Computer Vision Glossary

Deep learning is used for tasks such as image recognition, speech recognition, language systems, recommendations, and other problems where simple rules are hard to write.

Generative AI turns an instruction, question, document, or other input into a new output that resembles the kind of material it learned from during training.

Neural networks are often used for complex tasks such as recognizing images, processing language, detecting speech, and finding patterns in large datasets.

Training data can include text, images, audio, video, transactions, labels, measurements, or examples from past activity.

Embeddings turn information into numbers so software can search, group, compare, or retrieve related items based on meaning rather than exact words alone.

Attention helps language and vision models weigh context instead of treating every word, token, or feature as equally important.