SECAI Core · phase 2 of 15 · noun

Pre-trained model

An AI model that has already been trained on a large, general dataset and can be reused or fine-tuned for specific tasks

The Explain card

Plain English: A pre-trained model has already been trained on a large, general dataset. You download it and reuse it as-is or fine-tune it for your own task instead of training from scratch.
Example: A team downloads an open-source image classifier from a public model hub and fine-tunes it to spot tampered identity documents.
Why it matters: You inherit everything baked into that model: its biases, memorised data and any backdoor planted upstream. Pre-trained models are a supply chain, and they deserve the same scrutiny as third-party code.
Hook: Someone else's homework, with someone else's mistakes inside.

How the word is built, where it came from, and what it sits beside in memory.

In a sentence

Before the ID scanner goes live, the vision job loads a pre-trained model from the public hub and continues training on labeled passport photos.

Why these words

pre-trained Latin prae, before, plus Old French trainer, to drag, as a past participle marks the teaching as already finished before this job begins
model French modele, from Latin modulus, a small measure, diminutive of modus, measure supplies the head noun, the mathematical stand-in that stores those finished weights

Where it came from

Origin: English machine-learning compound, from Latin prae (before) plus Old French trainer (to drag, from Latin trahere) as a past participle, and French modele (from Latin modulus, a small measure)
Entered the language: early 2010s
What changed: Pre- had meant before since Latin. Train had meant to drag, then to drill troops, then to teach a skill. Model had meant a small measure, then a miniature to copy, then a mathematical stand-in. Deep-net papers around 2006 named a first teaching pass pretraining. By the early 2010s the hyphenated pair named the checkpoint fetched from a hub.

How it is spelled

Pattern: pre- keeps its hyphen before a free English stem, so train stays visible
Pattern: -ed marks the past participle used as an adjective
Pattern: Latin modulus drops the u in English model
Breaks the pattern: papers swing between pre-trained with a hyphen and the closed form pretrained; the hyphen is the catalog spelling
Breaks the pattern: the noun phrase stays two words; a second hyphen appears only when the pair modifies another noun

Spelled like

Broken into chunks

What it sits beside

Same subject

Same shape

Training-stage labels

Open model compounds

Hyphenated pre- participles

Phase 2: How Models Are Built: Architectures and Learning Mechanics

With the paradigms named, zoom in on the architectural building blocks that turn data into trained artifacts.