Glossary Terms

Data Augmentation

Data Augmentation – Techniques for increasing data diversity, often by modifying existing samples.

Decoder

Decoder – The part of a model responsible for generating output, especially in sequence models.

Deepfake

Deepfake – AI-generated videos or images that convincingly replace one person’s face with another’s, often used in media or entertainment.

Dropout

Dropout – A regularization method where random neurons are ignored during training to prevent overfitting.

Dynamic Memory Network (DMN)

Dynamic Memory Network (DMN) – A model architecture designed to understand and answer questions about sequential data.