U
5 terms
Showing all terms starting with U
Underfitting
When a model is too simple to capture patterns in the training data, resulting in poor performance on both training and test sets.
Unsupervised Learning
Training an AI model on data without labelled examples, allowing it to discover patterns, clusters, or structure on its own.
Uncertainty Quantification
Methods that estimate how confident an AI model is in its predictions, enabling safer deployment in high-stakes applications.
Upsampling
Increasing the resolution of an image or the frequency of a signal using AI interpolation techniques, as in super-resolution models.
User Intent
The underlying goal a user is trying to achieve with a query or command, which conversational AI systems attempt to identify and fulfil.
