Test 1 · ★★ · In the news: tricked bots, real discoveries, careful words

"AI won at Go so it's smarter than every human at everything now" Best reply?

Mova's choice: option 1 · The right answer: option 2 · wrong ✗ · Who acts: AI · What happens: win, be · To what, where: Go, everything

This is test 1 · ★★ of lesson 23 «In the news: tricked bots, real discoveries, careful words» of the AI course of Mova's primary school. 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	"	"	PUNCT	_	3 won	punct
2	AI	ai	NOUN	Number=Sing	3 won	nsubj
3	won	win	VERB	Mood=Ind Number=Sing Person=3 Tense=Past VerbForm=Fin	9 smarter	advcl
4	at	at	ADP	_	5 Go	case
5	Go	Go	PROPN	Number=Sing	3 won	obl
6	so	so	ADV	_	9 smarter	advmod
7	it	it	PRON	Case=Nom Gender=Neut Number=Sing Person=3 PronType=Prs	9 smarter	nsubj
8	's	be	AUX	Mood=Ind Number=Sing Person=3 Tense=Pres VerbForm=Fin	9 smarter	cop
9	smarter	smart	ADJ	Degree=Cmp	root	root
10	than	than	ADP	_	12 human	case
11	every	every	DET	PronType=Tot	12 human	det
12	human	human	NOUN	Number=Sing	9 smarter	obl
13	at	at	ADP	_	14 everything	case
14	everything	everything	PRON	Number=Sing PronType=Tot	9 smarter	obl
15	now	now	ADV	PronType=Dem	18 reply	advmod
16	"	"	PUNCT	_	18 reply	punct
17	Best	good	ADJ	Degree=Sup	18 reply	amod
18	reply	reply	NOUN	Number=Sing	14 everything	nmod
19	?	?	PUNCT	_	9 smarter	punct

#	verb	tense, mood	who	what, whom	to whom	where	when	other	roles by
1	win in a quote	past	AI			Go			vn succeed-74-2 «NP V PP.eventuality S_ING»: agent=Agent at go=Eventuality
2	is in a quote	present	AI			everything		than every human; smart property	tree rules

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

Yes, so people should stop learning

The words, one row each (CoNLL-U)

#	word	lemma	UPOS	features	head	relation
1	Yes	yes	INTJ	Polarity=Pos	6 stop	discourse
2	,	,	PUNCT	_	1 Yes	punct
3	so	so	ADV	_	6 stop	advmod
4	people	people	NOUN	Number=Plur	6 stop	nsubj
5	should	should	AUX	VerbForm=Fin	6 stop	aux
6	stop	stop	VERB	VerbForm=Inf	root	root
7	learning	learn	VERB	VerbForm=Ger	6 stop	xcomp

#	verb	tense, mood	who	what, whom	to whom	where	when	other	roles by
1	stop	should	people						vn stop-55.4-1-1 «NP V S_ING»: agent=Affector clause=Eventuality
2	learn	gerund	people not said here						vn learn-14-1 «NP V»: agent=Agent

Option 2 · ✓ the right answer

It's brilliant at Go; that doesn't make it good at everything

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 brilliant	nsubj
2	's	be	AUX	Mood=Ind Number=Sing Person=3 Tense=Pres VerbForm=Fin	3 brilliant	cop
3	brilliant	brilliant	ADJ	Degree=Pos	root	root
4	at	at	ADP	_	5 Go	case
5	Go	Go	PROPN	Number=Sing	3 brilliant	obl
6	;	;	PUNCT	_	10 make	punct
7	that	that	PRON	Number=Sing PronType=Dem	10 make	nsubj
8	does	do	AUX	Mood=Ind Number=Sing Person=3 Tense=Pres VerbForm=Fin	10 make	aux
9	n't	not	PART	Polarity=Neg	10 make	advmod
10	make	make	VERB	VerbForm=Inf	3 brilliant	parataxis
11	it	it	PRON	Case=Acc Gender=Neut Number=Sing Person=3 PronType=Prs	10 make	obj
12	good	good	ADJ	Degree=Pos	10 make	xcomp
13	at	at	ADP	_	14 everything	case
14	everything	everything	PRON	Number=Sing PronType=Tot	10 make	obl

#	verb	tense, mood	who	what, whom	to whom	where	when	other	roles by
1	is	present	It			Go		brilliant property	tree rules
2	not make	present · do-support	that	it		everything		good property	vn build-26.1-1 «NP.asset V NP»: agent=Asset it=Product

Option 3

No, it lost every game it played

The words, one row each (CoNLL-U)

#	word	lemma	UPOS	features	head	relation
1	No	no	INTJ	Polarity=Neg	4 lost	discourse
2	,	,	PUNCT	_	1 No	punct
3	it	it	PRON	Case=Nom Gender=Neut Number=Sing Person=3 PronType=Prs	4 lost	nsubj
4	lost	lose	VERB	Mood=Ind Number=Sing Person=3 Tense=Past VerbForm=Fin	root	root
5	every	every	DET	PronType=Tot	6 game	det
6	game	game	NOUN	Number=Sing	4 lost	obj
7	it	it	PRON	Case=Nom Gender=Neut Number=Sing Person=3 PronType=Prs	8 played	nsubj
8	played	play	VERB	Mood=Ind Number=Sing Person=3 Tense=Past VerbForm=Fin	6 game	acl:relcl

#	verb	tense, mood	who	what, whom	to whom	where	when	other	roles by
1	lose	past	it	every game					vn succeed-74-3-1 «NP V NP»: agent=Agent game=Eventuality
2	play	past	game	every game not said here					vn use-105.1 «NP V NP»: agent=Agent game=Theme

Option 4

Yes, Go is the hardest thing in the world, so it can do anything

The tree, 14 words

The words, one row each (CoNLL-U)

#	word	lemma	UPOS	features	head	relation
1	Yes	yes	INTJ	Polarity=Pos	7 thing	discourse
2	,	,	PUNCT	_	1 Yes	punct
3	Go	Go	PROPN	Number=Sing	7 thing	nsubj
4	is	be	AUX	Mood=Ind Number=Sing Person=3 Tense=Pres VerbForm=Fin	7 thing	cop
5	the	the	DET	Definite=Def PronType=Art	7 thing	det
6	hardest	hard	ADJ	Degree=Sup	7 thing	amod
7	thing	thing	NOUN	Number=Sing	root	root
8	in	in	ADP	_	10 world	case
9	the	the	DET	Definite=Def PronType=Art	10 world	det
10	world	world	NOUN	Number=Sing	7 thing	nmod
11	,	,	PUNCT	_	15 do	punct
12	so	so	ADV	_	15 do	advmod
13	it	it	PRON	Case=Nom Gender=Neut Number=Sing Person=3 PronType=Prs	15 do	nsubj
14	can	can	AUX	VerbForm=Fin	15 do	aux
15	do	do	VERB	VerbForm=Inf	7 thing	parataxis
16	anything	anything	PRON	Number=Sing PronType=Ind	15 do	obj

#	verb	tense, mood	who	what, whom	to whom	where	when	other	roles by
1	is a	present	Go	the hardest thing		the world			tree rules
2	do	can	world	anything					vn act-114-1 «NP V NP»: agent=Agent anything=Eventuality

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/23-news-tricked-bots-discoveries/t1/