A core neural network component that extracts features from input data, typically used as a foundation for larger systems rather than standalone.
Quality of vision, audio, and image understanding (distinct from modality support)