Mixture of experts

An architecture that splits a large model into many smaller sub-networks and sends each input to only a few of them, so the model can hold far more parameters without costing proportionally more to run. The name misleads people. The experts are partitions the model worked out during training, not specialists in any recognizable sense, and none of them corresponds to a clinical specialty. It matters mainly as the explanation for why some very large models are cheap enough to serve at scale. If a product is marketed as having a radiology expert or a cardiology expert inside it because it uses a mixture of experts, that is a sales line, not an architecture.

Terms like this come up in real clinical scenarios across the HelloAI courses: bite-sized modules with verifiable certificates. An account takes one minute, no password needed.