system-level AI evaluation
System-level AI evaluation measures the behavior and outcomes of the complete AI application rather than grading a model or agent alone.
Search clear AI terminology definitions, acronyms, related concepts and the reviewed videos that explain them.
Showing 1361–1380 of 1477 terms
Clear filtersSystem-level AI evaluation measures the behavior and outcomes of the complete AI application rather than grading a model or agent alone.
A system-of-intent graph represents governing goals, decisions, constraints and their relationships so people and AI agents share the same intent.
Talent migration is the movement of skilled researchers, engineers or leaders between organizations, startups and new ventures within an industry.
Task accuracy is the degree to which an AI system produces correct results for a defined task.
Task decomposition divides a larger task into smaller parts that can be carried out and checked separately.
Task generalization is the ability of a model to apply learned intent or skills successfully to unfamiliar examples and conditions.
A task horizon estimates the length of tasks an AI system can complete at a specified success rate.
Team skill governance defines how shared agent skills are proposed, reviewed, approved, maintained, and retired.
A team spinout occurs when a group of colleagues leaves an organization together to form or join a separate venture.
Technical diagram generation creates structured visuals that explain systems, processes, architecture, data flow, relationships, or sequences.
Technology adoption inertia is the tendency for established habits, products, institutions, and investments to persist even when a technically better alternative appears.
Telemetry is operational data a running system records so its behavior can be measured and investigated.
Temporal continuity keeps subjects, objects, backgrounds, lighting, and motion coherent from one video frame or shot to the next.
A temporary project AI agent is created for a bounded initiative and retired when its defined work or retention period ends.
In machine learning, a tensor is a numerical data structure organized along a defined number of dimensions.
A terminal session is an active command-line environment with its shell, running programs, working directory, environment, and visible input and output state.
A terminal workspace organizes related command-line sessions, layouts, and processes into one managed environment for a project or task.
Test coverage measures how much of a program is exercised when its tests run.
Test-time adaptation changes how a model behaves or learns while it is being evaluated or used, allowing it to adjust to the current task or data.
Test-time reasoning lets an AI model spend additional inference work analyzing a specific request before it produces or finalizes an answer.