Mova0.5 pre-release⌕

☰

home / corpora / questions / AI / How a language model learns: guessing… / e4

Exam question 4 · How a language model learns: guessing the next word

It is given that numerous distinct programs, when requested to select a number between one and ten, select seven. Ascertain the most probable explanation.

Mova's choice: option 1 · The right answer: option 1 · right ✓ · Who acts: numerous distinct programs, you · What happens: give, request, select, ascertain · To what, where: It, number, seven, probable explanation

This is exam question 4 of lesson 2 «How a language model learns: guessing the next word» of the AI 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

Sentence 1: It is given that numerous distinct programs, when requested to select a number between one and ten, select seven.

The words, one row each (CoNLL-U)

#	word	lemma	UPOS	features	head	relation
1	It	it	PRON	Case=Nom Gender=Neut Number=Sing Person=3 PronType=Prs	3 given	nsubj:pass
2	is	be	AUX	Mood=Ind Number=Sing Person=3 Tense=Pres VerbForm=Fin	3 given	aux:pass
3	given	give	VERB	Tense=Past VerbForm=Part Voice=Pass	root	root
4	that	that	SCONJ	_	20 select	mark
5	numerous	numerous	ADJ	Degree=Pos	7 programs	amod
6	distinct	distinct	ADJ	Degree=Pos	7 programs	amod
7	programs	program	NOUN	Number=Plur	20 select	nsubj
8	,	,	PUNCT	_	10 requested	punct
9	when	when	ADV	PronType=Int	10 requested	advmod
10	requested	request	VERB	Tense=Past VerbForm=Part Voice=Pass	7 programs	acl:relcl
11	to	to	PART	_	12 select	mark
12	select	select	VERB	VerbForm=Inf	10 requested	xcomp
13	a	a	DET	Definite=Ind PronType=Art	14 number	det
14	number	number	NOUN	Number=Sing	12 select	obj
15	between	between	ADP	_	16 one	case
16	one	one	NUM	NumForm=Word NumType=Card	14 number	nmod
17	and	and	CCONJ	_	18 ten	cc
18	ten	ten	NUM	NumForm=Word NumType=Card	16 one	conj
19	,	,	PUNCT	_	7 programs	punct
20	select	select	VERB	VerbForm=Inf	3 given	advcl
21	seven	seven	NUM	NumForm=Word NumType=Card	20 select	obj
22	.	.	PUNCT	_	3 given	punct

#	verb	tense, mood	who	what, whom	to whom	where	when	other	roles by
1	give	present · passive		It					vn give-13.1-1 «NP V NP»: subject=Theme
2	request	participle							vn beg-58.2 «NP V S_INF»: agent=Agent clause=Topic
3	select	to-infinitive		a number				between one	vn obtain-13.5.2 «NP V NP»: agent=Agent number=Theme
4	select	infinitive	numerous distinct programs	seven					vn obtain-13.5.2 «NP V NP»: agent=Agent seven=Theme

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.

Sentence 2: Ascertain the most probable explanation.

The words, one row each (CoNLL-U)

#	word	lemma	UPOS	features	head	relation
1	Ascertain	ascertain	VERB	Mood=Imp VerbForm=Fin	root	root
2	the	the	DET	Definite=Def PronType=Art	5 explanation	det
3	most	most	ADV	Degree=Sup	4 probable	advmod
4	probable	probable	ADJ	Degree=Pos	5 explanation	amod
5	explanation	explanation	NOUN	Number=Sing	1 Ascertain	obj
6	.	.	PUNCT	_	1 Ascertain	punct

#	verb	tense, mood	who	what, whom	to whom	where	when	other	roles by
1	ascertain	imperative	you not said here	the probable explanation					vn discover-84 «NP V NP»: agent=Agent explanation=Topic

The options

Option 1 · ← Mova's choice · ✓ the right answer

Such programs tend to reproduce the most frequent human preference

The words, one row each (CoNLL-U)

#	word	lemma	UPOS	features	head	relation
1	Such	such	ADJ	Degree=Pos	2 programs	amod
2	programs	program	NOUN	Number=Plur	3 tend	nsubj
3	tend	tend	VERB	Mood=Ind Number=Plur Person=3 Tense=Pres VerbForm=Fin	root	root
4	to	to	PART	_	5 reproduce	mark
5	reproduce	reproduce	VERB	VerbForm=Inf	3 tend	xcomp
6	the	the	DET	Definite=Def PronType=Art	10 preference	det
7	most	most	ADV	Degree=Sup	8 frequent	advmod
8	frequent	frequent	ADJ	Degree=Pos	10 preference	amod
9	human	human	ADJ	Degree=Pos	10 preference	amod
10	preference	preference	NOUN	Number=Sing	5 reproduce	obj

#	verb	tense, mood	who	what, whom	to whom	where	when	other	roles by
1	tend	present	Such programs						tree rules
2	reproduce	to-infinitive	Such programs not said here	the frequent human preference					vn other_cos-45.4 «NP V NP.patient»: agent=Agent preference=Patient

Option 2

The programs have conferred with one another in advance of the request

The words, one row each (CoNLL-U)

#	word	lemma	UPOS	features	head	relation
1	The	the	DET	Definite=Def PronType=Art	2 programs	det
2	programs	program	NOUN	Number=Plur	4 conferred	nsubj:pass
3	have	have	AUX	Mood=Ind Number=Plur Person=3 Tense=Pres VerbForm=Fin	4 conferred	aux
4	conferred	confer	VERB	Tense=Past VerbForm=Part Voice=Pass	root	root
5	with	with	ADP	_	6 one	case
6	one	one	NUM	NumForm=Word NumType=Card	4 conferred	obl
7	another	another	DET	PronType=Ind	6 one	fixed
8	in	in	ADP	_	9 advance	case
9	advance	advance	NOUN	Number=Sing	4 conferred	obl
10	of	of	ADP	_	12 request	case
11	the	the	DET	Definite=Def PronType=Art	12 request	det
12	request	request	NOUN	Number=Sing	9 advance	nmod

#	verb	tense, mood	who	what, whom	to whom	where	when	other	roles by
1	confer	present · perfect · passive		the programs		advance of the request		with one	tree rules
2	advance		the request						nom advance «event intransitive»: of=agent

Option 3

Seven is the only integer such programs are able to produce

The words, one row each (CoNLL-U)

#	word	lemma	UPOS	features	head	relation
1	Seven	seven	NUM	NumForm=Word NumType=Card	9 able	nsubj
2	is	be	AUX	Mood=Ind Number=Sing Person=3 Tense=Pres VerbForm=Fin	9 able	cop
3	the	the	DET	Definite=Def PronType=Art	5 integer	det
4	only	only	ADJ	Degree=Pos	5 integer	amod
5	integer	integer	NOUN	Number=Sing	9 able	nsubj
6	such	such	ADJ	Degree=Pos	7 programs	amod
7	programs	program	NOUN	Number=Plur	9 able	nsubj
8	are	be	AUX	Mood=Ind Number=Plur Person=3 Tense=Pres VerbForm=Fin	9 able	cop
9	able	able	ADJ	Degree=Pos	root	root
10	to	to	PART	_	11 produce	mark
11	produce	produce	VERB	VerbForm=Inf	9 able	xcomp

#	verb	tense, mood	who	what, whom	to whom	where	when	other	roles by
1	is	present	Seven					able property	tree rules
2	produce	to-infinitive	Seven not said here						vn create-26.4 «NP V NP»: agent=Agent

Option 4

The selection is governed by the date of the request

The words, one row each (CoNLL-U)

#	word	lemma	UPOS	features	head	relation
1	The	the	DET	Definite=Def PronType=Art	2 selection	det
2	selection	selection	NOUN	Number=Sing	4 governed	nsubj:pass
3	is	be	AUX	Mood=Ind Number=Sing Person=3 Tense=Pres VerbForm=Fin	4 governed	aux:pass
4	governed	govern	VERB	Tense=Past VerbForm=Part Voice=Pass	root	root
5	by	by	ADP	_	7 date	case
6	the	the	DET	Definite=Def PronType=Art	7 date	det
7	date	date	NOUN	Number=Sing	4 governed	obl
8	of	of	ADP	_	10 request	case
9	the	the	DET	Definite=Def PronType=Art	10 request	det
10	request	request	NOUN	Number=Sing	7 date	nmod

#	verb	tense, mood	who	what, whom	to whom	where	when	other	roles by
1	govern	present · passive	the date of the request	the selection					vn supervision-95.2.2 «NP V NP.theme»: agent=Agent subject=Theme

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/ai/2-guessing-the-next-word/e4/