Zero-Shot
A scenario in which a model performs inference directly on a target task without fine-tuning for that task or providing labeled examples.
Core Facts
Zero-shot learning refers to a model making predictions at inference time without relying on any labeled samples for the task, directly based on pre-trained knowledge or external structural information.
Zero-shot refers to a model completing a task directly without having seen any examples of that specific task.
Timeline (last 90 days)
Zero-shot learning refers to a model making predictions at inference time without relying on any labeled samples for the task, directly based on pre-trained knowledge or external structural information.
The research team believes that true physical AI must be able to work under zero-shot conditions, i.e., managing long-horizon physical tasks without task-specific training.
All Facts (3)
Zero-shot learning refers to a model making predictions at inference time without relying on any labeled samples for the task, directly based on pre-trained knowledge or external structural information.
75%VerifiedZero-shot refers to a model completing a task directly without having seen any examples of that specific task.
65%UnverifiedThe research team believes that true physical AI must be able to work under zero-shot conditions, i.e., managing long-horizon physical tasks without task-specific training.
50%