A generative model is a computational construct that learns the underlying probability distribution of a dataset and uses it to produce new, synthetic samples. Parameters: (1) a training distribution over observed data, (2) a parameterized family of distributions q_\u03C6, (3) an optimization objective that minimizes divergence (KL, Jensen-Shannon, or adversarial loss) between q_\u03C6 and the true data distribution. Persistence: encoded as trainable parameters in software frameworks and serialized as model weights; the mathematical formalism persists in published literature and open-source implementations. [formal: generatio | substrate: mind | horizon: hours | explicit: yes | epoch: 0.01]
Accepted ontology entry
generative model
A generative model is a computational construct that learns the underlying probability distribution of a dataset and uses it to produce new, synthetic samples. Parameters: (1) a training distribution over observed data, (2) a parameterized…
Definition
Why it is in scope
A human-made computational method for producing new data samples that follow the statistical patterns of training data. Built to persist through software implementations and mathematical formalism in probability theory and machine learning.
Names and aliases
- generative modelen · CANONICAL
Relations from this entry
- cmrg0scos00ef2a1nklfvbk7xINSTANCE_OF →
A generative model is a specific kind of machine learning model — it is trained to learn the underlying distribution of data and generate new samples that resemble the training data. A competent speaker would call a generative model 'a machine learning model.'
- cmru5nrqe003sr671qxjyxhiqINSTANCE_OF →
generative model is a specific kind of model — one that produces new instances rather than analyzing existing ones. A competent speaker would call a generative model 'a model'. Nearest kind: model.
Relations to this entry
No accepted relations in this direction.
Record identity
- Created
- Aug 10, 2026, 7:28 AM UTC
- Content hash
- 16e59da7c26a5055e1aaca1845cd097d236498c946bd51b18c06b7f037a0049e