We use logical inference techniques for recognising textual entailment. As the performance of theorem proving turns out to be highly dependent on not readily available background knowledge, we incorporate model building, a technique borrowed from automated reasoning, and show that it is a useful robust method to approximate entailment. Finally, we use machine learning to combine these deep semantic analysis techniques with simple shallow word overlap; the resulting hybrid model achieves high accuracy on the RTE testset, given the state of the art. Our results also show that the different techniques that we employ perform very differently on some of the subsets of the RTE corpus and as a result, it is useful to use the nature of the dataset as a feature. 1

Exhaustive extraction of semantic information from text is one of the formidable goals of state-of-the-art NLP systems. In this paper, we take a step closer to this objective. We combine the semantic information provided by different resources and extract new semantic knowledge to improve the performance of a recognizing textual entailment system. 1 Recognizing Textual Entailment While communicating, humans use different expressions to convey the same meaning. Therefore, numerous NLP applications, such as, Question Answering, Information Extraction, or Summarization require computational models of language that recognize if two texts semantically overlap. Trying to capture the major inferences needed to understand equivalent semantic expressions, the PASCAL Network proposed the Recognizing Textual Entailment (RTE) challenge (Dagan et al., 2005). Given two text fragments, the task is to determine if the meaning of one text (the entailed hypothesis, H) can be inferred from the meaning of the other text (the entailing text, T). Given the wide applicability of this task, there is an increased interest in creating systems which detect the semantic entailment between two texts. The systems that participated in the Pascal RTE challenge competition exploit various inference elements which, later, they combine within statistical models, scoring methods, or machine learning frameworks. Several systems (Bos and Markert, 2005; Herrera et al., 2005; Jijkoun and de Rijke, 2005; Kouylekov and Magnini, 2005; Newman et al., 2005) measured the word overlap between the two text strings. Using either statistical or Word-Net’s relations, almost all systems considered lexical relationships that indicate entailment. The degree of similarity between the syntactic parse trees of the two texts was also used as a clue for entailment by several systems (Herrera et al., 2005; Kouylekov and Magnini, 2005; de Salvo Braz et al., 2005; Raina et al., 2005). Several groups used logic provers to show the entailment between T and H (Bayer et

Recognizing textual entailment is a challenging problem and a fundamental component of many applications in natural language processing. We present a novel framework for recognizing textual entailment that focuses on the use of syntactic heuristics to recognize false entailment. We give a thorough analysis of our system, which demonstrates state-of-the-art performance on a widely-used test set. 1

In this paper we define a novel similarity measure between examples of textual entailments and we use it as a kernel function in Support Vector Machines (SVMs). This allows us to automatically learn the rewrite rules that describe a non trivial set of entailment cases. The experiments with the data sets of the RTE 2005 challenge show an improvement of 4.4 % over the state-of-the-art methods. 1

Abstract. We use logical inference techniques for recognising textual entailment, with theorem proving operating on deep semantic interpretations as the backbone of our system. However, the performance of theorem proving on its own turns out to be highly dependent on a wide range of background knowledge, which is not necessarily included in publically available knowledge sources. Therefore, we achieve robustness via two extensions. Firstly, we incorporate model building, a technique borrowed from automated reasoning, and show that it is a useful robust method to approximate entailment. Secondly, we use machine learning to combine these deep semantic analysis techniques with simple shallow word overlap. The resulting hybrid model achieves high accuracy on the RTE testset, given the state of the art. Our results also show that the various techniques that we employ perform very differently on some of the subsets of the RTE corpus and as a result, it is useful to use the nature of the dataset as a feature. 1

In this paper we present a novel approach for learning entailment relations from positive and negative examples. We define a similarity between two text-hypothesis pairs based on a syntactic and lexical information. We experimented our model within the RTE 2006 challenge obtaining the accuracy of 63.88 % and 62.50 % for the two submissions. 1

Traditional approaches to Natural Language Text Processing limit performance and flexibility by committing to canonical reprentations of input text, while many NLP applications for general tasks such as Textual Entailment use ad-hoc architectures with limited flexibility, and which limit the expressiveness of inference procedures over components. We present a Modular Representation and Comparison Scheme (MRCS) that addresses these problems by combining a modular representation with a modular, unification-like inference algorithm that allows the system architect to defer appropriate disambiguation decisions until run-time. 1

This paper analyses the contribution of shallow syntactic matching and thesaurus based equivalence in determining semantic equivalence of a pair of sentences. The performance of this approach is evaluated on two data sets and compared to other systems, as well as to manual evaluation results. We conclude that shallow semantics can model equivalence and entailment for pairs of syntactically similar sentences but it is not sufficient for reliable recognition of these relations in more complex cases. 1

1. Program goal: incorporating content words into a model of entailment One of the most important properties of human languages is their ability to convey intricate meanings. The vastness and effectiveness of those meanings for everyday communication and reasoning transcends all known non-human languages, including other animal languages and artificial languages. For studying communication and reasoning in language, an indispensable empirical concept is entailment: the relation between premises and valid conclusions expressed as natural language sentences. Entailment relations may appear between sentences due to the presence of words and expressions like other, either…or, not or exactly five, whose meanings have been studied in logical frameworks since antiquity. However, entailments may also appear due to semantic properties of “non-logical ” words like parrot, hug, far or knowledge. Such content words constitute the bulk of the lexicon in all natural languages. Without considering them, it is simply impossible to understand entailment phenomena and human reasoning in general. However, while content words have played an important role in cognitive psychology and artificial intelligence, their meanings have turned out to be richer

Paraphrasing methods recognize, generate, or extract phrases, sentences, or longer natural language expressions that convey almost the same information. Textual entailment methods, on the other hand, recognize, generate, or extract pairs of natural language expressions, such that a human who reads (and trusts) the first element of a pair would most likely infer that the other element is also true. Paraphrasing can be seen as bidirectional textual entailment and methods from the two areas are often very similar. Both kinds of methods are useful, at least in principle, in a wide range of natural language processing applications, including question answering, summarization, text generation, and machine translation. In this thesis, we focus on paraphrase and textual entailment recognition, as well as paraphrase generation. We propose three paraphrase and textual entailment recognition methods, experimentally evaluated on existing benchmarks. The key idea is that by capturing similarities at various abstractions of the inputs, we can recognize paraphrases and textual entailment reasonably well. Additionally, we exploit WordNet and use features that operate on the syntactic level of the language expressions. The best