Model
A trained system containing learned parameters and behavior.
Seven Layers · Layer 4
Model Architecture
About this section
Model architecture is the internal organization that determines how an AI system represents information, transforms it, learns from data, remembers context, combines modalities, and produces an answer or action. Architecture is not the same thing as a finished model. It is the design pattern from which models are built.
A model’s training data teaches it what to know. Its architecture helps determine what it can learn, how it reasons, how efficiently it operates, and where its limits appear.
This layer sits beneath domain models and above infrastructure. Domain models adapt architectural ideas to specialized fields. Infrastructure supplies the computing systems that train and run them.
Architecture
The structural design that organizes computation, information flow, memory, and learning.