---
title: Picture-making AI learned from millions of pictures with description… — How a language model learns: guessing the next w… · tree and pragmatics
description: Picture-making AI learned from millions of pictures with descriptions and builds new pictures that fit the words. Who acts: making AI, new pictures. A sentence of the lesson «How a language model learns: guessing the next word» of the ai course of Mova's primary school with its dependency tree and t
---

<!-- md -->
[Mova](https://md.mova.elamur.ai/free/en/) · [this page in HTML](https://mova.elamur.ai/free/en/corpora/school/ai/2-guessing-the-next-word/15/)

<!-- crumbs -->
[home](https://md.mova.elamur.ai/free/en/) / [corpora](https://md.mova.elamur.ai/free/en/corpora/) / [school](https://md.mova.elamur.ai/free/en/corpora/school/) / [AI](https://md.mova.elamur.ai/free/en/corpora/school/ai/) / [How a language model learns: guessing…](https://md.mova.elamur.ai/free/en/corpora/school/ai/2-guessing-the-next-word/) / 15

# How a language model learns: guessing the next word, sentence 15

<!-- .ven lang=en -->
Picture-making AI learned from millions of pictures with descriptions and builds new pictures that fit the words.

<!-- .vintro -->
**Who acts:** making AI, new pictures · **What happens:** make, learn, build, fit · **To what, where:** millions, new pictures, words

<!-- .vabout -->
This is a sentence of the lesson «How a language model learns: guessing the next word» of the ai course of Mova's primary school. The page gives its syntax tree — a dependency tree in the style of Universal Dependencies — and its pragmatics clause by clause: who acts (making AI, new pictures), what happens, when and where, as Mova's frames read it from the tree.

## The tree {#tree}

```conllu
# sent_id = school:ai:002-guessing-the-next-word.lesson:15
# text = Picture-making AI learned from millions of pictures with descriptions and builds new pictures that fit the words.
# source = school lesson (ours)
# course = ai
# lesson = 002-guessing-the-next-word.lesson
# annotation = Mova annot: Rust (en) → Haiku 5.5 (medium, Batches API) patches; gates UD 2.18; agreed by treeagree rules (Stanza, Trankit, spaCy trf): no change
1	Picture	picture	NOUN	NN	Number=Sing	3	compound	_	_
2	-	-	PUNCT	HYPH	_	1	punct	_	_
3	making	make	VERB	VBG	VerbForm=Ger	4	amod	_	_
4	AI	ai	NOUN	NN	Number=Sing	5	nsubj	_	_
5	learned	learn	VERB	VBD	Mood=Ind|Number=Sing|Person=3|Tense=Past|VerbForm=Fin	0	root	_	_
6	from	from	ADP	IN	_	7	case	_	_
7	millions	million	NOUN	NNS	Number=Plur	5	obl	_	_
8	of	of	ADP	IN	_	9	case	_	_
9	pictures	picture	NOUN	NNS	Number=Plur	7	nmod	_	_
10	with	with	ADP	IN	_	11	case	_	_
11	descriptions	description	NOUN	NNS	Number=Plur	9	nmod	_	_
12	and	and	CCONJ	CC	_	13	cc	_	_
13	builds	build	VERB	VBZ	Mood=Ind|Number=Sing|Person=3|Tense=Pres|VerbForm=Fin	5	conj	_	_
14	new	new	ADJ	JJ	Degree=Pos	15	amod	_	_
15	pictures	picture	NOUN	NNS	Number=Plur	13	obj	_	_
16	that	that	PRON	WDT	PronType=Rel	17	nsubj	_	_
17	fit	fit	VERB	VBP	Mood=Ind|Number=Plur|Person=3|Tense=Pres|VerbForm=Fin	15	acl:relcl	_	_
18	the	the	DET	DT	Definite=Def|PronType=Art	19	det	_	_
19	words	word	NOUN	NNS	Number=Plur	17	obj	_	_
20	.	.	PUNCT	.	_	5	punct	_	_
```

## Pragmatics {#pragmatics}

<!-- .note -->
No stored speech act for this sentence.

<!-- .fr -->
| # | verb | tense, mood | who | what, whom | to whom | where | when | other | roles by |
| --- | --- | --- | --- | --- | --- | --- | --- | --- | --- |
| 1 | **make** | gerund |  |  |  |  |  |  | vn build-26.1-1 «NP.asset V NP»: agent=Asset |
| 2 | **learn** | past | making AI |  |  | millions of pictures with descriptions |  |  | vn learn-14-2-1 «NP V PP.source»: agent=Agent from million=Source |
| 3 | **build** | present | making AI *not said here* | new pictures |  |  |  |  | vn build-26.1-1 «NP.asset V NP»: agent=Asset picture=Product |
| 4 | **fit** | present | new pictures | the words |  |  |  |  | vn adjust-26.9 «NP V NP»: agent=Agent word=Patient |

<!-- .note -->
One row per clause the reader finds. «not said here»: a role filled from the clause above. «roles by»: a VerbNet class and frame when one fits the verb, else the reader's rules on the tree.

<!-- div src -->

**Licence: [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/)**, as the whole of Mova corpus 0.4, of which this treebank (the agreed school, `mova-school-agreed-en.conllu`) is part. The lessons were written for Mova's primary school. Sentence `school:ai:002-guessing-the-next-word.lesson:15`.

**The tree:** Mova annot: Rust (en) → Haiku 5.5 (medium, Batches API) patches; gates UD 2.18; agreed by treeagree rules (Stanza, Trankit, spaCy trf): no change. **The frames** are read from the tree by Mova's frames reader (`prag`) when this page is built, the lesson read in order, so a pronoun may be resolved from an earlier sentence.

See the [release 0.4](https://md.mova.elamur.ai/free/en/releases/0.4/).

<!-- /div -->

<!-- pn -->
- ← previous: [How a language model learns: guessing…, 14](https://md.mova.elamur.ai/free/en/corpora/school/ai/2-guessing-the-next-word/14/)
- next →: [How a language model learns: guessing…, 16](https://md.mova.elamur.ai/free/en/corpora/school/ai/2-guessing-the-next-word/16/)
- ↑ back to [the lesson](https://md.mova.elamur.ai/free/en/corpora/school/ai/2-guessing-the-next-word/)