7 terms
Showing all terms starting with V
A database optimised for storing and querying high-dimensional embedding vectors, enabling fast semantic similarity search for RAG pipelines.
An AI model that combines visual and language understanding, enabling tasks like image captioning, visual question answering, and document parsing.
A held-out subset of data used during training to tune hyperparameters and monitor for overfitting without touching the final test set.
Ensuring AI systems pursue goals and behave in ways that reflect human preferences, avoiding unintended or harmful side effects.
A generative model that learns a probabilistic latent space, enabling smooth interpolation between and generation of new data samples.
A transformer architecture applied directly to images by treating patches as tokens, achieving state-of-the-art results on image classification.
AI technology that replicates a specific person's voice from audio samples to generate new speech in that voice.
