Artificial Intelligence is the field of computer science dedicated to creating systems capable of performing tasks that typically require human intelligence, such as reasoning, learning, perception, and decision-making.
Launch interactive study modes or print uniform flashcard sheets.
Review definitions, formulas, and aliases included in this glossary.
A framework or set of guidelines, often used in AI ethics and responsible development, to ensure AI systems are Accountable, Consistent, and Transparent. It helps developers and organizations evaluate and mitigate potential risks and biases.
An AI agent is an autonomous entity that perceives its environment through sensors and acts upon that environment using actuators to achieve specific goals. It can be software-based, like a chatbot, or physical, like a robot.
The strategic use of AI systems to enhance human capabilities and decision-making, rather than replacing humans entirely. AI tools assist individuals by providing insights, automating routine tasks, and significantly improving overall efficiency and effectiveness.
A computer program or algorithm trained on a specific dataset to recognize patterns, make predictions, or generate new content based on its learned knowledge. It embodies the intelligence extracted from the data, enabling it to perform specific tasks.
A set of well-defined, step-by-step instructions or rules designed to solve a specific problem or perform a task efficiently. In AI, algorithms are the fundamental foundation for learning, reasoning, and decision-making processes.
The simulation of human intelligence processes by machines, especially computer systems. These processes include learning, reasoning, problem-solving, perception, and decision-making, enabling machines to perform tasks typically requiring human cognition.
Systems that can operate independently without continuous human oversight, making their own decisions and adapting to dynamic environments. They often integrate various AI technologies for perception, planning, and control to achieve specific goals.
A systematic error or prejudice in an AI system's output that leads to unfair or inaccurate results for certain groups or categories. It often stems from biased training data, algorithm design, or deployment context, leading to discrimination.
A central concept in machine learning that describes the conflict between a model's ability to fit the training data (low bias) and its ability to generalize to new data (low variance). Reducing one often increases the other.
The provision of AI capabilities and services through cloud computing platforms. Users can access powerful AI models, tools, and infrastructure remotely, without needing to manage their own hardware, enabling scalability, accessibility, and cost-effectiveness.
Computer Vision is a field of AI that enables computers to "see," interpret, and understand visual information from the real world, such as images and videos. It involves tasks like object detection, image recognition, facial recognition, and scene understanding.
A data set is a collection of related data points, often organized in a structured format, used to train, validate, and test machine learning models. It typically consists of features (inputs) and, for supervised learning, corresponding labels (outputs).
A subfield of machine learning that uses artificial neural networks with multiple layers (deep neural networks) to learn complex patterns from large amounts of data. It's particularly effective for tasks like image recognition, natural language processing, and speech synthesis.
The deployment of AI models directly on edge devices (e.g., smartphones, IoT sensors, smart cameras) rather than in centralized cloud servers. This enables faster processing, lower latency, enhanced data privacy, and reduced bandwidth usage.
The branch of AI concerned with the moral principles and values that should guide the design, development, and deployment of AI systems. It addresses issues like fairness, privacy, accountability, transparency, and potential societal impact.
Early AI systems designed to mimic the decision-making ability of a human expert within a specific domain. They use a knowledge base of facts and heuristic rules, along with an inference engine, to solve complex problems.
Explainable AI (XAI) is a set of techniques and methods that allow humans to understand the output, capabilities, and decision-making process of AI models. It aims to make AI systems more transparent, trustworthy, and interpretable, especially for complex deep learning models.
The phenomenon where an AI model, over time, gradually deviates from factual accuracy or its original intended behavior, often due to exposure to new, potentially biased, or low-quality data during continuous learning or fine-tuning processes.
The principle that AI systems should treat all individuals and groups equitably, avoiding discrimination and ensuring impartial outcomes. It involves identifying and mitigating biases in data, algorithms, and deployment to promote justice.
The process of selecting, transforming, or creating new variables (features) from raw data to improve the performance of a machine learning model. It helps the model better understand the underlying patterns and relationships for more accurate predictions.
A foundation model is a large AI model, typically pre-trained on a vast and diverse dataset at scale, that can be adapted or fine-tuned for a wide range of downstream tasks. Its pre-training captures broad knowledge and capabilities, making it a versatile base for many applications.
General AI (AGI), also known as Strong AI, refers to hypothetical AI systems that possess human-like intelligence across a wide range of cognitive tasks, including reasoning, problem-solving, learning from experience, and understanding complex concepts, rather than being limited to a specific domain.
Generative AI refers to AI models capable of producing novel and realistic content, such as text, images, audio, or video, that resembles data they were trained on but is not an exact copy. These models learn patterns and structures to create entirely new outputs.
In AI, hallucinations refer to instances where a generative model, especially a large language model, produces outputs that are plausible-sounding but factually incorrect, nonsensical, or entirely fabricated. These outputs are not grounded in the training data or real-world facts.
An approach to AI development and deployment where humans are actively involved in supervising, correcting, or improving AI systems, especially during the training, validation, or anomaly resolution phases. This ensures accuracy, fairness, and safety.
The external configuration settings that the practitioner choose to control the learning process of a machine learning model.
Implicit bias in AI refers to unconscious prejudices or stereotypes present in the training data or algorithm design that can lead to unfair or discriminatory outcomes when the AI system makes decisions or predictions. These biases are often unintentional and reflect societal inequalities.
The field of AI dedicated to representing information about the world in a form that a computer system can use to solve complex tasks. It involves formalizing concepts, objects, and relationships, enabling logical inference.
A Large Language Model (LLM) is a type of deep learning model, typically based on the Transformer architecture, that has been trained on a massive amount of text data to understand, generate, and process human language. LLMs can perform a wide range of NLP tasks, from translation to creative writing.
A subset of Artificial Intelligence that enables systems to learn from data, identify patterns, and make decisions with minimal human intervention. It uses various algorithms to process data and improve performance over time, without explicit programming.
The degree to which a human can understand the reasons behind an AI model's decision or prediction. High interpretability is crucial for building trust, debugging errors, and ensuring fairness and accountability in AI applications.
Narrow AI (ANI), also known as Weak AI, refers to AI systems designed and trained for a specific task or limited domain, such as playing chess, recognizing faces, or providing recommendations. It excels at its designated task but lacks broader cognitive abilities or general intelligence.
Natural Language Processing (NLP) is a field of AI that focuses on enabling computers to understand, interpret, generate, and interact with human language in both written and spoken forms. It involves tasks like sentiment analysis, machine translation, and text summarization.
A phenomenon where a machine learning model learns the training data too well, including its noise and specific patterns, leading to poor performance on new, unseen data. The model becomes overly complex.
Prompt engineering is the art and science of designing and refining input prompts for generative AI models, especially large language models, to elicit desired and accurate outputs. It involves crafting clear, specific, and effective instructions to guide the model's generation process.
A type of machine learning where an agent learns to make decisions by performing actions in an environment to maximize a cumulative reward. It learns through trial and error, receiving positive or negative feedback for its actions over time.
An advanced technique in reinforcement learning where human preferences and feedback are used to train a reward model, which then guides the policy optimization of a language model or other agent. It helps align model behavior with human values.
The interdisciplinary field of engineering and computer science that deals with the design, construction, operation, and application of robots. AI often provides the intelligence for robots to perceive, reason, and act autonomously in complex environments.
Supervised learning is a type of machine learning where an algorithm learns from a labeled dataset, meaning each input data point is paired with its corresponding correct output. The goal is for the model to learn a mapping function from inputs to outputs to make predictions on new, unseen data.
Training data is the portion of a dataset used to teach a machine learning model to recognize patterns, features, and relationships. During this phase, the model adjusts its internal parameters to minimize errors between its predictions and the actual labels in the data.
Transfer learning is a machine learning technique where a model trained on one task is re-purposed or adapted for a second, related task. This often involves using a pre-trained model as a starting point and fine-tuning it with a smaller, task-specific dataset, saving time and computational resources.
The Transformer is a neural network architecture introduced in 2017, widely used in natural language processing tasks, particularly for large language models. It revolutionized sequence-to-sequence modeling by relying on a self-attention mechanism, allowing parallel processing of input data.
Unsupervised learning is a type of machine learning where an algorithm learns from unlabeled data, seeking to find hidden patterns, structures, or relationships within the input without any predefined output targets. Common tasks include clustering, dimensionality reduction, and anomaly detection.