Glossary
Quantum Computing
Qubit - The quantum equivalent of a classical bit. Unlike classical bits that are either 0 or 1, qubits can exist in superposition, representing both states simultaneously until measured.
Superposition - A fundamental quantum principle where a quantum system exists in multiple states at once. A qubit in superposition is both 0 and 1 until observation collapses it to a definite state.
Entanglement - A quantum phenomenon where two or more qubits become correlated in such a way that the state of one instantly affects the others, regardless of distance. Einstein called this “spooky action at a distance.”
Quantum Gate - The quantum equivalent of classical logic gates. Quantum gates manipulate qubits through unitary operations, forming the building blocks of quantum algorithms.
Quantum Algorithm - A step-by-step procedure designed to run on quantum computers. Examples include Shor’s algorithm (factoring) and Grover’s algorithm (search).
Quantum Advantage - The point at which a quantum computer solves a practical problem faster or more efficiently than the best known classical approach. Also called “quantum supremacy.”
NISQ Era - Noisy Intermediate-Scale Quantum. The current era of quantum computing characterized by 50-1000 qubits with significant error rates, requiring error mitigation rather than full error correction.
Decoherence - The loss of quantum properties due to environmental interference. Qubits lose their quantum state when they interact with their surroundings, limiting computation time.
Quantum Annealing - A quantum computing approach optimized for solving optimization problems by finding the lowest energy state of a system. Used in D-Wave systems.
Quantum Circuit - A model for quantum computation where a sequence of quantum gates is applied to qubits, analogous to classical circuit diagrams.
Post-Quantum Cryptography
Post-Quantum Cryptography (PQC) - Cryptographic algorithms designed to be secure against attacks from both classical and quantum computers. NIST standardized several PQC algorithms in 2024.
Shor’s Algorithm - A quantum algorithm that can factor large numbers exponentially faster than classical algorithms, threatening RSA and ECC encryption when large quantum computers exist.
Harvest Now, Decrypt Later (HNDL) - Also known as Store Now, Decrypt Later (SNDL). An attack strategy where adversaries collect encrypted data today to decrypt it once quantum computers become available.
Quantum-Resistant - Algorithms believed to be secure against attacks from quantum computers. Includes lattice-based, hash-based, code-based, and multivariate cryptography.
Y2Q (Year to Quantum) - The uncertain date when quantum computers will become powerful enough to break current public-key cryptography. Analogous to Y2K but with unknown timing.
RSA - Rivest-Shamir-Adleman. A widely-used public-key cryptosystem that relies on the difficulty of factoring large numbers. Vulnerable to Shor’s algorithm.
Elliptic Curve Cryptography (ECC) - Public-key cryptography based on elliptic curves over finite fields. More efficient than RSA but also vulnerable to quantum attacks.
Lattice-Based Cryptography - Post-quantum cryptographic systems based on the hardness of lattice problems. Believed resistant to quantum attacks and forms basis of several NIST standards.
Artificial Intelligence & Machine Learning
Machine Learning (ML) - A subset of AI where systems learn from data without explicit programming, improving performance through experience.
Deep Learning - Machine learning using neural networks with multiple layers (deep neural networks) to learn hierarchical representations of data.
Neural Network - A computing system inspired by biological neural networks, consisting of interconnected nodes (neurons) that process and transmit information.
Training Data - The dataset used to teach a machine learning model. Quality and quantity of training data significantly impact model performance.
Supervised Learning - Machine learning where models learn from labeled data, with known input-output pairs. Used for classification and regression tasks.
Unsupervised Learning - Learning from unlabeled data to discover hidden patterns or structures. Includes clustering and dimensionality reduction.
Reinforcement Learning - Learning through interaction with an environment, receiving rewards or penalties for actions. Used in game playing and robotics.
Overfitting - When a model learns training data too well, including noise and outliers, resulting in poor performance on new data.
Feature Engineering - The process of selecting, modifying, or creating input variables (features) to improve machine learning model performance.
Natural Language Processing (NLP) - AI techniques for understanding, interpreting, and generating human language. Powers chatbots, translation, and text analysis.
Computer Vision - AI field focused on enabling computers to derive meaningful information from visual inputs like images and videos.
Transformer - A neural network architecture using self-attention mechanisms. Foundation of large language models like GPT and BERT.
Large Language Model (LLM) - Massive neural networks trained on vast text datasets to understand and generate human language. Examples include GPT-4, Claude, and BARD.
Generative AI - AI systems that create new content (text, images, code, audio) rather than just analyzing or classifying existing data. Powered by models like GPT, DALL-E, and Stable Diffusion.
Gradient Descent - An optimization algorithm that iteratively adjusts model parameters to minimize error, moving “downhill” toward the optimal solution.
Backpropagation - The method neural networks use to learn, calculating gradients of the loss function and propagating errors backward through the network to update weights.
Activation Function - Mathematical functions in neural networks (ReLU, sigmoid, tanh) that introduce non-linearity, enabling networks to learn complex patterns.
Convolutional Neural Network (CNN) - Neural networks specialized for processing grid-like data such as images, using convolutional layers to detect spatial patterns.
Recurrent Neural Network (RNN) - Neural networks designed for sequential data, maintaining internal memory to process sequences like text or time series. Includes LSTM and GRU variants.
Attention Mechanism - A technique allowing models to focus on relevant parts of input when making predictions. Core innovation behind transformer architecture.
Embedding - Dense vector representations of data (words, images, concepts) that capture semantic relationships in a continuous space. Words with similar meanings have similar embeddings.
Fine-tuning - Adapting a pre-trained model to a specific task by training it further on task-specific data. More efficient than training from scratch.
Hyperparameter - Configuration settings for training machine learning models (learning rate, batch size, number of layers) set before training begins, not learned from data.
Inference - The process of using a trained model to make predictions on new, unseen data. Distinct from the training phase.
Bias (AI) - Systematic errors in AI systems arising from training data or algorithm design, leading to unfair or prejudiced outcomes for certain groups.
Explainable AI (XAI) - Techniques and methods for making AI decisions interpretable and understandable to humans. Critical for regulated industries and high-stakes decisions.
Hallucination - When AI models generate plausible-sounding but factually incorrect or nonsensical information, particularly common in large language models.
Prompt Engineering - The practice of carefully designing input text (prompts) to elicit desired responses from language models. Critical skill for working with LLMs.
Few-shot Learning - Training models to learn new tasks from only a few examples, mimicking human ability to generalize from limited data.
Zero-shot Learning - The ability of models to perform tasks they weren’t explicitly trained for, using only instructions or descriptions without examples.
Transfer Learning - Leveraging knowledge learned from one task to improve performance on related tasks. Reduces data and compute requirements.
Model Drift - Gradual degradation of model performance over time as real-world data distributions change from training data. Requires monitoring and retraining.
Edge AI - Running AI models directly on local devices (phones, IoT sensors) rather than cloud servers, enabling faster responses and privacy protection.
Ensemble Learning - Combining predictions from multiple models to achieve better performance than any single model. Methods include bagging, boosting, and stacking.
Quantum Machine Learning
Quantum Machine Learning (QML) - The integration of quantum computing with machine learning, using quantum algorithms to accelerate or enhance ML tasks.
Variational Quantum Eigensolver (VQE) - A hybrid quantum-classical algorithm for finding ground state energies of molecules, used in quantum chemistry and optimization.
Quantum Neural Network (QNN) - A neural network model that runs on quantum computers, using quantum circuits to process information.
Quantum Kernel - A method for embedding classical data into quantum Hilbert space, enabling quantum computers to identify patterns difficult for classical algorithms.
QAOA (Quantum Approximate Optimization Algorithm) - A variational quantum algorithm for solving combinatorial optimization problems on NISQ devices.
General Computing & Optimization
Algorithm - A precise step-by-step procedure for solving a problem or performing a computation. Can be implemented on classical or quantum computers.
Optimization - The process of finding the best solution from a set of possible alternatives. Common in logistics, finance, and resource allocation.
Combinatorial Optimization - Finding optimal solutions from a finite set of objects. Examples include traveling salesman problem and scheduling.
Heuristic - A practical approach to problem-solving that may not be perfect but provides good-enough solutions efficiently. Often used when exact solutions are computationally infeasible.
NP-Hard - Computational problems for which no known polynomial-time algorithm exists. Many real-world optimization problems are NP-hard.
Classical Computing - Traditional computing using binary bits (0 or 1) and Boolean logic, as opposed to quantum computing.
Turing Machine - A theoretical computing model that defines what can be computed. Classical computers are physical implementations of Turing machines.
Business & Strategy Terms
Digital Transformation - The integration of digital technology into all areas of business, fundamentally changing operations and value delivery.
Technical Debt - The implied cost of additional work caused by choosing quick solutions now instead of better approaches that would take longer.
Proof of Concept (PoC) - A demonstration to verify that certain concepts or theories have practical potential. Often the first step in technology adoption.
Return on Investment (ROI) - A performance measure evaluating the efficiency of an investment, calculated as net profit divided by cost.
Scalability - The capability of a system to handle growing amounts of work by adding resources. Critical for production AI/ML systems.
Vendor Lock-in - Dependence on a specific vendor’s products or services, making it difficult to switch to alternatives without substantial costs.