Exam question 1 · AI pictures and AI music, as reported

Identify the principal manner in which AI image generators acquire their capacity to produce pictures.

Mova's choice: option 1 · The right answer: option 2 · wrong ✗ · Who acts: you, image generators · What happens: identify, acquire, produce · To what, where: principal manner, capacity, pictures

This is exam question 1 of lesson 36 «AI pictures and AI music, as reported» of the arts course of Mova's primary school, asked in stiff, formal language. The page gives the question's syntax tree — a dependency tree in the style of Universal Dependencies — and its pragmatics clause by clause, then the 4 options as the student saw them, each with its tree and pragmatics. Every text is parsed by Mova itself. How the student answered it, with the score of every option and why: the test at school.

The question

The words, one row each (CoNLL-U)

#	word	lemma	UPOS	features	head	relation
1	Identify	identify	VERB	Mood=Imp VerbForm=Fin	root	root
2	the	the	DET	Definite=Def PronType=Art	4 manner	det
3	principal	principal	ADJ	Degree=Pos	4 manner	amod
4	manner	manner	NOUN	Number=Sing	1 Identify	obj
5	in	in	ADP	_	6 which	case
6	which	which	DET	PronType=Int	9 generators	det
7	AI	ai	NOUN	Number=Sing	8 image	compound
8	image	image	NOUN	Number=Sing	9 generators	compound
9	generators	generator	NOUN	Number=Plur	10 acquire	nsubj
10	acquire	acquire	VERB	Mood=Ind Number=Plur Person=3 Tense=Pres VerbForm=Fin	1 Identify	parataxis
11	their	their	PRON	Case=Gen Number=Plur Person=3 Poss=Yes PronType=Prs	12 capacity	nmod:poss
12	capacity	capacity	NOUN	Number=Sing	10 acquire	obj
13	to	to	PART	_	14 produce	mark
14	produce	produce	VERB	VerbForm=Inf	10 acquire	advcl
15	pictures	picture	NOUN	Number=Plur	14 produce	obj
16	.	.	PUNCT	_	1 Identify	punct

#	verb	tense, mood	who	what, whom	roles by
1	identify	imperative	you not said here	the principal manner	vn characterize-29.2-1-1 «NP V NP S_ING»: agent=Agent manner=Theme
2	acquire	present	which image generators	generators's capacity	vn obtain-13.5.2-1 «NP.asset V NP»: agent=Asset capacity=Theme
3	produce	to-infinitive	which image generators not said here	pictures	vn create-26.4 «NP V NP»: agent=Agent picture=Result

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.

The options

Option 1 · ← Mova's choice

By inventing images without any exposure whatsoever to prior work

The words, one row each (CoNLL-U)

#	word	lemma	UPOS	features	head	relation
1	By	by	SCONJ	_	2 inventing	mark
2	inventing	invent	VERB	Tense=Pres VerbForm=Part	root	root
3	images	image	NOUN	Number=Plur	2 inventing	obj
4	without	without	ADP	_	6 exposure	case
5	any	any	DET	PronType=Ind	6 exposure	det
6	exposure	exposure	NOUN	Number=Sing	2 inventing	obl
7	whatsoever	whatsoever	ADV	_	2 inventing	advmod
8	to	to	ADP	_	10 work	case
9	prior	prior	ADJ	Degree=Pos	10 work	amod
10	work	work	NOUN	Number=Sing	2 inventing	obl

#	verb	tense, mood	who	what, whom	to whom	where	when	other	roles by
1	invent	gerund		images				without any exposure; to prior work; whatsoever property	vn create-26.4-1 «NP V NP»: agent=Agent image=Result

Option 2 · ✓ the right answer

Training on many images made by people

The words, one row each (CoNLL-U)

#	word	lemma	UPOS	features	head	relation
1	Training	train	VERB	Tense=Pres VerbForm=Part	root	root
2	on	on	ADP	_	4 images	case
3	many	many	ADJ	Degree=Pos	4 images	amod
4	images	image	NOUN	Number=Plur	1 Training	obl
5	made	make	VERB	Tense=Past VerbForm=Part Voice=Pass	4 images	acl
6	by	by	ADP	_	7 people	case
7	people	people	NOUN	Number=Plur	5 made	obl:agent

#	verb	tense, mood	who	what, whom	to whom	where	when	other	roles by
1	train	gerund				many images			vn hire-13.5.3 «NP V NP»: agent=Agent
2	make	participle		many images not said here					vn build-26.1-1 «NP.asset V NP»: agent=Asset

Option 3

Copying one artist each time

The words, one row each (CoNLL-U)

#	word	lemma	UPOS	features	head	relation
1	Copying	copy	VERB	Tense=Pres VerbForm=Part	root	root
2	one	one	NUM	NumForm=Word NumType=Card	3 artist	nummod
3	artist	artist	NOUN	Number=Sing	1 Copying	obj
4	each	each	DET	PronType=Tot	5 time	det
5	time	time	NOUN	Number=Sing	1 Copying	obl:unmarked

#	verb	tense, mood	who	what, whom	to whom	where	when	other	roles by
1	copy	gerund		one artist			time		vn scribble-25.2 «NP V NP»: agent=Agent artist=Theme

Option 4

By employing concealed human illustrators to fulfil every request

The words, one row each (CoNLL-U)

#	word	lemma	UPOS	features	head	relation
1	By	by	SCONJ	_	2 employing	mark
2	employing	employ	VERB	Tense=Pres VerbForm=Part	root	root
3	concealed	conceal	VERB	Tense=Past VerbForm=Part Voice=Pass	5 illustrators	amod
4	human	human	ADJ	Degree=Pos	5 illustrators	amod
5	illustrators	illustrator	NOUN	Number=Plur	2 employing	obj
6	to	to	PART	_	7 fulfil	mark
7	fulfil	fulfil	VERB	VerbForm=Inf	2 employing	advcl
8	every	every	DET	PronType=Tot	9 request	det
9	request	request	NOUN	Number=Sing	7 fulfil	obj

#	verb	tense, mood	what, whom	roles by
1	employ	gerund	concealed human illustrators	vn hire-13.5.3 «NP V NP»: agent=Agent illustrator=Theme
2	conceal	participle		vn concealment-16 «NP V NP»: agent=Agent
3	fulfil	to-infinitive	every request	tree rules

Text: the tests of Mova's primary school, written for Mova (world/data/school/), as the course runner gave them to the student. Trees: Mova's own parse — the en model on the CPU, when the site is built — not checked by a teacher; they may be wrong. Frames: Mova's frames reader (prag), each text read on its own. Elamur Cosmic License.

This page in Markdown → https://md.mova.elamur.ai/free/en/corpora/questions/arts/36-ai-pictures-and-ai-music-as-reported/e1/