Word Sense and Subjectivity Jan Wiebe Rada Mihalcea

Word Sense and Subjectivity Jan Wiebe Rada Mihalcea University of Pittsburgh University of North Texas

Introduction Growing interest in the automatic extraction of opinions, emotions, and sentiments in text (subjectivity)

Subjectivity Analysis: Applications Opinion-oriented question answering: How do the Chinese regard the human rights record of the United States? Product review mining: What features of the Think. Pad T 43 do customers like and which do they dislike? Review classification: Is a review positive or negative toward the movie? Tracking emotions toward topics over time: Is anger ratcheting up or cooling down toward an issue or event? Etc.

Introduction Continuing interest in word sense – Sense annotated resources being developed for many languages » www. globalwordnet. org – Active participation in evaluations such as SENSEVAL

Word Sense and Subjectivity Though both are concerned with text meaning, they have mainly been investigated independently

Subjectivity Labels on Senses S Alarm, dismay, consternation – (fear resulting from the awareness of danger) O Alarm, warning device, alarm system – (a device that signals the occurrence of some undesirable event)

Subjectivity Labels on Senses S Interest, involvement -- (a sense of concern with and curiosity about someone or something; "an interest in music") O Interest -- (a fixed charge for borrowing money; usually a percentage of the amount borrowed; "how much interest do you pay on your mortgage? ")

WSD using Subjectivity Tagging He spins a riveting plot which grabs and holds the reader’s interest. Sense 4 “a sense of concern with and curiosity about someone or something” S Sense 1 “a fixed charge for borrowing money” O Sense 4 Sense 1? WSD System Sense 1 Sense 4? The notes do not pay interest.

WSD using Subjectivity Tagging He spins a riveting plot which grabs and holds the reader’s interest. S Subjectivity Classifier Sense 4 “a sense of concern with and curiosity about someone or something” S Sense 1 “a fixed charge for borrowing money” O O Sense 4 Sense 1? WSD System Sense 1 Sense 4? The notes do not pay interest.

WSD using Subjectivity Tagging He spins a riveting plot which grabs and holds the reader’s interest. S Subjectivity Classifier Sense 4 “a sense of concern with and curiosity about someone or something” S Sense 1 “a fixed charge for borrowing money” O O Sense 4 Sense 1? WSD System Sense 1 Sense 4? The notes do not pay interest.

Subjectivity Tagging using WSD S O? He spins a riveting plot which grabs and holds the reader’s interest. Subjectivity Classifier O S? The notes do not pay interest.

Subjectivity Tagging using WSD S O? He spins a riveting plot which grabs and holds the reader’s interest. Sense 4 S Sense 4 “a sense of Subjectivity Classifier concern with and curiosity about someone or something” O Sense 1 “a fixed charge WSD System for borrowing money” O S? Sense 1 The notes do not pay interest.

Subjectivity Tagging using WSD S O? He spins a riveting plot which grabs and holds the reader’s interest. Sense 4 S Sense 4 “a sense of Subjectivity Classifier concern with and curiosity about someone or something” O Sense 1 “a fixed charge WSD System for borrowing money” O S? Sense 1 The notes do not pay interest

Goals Explore interactions between word sense and subjectivity – Can subjectivity labels be assigned to word senses? » Manually » Automatically – Can subjectivity analysis improve word sense disambiguation? – Can word sense disambiguation improve subjectivity analysis? Future work

Outline Motivation and Goals Assigning Subjectivity Labels to Word Senses – Manually – Automatically Word Sense Disambiguation using Automatic Subjectivity Analysis Conclusions

Prior Work on Subjectivity Tagging Identifying subjectivity words and phrases associated with – Think ~ private state; Beautiful ~ positive sentiment » Hatzivassiloglou & Mc. Keown 1997; Wiebe 2000; Kamps & Marx 2002; Turney 2002; Esuli & Sabastiani 2005; Etc Subjectivity classification of sentences, clauses, phrases, or word instances in context – subjective/objective; positive/negative/neutral » Riloff & Wiebe 2003; Yu & Hatzivassiloglou 2003; Dave et al 2003; Hu & Liu 2004; Kim & Hovy 2004; Etc. Here: subjectivity labels are applied to word senses

Outline Motivation and Goals Assigning Subjectivity Labels to Word Senses – Manually – Automatically Word Sense Disambiguation using Automatic Subjectivity Analysis Conclusions

Annotation Scheme Assigning subjectivity labels to Word. Net senses – S: subjective – O: objective – B: both

Annotators are given the synset and its hypernym S Alarm, dismay, consternation – (fear resulting form the awareness of danger) – Fear, fearfulness, fright – (an emotion experiences in anticipation of some specific pain or danger (usually accompanied by a desire to flee or fight))

Subjective Sense Definition When the sense is used in a text or conversation, we expect it to express subjectivity, and we expect the phrase/sentence containing it to be subjective.

Objective Senses: Observation We don’t necessarily expect phrases/sentences containing objective senses to be objective – Would you actually be stupid enough to pay that rate of interest? – Will someone shut that darn alarm off? Subjective, but not due to interest or alarm

Objective Sense Definition When the sense is used in a text or conversation, we don’t expect it to express subjectivity and, if the phrase/sentence containing it is subjective, the subjectivity is due to something else.

Senses that are Both Covers both subjective and objective usages Example: absorb, suck, imbibe, soak up, sop up, suck up, draw, take in, take up – (take in, also metaphorically; “The sponge absorbs water well”; “She drew strength from the Minister’s Words”)

Annotated Data 64 words; 354 senses – Balanced subset [32 words; 138 senses]; 2 judges – The ambiguous nouns of the SENSEVAL-3 English Lexical Task [20 words; 117 senses]; 2 judges » [Mihalcea, Chklovski & Kilgarriff, 2004] – Others [12 words; 99 senses]; 1 judge

Annotated Data: Agreement Study 64 words; 354 senses – Balanced subset [32 words; 138 senses]; 2 judges » 16 words have both S and O senses » 16 words do not (8 only S and 8 only O) » All subsets balanced between nouns and verbs » Uncertain tags also permitted

Inter-Annotator Agreement Results Overall: – Kappa=0. 74 – Percent Agreement=85. 5%

Inter-Annotator Agreement Results Overall: – Kappa=0. 74 – Percent Agreement=85. 5% Without the 12. 3% cases when a judge is U: – Kappa=0. 90 – Percent Agreement=95. 0%

Inter-Annotator Agreement Results Overall: – Kappa=0. 74 – Percent Agreement=85. 5% 16 words with S and O senses: Kappa=0. 75 16 words with only S or O: Kappa=0. 73 Comparable difficulty

Inter-Annotator Agreement Results 64 words; 354 senses – The ambiguous nouns of the SENSEVAL-3 English Lexical Task [20 words; 117 senses] 2 judges » U tags not permitted » Even so, Kappa=0. 71

Outline Motivation and Goals Assigning Subjectivity Labels to Word Senses – Manually – Automatically Word Sense Disambiguation using Automatic Subjectivity Analysis Conclusions

Related Work unsupervised word-sense ranking algorithm of [Mc. Carthy et al 2004] – That task: approximate corpus frequencies of word senses – Our task: predict a word-sense property (subjectivity) method 2000] for learning subjective adjectives of [Wiebe – That task: label words – Our task: label word senses

Overview Main idea: assess the subjectivity of a word sense based on information about the subjectivity of – a set of distributionally similar words – in a corpus annotated with subjective expressions

MPQA Opinion Corpus 10, 000 sentences from the world press annotated for subjective expressions » [Wiebe at al. , 2005] » www. cs. pitt. edu/mpqa

Subjective Expressions Subjective expressions: opinions, sentiments, speculations, etc. (private states) expressed in language

Examples His alarm grew. The leaders roundly condemned the Iranian President’s verbal assault on Israel. He would be quite a catch. That doctor is a quack.

Preliminaries: subjectivity of word w Unannotated Corpus (BNC) Lin 1998 Annotated Corpus (MPQA) subj(w) = #insts(DSW) in SE - #insts(DSW) not in SE #insts (DSW) DSW = {dsw 1, …, dswj}

Subjectivity of word w Unannotated Corpus (BNC) Annotated Corpus (MPQA) subj(w) = #insts(DSW) in SE - #insts(DSW) not in SE #insts (DSW) [-1, 1] [highly objective, highly subjective] DSW = {dsw 1, …, dswj}

Subjectivity of word w Unannotated Corpus (BNC) Annotated Corpus (MPQA) dsw 1 inst 1 +1 dsw 1 inst 2 -1 dsw 2 inst 1 +1 subj(w) = +1 -1 +1 3 DSW = {dsw 1, dsw 2} = 1/3

Subjectivity of word sense wi Annotated Corpus (MPQA) [-1, 1] subj(wi) = Rather than 1, add or subtract sim(wi, dswj) dsw 1 inst 1 +sim(wi, dsw 1) dsw 1 inst 2 -sim(wi, dsw 1) dsw 2 inst 1 +sim(wi, dsw 2) +sim(wi, dsw 1) - sim(wi, dsw 1) + sim(wi, dsw 2) 2 * sim(wi, dsw 1) + sim(wi, dsw 2)

Method –Step 1 Given word w Find distributionally similar words [Lin 1998] – DSW = {dswj | j = 1. . n} – Experiment with top 100 and 160

Method –Step 2 word w = Alarm Sense w 1 “fear resulting from the awareness of danger” Sense w 2 “a device that signals the occurrence of some undesirable DSW 1 Panic DSW 2 Detector sim(w 1, panic) sim(w 1, detector ) sim(w 2, panic) sim(w 2, detector)

Method – Step 2 Find the similarity between each word sense and each distributionally similar word wnss can be any concept-based similarity measure between word senses we use Jiang & Conrath 1997

Method – Step 2 Find the similarity between each word sense and each distributionally similar word wnss can be any concept-based similarity measure between word senses we use Jiang & Conrath 1997

Method – Step 2 Find the similarity between each word sense and each distributionally similar word wnss can be any concept-based similarity measure between word senses we use Jiang & Conrath 1997

Method – Step 2 Find the similarity between each word sense and each distributionally similar word wnss can be any concept-based similarity measure between word senses we use Jiang & Conrath 1997

Method – Step 2 Find the similarity between each word sense and each distributionally similar word wnss can be any concept-based similarity measure between word senses we use Jiang & Conrath 1997

Method –Step 3 Input: word sense wi of word w DSW = {dswj | j = 1. . n} sim(wi, dswj) MPQA Opinion Corpus Output: subjectivity score subj(wi)

Method –Step 3 totalsim = #insts(dswj) * sim(wi, dswj) subj = 0 for each dswj in DSW: for each instance k in insts(dswj): if k is in a subjective expression: subj += sim(wi, dswj) else: subj -= sim(wi, dswj) subj(wi) = subj / totalsim

Method – Optional Variation if k is in a subjective expression: subj += sim(wi, dswj) else: subj -= sim(wi, dswj) w 1 w 2 dsw 1 dsw 2 dsw 3 w 3 dsw 1 dsw 2 dsw 3 “Selected”

Evaluation Calculate subj scores for all word senses, and sort them While 0 is a natural candidate for division between S and O, we perform the evaluation for different thresholds in [-1, +1] Calculate the precision of the algorithm at different points of recall

Evaluation Automatic assignment of subjectivity for 272 word senses (no DSW instances for 82 senses) Baseline: random selection of S labels » Number of assigned S labels matches number of S labels in the gold standard (recall = 1. 0)

Evaluation: precision/recall curves Number of distributionally similar words = 160

Evaluation Break-even point » Point where precision and recall are equal

Outline Motivation and Goals Assigning Subjectivity Labels to Word Senses – Manually – Automatically Word Sense Disambiguation using Automatic Subjectivity Analysis Conclusions

Overview Augment an existing WSD system with a feature reflecting the subjectivity of the context of the ambiguous word Compare the performance of original and subjectivity-aware WSD systems The ambiguous nouns of the SENSEVAL-3 English Lexical Task SENSEVAL-3 data

Original WSD System Integrates local and topical features: » Local: context of three words to the left and right, their part-of-speech » Topical: top five words occurring at least three times in the context of a word sense » [Ng & Lee, 1996], [Mihalcea, 2002] Naïve Bayes classifier » [Lee & Ng, 2003]

Automatic Subjectivity Classifier Rule-based automatic sentence classifier from [Wiebe & Riloff 2005] Included in Opinion. Finder; available at: – www. cs. pitt. edu/mpqa/

Subjectivity Tagging for WSD Used to tag sentences of the SENSEVAL-3 data that contain target nouns Sentencej Sentencek “interest” “atmosphere” S Subjectivity Classifier O S … … Sentencei

WSD using Subjectivity Tagging Sentencei Original WSD System “interest” Sense 4 Sense 1 S Subjectivity Classifier S, O, or B Subjectivity Aware WSD System Sense 4 “a fixed charge for borrowing money” Sense 1 “a sense of concern with and curiosity about someone or something”

Words with S and O Senses < < < < = 4. 3% error reduction; significant (p < 0. 05 paired t-test)

Words with Only O Senses > = = = < = > = = =

Conclusions Can subjectivity labels be assigned to word senses? – Manually » Good agreement; Kappa=0. 74 » Very good when uncertain cases removed; Kappa=0. 90 – Automatically » Method substantially outperforms baseline » Showed feasibility of assigning subjectivity labels to the fine-grained level of word senses

Conclusions Can subjectivity analysis improve word sense disambiguation? – Improves performance, but mainly for words with both S and O senses (4. 3% error reduction; significant (p < 0. 05)) – Performance largely remains the same or degrades for words that don’t – Assign subjectivity labels to Word. Net; WSD system should consult Word. Net tags to decide when to pay attention to the contextual subjectivity feature.

Thank You

Refining Word. Net Semantic Richness Find inconsistencies and gaps – Verb assault – attack, round, assail, last out, snipe, assault (attack in speech or writing) “The editors of the left-leaning paper attacked the new House Speaker” – But no sense for the noun as in “His verbal assault was vicious”

Observation MPQA corpus Corpus somewhat noisy for our task » MPQA annotates subjective expressions » Objective senses can appear in subjective expressions Hypothesis: subjective senses tend to appear more often in subjective expressions than objective senses do, and so the appearance of words in subjective expressions is evidence of sense subjectivity

WSD using Subjectivity Tagging Hypothesis: instances of subjective senses are more likely to be in subjective sentences, so sentence subjectivity is an informative feature for WSD of words with both subjective and objective senses

Subjective Sense Examples He was boiling with anger Seethe, boil – (be in an agitated emotional state; “The customer was seething with anger”) – Be – (have the quality of being; (copula, used with an adjective or a predicate noun); “John is rich”; “This is not a good answer”)

Subjective Sense Examples What’s the catch? Catch – (a hidden drawback; “it sounds good but what’s the catch? ”) Drawback – (the quality of being a hindrance; “he pointed out all the drawbacks to my plan”) That doctor is a quack. Quack – (an untrained person who pretends to be a physician and who dispenses medical advice) – Doctor, doc, physician, MD, Dr. , medico

Objective Sense Examples The alarm went off Alarm, warning device, alarm system – (a device that signals the occurrence of some undesirable event) – Device – (an instrumentality invented for a particular purpose; “the device is small enough to wear on your wrist”; “a device intended to conserve water” The water boiled Boil – (come to the boiling point and change from a liquid to vapor; “Water boils at 100 degrees Celsius”) – Change state, turn – (undergo a transformation or a change of position or action)

Objective Sense Examples He sold his catch at the market Catch, haul – (the quantity that was caught; “the catch was only 10 fish”) – Indefinite quantity – (an estimated quantity) The duck’s quack was loud and brief Quack – (the harsh sound of a duck) – Sound – (the sudden occurrence of an audible event)