Antonymy and Conceptual Vectors Didier Schwab Mathieu Lafourcade

Antonymy and Conceptual Vectors Didier Schwab, Mathieu Lafourcade, Violaine Prince presented by Ch. Boitet (works with M. Lafourcade on conceptual vectors & UNL) Schwab, Lafourcade, Prince, pres. by Ch. Boitet Laboratoire d’informatique, de robotique Et de microélectronique de Montpellier CNRS - Université Montpellier II Antonymy and Conceptual Vectors

Outline n n n The main idea Background on conceptual vectors How we use CVs & why we need to distinguish CVs of antonyms n n n Brief study of antonymies Representation of antonymies Measure for « antonymousness » Schwab, Lafourcade, Prince, pres. by Ch. Boitet Antonymy and Conceptual Vectors

The main idea n Work on meaning representation in NLP, using conceptual vectors (CV) n n applications = WSD & thematic indexing but V(existence) = V(non-existence) ! basic « concepts » activated the same n Idea: n n use lexical functions to improve the adequacy For this, « transport » the lexical functions in the vector space Schwab, Lafourcade, Prince, pres. by Ch. Boitet Antonymy and Conceptual Vectors

Background on conceptual vectors n Lexical Item = ideas = combination of concepts = Vector V n Ideas space = vector space (generator space) n Concept = idea = vector Vc n Vc taken from a thesaurus hierarchy (Larousse) n translation of Roget’s thesaurus, 873 leaf nodes n the word ‘peace’ has non zero values for concept PEACE and other concepts Schwab, Lafourcade, Prince, pres. by Ch. Boitet Antonymy and Conceptual Vectors

Our conceptual vectors Thesaurus • H : thesaurus hierarchy — K concepts Thesaurus Larousse = 873 concepts • V(Ci) : <a 1, …, ai, … , a 873> aj = 1/ (2 ** Dum(H, i, j)) 1/16 1/4 4 Schwab, Lafourcade, Prince, pres. by Ch. Boitet 1 2 1/4 1/64 6 Antonymy and Conceptual Vectors

Conceptual vectors Concept c 4: ‘PEACE’ peace conflict relations hierarchical relations The world, manhood Schwab, Lafourcade, Prince, pres. by Ch. Boitet society Antonymy and Conceptual Vectors

Conceptual vectors Term “peace” c 4: ’PEACE’ Schwab, Lafourcade, Prince, pres. by Ch. Boitet Antonymy and Conceptual Vectors

exchange profit finance Schwab, Lafourcade, Prince, pres. by Ch. Boitet Antonymy and Conceptual Vectors

Angular or « thematic » distance n Da(x, y) = angle(x, y) = acos(sim(x, y)) = acos(x. y /|x ||y |) n 0 ≤ D(x, y) ≤ (positive components) n n If 0 then x and y are colinear : same idea. If /2 : nothing in common. x y Schwab, Lafourcade, Prince, pres. by Ch. Boitet Antonymy and Conceptual Vectors

Thematic Distance (examples) n n n Da(anteater , anteater ) Da(anteater , animal ) Da(anteater , train ) Da(anteater , mammal ) Da(anteater , quadruped ) Da(anteater , ant ) = 0 (0°) = 0, 45 (26°) = 1, 18 (68°) = 0, 36 (21°) = 0, 42 (24°) = 0, 26 (15°) thematic distance ≠ ontological distance Schwab, Lafourcade, Prince, pres. by Ch. Boitet Antonymy and Conceptual Vectors

Vector Proximity n Function V gives the vectors closest to a lexical item. V (life) = life, alive, birth… V (death) = death, to die, to kill… Schwab, Lafourcade, Prince, pres. by Ch. Boitet Antonymy and Conceptual Vectors

How we build & use conceptual vectors n Conceptual vectors give thematic representations n n New CVs for word senses are permanently learned from NL definitions n n of word senses of words (averaging CVs of word senses) of the content ( « ideas » ) of any textual segment coming from electronic dictionaries CVs of word senses are permanently recomputed n for French, 3 years, 100000 words, 300000 CVs Schwab, Lafourcade, Prince, pres. by Ch. Boitet Antonymy and Conceptual Vectors

Continuous building of the conceptual vectors database Definitions SYGMART Human usage dictionaries Morphosyntactic analysis Conceptual vectors base New Vector Schwab, Lafourcade, Prince, pres. by Ch. Boitet Antonymy and Conceptual Vectors

We should distinguish CVs of different but related words… Non-existent : who or which does not exist cold : #ant# warm, hot n Without a specific treatment, we get V(non-existence) = V(existence) V(cold) = V(hot) n We want to obtain V(non-existence) ≠ V(existence) V(cold) ≠ V(hot) Schwab, Lafourcade, Prince, pres. by Ch. Boitet Antonymy and Conceptual Vectors

…in order to improve applications and resources n Applications: more precision n Thematic analysis of texts Thematic analysis of definitions Resources: coherence & adequacy n n General coherence of the CV data base Conceptual Vector quality (adequacy) Schwab, Lafourcade, Prince, pres. by Ch. Boitet Antonymy and Conceptual Vectors

Lexical functions may help! n Lexical function (Mel’tchuk): WS {WS 1…WSn} synonymy (#Syn#), antonymy (#Anti#), intensification (#Magn#)… n Examples : #Syn# (car) = {automobile} #Anti# (respect) = {disrespect; disdain} #Sing# (fleet) = {boat, ship; embarcation} Schwab, Lafourcade, Prince, pres. by Ch. Boitet Antonymy and Conceptual Vectors

Method: transport the LFs as functions on the CV space n e. g. for antonymy, to get n V(non-existence) ≠ V(existence) find vector function Anti such that: V(non-existence) = V(#Anti#(existence)) = Anti (V(existence)) n similarly for other lexical functions we simply began by studying antinomy Schwab, Lafourcade, Prince, pres. by Ch. Boitet Antonymy and Conceptual Vectors

Brief study of antonymy Definition : Two lexical items are in antonymy relation if there is a symmetry between their semantic components relatively to an axis Antonymy relations depend on the type of medium that supports symmetry There are several types of antonymy On the axis, there are fixed points: Anti (V(car)) = V(car) because #Anti# (car) = Schwab, Lafourcade, Prince, pres. by Ch. Boitet Antonymy and Conceptual Vectors

1 - Complementary antonymy Values are boolean & symmetric (0 1) Examples : event/non-event dead/alive existence/non-existence He is present He is not absent He is absent He is not present Schwab, Lafourcade, Prince, pres. by Ch. Boitet Antonymy and Conceptual Vectors

2 - Scalar antonymy n n Values are scalar Symmetry is relative to a reference value Examples : cold/hot, small/tall This man is small This man is not tall This man is tall This man is not small This man is neither tall nor small reference value = « of medium height » Schwab, Lafourcade, Prince, pres. by Ch. Boitet Antonymy and Conceptual Vectors

3 - Dual Antonymy (1) n Conversive duals same semantics but inversion of roles Examples : sell/buy, husband/wife, father/son Jack is John’s son John is Jack’s father Jack sells a car to John buys a car from Jack Schwab, Lafourcade, Prince, pres. by Ch. Boitet Antonymy and Conceptual Vectors

3 - Dual Antonymy (2) n Contrastive duals contrastive expressions accepted by usage n Cultural : sun/moon, yin/yang n Associative : question/answer n Spatio-temporal : birth/death, start/finish Schwab, Lafourcade, Prince, pres. by Ch. Boitet Antonymy and Conceptual Vectors

Coherence and adequacy of the base n n Learning bootstrap based on a kernel composed of pre-computed vectors considered as adequate Learning must be coherent = preserve adequacy Adequacy = judgement that activations of concepts (coordinates) make sense for the meaning corresponding to a definition For coherence improvement, we use semantic relations between terms Schwab, Lafourcade, Prince, pres. by Ch. Boitet Antonymy and Conceptual Vectors

Antonymy function n Based on the antonym vectors of concepts : one list for each kind of antonymy Antic (EXISTENCE) Antis (HOT) Antic (GAME) n = V (NON-EXISTENCE) = V (COLD) = V (GAME) Anti (X, C) builds the vector « opposite » of vector X in context C Schwab, Lafourcade, Prince, pres. by Ch. Boitet Antonymy and Conceptual Vectors

Construction of the antonym vector of X in context C n n n The method is to focus on the salient notions in V(X) and V(C) If the notions can be opposed, then the antonym should have the inverse ideas in the same proportions The following formula was obtained after several experiments Schwab, Lafourcade, Prince, pres. by Ch. Boitet Antonymy and Conceptual Vectors

Construction of the antonym vector (2) n N Anti. R (V(X), V(C)) = Pi *Anti. C (Ci, V(C)) i=1 n 1+CV(V(X)) Pi = V * max (V(X), V(Ci)) Xi n n n Not symmetrical Stress more on vector X than on context C Consider an important idea of the vector to oppose even if it is not in the referent Schwab, Lafourcade, Prince, pres. by Ch. Boitet Antonymy and Conceptual Vectors

Results n n n V (#Anti# (death, life & death)) = (LIFE 0, 3), (birth 0, 48), (alive 0, 54)… V (#Anti# (life, life & death)) = (death 0, 336), (killer 0, 45), (murdered 0, 53)… V (#Anti# (LIFE)) = (DEATH 0, 034), (death 0, 43), (killer 0, 53). . . Schwab, Lafourcade, Prince, pres. by Ch. Boitet Antonymy and Conceptual Vectors

Antonymy evaluation measure n n Assess « how much » two lexical items are antonymous Manti(A, B) = DA(A B, Anti(A, C) Anti(B, C)) A Anti(B) Anti(A) B Schwab, Lafourcade, Prince, pres. by Ch. Boitet Antonymy and Conceptual Vectors

Examples n n n Manti (EXISTENCE, NON-EXISTENCE) = 0, 03 Manti (existence, non-existence) = 0, 44 Manti (EXISTENCE, CAR) = 1, 45 Manti (existence, car) = 1, 06 Manti (CAR, CAR) = 0, 006 Manti (car, car) = 0, 407 Schwab, Lafourcade, Prince, pres. by Ch. Boitet Antonymy and Conceptual Vectors

Conclusion and perspectives n Progress so far : n n Antonymy definition based on a notion of symmetry Implemented formula to compute an antonym vector Implemented measure to assess the level of antonymy between two items Perspectives : n n n Use of the symbolic opposition found in dictionaries Search the opposite meaning of a word Study of the other semantic relations (hyperonymy/hyponymy, meronymy/holonymy…) Schwab, Lafourcade, Prince, pres. by Ch. Boitet Antonymy and Conceptual Vectors