File Information
File: 05-lr/acl_arc_1_sum/cleansed_text/xml_by_section/metho/03/w03-1207_metho.xml
Size: 17,091 bytes
Last Modified: 2025-10-06 14:08:35
<?xml version="1.0" standalone="yes"?> <Paper uid="W03-1207"> <Title>Discovery of Manner Relations and their Applicability to Question Answering</Title> <Section position="3" start_page="0" end_page="0" type="metho"> <SectionTitle> 2 Lexico-syntactic patterns expressing </SectionTitle> <Paragraph position="0"/> <Section position="1" start_page="0" end_page="0" type="sub_section"> <SectionTitle> manner 2.1 Manner as semantic role </SectionTitle> <Paragraph position="0"> The most frequently occurring form of manner is as a semantic role (Quirk et al., 1985). In this case, manner is encoded as a relationship between a verb and one of its arguments which can be represented by various parts of speech, the most common ones being adverb, adverbial phrase, prepositional phrase, noun phrase, and clause.</Paragraph> <Paragraph position="1"> Verb-adverb patterns One of the most frequently used patterns expressing manner is verb-adverb. In English, there are different kinds of adverbs (Quirk et al., 1985): adverbs of time, manner, degree, location, direction, frequency, transition and hedges.</Paragraph> <Paragraph position="2"> Based on the classification provided by Quirk et al. (Quirk et al., 1985) and our statistics of English texts, we present below the adverbial patterns in order of their frequency of occurrence: a) Adverbs of manner that end in &quot;-ly&quot; This manner adverbs are the most frequently used.</Paragraph> <Paragraph position="3"> Their position is not fixed, as they can be placed either before or after the verb they modify. These adverbs can be modified by other adverbs forming this way adverbial expressions. Examples: slowly, heavily, angrily, etc.</Paragraph> <Paragraph position="4"> b) Adverbs of manner that do not end in &quot;-ly&quot; These adverbs also called Quality description adverbs provide a description of a particular quality. Example: fast, good, well, etc.</Paragraph> <Paragraph position="5"> c) Adverbial expressions These are expressions that modify the underlying verb and refer along with the verb to a manner relation. Examples of such patterns are: a4 as adv manner as a0a2a1a4a3a6a5a8a7a10a9a11a3a13a12 a5 , a4 NP as adv manner a5 , a4 as adv manner S a5 .</Paragraph> <Paragraph position="6"> Examples: several times as fast, as much as 60% faster, louder than ever, all around, etc.</Paragraph> <Paragraph position="7"> d) Compound adverbs of manner These adverbs are usually formed with words linked by hypens. Examples: radio-style, tax-free, flat-out, first-hand, etc e) Foreign adverbial expressions There are expressions boroughed from other languages that are in a manner relationship with the underlying verb. Examples: in flagrante, a la Gorbachev, en masse, etc.</Paragraph> </Section> <Section position="2" start_page="0" end_page="0" type="sub_section"> <SectionTitle> 2.2 Other forms of manner relations </SectionTitle> <Paragraph position="0"> In addition to the manner roles expressed as verb-adverb pairs, manner relations are also expressed as (1) complex nominals (fast car), (2) verbs of implicit manner (for example whisper is a manner of speaking), (3) verb-PP (I took your coat by mistake), (4) verb-NP (He breathed a deep breath), (5) verb clauses (I cook vegetables as Chinese do), and others. null All these lexico-syntactic patterns are ambiguous. Thus we need some syntactic and semantic constraints to differentiate the manner relations from the other possible meanings these patterns may have. In this paper we focus only on the discovery of manner semantic roles expressed as verb- adverb pairs. The method, however, is extendable to many other manner forms and even to other semantic relations. null</Paragraph> </Section> </Section> <Section position="4" start_page="0" end_page="0" type="metho"> <SectionTitle> 3 Approach </SectionTitle> <Paragraph position="0"> The learning procedure proposed here is supervised, for the learning algorithm is provided with a set of inputs along with the corresponding set of correct outputs. In this paper we use the Naive Bayes Classifier approach to determine whether or not a verb-adverb pair indicates a manner relation. This method is similar with the basic algorithm for Document Classification (Mitchell, 1997).</Paragraph> <Paragraph position="1"> Nr. Feature 1 Specific adverb statistics 2 Parent phrase type 3 Present or not in the Adverb Dictionary 4 Distance between verb and adverb 5 Component before adverb 6 Component after the adverb 7 Adverbs ends or not with 'ly This approach requires a decision on how to represent an arbitrary text in terms of attribute (or features) values and how to estimate their probabilities as required by the Naive Bayes Classifier.</Paragraph> </Section> <Section position="5" start_page="0" end_page="0" type="metho"> <SectionTitle> 4 Selecting features </SectionTitle> <Paragraph position="0"> Many researchers ((Blaheta-Charniak, 2000), (Gildea-Jurafsky, 2000), (Gildea-Palmer, 2002)) showed that lexical and syntactic information is very useful for predicate-argument recognition tasks. Their systems are statistical-based and have been trained to automatically label semantic roles only from the output of syntactic parsers.</Paragraph> <Paragraph position="1"> However, lexical and syntactic information alone is not sufficient for the detection of the manner semantic roles, semantic information is necessary as well.</Paragraph> <Paragraph position="2"> To represent the text for the discovery of manner relations, seven features which contribute the most to the classification were chosen. These features capture the context of the adverb and help in deciding the presence of the manner (MNR) component. We have developed an Adverb Dictionary that is a source for some of the features. The Adverb Dictionary is created with adverbs from WordNet and TreeBank. The adverbs that contain the pattern &quot;in a --- manner&quot; in their gloss were extracted from WordNet. The adverbs that are annotated in Tree-Bank as MNR adverb-verb pairs are also included in the Dictionary. A total of 2183 adverbs were included in the Dictionary.</Paragraph> <Paragraph position="3"> The features are explained with the help of the following example:</Paragraph> <Paragraph position="5"> Feature 1 checks if a specific adverb is present in the Dictionary or not. For example, aggressively is part of the Dictionary, where as now is not. The positive frequency calculated from this feature is the total number of times that adverb was encountered in the training corpus. In the case the adverb of a sentence in the testing corpus is part of the Dictionary, this feature helps in deciding what are its chances of being a Positive/Negative Indicator of Manner. This is a good feature as long as the training corpus is very rich (i.e it covers all adverbs).</Paragraph> <Paragraph position="6"> (2) Parent phrase type The second feature is the phrase type to which the adverb attaches. Here both now and aggressively attach to &quot;VP&quot;. Most of the MNR indicating adverbs attach to verbs. This feature helps eliminate adverbs, which modify nouns or adjectives.</Paragraph> <Paragraph position="7"> (3) Whether or not Adverb is present in the Feature 3, like feature 1 checks whether or not an adverb is present in the Adverb Dictionary. The difference is that its statistics are not calculated on the training corpus like in feature 1, but instead it takes the probability of being a manner adverb in the Adverb Dictionary.</Paragraph> <Paragraph position="8"> The usefulness of feature 3 is realized when the test corpus has an adverb which was not encountered in the training corpus. The estimates from feature 1 fail to be of any use at such a point because it is a missing value and both positive and negative frequencies are the same. However, feature 3 assigns the probabilities of that adverb being a manner adverb in the Adverb Dictionary. So, we still have a good estimate from this feature to decide if it is a potential MNR indicator or not (which would have been nullified, had we relied only on feature 1).</Paragraph> <Paragraph position="9"> For example, let's say we encounter the adverb excitedly in the test corpus and it is present in the Adverb Dictionary but not in the training corpus.</Paragraph> <Paragraph position="10"> Feature 1 will not contribute to the decision while feature 3 will help. We can use the lookup table for feature 3 and it is evident that an adverb present in the Dictionary has a higher probability of indicating manner.</Paragraph> <Paragraph position="11"> (4) Distance between verb and adverb The fourth feature is the distance between verb and adverb. This doesn't take into consideration whether the adverb precedes or succeeds the verb. Distance refers to the number of English words that separate them. For example, there are no words between aggressively and marketing, thus the distance is 0.</Paragraph> <Paragraph position="12"> Similarly, the distance between now and marketing is 1. The rational of this feature is based on the observation that most frequently a MNR indicating adverb appears immediately next to a VB.</Paragraph> <Paragraph position="13"> (5) Component before the adverb The fifth feature concerns the POS of the word preceding the adverb. This captures the context of the adverb. This is based on the observation that an adverb that succeeds an AUX is usually not a MNR indicator. For example now is preceeded by &quot;AUX&quot; and aggressively is preceded by an &quot;ADVP&quot;.</Paragraph> <Paragraph position="14"> (6) Component after the adverb The sixth feature concerns the POS of the word after the RB. For example now is succeeded by an &quot;AUX&quot; and aggressively by an &quot;VBG&quot;.</Paragraph> <Paragraph position="15"> (7) Adverb ends in &quot;ly&quot; This feature is 1 when the adverb ends in &quot;ly&quot; and 0 otherwise. The rational for this feature is that many adverbs in manner roles end in &quot;ly&quot;.</Paragraph> <Section position="1" start_page="0" end_page="0" type="sub_section"> <SectionTitle> Estimating Probabilities </SectionTitle> <Paragraph position="0"> The next step is to calculate the probabilities required by the Naive Bayes Classifier.</Paragraph> <Paragraph position="1"> a. Class prior probabilities. This is the ratio between the number of adverbs of each class over the total number of adverbs in the training examples. In our case the classes are positive (or Manner) and negative (not Manner). This is defined as: a1a1a0a3a2a5a4a7a6a9a8a11a10a12a4 a3a13a10 where a10a14a4 is the total number of examples for which the target value is a2a15a4 and a10 is the total number of examples.</Paragraph> <Paragraph position="2"> b. Class conditional probability. This is the probability that any of the seven features drawn from the parsed text tagged positive or negative will belong to the domain of the corresponding features. We use the m-estimate to avoid the cases when a16a18a17a20a19 a21a23a22a25a24a27a26</Paragraph> <Paragraph position="4"> where a16a23a17a35a19 a21a23a22a25a24a32a26 is the number of times the feature occurred in the Positive class, a28a29a24a58a30 a21a23a22a25a24a27a26 is the number of times the feature occurred in the Negative class, a2a43a42a63a44a33a46a48a47 is the distinct number of positive and negative instances for a given feature, and a49a51a10a63a52a62a49 is the total number of all positive and negative instances in the examples.</Paragraph> </Section> <Section position="2" start_page="0" end_page="0" type="sub_section"> <SectionTitle> 4.1 Learning Algorithm </SectionTitle> <Paragraph position="0"> The algorithm learns the probability that a given adverb indicates manner (i.e. how many times the adverb occurred in the positive class and how many times in the negative class). Similarly, it learns the probability that it attaches to a VP/NP/... in each of the positive and negative classes. The same is true for all features.</Paragraph> <Paragraph position="1"> At the end of the learning process, the algorithm creates look-up tables for all the features. These are used by the classifier. The learning step along with the output are explained in the next section.</Paragraph> <Paragraph position="3"> where a2a65a64a13a66 is the output of the Naive Bayes Classifier, a9a27a4 is the class in the target set a2 , and a21a25a74 are the individual features from the set a77 of the seven features.</Paragraph> </Section> </Section> <Section position="6" start_page="0" end_page="0" type="metho"> <SectionTitle> 5 Experimental Setting </SectionTitle> <Paragraph position="0"/> <Section position="1" start_page="0" end_page="0" type="sub_section"> <SectionTitle> 5.1 Building the Training and Test Corpus </SectionTitle> <Paragraph position="0"> In order to learn the constraints, we used the Treebank2 (Marcus, 1994) text collection and LA Times Corpus. Treebank2 is a corpus featuring one million words of 1989 Wall Street Journal material annotated with several predicate-argument structures.</Paragraph> <Paragraph position="1"> It is annotated with the following semantic roles: BNF (beneficiary), DIR (direction), EXT (spatial extent), LOC (location), MNR (manner), PRP (purpose and reason), and TMP (temporal). Treebank2 contains different types of manner annotations: ADVP-</Paragraph> <Paragraph position="3"> MNR(1). For the work in this paper we used the ADVP-MNR annotations from Treebank2.</Paragraph> <Paragraph position="4"> The input to the program is a parsed text. For training and testing the Treebank2 corpus is split in the 3:1 ratio. The algorithm doesn't work on the parsed text directly. Instead, the parsed text is converted into the 7-feature format augmented with the value of the target function as shown in Table 2 Creation of the Look-Up table Given this format as input, the learning algorithm creates LookUp tables using the Class Conditional Probability and Reference files. These files contain the domain of the features. Each feature can take a set of legal values encountered during training. Table 3 exemplifies the lookup entries for some feature examples.</Paragraph> </Section> </Section> <Section position="7" start_page="0" end_page="0" type="metho"> <SectionTitle> 6 Results for discovering manner relations </SectionTitle> <Paragraph position="0"> Let us define the precision and recall performance metrics in this context.</Paragraph> <Paragraph position="1"> The experiments were conducted with the annotations in UPenn's Treebank2. The results of the first experiment are shown in Tables 4.</Paragraph> <Paragraph position="2"> amples.</Paragraph> <Paragraph position="3"> Output of the program:</Paragraph> <Paragraph position="5"> Second experiment Based on the results from the previous set of results it is observed that considering adverbs like moreover, then, thus which can never indicate MNR reduces both the precision and recall. Therefore they were removed from the set of negative examples.</Paragraph> <Paragraph position="6"> Similarly the intensifiers like much, very, so were also removed from the positive examples.</Paragraph> </Section> <Section position="8" start_page="0" end_page="0" type="metho"> <SectionTitle> 7 Application to Question Answering </SectionTitle> <Paragraph position="0"> The manner semantic relation occurs with high frequency in open text. Its discovery is paramount for many applications, such as Information Extraction, Text Mining, Knowledge Base construction, etc. In this section we mentioned only Question Answering. null The concepts and manner relations acquired from a collection of documents can be useful in answering difficult questions that normally can not be handled based solely on keywords matching and proximity. As the level of difficulty increases, Question Answering systems need richer semantic resources, including the discovery of semantic relations in open texts. In the case of a manner question, the answer type of that question may be tagged as MNR. To provide the correct answer, often it is sufficient to locate first the paragraph where the potential answer is and then identify the MNR tag in that paragraph. In case when several such MNR tags exist, more reasoning is necessary. Consider the following examples which show the MNR tag in the answer sentence.</Paragraph> <Paragraph position="1"> Q: How did Bob Marley die? A1: Bob Marley died a0 of Melanomaa1a2a0 MNRa1 .</Paragraph> <Paragraph position="2"> Q: How was little Johnny dressed last night? A1: Dressed a0 in a cowboy stylea1a2a0 MNRa1 , Johnny walked proudly on the street.</Paragraph> <Paragraph position="3"> Q: How does Marry dance? A1: Marry danced a0 as well as Billa1a2a0 MNRa1 .</Paragraph> <Paragraph position="4"> Q: How does Lina Mayors charms her audience? A1: Countering every unfruitful description, her work communicates and a0 impresses through the rhythm of the colorsa1a2a0 MNRa1 .</Paragraph> </Section> class="xml-element"></Paper>