AI agent
A system with capabilities such as perception, memory, planning and tool execution. The precise scope depends on the relevant record.
GLOSSARY · WORDS & CONNECTIONS
Connect concepts, English terms and abbreviations. These candidate terms and editorial introductions await verification against the original standards.
27 candidate terms · editorial explanations pending review
A system with capabilities such as perception, memory, planning and tool execution. The precise scope depends on the relevant record.
An interface boundary connecting capabilities of applications, platforms or devices.
A technology and application concept related to AI-generated content. Consult Q5's official mandate for its exact responsibilities.
A way to compare system performance using agreed tasks, data and measurement methods.
A way to evaluate AI computing systems using agreed workloads and measurement methods.
Evaluating participants' contributions to a collaborative process or its results.
Adding labels or structured descriptions to data for training, evaluation and subsequent processing.
A model with multiple computational layers used to learn features from data.
An architectural approach distributing tasks and collaborative capabilities across devices, edge and cloud.
Systems built around digital human appearance, voice, motion and interaction capabilities.
A collection of data organized around a specific domain or business task.
A research and systems field connecting intelligence with physical entities, environmental perception and action feedback.
An evaluation perspective relating completed work to energy consumption.
The process by which a system acquires and interprets information about its surroundings.
A family of methods in which multiple participants train a model collaboratively. Data and model information flows depend on the architecture.
A model trained on broad data that can be adapted to various downstream tasks. This is an introductory summary; verify formal definitions in the relevant standard and edition.
Operational characteristics such as latency and throughput when a model responds to tasks.
The ability of different systems to exchange information and collaborate through agreed interfaces.
Engineering work to acquire, organize, represent and apply domain knowledge.
A way to organize stages such as data preparation, training, deployment and ongoing operation.
Handling multiple forms of information, such as text, audio and images, within one task or system.
The engineering capability to observe system operation through logs, metrics and related information.
A system approach that retrieves relevant external material and supplies it as context to a generative model.
An automation method that uses software to execute repeatable business processes.
Generating or controlling tactile feedback based on computational results.
Organizing steps, resources and their execution order around a goal.
An operating method in which a remote operator controls equipment and receives feedback.