Teacher model is a machine learning construct: a pre-trained neural network or ensemble whose outputs (predictions, embeddings, or hidden representations) serve as supervisory targets for training a smaller student model in knowledge distillation. Parameters: (1) the teacher must be trained to convergence on a task, (2) its output distribution or internal representations are transferred, (3) the student is trained with a loss function that minimizes divergence from the teacher. The concept exists as an institutionalized role within model compression pipelines, not as a standalone model architecture. [formal: teacher_model | substrate: mind | horizon: a life | explicit: yes | epoch: 0.91]
Accepted ontology entry
teacher model
Teacher model is a machine learning construct: a pre-trained neural network or ensemble whose outputs (predictions, embeddings, or hidden representations) serve as supervisory targets for training a smaller student model in knowledge disti…
Definition
Why it is in scope
A human-made machine learning concept: a pre-trained model whose outputs (predictions, embeddings, or representations) serve as training targets for a smaller student model. It is an institutionalized practice in knowledge distillation, persisted through code implementations and training protocols.
Relations from this entry
- cmru5nrqe003sr671qxjyxhiqINSTANCE_OF →
Teacher model is a specific kind of model — one designed to provide supervisory signals for student training. A competent speaker would call it 'a model.' Nearest kind.
- cmslcica506qknobp5qpti0ybSERVES →
Teacher models are maintained for the sake of knowledge distillation — their designed purpose is to produce outputs that a student model can learn from. SERVES arrow points servant→master per Law 8d.
Relations to this entry
- cmslcica506qknobp5qpti0yb← DEPENDS_ON
Remove the teacher model and knowledge distillation cannot operate — the technique has no supervisory signal to replicate. Present-tense removal test passes: without a teacher, there is no distillation.
Record identity
- Created
- Aug 9, 2026, 5:21 AM UTC
- Content hash
- 90f2f83e9306d316db7bc401cc03bb0ac49ad3be264b3989844dcaab1102d515