Model
In machine learning, a model is a system learned from data that produces predictions or other outputs from inputs. Its learned parameters are part of the model; the application around it supplies the interface, data access and controls.
A support team can train a model on labeled tickets to predict whether a new ticket concerns billing, sales or a technical problem. Training fits the model to examples. Inference uses the trained model on a new ticket. The routing application then decides what to do with that prediction.
A model includes more than its parameters: the structure and computations determine how those parameters are used. A saved checkpoint may need matching code and input preparation to run correctly. The same model can also produce different outputs when its inputs or generation settings change, without learning new parameters.
This distinction matters when debugging. A wrong answer might come from the learned model, missing evidence, a bad prompt or a broken tool. Replacing the model will not repair every failure in the surrounding system. Compare changes on representative tasks using evaluations.
Sources
- Google: Machine Learning Glossary — model — Defines a model and distinguishes training from prediction on new examples.
Go deeper
- How do large language models actually work? AI, decoded · How Large Language Models Actually Work
- How do you test an AI system when the output isn't deterministic? AI, decoded · How to Test an AI System
- Google: Linear regression course
Build intuition for a simple model by following inputs, weights and predictions.
