Recognizing Implied Predicate-Argument Relationships in Textual Inference

We investigate recognizing implied predicate-argument relationships which are not explicitly expressed in syntactic structure. While prior works addressed such relationships as an extension to semantic role labeling, our work investigates them in the context of textual inference scenarios. Such scenarios provide prior information, which substantially eases the task. We provide a large and freely available evaluation dataset for our task setting, and propose methods to cope with it, while obtaining promising results in empirical evaluations. 1 Motivation and Task This paper addresses a typical sub-task in textual inference scenarios, of recognizing implied predicate-argument relationships which are not expressed explicitly through syntactic structure. Consider the following example: (i) The crucial role Vioxx plays in Merck’s portfolio was apparent last week when Merck’s shares plunged 27 percent to 33 dollars after the withdrawal announcement. While a human reader understands that the withdrawal refers to Vioxx, and hence an implied predicate-argument relationship holds between them, this relationship is not expressed in the syntactic structure, and will be missed by syntactic parsers or standard semantic role labelers. This paper targets such types of implied relationships in textual inference scenarios. Particularly, we investigate the setting of Recognizing Textual Entailment (RTE) as a typical scenario of textual inference. We suggest, however, that the same challenge, as well as the solutions proposed in our work, are applicable, with proper adaptations, to other textual-inference scenarios, like Question Answering, and Information Extraction (see Section 6). An RTE problem instance is composed of two text fragments, termed Text and Hypothesis, as input. The task is to recognize whether a human reading the Text would infer that the Hypothesis is most likely true (Dagan et al., 2006). For our problem, consider a positive Text Hypothesis pair, where the Text is example (i) above and the Hypothesis is: (ii) Merck withdrew Vioxx. A common approach for recognizing textual entailment is to verify that all the textual elements of the Hypothesis are covered, or aligned, by elements of the Text. These elements typically include lexical terms as well as relationships between them. In our example, the Hypothesis lexical terms (“Merck”, “withdrew” and “Vioxx”) are indeed covered by the Text. Yet, the predicateargument relationships (e.g., “withdrawal-Vioxx”) are not expressed in the text explicitly. In such a case, an RTE system has to verify that the predicate-argument relationships which are explicitly expressed in the Hypothesis, are implied from the Text discourse. Such cases are quite frequent (∼17%) in the settings of our dataset, described in Section 3. Consequently, we define the task of recognizing implied predicate-argument relationships, with illustrating examples in Table 1, as follows. The input includes a Text and a Hypothesis. Two terms in the Hypothesis, predicate and argument, are marked, where a predicate-argument relationship between them is explicit in the Hypothesis syntactic structure. Two terms in the Text, candidatepredicate and candidate-argument, aligned to the Hypothesis predicate and argument, are marked as well. However, no predicate-argument relationship between them is expressed syntactically. The task is to recognize whether the predicate# Hypothesis Text Y/N 1 Merck [withdrew]pred [Vioxx]arg from the market. The crucial role [Vioxx]cand-arg plays in Merck’s portfolio was apparent last week when Merck’s shares plunged 27 percent to 33 dollars after the [withdrawal]cand-pred announcement. Y 2 Barbara Cummings heard the tale of a woman who was coming to Crawford to [join]pred Cindy Sheehans [protest]arg. Sheehan’s [protest]cand-arg is misguided and is hurting troop morale. . . . Sheehan never wanted Casey to [join]cand-pred the military. N 3 Casey Sheehan was [killed]pred in [Iraq]arg. 5 days after he arrived in [Iraq]cand-arg last year, Casey Sheehan was [killed]cand-pred. Y 4 Hurricane Rita [threatened]pred [New Orleans]arg. Hurricane Rita was upgraded from a tropical storm as it [threatened]cand-pred the southeastern United States, forcing an alert in southern Florida and scuttling plans to repopulate [New Orleans]cand-arg after Hurricane Katrina turned it into a ghost city 3 weeks earlier. Y 5 Alberto Gonzales defends [renewal]pred of the [Patriot Act]arg to Congress. A senior official defended the [Patriot Act]cand-arg . . . . . . President Bush has urged Congress to [renew]cand-pred the law . . . Y 6 The [train]arg [crash]pred injured nearly 200 people. At least 10 people were killed . . . in the [crash]cand-pred . . . Alvarez is accused of . . . causing the derailment of one [train]cand-arg . . . Y Table 1: Example task instances from our dataset. The last column specifies the Yes/No annotation, indicating whether the sought predicate-argument relationship is implied in the Text. For illustration, a dashed line indicates an explicit argument that is related to the candidate argument through some kind of discourse reference. Pred, arg and cand abbreviate predicate, argument and candidate respectively. argument relationship, as expressed in the Hypothesis, holds implicitly also in the Text. To address this task, we provide a large and freely available annotated dataset, and propose methods for coping with it. A related task, described in the next section, deals with such implied predicate-argument relationships as an extension to Semantic Role Labeling. While the results reported so far on that annotation task were relatively low, we suggest that the task itself may be more complicated than what is actually required in textual inference scenarios. On the other hand, the results obtained for our task, which does fit textual inference scenarios, are promising, and encourage utilizing algorithms for this task in actual inference systems.