate them with commas.

Your email

Only to tell you when it’s live.

Your name (optional)

Website

How we use your email: privacy.