emnlp emnlp2011 emnlp2011-128 knowledge-graph by maker-knowledge-mining

128 emnlp-2011-Structured Relation Discovery using Generative Models

Source: pdf

Author: Limin Yao ; Aria Haghighi ; Sebastian Riedel ; Andrew McCallum

Abstract: We explore unsupervised approaches to relation extraction between two named entities; for instance, the semantic bornIn relation between a person and location entity. Concretely, we propose a series of generative probabilistic models, broadly similar to topic models, each which generates a corpus of observed triples of entity mention pairs and the surface syntactic dependency path between them. The output of each model is a clustering of observed relation tuples and their associated textual expressions to underlying semantic relation types. Our proposed models exploit entity type constraints within a relation as well as features on the dependency path between entity mentions. We examine effectiveness of our approach via multiple evaluations and demonstrate 12% error reduction in precision over a state-of-the-art weakly supervised baseline.

Reference: text

Summary: the most important sentenses genereted by tfidf model

sentIndex sentText sentNum sentScore

1 udum + Abstract We explore unsupervised approaches to relation extraction between two named entities; for instance, the semantic bornIn relation between a person and location entity. [sent-6, score-0.851]

2 Concretely, we propose a series of generative probabilistic models, broadly similar to topic models, each which generates a corpus of observed triples of entity mention pairs and the surface syntactic dependency path between them. [sent-7, score-0.761]

3 The output of each model is a clustering of observed relation tuples and their associated textual expressions to underlying semantic relation types. [sent-8, score-0.941]

4 Our proposed models exploit entity type constraints within a relation as well as features on the dependency path between entity mentions. [sent-9, score-1.246]

5 For instance, knowing that the entity Steve Balmer bears the leaderOf relation to the entity Microsoft, would facilitate question answering (Ravichandran and Hovy, 2002), data mining, and a host of other end-user applications. [sent-12, score-0.879]

6 Due to these many potential applications, relation extraction has gained much attention in information extraction (Kambhatla, 2004; Culotta and Sorensen, 2004; Mintz et al. [sent-13, score-0.459]

7 We propose a series of generative probabilistic models, broadly similar to standard topic models, which generate a corpus of observed triples of entity mention pairs and the surface syntactic dependency path between them. [sent-17, score-0.761]

8 Our proposed models exploit entity type constraints within a relation as well as features on the dependency path between entity mentions. [sent-18, score-1.246]

9 The output of our approach is a clustering over observed relation paths (e. [sent-19, score-0.534]

10 “X was born in Y” and “X is from Y”) such that expressions in the same cluster bear the same semantic relation type between entities. [sent-21, score-0.598]

11 Past work has shown that standard supervised techniques can yield high-performance relation detection when abundant labeled data exists for a fixed inventory of individual relation types (e. [sent-22, score-0.69]

12 However, less explored are open-domain approaches where the set of possible relation types are not fixed and little to no labeled is given for each relation type (Banko et al. [sent-25, score-0.738]

13 For example, DIRT (Lin and Pantel, 2001) aims to discover different representations of the same semantic relation using distributional similarity of dependency paths. [sent-28, score-0.505]

14 Our work is similar to DIRT and USP in that we induce relation types from observed dependency paths, but our approach is a straightforward and principled generative model which can be efficiently learned. [sent-33, score-0.582]

15 1 The former evaluation is performed using subset of induced clusters against Freebase relations, a large manually-built entity and relational database. [sent-36, score-0.543]

16 We also show some clusters which are not included as Freebase relations, as well as some entity clusters found by our approach. [sent-37, score-0.701]

17 The latter evaluation uses the clustering induced by our models as features for relation extraction in distant supervision framework. [sent-38, score-0.839]

18 In relation extraction, we can achieve 12% error reduction in precision over a state-of-the-art weakly supervised baseline and we show that using features from our proposed models can find more facts for a relation without significant accuracy loss. [sent-40, score-0.861]

19 2 Problem and Experimental Setup The task of relation extraction is mapping surface textual relations to underlying semantic relations. [sent-41, score-0.574]

20 For instance, the textual expression “X was born in Y” indicates a semantic relation bornIn between entities “X” and “Y”. [sent-42, score-0.572]

21 There are several components to a coherent relation type, including a tight small number of textual expressions as well as constraints on the entities involved in the relation. [sent-44, score-0.483]

22 For instance, in the bornIn relation “X” must be a person entity and “Y” a location (typically a city or nation). [sent-45, score-0.649]

23 In this work, we present an unsupervised probabilistic generative model for inducing clusters of relation types and recognizing their textual expressions. [sent-46, score-0.726]

24 The set of relation types is not pre-specified but induced from observed unlabeled data. [sent-47, score-0.378]

25 Each tuple represents an observed syntactic relationship between two Named Entities (NE) and consists of three components: the dependency path between two NE mentions, the source argument NE, and the destination argument NE. [sent-51, score-0.745]

26 A dependency path is a concatenation of dependency relations (edges) and words (nodes) along a path in a dependency tree. [sent-52, score-0.692]

27 For instance, the sentence “John Lennnon was born in Liverpool” would yield the relation tuple (Lennon, [↑ −nsubjpass, born, −in] , Liverpool). [sent-53, score-0.651]

28 The dependency path in this example corresponds to the “X was born in Y” textual expression given earlier. [sent-56, score-0.38]

29 Note that for the above example, the bornIn relation can only occur between a person and a location. [sent-57, score-0.382]

30 The relation tuple is the primary observed random variable in our model and we construct our models (see Section 3) so that clusters consist of textual expressions representing the same underlying relation type. [sent-58, score-1.24]

31 ↓ 3 Models We propose three generative models for modeling tuples of entity mention pairs and the syntactic dependency path between them (see Section 2). [sent-59, score-0.816]

32 At the document level, our model is identical to standard LDA; a multinomial distribution is drawn over a fixed number of relation types R. [sent-62, score-0.463]

33 In our approach, a document consists of an exchangeable set of relation tuples. [sent-65, score-0.378]

34 Each relation tuple is drawn from a relation type ‘topic’ distribution selected by a latent relation type indicator variable. [sent-66, score-1.433]

35 Relation tuples are generated using a collection of independent features drawn from the underlying relation type distribution. [sent-67, score-0.593]

36 These changes to standard LDA are intended to have the effect that instead of representing semantically related words, the ‘topic’ latent variable represents a relation type. [sent-68, score-0.38]

37 Our third model exploits entity type constraints within a relation and induces clusters of relations and entities jointly. [sent-69, score-1.085]

38 To generate a relation tuple, we first draw a relation ‘topic’ r from Multi(θ). [sent-77, score-0.69]

40 Each feature of a tuple is generated independently from a hidden relation variable r. [sent-91, score-0.593]

41 Adding named entity tag pair can refine the cluster too. [sent-99, score-0.45]

42 Besides p, s, and d, we introduce trigger words, lexical pattern, POS tag pattern, the named entity pair and the syntactic category pair features for each tuple. [sent-104, score-0.479]

43 Lexical pattern is the word sequence between the two arguments of a tuple and POS tag pattern is the POS tag sequence of the lexical pattern. [sent-105, score-0.361]

44 3 Type-LDA model We know that relations can only hold between certain entity types, known as selectional preferences (Ritter et al. [sent-114, score-0.437]

45 In the mean time, it clusters tuples into relational clusters, and arguments into different entity clusters. [sent-118, score-0.724]

46 The entity clusters could be interesting in many ways, for example, defining fine-grained entity types and finding new concepts. [sent-119, score-0.751]

47 We split the features of a tuple into relation level features and entity level features. [sent-120, score-0.995]

48 Relation level features include the dependency path, trigger, lex and POS features; entity level features include the entity mention itself and its named entity tag. [sent-121, score-1.186]

49 Each tuple has a set of features, relation level features f and entity level features of source argument fs and destination argument fd. [sent-125, score-1.283]

50 Relation level features and two hidden entity types T1 and T2 are generated from hidden relation variable r independently. [sent-126, score-0.695]

51 Source entity features are generated from T1 and destination features are generated from T2. [sent-127, score-0.498]

52 Each tuple is represented by relation level features (f) and entity level features of source argument (fs) and destination argument (fd). [sent-130, score-1.25]

53 For each tuple, a relation r is drawn from Multi(θdoc). [sent-131, score-0.395]

54 The relation level features and two hidden entity types T1 and T2 are independently generated from r. [sent-132, score-0.726]

55 We sample the relation variable (r) and two mention types variables (T1,T2) iteratively for each tuple. [sent-141, score-0.406]

56 , 2004) and extract dependency paths for each pair of named entity mentions in one sentence. [sent-158, score-0.57]

57 Second, the dependency relations in the path should connect two content words, i. [sent-161, score-0.361]

58 This selection is based on the observation that most of the times the corresponding dependency relations do not explicitly state a relation between two candidate arguments. [sent-166, score-0.557]

59 After all entity mentions are generated and paths are extracted, we have nearly 2. [sent-167, score-0.412]

60 In our experiments, we cluster the tuples into 100 relation clusters for all three models. [sent-171, score-0.781]

61 The second seeks to assess the utility of our predicted clusters as features for relation extraction. [sent-175, score-0.609]

62 We discover clusters that roughly correspond to the parentCom (parent company relation), filmDirector, authorOf, comBase (base of a company relation) and dieIn relations in Freebase. [sent-178, score-0.487]

63 We count each tuple in a cluster as true positive if Freebase states the corresponding relation between its argument pair. [sent-180, score-0.737]

64 This relation subsumes all types of employment relationships, irrespective of the role the employee plays for the employer. [sent-203, score-0.379]

65 By contrast, our models discover clusters such as leaderOf, editorOf that correspond to more specific roles an employee can have. [sent-204, score-0.352]

66 In Table 5 some entity clusters found by Type-LDA are shown. [sent-211, score-0.484]

67 This may be because this entity cluster is affected by many relations. [sent-215, score-0.383]

68 2 Distant Supervision based Relation Extraction Our generative models detect clusters ofdependency paths and their arguments. [sent-217, score-0.454]

69 Such clusters are interesting in their own right, but we claim that they can also be used to help a supervised relation extractor. [sent-218, score-0.562]

70 We validate this hypothesis in the context of relation extraction with distant supervision using predicted clusters as features. [sent-219, score-0.913]

71 , 2009), we use Freebase as our distant supervision source, and align related entity pairs to the New York Times articles discussed earlier. [sent-221, score-0.561]

72 The classifiers with additional features from generative models are named after the generative models. [sent-236, score-0.372]

73 AgencyCongress, European Union, NATO, Federal Reserve, United States Court of Appeals News/AgencyThe New York Times, The Times, Supreme Court, Security Council, Book Review Table 5: The entity clusters found by Type-LDA these against Distant and the DIRT database. [sent-248, score-0.484]

74 For the latter we parse our data using Minipar (Lin, 1998) and extract dependency paths between pairs of named entity mentions. [sent-249, score-0.517]

75 We report precision oftop ranked 50 instances for each relation 1462 in table 6. [sent-256, score-0.481]

76 Why do generative model features help to im- Table 6: Precision (%) of some frequent relations prove relation extraction? [sent-266, score-0.626]

77 For example, given “Sidney Mintz, the great food anthropologist at Johns Hopkins University”, we want to predict the relation between ‘Sidney Mintz’ and ‘Johns Hopkins University’ . [sent-268, score-0.389]

78 The distant supervision system incorrectly predicts the pair as ‘O’ since it has not seen the path ‘X, the anthropologist at Y’ in the training data. [sent-269, score-0.487]

79 By contrast, Rel-LDA can predict this pair correctly as worksFor because the dependency path of this pair is in a cluster which contains the path ‘X, a professor at Y’. [sent-270, score-0.505]

80 We also notice that clustering quality is not consistent with distant supervision performance. [sent-280, score-0.358]

81 One possible reason for the inconsistency is that the baseline distant supervision system already employs features that are used in RelLDA1. [sent-285, score-0.381]

82 2 For distant supervision based relation extraction, we only match about 500 Freebase instances to this small data set. [sent-291, score-0.714]

83 Since USP provides clusters of words and phrases, we use the USP clusterID associated with the words on the path as binary features in the clas- sifier. [sent-293, score-0.413]

84 5 Related Work Many approaches have been explored in relation extraction, including bootstrapping, supervised classification, distant supervision, and unsupervised approaches. [sent-303, score-0.516]

85 Distant supervision for relation extraction requires no labeled data. [sent-309, score-0.525]

86 Unsupervised semantic parsing (USP) (Poon and Domingos, 2008) discovers relations by merging 1464 predicates which have similar meanings; it proceeds to recursively cluster dependency tree fragments (or “parts”) to best explain the observed sentence. [sent-322, score-0.361]

87 Our approach differs in that we are focused on capturing a narrow range of binary relations between named entities; some of our models (see Section 3) utilize entity type information to constraint relation type induction. [sent-324, score-0.928]

88 , 2010) employs co-clustering to find clusters of entity pairs and patterns. [sent-328, score-0.524]

89 They identify each cluster of entity pairs as a relation by selecting representative patterns for that relation. [sent-329, score-0.768]

90 Generative probabilistic models are widely employed in relation extraction. [sent-331, score-0.377]

91 For example, they are used for in-domain relation discovery while incorporating constraints via posterior regularization (Chen et al. [sent-332, score-0.387]

92 In this scenario, the authors assume relation labels are given while we automatically discover relations. [sent-337, score-0.414]

93 Researchers also use topic models to perform dimension reduction on features when they cluster relations (Hachey, 2009). [sent-343, score-0.363]

94 Open information extraction aims to discover relations independent of specific domains and relations (Banko et al. [sent-345, score-0.368]

95 A self-learner is employed to extract relation instances but the systems do not cluster the instances into relations. [sent-347, score-0.611]

96 Our approach integrates entity and relation discovery in a probabilistic model. [sent-349, score-0.654]

97 6 Conclusion We have presented an unsupervised probabilistic generative approach to relation extraction between two named entities. [sent-350, score-0.582]

98 Our proposed models exploit entity type constraints within a relation as well as features on the dependency path between entity mentions to cluster equivalent textual expressions. [sent-351, score-1.466]

99 We demonstrate the effectiveness of this approach by comparing induced relation clusters against a large knowledge base. [sent-352, score-0.562]

100 We also show that using clusters of our models as features in distant supervised framework yields 12% error reduction in precision over a weakly supervised baseline and outperforms other state-of-the art relation extraction techniques. [sent-353, score-0.961]

similar papers computed by tfidf model

tfidf for this paper:

wordName wordTfidf (topN-words)

[('relation', 0.345), ('usp', 0.307), ('freebase', 0.269), ('entity', 0.267), ('tuple', 0.217), ('clusters', 0.217), ('distant', 0.171), ('dirt', 0.151), ('path', 0.149), ('destination', 0.137), ('bornin', 0.132), ('supervision', 0.123), ('relations', 0.121), ('cluster', 0.116), ('generative', 0.113), ('tuples', 0.103), ('mintz', 0.098), ('trigger', 0.098), ('paths', 0.092), ('dependency', 0.091), ('born', 0.089), ('cisco', 0.088), ('entities', 0.087), ('arguments', 0.078), ('instances', 0.075), ('yao', 0.074), ('poon', 0.074), ('banko', 0.074), ('discover', 0.069), ('doc', 0.068), ('named', 0.067), ('leaderof', 0.066), ('parentcom', 0.066), ('clustering', 0.064), ('companies', 0.064), ('domingos', 0.063), ('mention', 0.061), ('precision', 0.061), ('argument', 0.059), ('relational', 0.059), ('riedel', 0.057), ('extraction', 0.057), ('culotta', 0.054), ('mentions', 0.053), ('lda', 0.052), ('multi', 0.051), ('kambhatla', 0.051), ('limin', 0.051), ('textual', 0.051), ('drawn', 0.05), ('selectional', 0.049), ('type', 0.048), ('gamma', 0.048), ('topic', 0.047), ('features', 0.047), ('sorensen', 0.045), ('anthropologist', 0.044), ('clusterid', 0.044), ('elekta', 0.044), ('knife', 0.044), ('linksys', 0.044), ('liverpool', 0.044), ('minipar', 0.044), ('obituary', 0.044), ('seaghdha', 0.044), ('sidney', 0.044), ('typelda', 0.044), ('worksfor', 0.044), ('discovery', 0.042), ('patterns', 0.04), ('company', 0.04), ('employs', 0.04), ('rf', 0.038), ('oren', 0.038), ('sandhaus', 0.038), ('person', 0.037), ('level', 0.036), ('ritter', 0.035), ('york', 0.035), ('multinomial', 0.035), ('latent', 0.035), ('bollegala', 0.034), ('hasegawa', 0.034), ('qf', 0.034), ('duality', 0.034), ('tau', 0.034), ('nf', 0.034), ('court', 0.034), ('employee', 0.034), ('document', 0.033), ('fs', 0.033), ('observed', 0.033), ('pattern', 0.033), ('models', 0.032), ('coreference', 0.032), ('massachusetts', 0.032), ('fy', 0.032), ('weakly', 0.031), ('etzioni', 0.031), ('independently', 0.031)]

similar papers list:

simIndex simValue paperId paperTitle

same-paper 1 0.99999821 128 emnlp-2011-Structured Relation Discovery using Generative Models

Author: Limin Yao ; Aria Haghighi ; Sebastian Riedel ; Andrew McCallum

2 0.29786313 14 emnlp-2011-A generative model for unsupervised discovery of relations and argument classes from clinical texts

Author: Bryan Rink ; Sanda Harabagiu

Abstract: This paper presents a generative model for the automatic discovery of relations between entities in electronic medical records. The model discovers relation instances and their types by determining which context tokens express the relation. Additionally, the valid semantic classes for each type of relation are determined. We show that the model produces clusters of relation trigger words which better correspond with manually annotated relations than several existing clustering techniques. The discovered relations reveal some of the implicit semantic structure present in patient records.

3 0.2853348 98 emnlp-2011-Named Entity Recognition in Tweets: An Experimental Study

Author: Alan Ritter ; Sam Clark ; Mausam ; Oren Etzioni

Abstract: People tweet more than 100 Million times daily, yielding a noisy, informal, but sometimes informative corpus of 140-character messages that mirrors the zeitgeist in an unprecedented manner. The performance of standard NLP tools is severely degraded on tweets. This paper addresses this issue by re-building the NLP pipeline beginning with part-of-speech tagging, through chunking, to named-entity recognition. Our novel T-NER system doubles F1 score compared with the Stanford NER system. T-NER leverages the redundancy inherent in tweets to achieve this performance, using LabeledLDA to exploit Freebase dictionaries as a source of distant supervision. LabeledLDA outperforms cotraining, increasing F1 by 25% over ten common entity types. Our NLP tools are available at: http : / / github .com/ aritt er /twitte r_nlp

4 0.24430937 114 emnlp-2011-Relation Extraction with Relation Topics

Author: Chang Wang ; James Fan ; Aditya Kalyanpur ; David Gondek

Abstract: This paper describes a novel approach to the semantic relation detection problem. Instead of relying only on the training instances for a new relation, we leverage the knowledge learned from previously trained relation detectors. Specifically, we detect a new semantic relation by projecting the new relation’s training instances onto a lower dimension topic space constructed from existing relation detectors through a three step process. First, we construct a large relation repository of more than 7,000 relations from Wikipedia. Second, we construct a set of non-redundant relation topics defined at multiple scales from the relation repository to characterize the existing relations. Similar to the topics defined over words, each relation topic is an interpretable multinomial distribution over the existing relations. Third, we integrate the relation topics in a kernel function, and use it together with SVM to construct detectors for new relations. The experimental results on Wikipedia and ACE data have confirmed that backgroundknowledge-based topics generated from the Wikipedia relation repository can significantly improve the performance over the state-of-theart relation detection approaches.

5 0.17157678 116 emnlp-2011-Robust Disambiguation of Named Entities in Text

Author: Johannes Hoffart ; Mohamed Amir Yosef ; Ilaria Bordino ; Hagen Furstenau ; Manfred Pinkal ; Marc Spaniol ; Bilyana Taneva ; Stefan Thater ; Gerhard Weikum

Abstract: Disambiguating named entities in naturallanguage text maps mentions of ambiguous names onto canonical entities like people or places, registered in a knowledge base such as DBpedia or YAGO. This paper presents a robust method for collective disambiguation, by harnessing context from knowledge bases and using a new form of coherence graph. It unifies prior approaches into a comprehensive framework that combines three measures: the prior probability of an entity being mentioned, the similarity between the contexts of a mention and a candidate entity, as well as the coherence among candidate entities for all mentions together. The method builds a weighted graph of mentions and candidate entities, and computes a dense subgraph that approximates the best joint mention-entity mapping. Experiments show that the new method significantly outperforms prior methods in terms of accuracy, with robust behavior across a variety of inputs.

6 0.14997038 57 emnlp-2011-Extreme Extraction - Machine Reading in a Week

7 0.14971466 113 emnlp-2011-Relation Acquisition using Word Classes and Partial Patterns

8 0.1268203 145 emnlp-2011-Unsupervised Semantic Role Induction with Graph Partitioning

9 0.10855874 70 emnlp-2011-Identifying Relations for Open Information Extraction

10 0.1055638 59 emnlp-2011-Fast and Robust Joint Models for Biomedical Event Extraction

11 0.10212965 109 emnlp-2011-Random Walk Inference and Learning in A Large Scale Knowledge Base

12 0.094667211 84 emnlp-2011-Learning the Information Status of Noun Phrases in Spoken Dialogues

13 0.089935079 141 emnlp-2011-Unsupervised Dependency Parsing without Gold Part-of-Speech Tags

14 0.088902123 108 emnlp-2011-Quasi-Synchronous Phrase Dependency Grammars for Machine Translation

15 0.088445693 4 emnlp-2011-A Fast, Accurate, Non-Projective, Semantically-Enriched Parser

16 0.08751528 67 emnlp-2011-Hierarchical Verb Clustering Using Graph Factorization

17 0.083099961 7 emnlp-2011-A Joint Model for Extended Semantic Role Labeling

18 0.082915619 15 emnlp-2011-A novel dependency-to-string model for statistical machine translation

19 0.08015386 143 emnlp-2011-Unsupervised Information Extraction with Distributional Prior Knowledge

20 0.078439049 144 emnlp-2011-Unsupervised Learning of Selectional Restrictions and Detection of Argument Coercions

similar papers computed by lsi model

lsi for this paper:

topicId topicWeight

[(0, 0.289), (1, -0.217), (2, -0.267), (3, 0.06), (4, -0.234), (5, -0.185), (6, 0.202), (7, -0.059), (8, -0.047), (9, 0.094), (10, 0.082), (11, 0.217), (12, 0.005), (13, -0.041), (14, 0.086), (15, 0.065), (16, 0.053), (17, 0.023), (18, -0.024), (19, -0.013), (20, -0.11), (21, 0.039), (22, -0.053), (23, -0.218), (24, 0.045), (25, 0.033), (26, 0.014), (27, 0.041), (28, -0.014), (29, 0.04), (30, 0.071), (31, -0.116), (32, 0.067), (33, -0.007), (34, 0.028), (35, 0.014), (36, 0.007), (37, -0.04), (38, -0.097), (39, 0.055), (40, 0.016), (41, -0.039), (42, 0.026), (43, -0.064), (44, -0.087), (45, -0.018), (46, -0.038), (47, -0.044), (48, 0.012), (49, 0.018)]

similar papers list:

simIndex simValue paperId paperTitle

same-paper 1 0.9770577 128 emnlp-2011-Structured Relation Discovery using Generative Models

Author: Limin Yao ; Aria Haghighi ; Sebastian Riedel ; Andrew McCallum

2 0.88543439 14 emnlp-2011-A generative model for unsupervised discovery of relations and argument classes from clinical texts

Author: Bryan Rink ; Sanda Harabagiu

3 0.8205325 114 emnlp-2011-Relation Extraction with Relation Topics

Author: Chang Wang ; James Fan ; Aditya Kalyanpur ; David Gondek

4 0.63979769 70 emnlp-2011-Identifying Relations for Open Information Extraction

Author: Anthony Fader ; Stephen Soderland ; Oren Etzioni

Abstract: Open Information Extraction (IE) is the task of extracting assertions from massive corpora without requiring a pre-specified vocabulary. This paper shows that the output of state-ofthe-art Open IE systems is rife with uninformative and incoherent extractions. To overcome these problems, we introduce two simple syntactic and lexical constraints on binary relations expressed by verbs. We implemented the constraints in the REVERB Open IE system, which more than doubles the area under the precision-recall curve relative to previous extractors such as TEXTRUNNER and WOEpos. More than 30% of REVERB’s extractions are at precision 0.8 or higher— compared to virtually none for earlier systems. The paper concludes with a detailed analysis of REVERB’s errors, suggesting directions for future work.1 1 Introduction and Motivation Typically, Information Extraction (IE) systems learn an extractor for each target relation from labeled training examples (Kim and Moldovan, 1993; Riloff, 1996; Soderland, 1999). This approach to IE does not scale to corpora where the number of target relations is very large, or where the target relations cannot be specified in advance. Open IE solves this problem by identifying relation phrases—phrases that denote relations in English sentences (Banko et al., 2007). The automatic identification of rela1The source code for REVERB is available at reverb . cs .washingt on .edu/ http : // 1535 tion phrases enables the extraction of arbitrary relations from sentences, obviating the restriction to a pre-specified vocabulary. Open IE systems have achieved a notable measure of success on massive, open-domain corpora drawn from the Web, Wikipedia, and elsewhere. (Banko et al., 2007; Wu and Weld, 2010; Zhu et al., 2009). The output of Open IE systems has been used to support tasks like learning selectional preferences (Ritter et al., 2010), acquiring common sense knowledge (Lin et al., 2010), and recognizing entailment (Schoen- mackers et al., 2010; Berant et al., 2011). In addition, Open IE extractions have been mapped onto existing ontologies (Soderland et al., 2010). We have observed that two types of errors are frequent in the output of Open IE systems such as TEXTRUNNER and WOE: incoherent extractions and uninformative extractions. Incoherent extractions are cases where the extracted relation phrase has no meaningful interpretation (see Table 1 for examples). Incoherent extractions arise because the learned extractor makes a sequence of decisions about whether to include each word in the relation phrase, often resulting in incomprehensible predictions. To solve this problem, we introduce a syntactic constraint: every multi-word relation phrase must begin with a verb, end with a preposition, and be a contiguous sequence of words in the sentence. Thus, the identification of a relation phrase is made in one fell swoop instead of on the basis of multiple, word-by-word decisions. Uninformative extractions are extractions that omit critical information. For example, consider the sentence “Faust made a deal with the devil.” PreviProce dEindgisnb oufr tgh e, 2 S0c1o1tl Canodn,f eUrKen,c Jeuol yn 2 E7m–3p1ir,ic 2a0l1 M1.e ?tc ho2d0s1 in A Nsasotucira tlio Lnan fogru Cagoem Ppruotcaetisosninagl, L pinag uesis 1ti5c3s5–1545, ous Open IE systems return the uninformative (Faust, made, a deal) instead of (Faust, made a deal with, the devil). This type of error is caused by improper handling of relation phrases that are expressed by a combination of a verb with a noun, such as light verb constructions (LVCs). An LVC is a multi-word expression composed of a verb and a noun, with the noun carrying the semantic content of the predicate (Grefenstette and Teufel, 1995; Stevenson et al., 2004; Allerton, 2002). Table 2 illustrates the wide range of relations expressed this way, which are not captured by existing open extractors. Our syntactic constraint leads the extractor to include nouns in the relation phrase, solving this problem. Although the syntactic constraint significantly reduces incoherent and uninformative extractions, it allows overly-specific relation phrases such as is offering only modest greenhouse gas reduction targets at. To avoid overly-specific relation phrases, we introduce an intuitive lexical constraint: a binary relation phrase ought to appear with at least a minimal number of distinct argument pairs in a large corpus. In summary, this paper articulates two simple but surprisingly powerful constraints on how binary relationships are expressed via verbs in English sentences, and implements them in the REVERB Open IE system. We release REVERB and the data used in our experiments to the research community. The rest of the paper is organized as follows. Section 2 analyzes previous work. Section 3 defines our constraints precisely. Section 4 describes REVERB, our implementation of the constraints. Section 5 reports on our experimental results. Section 6 concludes with a summary and discussion of future work. 2 Previous Work Open IE systems like TEXTRUNNER (Banko et al., 2007), WOEpos, and WOEparse (Wu and Weld, 2010) focus on extracting binary relations of the form (arg1, relation phrase, arg2) from text. These systems all use the following three-step method: 1. Label: Sentences are automatically labeled with extractions using heuristics or distant supervision. 1536 tTSblaoenh rgdpe atyoneMgdrm.uoahcietrsdkcaes1lci4neatodrwnsea.tlrhoianfctsehndNaterufplndergtcoli.sn hkcetrIwnoca stlhceindrstbeoanmgltiRonserplatdion Table 1: Examples of incoherent extractions. Incoherent extractions make up approximately 13% of TEXTRUNNER’s output, 15% of WOEpos’s output, and 30% of WOEparse’s output. tighmosatvdkeihgmosa tvdnkeaicbp lkaroedpucthmsalitonbw,yigtoanhivst,oekamfh,cedotanlusetkroahlntPpr,ohgv.aeDomfvrt,.isonue,atdkhowcainstmdygaevifncromtnaeg oinf Table 2: Examples of uninformative relations (left) and their completions (right). Uninformative relations occur in approximately 4% of WOEparse’s output, 6% of WOEpos’s output, and 7% of TEXTRUNNER’s output. 2. Learn: A relation phrase extractor is learned using a sequence-labeling (e.g., CRF). graphical model 3. Extract: the system takes a sentence as input, identifies a candidate pair of NP arguments (arg1, arg2) from the sentence, and then uses the learned extractor to label each word between the two arguments as part of the relation phrase or not. The extractor is applied to the successive sentences in the corpus, and the resulting extractions are collected. This method faces several challenges. First, the training phase requires a large number of labeled training examples (e.g., 200, 000 heuristicallylabeled sentences for TEXTRUNNER and 300, 000 for WOE). Heuristic labeling of examples obviates hand labeling but results in noisy labels and distorts the distribution of examples. Second, the extraction step is posed as a sequence-labeling problem, where each word is assigned its own label. Because each assignment is uncertain, the likelihood that the extracted relation phrase is flawed increases with the length of the sequence. Finally, the extractor chooses an extraction’s arguments heuristically, and cannot backtrack over this choice. This is problematic when a word that belongs in the relation phrase is chosen as an argument (for example, deal from the “made a deal with” sentence). Because of the feature sets utilized in previous work, the learned extractors ignore both “holistic” aspects of the relation phrase (e.g., is it contiguous?) as well as lexical aspects (e.g., how many instances of this relation are there?). Thus, as we show in Section 5, systems such as TEXTRUNNER are unable to learn the constraints embedded in REVERB. Of course, a learning system, utilizing a different hypothesis space, and an appropriate set of training examples, could potentially learn and refine the constraints in REVERB. This is a topic for future work, which we consider in Section 6. The first Open IE system was TEXTRUNNER (Banko et al., 2007), which used a Naive Bayes model with unlexicalized POS and NP-chunk features, trained using examples heuristically generated from the Penn Treebank. Subsequent work showed that utilizing a linear-chain CRF (Banko and Etzioni, 2008) or Markov Logic Network (Zhu et al., 2009) can lead to improved extraction. The WOE systems introduced by Wu and Weld make use of Wikipedia as a source of training data for their extractors, which leads to further improvements over TEXTRUNNER (Wu and Weld, 2010). Wu and Weld also show that dependency parse features result in a dramatic increase in precision and recall over shallow linguistic features, but at the cost of extraction speed. Other approaches to large-scale IE have included Preemptive IE (Shinyama and Sekine, 2006), OnDemand IE (Sekine, 2006), and weak supervision for IE (Mintz et al., 2009; Hoffmann et al., 2010). Preemptive IE and On-Demand IE avoid relationspecific extractors, but rely on document and entity clustering, which is too costly for Web-scale IE. Weakly supervised methods use an existing ontology to generate training data for learning relationspecific extractors. While this allows for learning relation-specific extractors at a larger scale than what was previously possible, the extractions are still restricted to a specific ontology. Many systems have used syntactic patterns based on verbs to extract relation phrases, usually rely1537 ing on a full dependency parse of the input sentence (Lin and Pantel, 2001 ; Stevenson, 2004; Specia and Motta, 2006; Kathrin Eichler and Neumann, 2008). Our work differs from these approaches by focusing on relation phrase patterns expressed in terms of POS tags and NP chunks, instead of full parse trees. Banko and Etzioni (Banko and Etzioni, 2008) showed that a small set of POS-tag patterns cover a large fraction of relationships in English, but never incorporated the patterns into an extractor. This paper reports on a substantially improved model of binary relation phrases, which increases the recall of the Banko-Etzioni model (see Section 3.3). Further, while previous work in Open IE has mainly focused on syntactic patterns for relation extraction, we introduce a lexical constraint that boosts precision and recall. Finally, Open IE is closely related to semantic role labeling (SRL) (Punyakanok et al., 2008; Toutanova et al., 2008) in that both tasks extract relations and arguments from sentences. However, SRL systems traditionally rely on syntactic parsers, which makes them susceptible to parser errors and substantially slower than Open IE systems such as REVERB. This difference is particularly important when operating on the Web corpus due to its size and heterogeneity. Finally, SRL requires hand-constructed semantic resources like Propbank and Framenet (Martha and Palmer, 2002; Baker et al., 1998) as input. In contrast, Open IE systems require no relation-specific training data. ReVerb, in particular, relies on its explicit lexical and syntactic constraints, which have no correlate in SRL systems. For a more detailed comparison of SRL and Open IE, see (Christensen et al., 2010). 3 Constraints on Relation Phrases In this section we introduce two constraints on relation phrases: a syntactic constraint and a lexical constraint. 3.1 Syntactic Constraint The syntactic constraint serves two purposes. First, it eliminates incoherent extractions, and second, it reduces uninformative extractions by capturing relation phrases expressed by a verb-noun combination, including light verb constructions. PVW= v(p enr oVbupn |a prVta iPcdrtljiec|?lVaedaW|dv in|∗?fpP.rmonar|kder)t Figure 1: A simple part-of-speech-based regular expression reduces the number of incoherent extractions like was central torpedo and covers relations expressed via light verb constructions like gave a talk at. The syntactic constraint requires the relation phrase to match the POS tag pattern shown in Figure 1. The pattern limits relation phrases to be either a verb (e.g., invented), a verb followed immediately by a preposition (e.g., located in), or a verb followed by nouns, adjectives, or adverbs ending in a preposition (e.g., has atomic weight of). Ifthere are multiple possible matches in a sentence for a single verb, the longest possible match is chosen. Finally, if the pattern matches multiple adjacent sequences, we merge them into a single relation phrase (e.g., wants to extend). This refinement enables the model to readily handle relation phrases containing multiple verbs. A consequence ofthis pattern is that the relation phrase must be a contiguous span of words in the sentence. The syntactic constraint eliminates the incoherent relation phrases returned by existing systems. example, given the sentence For Extendicare agreed to buy Arbor Health Care for about US $432 million in cash and assumed debt. TEXTRUNNER returns the extraction (Arbor Health Care, for assumed, debt). The phrase for assumed is clearly not a valid relation phrase: it begins with a preposition and splices together two distant words in the sentence. The syntactic constraint prevents this type of error by simply restricting relation phrases to match the pattern in Figure 1. The syntactic constraint reduces uninformative extractions by capturing relation phrases expressed via LVCs. For example, the POS pattern matched against the sentence “Faust made a deal with the Devil,” would result in the relation phrase made a deal with, instead of the uninformative made. Finally, we require the relation phrase to appear between its two arguments in the sentence. This is a common constraint that has been implicitly enforced in other open extractors. 1538 3.2 Lexical Constraint While the syntactic constraint greatly reduces uninformative extractions, it can sometimes match relation phrases that are so specific that they have only a few possible instances, even in a Web-scale corpus. Consider the sentence: The Obama administration is offering only modest greenhouse gas reduction targets at the conference. The POS pattern will match the phrase: is offering only modest greenhouse gas reduction targets at (1) Thus, there are phrases that satisfy the syntactic constraint, but are not relational. To overcome this limitation, we introduce a lexical constraint that is used to separate valid relation phrases from overspecified relation phrases, like the example in (1). The constraint is based on the intuition that a valid relation phrase should take many distinct arguments in a large corpus. The phrase in (1) is specific to the argument pair (Obama administration, conference), so it is unlikely to represent a bona fide relation. We describe the implementation details of the lexical constraint in Section 4. 3.3 Limitations Our constraints represent an idealized model of relation phrases in English. This raises the question: How much recall is lost due to the constraints? To address this question, we analyzed Wu and Weld’s set of 300 sentences from a set of random Web pages, manually identifying all verb-based relationships between noun phrase pairs. This resulted in a set of 327 relation phrases. For each relation phrase, we checked whether it satisfies our constraints. We found that 85% of the relation phrases do satisfy the constraints. Of the remaining 15%, we identified some of the common cases where the constraints were violated, summarized in Table 3. Many of the example relation phrases shown in Table 3 involve long-range dependencies between words in the sentence. These types of dependencies are not easily representable using a pattern over POS tags. A deeper syntactic analysis of the input sentence would provide a much more general language for modeling relation phrases. For example, one could create a model of relations expressed in Table 3: Approximately 85% of the binary verbal relation phrases in a sample of Web sentences satisfy our constraints. terms of dependency parse features that would capture the non-contiguous relation phrases in Table 3. Previous work has shown that dependency paths do indeed boost the recall of relation extraction systems (Wu and Weld, 2010; Mintz et al., 2009). While using dependency path features allows for a more flexible model of relations, it significantly increases pro- cessing time, which is problematic for Web-scale extraction. Further, we have found that this increased recall comes at the cost of lower precision on Web text (see Section 5). The results in Table 3 are similar to Banko and Etzioni’s findings that a set of eight POS patterns cover a large fraction of binary verbal relation phrases. However, their analysis was based on a set of sentences known to contain either a company acquisition or birthplace relationship, while our results are on a random sample of Web sentences. We applied Banko and Etzioni’s verbal patterns to our random sample of 300 Web sentences, and found that they cover approximately 69% of the relation phrases in the corpus. The gap in recall between this and the 85% shown in Table 3 is largely due to LVC relation phrases (made a deal with) and phrases containing multiple verbs (refuses to return to), which their patterns do not cover. In sum, our model is by no means complete. However, we have empirically shown that the majority of binary verbal relation phrases in a sample of Web sentences are captured by our model. By focusing on this subset of language, our model can 1539 be used to perform Open IE at significantly higher precision than before. 4 REVERB This section introduces REVERB, a novel open extractor based on the constraints defined in the previous section. REVERB first identifies relation phrases that satisfy the syntactic and lexical constraints, and then finds a pair of NP arguments for each identified relation phrase. The resulting extractions are then assigned a confidence score using a logistic regression classifier. This algorithm differs in three important ways from previous methods (Section 2). First, the relation phrase is identified “holistically” rather than word-by-word. Second, potential phrases are filtered based on statistics over a large corpus (the implementation of our lexical constraint). Finally, REVERB is “relation first” rather than “arguments first”, which enables it to avoid a common error made by previous methods—confusing a noun in the relation phrase for an argument, e.g. the noun deal in made a deal with. 4.1 Extraction Algorithm REVERB takes as input a POS-tagged and NPchunked sentence and returns a set of (x, r, y) extraction triples.2 Given an input sentence s, REVERB uses the following extraction algorithm: 1. Relation Extraction: For each verb v in s, find the longest sequence of words rv such that (1) rv starts at v, (2) rv satisfies the syntactic constraint, and (3) rv satisfies the lexical constraint. If any pair of matches are adjacent or overlap in s, merge them into a single match. 2. Argument Extraction: For each relation phrase r identified in Step 1, find the nearest noun phrase x to the left of r in s such that x is not a relative pronoun, WHO-adverb, or existential “there”. Find the nearest noun phrase y to the right of r in s. If such an (x, y) pair could be found, return (x, r, y) as an extraction. We check whether a candidate relation phrase rv satisfies the syntactic constraint by matching it against the regular expression in Figure 1. 2REVERB uses OpenNLP for POS tagging and NP chunking: http : / / opennlp . s ource forge . net / To determine whether rv satisfies the lexical constraint, we use a large dictionary D of relation phrases that are known to take many distinct arguments. In an offline step, we construct D by finding all matches of the POS pattern in a corpus of 500 million Web sentences. For each matching relation phrase, we heuristically identify its arguments (as in Step 2 above). We set D to be the set of all relation phrases that take at least k distinct argument pairs in the set of extractions. In order to allow for minor variations in relation phrases, we normalize each relation phrase by removing inflection, auxiliary verbs, adjectives, and adverbs. Based on experiments on a held-out set of sentences, we found that a value of k = 20 works well for filtering out overspecified relations. This results in a set of approximately 1.7 million distinct normalized relation phrases, which are stored in memory at extraction time. As an example of the extraction algorithm in action, consider the following input sentence: Hudson was born in Hampstead, which is a suburb of London. Step 1 of the algorithm identifies three relation phrases that satisfy the syntactic and lexical constraints: was, born in, and is a suburb of. The first two phrases are adjacent in the sentence, so they are merged into the single relation phrase was born in. Step 2 then finds an argument pair for each relation phrase. For was born in, the nearest NPs are (Hudson, Hampstead). For is a suburb of, the extractor skips over the NP which and chooses the argument pair (Hampstead, London). The final output is e1: (Hudson, was born in, Hampstead) e2: (Hampstead, is a suburb of, London). 4.2 Confidence Function The extraction algorithm in the previous section has high recall, but low precision. Like with previous open extractors, we want way to trade recall for precision by tuning a confidence threshold. We use a logistic regression classifier to assign a confidence score to each extraction, which uses the features shown in Table 4. All of these features are efficiently computable and relation independent. We trained the confidence function by manually labeling the extractions from a set of 1, 000 sentences from the Web and Wikipedia as correct or incorrect. 1540 Table 4: REVERB uses these features to assign a confidence score to an extraction (x, r, y) from a sentence s using a logistic regression classifier. Previous open extractors require labeled training data to learn a model of relations, which is then used to extract relation phrases from text. In contrast, REVERB uses a specified model of relations for extraction, and requires labeled data only for assigning confidence scores to its extractions. Learning a confidence function is a much simpler task than learning a full model of relations, using two orders of magnitude fewer training examples than TEXTRUNNER or WOE. 4.3 TEXTRUNNER-R The model of relation phrases used by REVERB is specified, but could a TEXTRUNNER-like system learn this model from training data? While it is difficult to answer such a question for all possible permutations of features sets, training examples, and learning biases, we demonstrate that TEXTRUNNER itself cannot learn REVERB’s model even when re-trained using the output of REVERB as labeled training data. The resulting system, TEXTRUNNER-R, uses the same feature representation as TEXTRUNNER, but different parameters, and a different set of training examples. To generate positive instances, we ran REVERB on the Penn Treebank, which is the same dataset that TEXTRUNNER is trained on. To generate negative instances from a sentence, we took each noun phrase pair in the sentence that does not appear as arguments in a REVERB extraction. This process resulted in a set of 67, 562 positive instances, and 356, 834 negative instances. We then passed these labeled examples to TEXTRUNNER’s training procedure, which learns a linear-chain CRF using closedclass features like POS tags, capitalization, punctuation, etc.TEXTRUNNER-R uses the argument-first extraction algorithm described in Section 2. 5 Experiments We compare REVERB to the following systems: • • REVERB¬lex - The REVERB system described iRn the previous section, but without the lexical constraint. REVERB¬lex uses the same confidence function as REVERB. TEXTRUNNER - Banko and Etzioni’s 2008 extractor, which uses a second order linear-chain CRF trained on extractions heuristically generated from the Penn Treebank. TEXTRUNNER uses shallow linguistic features in its CRF, which come from the same POS tagger and NPchunker that REVERB uses. • • • TEXTRUNNER-R - Our modification to TEXTRUNNER, which uses the same extraction code, but with a model of relations trained on REVERB extractions. WOEpos - Wu and Weld’s modification to TEXTRUNNER, which uses a model of relations learned from extractions heuristically generated from Wikipedia. WOEparse - Wu and Weld’s parser-based extractor, which uses a large dictionary of dependency path patterns learned from heuristic extractions generated from Wikipedia. Each system is given a set of sentences as input, and returns a set of binary extractions as output. We created a test set of 500 sentences sampled from the Web, using Yahoo’s random link service.3 After run3http : / /random . yahoo .com/bin/ryl 1541 rAaneUudeCrv aPdRCAPreu0 0 . 543210 REV RB EV RBWOET XT-WOET XT¬¬lleexx ppaarrssee RRUUNNNNEERR--RR ppooss RRUUNNNNEERR Figure 2: REVERB outperforms state-of-the-art open extractors, with an AUC more than twice that of TEXTRUNNER or WOEpos, and 38% higher than WOEparse. CCoommppaarriissoonn ooff RREEVVEERRBB--BBaasseedd SSyysstteemmss RReeccaall l Figure 3: The lexical constraint gives REVERB a boost in precision and recall over REVERB¬lex. TEXTRUNNER-R is unable to learn the model used by REVERB, which results in lower precision and recall. ning each extractor over the input sentences, two human judges independently evaluated each extraction as correct or incorrect. The judges reached agreement on 86% of the extractions, with an agreement score of κ = 0.68. We report results on the subset of the data where the two judges concur. Thejudges labeled uninformative extractions conservatively. That is, if critical information was dropped from the relation phrase but included in the second argument, it is labeled correct. For example, both the extractions (Ackerman, is a professor of, biology) and (Ackerman, is, a professor of biology) are considered correct. Each system returns confidence scores for its extractions. For a given threshold, we can measure the precision and recall of the output. Precision is the fraction of returned extractions that are correct. Recall is the fraction of correct extractions in EExxttrraaccttiioonnss RReeccaall l Figure 4: REVERB achieves significantly higher precision than state-of-the-art Open IE systems, and comparable recall to WOEparse. the corpus that are returned. We use the total number of extractions labeled as correct by the judges as our measure of recall for the corpus. In order to avoid double-counting, we treat extractions that differ superficially (e.g., different punctuation or dropping inessential modifiers) as a single extraction. We compute a precision-recall curve by varying the confidence threshold, and then compute the area under the curve (AUC). 5.1 Results Figure 2 shows the AUC of each system. REVERB achieves an AUC that is 30% higher than WOEparse and is more than double the AUC of WOEpos or TEXTRUNNER. The lexical constraint provides a significant boost in performance, with REVERB achieving an AUC 23% higher than REVERB¬lex. REVERB proves to be a useful source of training data, with TEXTRUNNER-R having an AUC 71% higher than TEXTRUNNER and performing on par with WOEpos. From the training data, TEXTRUNNER-R was able to learn a model that predicts contiguous relation phrases, but still returned incoherent relation phrases (e.g., starting with a preposition) and overspecified relation phrases. These errors are due to TEXTRUNNER-R overfitting the training data and not having access to the lexical constraint. Figure 3 shows the precision-recall curves of the systems introduced in this paper. TEXTRUNNER-R has much lower precision than REVERB and 1542 RReellaattiioonnss OOnnllyy RReeccaall l Figure 5: On the subtask of identifying relations phrases, REVERB is able to achieve even higher precision and recall than other systems. REVERB¬lex at all levels of recall. The lexical constraint gives REVERB a boost in precision over REVERB¬lex, reducing overspecified extractions from 20% of REVERB¬lex’s output to 1% of REVERB’s. The lexical constraint also boosts recall over REVERB¬lex, since REVERB is able to find a correct relation phrase where REVERB¬lex finds an overspecified one. Figure 4 shows the precision-recall curves of REVERB and the external systems. REVERB has much higher precision than the other systems at nearly all levels of recall. In particular, more than 30% of REVERB’s extractions are at precision 0.8 or higher, compared to virtually none for the other systems. WOEparse achieves a slightly higher recall than REVERB (0.62 versus 0.64), but at the cost of lower precision. In order to highlight the role of the relational model of each system, we also evaluate their performance on the subtask of extracting just the relation phrases from the input text. Figure 5 shows the precision-recall curves for each system on the relation phrase-only evaluation. In this case, REVERB has both higher precision and recall than the other systems. REVERB’s biggest improvement came from the elimination of incoherent extractions. Incoherent extractions were a large fraction of the errors made by previous systems, accounting for approximately 13% of TEXTRUNNER’s extractions, 15% of WOEpos’s, and 30% of WOEparse’s. Uninformative 162857% IOCNmoRt-vhparenE -yrsVc,paoiEtenlRrvcetailBfgu oev-tdinoIur bsgecplroPahtOliroeStsni/c ,phEiunrxhcaoktsreincgtieoransgument Table 5: The majority of the incorrect extractions returned by REVERB are due to errors in argument extraction. extractions had a smaller effect on other systems’ precision, accounting for 4% of WOEparse’s extractions, 5% of WOEpos’s, and 7% of TEXTRUNNER’s, while only appearing in 1% of REVERB’s extractions. REVERB’s reduction in uninformative extractions resulted in a boost in recall, capturing many LVC relation phrases missed by other systems (like those shown in Table 2). To test the systems’ speed, we ran each extractor on a set of 100, 000 sentences using a Pentium 4 machine with 4GB of RAM. The processing times were 16 minutes for REVERB, 21 minutes for TEXTRUNNER, 21 minutes for WOEpos, and 11 hours for WOEparse. The times for REVERB, TEXTRUNNER, and WOEpos are all approximately the same, since they all use the same POS-tagging and NP-chunking software. WOEparse processes each sentence with a dependency parser, resulting in much longer processing time. 5.2 REVERB Error Analysis To better understand the limitations of REVERB, we performed a detailed analysis of its errors in precision (incorrect extractions returned by REVERB) and its errors in recall (correct extractions that REVERB missed). Table 5 summarizes the types of incorrect extractions that REVERB returns. We found that 65% of the incorrect extractions returned by REVERB were cases where a relation phrase was correctly identified, but the argument-finding heuristics failed. The remaining errors were cases where REVERB extracted an incorrect relation phrase. One common mistake that REVERB made was extracting a relation phrase that expresses an n-ary relationship via a ditransitive verb. For example, given the sentence 1543 521873% ICRPdOeolEunSaVt/dciEofhneRuodBnftiakl-dmneMrgnotdiersfoyresudpctroebEcyrxieftlcrxa icertlgaounimcosne tsrain Table 6: The majority of extractions that were missed by REVERB were cases where the correct relation phrase was found, but the arguments were not correctly identified. “I gave him 15 photographs,” REVERB extracts (I, gave, him). These errors are due to the fact that REVERB only models binary relations. Table 6 summarizes the correct extractions that were extracted by other systems and were not extracted by REVERB. As with the false positive extractions, the majority of false negatives (52%) were due to the argument-finding heuristics choosing the wrong arguments, or failing to extract all possible arguments (in the case of coordinating conjunctions). Other sources of failure were due to the lexical constraint either failing to filter out an overspecified relation phrase or filtering out a valid relation phrase. These errors hurt both precision and recall, since each case results in the extractor overlooking a correct relation phrase and choosing another. 5.3 Evaluation At Scale Section 5.1 shows that REVERB outperforms existing Open IE systems when evaluated on a sample of sentences. Previous work has shown that the frequency of an extraction in a large corpus is useful for assessing the correctness of extractions (Downey et al., 2005). Thus, it is possible a priori that REVERB’s gains over previous systems will diminish when extraction frequency is taken into account. In fact, we found that REVERB’s advantage over TEXTRUNNER when run at scale is qualitatively similar to its advantage on single sentences. We ran both REVERB and TEXTRUNNER on Banko and Etzioni’s corpus of 500 million Web sentences and examined the effect of redundancy on precision. As Downey’s work predicts, precision increased in both systems for extractions found multiple times, compared with extractions found only once. However, REVERB had higher precision than TEXTRUNNER at all frequency thresholds. In fact, REVERB’s frequency 1 extractions had a precision of 0.75, which TEXTRUNNER could not approach even with frequency 10 extractions, which had a precision of 0.34. Thus, REVERB is able to return more correct extractions at a higher precision than TEXTRUNNER, even when redundancy is taken into account. 6 Conclusions and Future Work The paper’s contributions are as follows: • We have identified and analyzed the problems oWfe ein hcaovhee riedennt tainfide dun ainndfo armnaaltyivzeed de txhtera pctrioobnlse mfors Open IE systems, and shown their prevalence for systems such as TEXTRUNNER and WOE. • We articulated general, easy-to-enforce consWtreain atrst on binary, nveerrabl,-b aesaseyd rteo-laetniofonr phrases in English that ameliorate these problems and yield richer and more informative relations (see, for example, Table 2). • Based on these constraints, we designed, implemented, haensde e cvoanlustartaeind tsh,e w wReE dVeEsRigBn eedx,tr iamc-tor, which substantially outperforms previous Open IE systems in both recall and precision. • We make REVERB and the data used in our experiments available to the research community.4 In future work, we plan to explore utilizing our constraints to improve the performance of learned CRF models. Roth et al. have shown how to incorporate constraints into CRF learners (Roth and Yih, 2005). It is natural, then, to consider whether the combination of heuristically labeled training examples, CRF learning, and our constraints will result in superior performance. The error analysis in Section 5.2 also suggests natural directions for future work. For instance, since many of REVERB’s errors are due to incorrect arguments, improved methods for argument extraction are in order. Acknowledgments We would like to thank Mausam, Dan Weld, Yoav Artzi, Luke Zettlemoyer, members of the KnowItAll 4http : / / reverb . cs .washingt on . edu 1544 group, and the anonymous reviewers for their helpful comments. This research was supported in part by NSF grant IIS-0803481, ONR grant N00014-08- 1-0431, and DARPA contract FA8750-09-C-0179, and carried out at the University of Washington’s Turing Center. References David J. Allerton. 2002. Stretched Verb Constructions in English. Routledge Studies in Germanic Linguistics. Routledge (Taylor and Francis), New York. Collin F. Baker, Charles J. Fillmore, and John B. Lowe. 1998. The berkeley framenet project. In Proceedings of the 1 international conference on Computational 7th linguistics, pages 86–90. Michele Banko and Oren Etzioni. 2008. The tradeoffs between open and traditional relation extraction. In Proceedings of ACL-08: HLT, pages 28–36, Columbus, Ohio, June. Association for Computational Linguistics. Michele Banko, Michael J. Cafarella, Stephen Soderland, Matt Broadhead, and Oren Etzioni. 2007. Open information extraction from the web. In In the Proceedings of the 20th International Joint Conference on Artificial Intelligence, pages 2670–2676, January. Jonathan Berant, Ido Dagan, and Jacob Goldberger. 2011. Global learning of typed entailment rules. In Proceedings of ACL, Portland, OR. Janara Christensen, Mausam, Stephen Soderland, and Oren Etzioni. 2010. Semantic role labeling for open information extraction. In Proceedings of the NAACL HLT 2010 First International Workshop on Formalisms and Methodology for Learning by Reading, FAM-LbR ’ 10, pages 52–60, Stroudsburg, PA, USA. Association for Computational Linguistics. Doug Downey, Oren Etzioni, and Stephen Soderland. 2005. A probabilistic model of redundancy in information extraction. In IJCAI, pages 1034–1041. Gregory Grefenstette and Simone Teufel. 1995. Corpusbased method for automatic identification of support verbs for nominalizations. In Proceedings of the seventh conference on European chapter of the Association for Computational Linguistics, pages 98–103, San Francisco, CA, USA. Morgan Kaufmann Publishers Inc. Raphael Hoffmann, Congle Zhang, and Daniel S. Weld. 2010. Learning 5000 relational extractors. In Proceedings ofthe 48thAnnual Meeting ofthe Association for Computational Linguistics, ACL ’ 10, pages 286– 295, Stroudsburg, PA, USA. Association for Computational Linguistics. Holmer Hemsen Kathrin Eichler and Gnter Neumann. 2008. Unsupervised relation extraction from web documents. In LREC. http://www.lrecconf.org/proceedings/lrec2008/. J. Kim and D. Moldovan. 1993. Acquisition of semantic patterns for information extraction from corpora. In Procs. of Ninth IEEE Conference on Artificial Intelligence for Applications, pages 171–176. Dekang Lin and Patrick Pantel. 2001. DIRT-Discovery of Inference Rules from Text. In Proceedings of ACM Conference on Knowledge Discovery and Data Mining(KDD-01), pages pp. 323–328. Thomas Lin, Mausam, and Oren Etzioni. 2010. Identifying Functional Relations in Web Text. In Proceedings of the 2010 Conference on Empirical Methods in Natural Language Processing, pages 1266–1276, Cambridge, MA, October. Association for Computational Linguistics. Paul Kingsbury Martha and Martha Palmer. 2002. From treebank to propbank. In In Proceedings of LREC2002. Mike Mintz, Steven Bills, Rion Snow, and Dan Jurafsky. 2009. Distant supervision for relation extraction without labeled data. In ACL-IJCNLP ’09: Proceedings of the Joint Conference of the 47th Annual Meeting of the ACL and the 4th International Joint Conference on Natural Language Processing of the AFNLP: Volume 2, pages 1003–101 1, Morristown, NJ, USA. Association for Computational Linguistics. V. Punyakanok, D. Roth, and W. Yih. 2008. The importance of syntactic parsing and inference in semantic role labeling. Computational Linguistics, 34(2). E. Riloff. 1996. Automatically constructing extraction patterns from untagged text. In Procs. of the Thirteenth National Conference on Artificial Intelligence (AAAI-96), pages 1044–1049. Alan Ritter, Mausam, and Oren Etzioni. 2010. A Latent Dirichlet Allocation Method for Selectional Preferences. In ACL. Dan Roth and Wen-tau Yih. 2005. Integer linear programming inference for conditional random fields. In Proceedings of the 22nd international conference on Machine learning, ICML ’05, pages 736–743, New York, NY, USA. ACM. Stefan Schoenmackers, Oren Etzioni, Daniel S. Weld, and Jesse Davis. 2010. Learning first-order horn clauses from web text. In Proceedings of the 2010 Conference on Empirical Methods in Natural Language Processing, EMNLP ’ 10, pages 1088–1098, Stroudsburg, PA, USA. Association for Computational Linguistics. Satoshi Sekine. 2006. On-demand information extraction. In Proceedings of the COLING/ACL on Main 1545 conference poster sessions, pages 73 1–738, Morristown, NJ, USA. Association for Computational Linguistics. Yusuke Shinyama and Satoshi Sekine. 2006. Preemptive Information Extraction using Unrestricted Relation Discovery. In Proceedings of the Human Language Technology Conference of the NAACL, Main Conference, pages 304–3 11, New York City, USA, June. Association for Computational Linguistics. Stephen Soderland, Brendan Roof, Bo Qin, Shi Xu, Mausam, and Oren Etzioni. 2010. Adapting open information extraction to domain-specific relations. AI Magazine, 3 1(3):93–102. S. Soderland. 1999. Learning Information Extraction Rules for Semi-Structured and Free Text. Machine Learning, 34(1-3):233–272. Lucia Specia and Enrico Motta. 2006. M.: A hybrid approach for extracting semantic relations from texts. In In. Proceedings of the 2 nd Workshop on Ontology Learning and Population, pages 57–64. Suzanne Stevenson, Afsaneh Fazly, and Ryan North. 2004. Statistical measures of the semi-productivity of light verb constructions. In 2nd ACL Workshop on Multiword Expressions, pages 1–8. M. Stevenson. 2004. An unsupervised WordNet-based algorithm for relation extraction. In Proceedings of the “Beyond Named Entity ” workshop at the Fourth International Conference on Language Resources and Evalutaion (LREC’04). Kristina Toutanova, Aria Haghighi, and Christopher D. Manning. 2008. A global joint model for semantic role labeling. Computational Linguistics, 34(2): 161– 191. Fei Wu and Daniel S. Weld. 2010. Open information ex- traction using Wikipedia. In Proceedings of the 48th Annual Meeting of the Association for Computational Linguistics, ACL ’ 10, pages 118–127, Morristown, NJ, USA. Association for Computational Linguistics. Jun Zhu, Zaiqing Nie, Xiaojiang Liu, Bo Zhang, and Ji-Rong Wen. 2009. StatSnowball: a statistical approach to extracting entity relationships. In WWW ’09: Proceedings of the 18th international conference on Worldwide web, pages 101–1 10, New York, NY, USA. ACM.

5 0.63447934 98 emnlp-2011-Named Entity Recognition in Tweets: An Experimental Study

Author: Alan Ritter ; Sam Clark ; Mausam ; Oren Etzioni

6 0.63095075 116 emnlp-2011-Robust Disambiguation of Named Entities in Text

7 0.61923212 57 emnlp-2011-Extreme Extraction - Machine Reading in a Week

8 0.49221629 143 emnlp-2011-Unsupervised Information Extraction with Distributional Prior Knowledge

9 0.48591536 113 emnlp-2011-Relation Acquisition using Word Classes and Partial Patterns

10 0.47847199 84 emnlp-2011-Learning the Information Status of Noun Phrases in Spoken Dialogues

11 0.37420368 23 emnlp-2011-Bootstrapped Named Entity Recognition for Product Attribute Extraction

12 0.3656899 145 emnlp-2011-Unsupervised Semantic Role Induction with Graph Partitioning

13 0.36216152 67 emnlp-2011-Hierarchical Verb Clustering Using Graph Factorization

14 0.34664342 37 emnlp-2011-Cross-Cutting Models of Lexical Semantics

15 0.34649572 109 emnlp-2011-Random Walk Inference and Learning in A Large Scale Knowledge Base

16 0.34532773 59 emnlp-2011-Fast and Robust Joint Models for Biomedical Event Extraction

17 0.32954535 40 emnlp-2011-Discovering Relations between Noun Categories

18 0.32121488 144 emnlp-2011-Unsupervised Learning of Selectional Restrictions and Detection of Argument Coercions

19 0.28990901 141 emnlp-2011-Unsupervised Dependency Parsing without Gold Part-of-Speech Tags

20 0.27350199 126 emnlp-2011-Structural Opinion Mining for Graph-based Sentiment Representation

similar papers computed by lda model

lda for this paper:

topicId topicWeight

[(3, 0.239), (23, 0.115), (36, 0.023), (37, 0.021), (45, 0.096), (53, 0.021), (54, 0.027), (57, 0.032), (62, 0.064), (64, 0.026), (66, 0.038), (69, 0.033), (75, 0.015), (79, 0.05), (82, 0.016), (90, 0.012), (96, 0.083), (98, 0.016)]

similar papers list:

simIndex simValue paperId paperTitle

same-paper 1 0.77671438 128 emnlp-2011-Structured Relation Discovery using Generative Models

Author: Limin Yao ; Aria Haghighi ; Sebastian Riedel ; Andrew McCallum

2 0.59223795 98 emnlp-2011-Named Entity Recognition in Tweets: An Experimental Study

Author: Alan Ritter ; Sam Clark ; Mausam ; Oren Etzioni

3 0.59185874 108 emnlp-2011-Quasi-Synchronous Phrase Dependency Grammars for Machine Translation

Author: Kevin Gimpel ; Noah A. Smith

Abstract: We present a quasi-synchronous dependency grammar (Smith and Eisner, 2006) for machine translation in which the leaves of the tree are phrases rather than words as in previous work (Gimpel and Smith, 2009). This formulation allows us to combine structural components of phrase-based and syntax-based MT in a single model. We describe a method of extracting phrase dependencies from parallel text using a target-side dependency parser. For decoding, we describe a coarse-to-fine approach based on lattice dependency parsing of phrase lattices. We demonstrate performance improvements for Chinese-English and UrduEnglish translation over a phrase-based baseline. We also investigate the use of unsupervised dependency parsers, reporting encouraging preliminary results.

4 0.58806437 28 emnlp-2011-Closing the Loop: Fast, Interactive Semi-Supervised Annotation With Queries on Features and Instances

Author: Burr Settles

Abstract: This paper describes DUALIST, an active learning annotation paradigm which solicits and learns from labels on both features (e.g., words) and instances (e.g., documents). We present a novel semi-supervised training algorithm developed for this setting, which is (1) fast enough to support real-time interactive speeds, and (2) at least as accurate as preexisting methods for learning with mixed feature and instance labels. Human annotators in user studies were able to produce near-stateof-the-art classifiers—on several corpora in a variety of application domains—with only a few minutes of effort.

5 0.58766359 116 emnlp-2011-Robust Disambiguation of Named Entities in Text

Author: Johannes Hoffart ; Mohamed Amir Yosef ; Ilaria Bordino ; Hagen Furstenau ; Manfred Pinkal ; Marc Spaniol ; Bilyana Taneva ; Stefan Thater ; Gerhard Weikum

6 0.58675861 37 emnlp-2011-Cross-Cutting Models of Lexical Semantics

7 0.58476049 33 emnlp-2011-Cooooooooooooooollllllllllllll!!!!!!!!!!!!!! Using Word Lengthening to Detect Sentiment in Microblogs

8 0.58443302 114 emnlp-2011-Relation Extraction with Relation Topics

9 0.58418828 53 emnlp-2011-Experimental Support for a Categorical Compositional Distributional Model of Meaning

10 0.58408999 81 emnlp-2011-Learning General Connotation of Words using Graph-based Algorithms

11 0.58329576 70 emnlp-2011-Identifying Relations for Open Information Extraction

12 0.58198029 123 emnlp-2011-Soft Dependency Constraints for Reordering in Hierarchical Phrase-Based Translation

13 0.58163661 101 emnlp-2011-Optimizing Semantic Coherence in Topic Models

14 0.5814501 59 emnlp-2011-Fast and Robust Joint Models for Biomedical Event Extraction

15 0.58098119 132 emnlp-2011-Syntax-Based Grammaticality Improvement using CCG and Guided Search

16 0.58090806 1 emnlp-2011-A Bayesian Mixture Model for PoS Induction Using Multiple Features

17 0.57842994 8 emnlp-2011-A Model of Discourse Predictions in Human Sentence Processing

18 0.5774222 78 emnlp-2011-Large-Scale Noun Compound Interpretation Using Bootstrapping and the Web as a Corpus

19 0.57641059 100 emnlp-2011-Optimal Search for Minimum Error Rate Training

20 0.57521647 57 emnlp-2011-Extreme Extraction - Machine Reading in a Week