AI glossary, plain English.

In AI, an agent is a program that can perceive its environment, make decisions, and take actions to achieve a goal. Think of it as an AI with a purpose.

Artificial General Intelligence, or AGI, refers to hypothetical AI that can understand, learn, and apply intelligence to any intellectual task a human being can.

The context window refers to the amount of information an AI model can 'see' and process at any given time when generating a response. It's like the AI's short-term memory.

CSP stands for Constraint Satisfaction Problem. It's a way to think about problems where you need to find values for variables that satisfy certain limits or rules.

CX stands for Customer Experience, which is all about how a customer perceives their interactions with a company.

Deep learning is a type of machine learning that uses multi-layered artificial neural networks to learn from data, often without direct human instruction.

DevOps is a set of practices that combines software development (Dev) and IT operations (Ops) to shorten the development lifecycle and provide continuous delivery of high-quality software.

Embeddings are numerical representations of words, phrases, or other data, capturing their meaning and relationships in a way computers can understand.

XAI, or Explainable AI, refers to methods and techniques that allow humans to understand why an AI model made a particular decision or prediction.

Multimodal refers to AI systems that can understand and work with more than one type of data, like text, images, or audio, all at the same time.

Natural Language Processing, or NLP, is an area of AI that focuses on enabling computers to understand, interpret, and generate human language.

Open source refers to software or projects where the original source code is made freely available and can be modified and distributed by anyone.

SSL stands for Secure Sockets Layer, a technology for securing communication over a computer network, like the internet. It helps keep your online activities private and safe.

SWE stands for Software Engineer, a professional who designs, develops, and maintains software systems.

AI stands for Artificial Intelligence. It's about designing computer systems that can do tasks that usually need human smarts, like understanding language or making decisions.

CPU stands for Central Processing Unit. It's the primary component of a computer that performs most processing functions, often called the 'brain' of the computer.

GPU stands for Graphics Processing Unit. It's a specialized electronic circuit designed to rapidly manipulate and alter memory to accelerate the creation of images in a frame buffer for output to a display device.

Machine learning is a field of artificial intelligence that lets computers learn from data without being explicitly programmed for every task.

Mean Squared Error, or MSE, is a common way to measure the average squared difference between predicted values and actual values in a statistical model.

A neural network is a type of computer program inspired by the structure and function of the human brain, designed to recognize patterns and make decisions.

In AI, a parameter is a setting within a model that it learns and adjusts during training. These settings determine how the model processes information and makes predictions.

PCA stands for Principal Component Analysis. It's a fundamental statistical technique used to simplify complex datasets by reducing the number of variables.

Reinforcement Learning (RL) is a type of machine learning where an AI learns to make decisions by trying things out and getting rewards or penalties.

SVD, or Singular Value Decomposition, is a powerful mathematical technique for breaking down a matrix into three simpler matrices. It helps us understand and simplify complex data.

Tokens are small pieces of text or code that large language models (LLMs) use to understand, process, and generate language, like the building blocks of words.

AIME stands for Artificial Intelligence in Medical Epidemiology, a technology that uses AI to analyze and predict disease patterns.

Cursor is an AI-first code editor designed to help developers write, debug, and understand code more efficiently by integrating large language models.

GitHub is a web-based platform that helps developers store, manage, and collaborate on software projects using Git, a version control system.

NAS stands for Network Attached Storage, which is a dedicated file storage device connected to your network that allows multiple users and devices to store and retrieve data from a central location.

SHAP, or SHapley Additive exPlanations, is a popular method for explaining the predictions of machine learning models by showing how each feature contributes to the final output.

TikTok is a popular short-form video app where users can create and share creative videos, often set to music or trending sounds. It's known for its powerful recommendation algorithm that helps users discover new content.

A vector database is a special kind of database designed to store and quickly search 'vector embeddings,' which are numerical representations of data like text or images.

Voiceflow is a visual design and prototyping tool that helps people build and launch conversational AI experiences, like chatbots and voice assistants, without needing to write code.

WEBP is an image format developed by Google that helps make websites load faster by creating smaller, high-quality images.

Anthropic is an AI safety and research company that develops large language models, including the Claude family of models.

AWS, or Amazon Web Services, is a cloud computing platform provided by Amazon that offers a vast range of on-demand services.

DeepSeek is an AI research and development company known for creating powerful open-source large language models and coding models.

ElevenLabs is an AI company that develops realistic text-to-speech and voice synthesis technology, allowing users to generate high-quality spoken audio from text.

Google is an American multinational technology company known for its search engine, Android mobile operating system, and a wide range of internet-related services and products.

NVIDIA is a technology company best known for designing graphics processing units (GPUs), which are crucial for rendering graphics and accelerating AI computations.

OpenAI is an artificial intelligence research and deployment company known for creating large language models like GPT-3.5, GPT-4, and the image generation tool DALL-E.

Attention is a powerful mechanism in AI, especially in models like Transformers, that helps the model focus on the most important parts of its input.

CNN stands for Convolutional Neural Network, a type of neural network especially good at processing images and other grid-like data.

Diffusion models are a type of generative AI model that learn to create new data, like images or audio, by progressively removing noise from an initial random input.

GAN stands for Generative Adversarial Network, a type of AI framework that learns to create new, realistic data like images, audio, or text.

Graph Neural Networks, or GNNs, are a type of neural network designed to work directly with data structured as graphs, like social networks or molecular structures.

GRU stands for Gated Recurrent Unit. It's a type of neural network layer that helps computers understand sequences of information, like words in a sentence.

Long Short-Term Memory, or LSTM, is a special kind of recurrent neural network designed to remember information for longer periods than standard networks.

A Recurrent Neural Network, or RNN, is a type of neural network that's really good at processing sequences of data, like text or time series, because it has an internal memory.

A neural network architecture, or blueprint, especially good at understanding sequences of data like text or speech. It processes all parts of a sequence at once, rather than one piece at a time.

A Variational Autoencoder, or VAE, is a type of neural network that learns to compress data into a smaller representation and then reconstruct it, often used for generating new, similar data.

BERT stands for Bidirectional Encoder Representations from Transformers. It's a popular natural language processing model developed by Google.

GPT stands for "Generative Pre-trained Transformer." It's a type of artificial intelligence model designed to generate human-like text.

LLM stands for Large Language Model, a type of AI program designed to understand and generate human-like text.

Chain of Thought is a prompting technique that encourages large language models to show their step-by-step reasoning process before giving a final answer.

Prompt engineering is the art and science of crafting inputs (prompts) to get the best possible results from AI models.

UMAP is a technique for visualizing high-dimensional data in a lower-dimensional space, often 2D or 3D, while trying to preserve the original data's structure.

CutMix is a data augmentation technique that helps train more robust image classification models by blending parts of two different images.

Fine-tuning is a process where a pre-trained AI model is further trained on a smaller, specific dataset to adapt its knowledge to a particular new task or domain.

Multi-Task Learning (MTL) is a machine learning technique where a single model learns to perform several related tasks at the same time.

RLHF is a way to train AI models using human preferences, helping the AI learn to produce outputs that are more helpful and safe.

SGD is a popular optimization algorithm used to train machine learning models, especially neural networks. It helps models learn by making small, iterative adjustments to their internal settings based on training data.

Training is the process where an AI model learns from data to identify patterns and make predictions or decisions.

GPQA is a challenging benchmark designed to evaluate advanced AI models on complex, expert-level questions requiring deep understanding and reasoning.

When an AI makes up information that sounds convincing but isn't true or based on its training data.

In AI, inference is the process where a trained model uses what it has learned to make predictions or generate outputs based on new, unseen data.

LIME stands for Local Interpretable Model-agnostic Explanations. It's a way to understand why a complex AI model made a specific prediction by explaining it with a simpler, local model.

A reasoning model refers to how an AI system processes information, draws connections, and makes logical conclusions, much like how humans think through problems.

Temperature is a setting in AI models that controls how predictable or creative their responses are. A higher temperature means more randomness and variety, while a lower temperature makes responses more focused and consistent.

Model Context Protocol. An open standard from Anthropic for connecting AI assistants to external tools and data.

A prompt is the input, usually text, that you give to an AI model to tell it what you want it to do or create.

RAG stands for Retrieval Augmented Generation, a technique that helps large language models (LLMs) access and use up-to-date or specific information outside of their original training data.