Bonaventura Coppola

Hierarchical Semantic Role Labeling

We present a four-step hierarchical SRL strategy which generalizes the classical two-level approach (boundary detection and classification). To achieve this, we have split the classification step by grouping together roles which share linguistic properties (e.g. Core Roles versus Adjuncts). The results show that the non-optimized hierarchical approach is computationally more efficient than the traditional systems and it preserves their accuracy.

Shallow Semantic Parsing for Spoken Language Understanding

Most Spoken Dialog Systems are based on speech grammars and frame/slot semantics. The semantic descriptions of input utterances are usually defined ad-hoc with no ability to generalize beyond the target application domain or to learn from annotated corpora. The approach we propose in this paper exploits machine learning of frame semantics, borrowing its theoretical model from computational linguistics. While traditional automatic Semantic Role Labeling approaches on written texts may not perform as well on spoken dialogs, we show successful experiments on such porting. Hence, we design and evaluate automatic FrameNet-based parsers both for English written texts and for Italian dialog utterances. The results show that disfluencies of dialog data do not severely hurt performance. Also, a small set of FrameNet-like manual annotations is enough for realizing accurate Semantic Role Labeling on the target domains of typical Dialog Systems.

Frame Detection over the Semantic Web

In the past, research in ontology learning from text has mainly focused on entity recognition, taxonomy induction and relation extraction. In this work we approach a challenging research issue: detecting semantic frames from texts and using them to encode web ontologies. We exploit a new generation Natural Language Processing technology for frame detection, and we enrich the frames acquired so far with argument restrictions provided by a super-sense tagger and domain specializations. The results are encoded according to a Linguistic MetaModel, which allows a complete translation of lexical resources and data acquired from text, enabling custom transformations of the enriched frames into modular ontology components.

Cross-Language Frame Semantics Transfer in Bilingual Corpora

Recent work on the transfer of semantic information across languages has been recently applied to the development of resources annotated with Frame information for different non-English European languages. These works are based on the assumption that parallel corpora annotated for English can be used to transfer the semantic information to the other target languages. In this paper, a robust method based on a statistical machine translation step augmented with simple rule-based post-processing is presented. It alleviates problems related to preprocessing errors and the complex optimization required by syntax-dependent models of the cross-lingual mapping. Different alignment strategies are here investigated against the Europarl corpus. Results suggest that the quality of the derived annotations is surprisingly good and well suited for training semantic role labeling systems.

Engineering of Syntactic Features for Shallow Semantic Parsing

Recent natural language learning research has shown that structural kernels can be effectively used to induce accurate models of linguistic phenomena. In this paper, we show that the above properties hold on a novel task related to predicate argument classification. A tree kernel for selecting the subtrees which encodes argument structures is applied. Experiments with Support Vector Machines on large data sets (i.e. the PropBank collection) show that such kernel improves the recognition of argument boundaries.

Automatic framenet-based annotation of conversational speech

Current Spoken Language Understanding technology is based on a simple concept annotation of word sequences, where the interdependencies between concepts and their compositional semantics are neglected. This prevents an effective handling of language phenomena, with a consequential limitation on the design of more complex dialog systems. In this paper, we argue that shallow semantic representation as formulated in the Berkeley FrameNet Project may be useful to improve the capability of managing more complex dialogs. To prove this, the first step is to show that a FrameNet parser of sufficient accuracy can be designed for conversational speech. We show that exploiting a small set of FrameNet-based manual annotations, it is possible to design an effective semantic parser. Our experiments on an Italian spoken dialog corpus, created within the LUNA project, show that our approach is able to automatically annotate unseen dialog turns with a high accuracy.

Towards entailment-based question answering: ITC-irst at CLEF 2006

This year, besides providing support to other groups participating in cross-language Question Answering (QA) tasks, and submitting runs both for the monolingual Italian and for the cross-language Italian/English tasks, the ITC-irst participation in the CLEF campaign concentrated on the Answer Validation Exercise (AVE). The participation in the AVE task, with an answer validation module based on textual entailment recognition, is motivated by our objectives of (i) creating a modular framework for an entailment-based approach to QA, and (ii) developing, in compliance with this framework, a stand-alone component for answer validation which implements different approaches to the problem.

Generalized Framework for Syntax-Based Relation Mining

Supervised approaches to data mining are particularly appealing as they allow for the extraction of complex relations from data objects. In order to facilitate their application in different areas, ranging from protein to protein interaction in bioinformatics to text mining in computational linguistics research, a modular and general mining framework is needed. The major constraint to the generalization process concerns the feature design for the description of relational data. In this paper, we present a machine learning framework for the automatic mining of relations, where the target objects are structurally organized in a tree. Object types are generalized by means of the use of roles, whereas the relation properties are described by means of the underlying tree structure. The latter is encoded in the learning algorithm thanks to kernel methods for structured data, which represent structures in terms of their all possible subparts. This approach can be applied to any kind of data disregarding their very nature. Experiments with support vector machines on two text mining datasets for relation extraction, i.e. the PropBank and FrameNet corpora, show both that our approach is general, and that it reaches state-of-the-art accuracy.

Unsupervised acquisition of entailment relations from the Web

Entailment recognition is a primary generic task in natural language inference, whose focus is to detect whether the meaning of one expression can be inferred from the meaning of the other. Accordingly, many NLP applications would benefit from high coverage knowledgebases of paraphrases and entailment rules. To this end, learning such knowledgebases from the Web is especially appealing due to its huge size as well as its highly heterogeneous content, allowing for a more scalable rule extraction of various domains. However, the scalability of state-of-the-art entailment rule acquisition approaches from the Web is still limited. We present a fully unsupervised learning algorithm for Webbased extraction of entailment relations. We focus on increased scalability and generality with respect to prior work, with the potential of a large-scale Web-based knowledgebase. Our algorithm takes as its input a lexical–syntactic template and searches the Web for syntactic templates that participate in an entailment relation with the input template. Experiments show promising results, achieving performance similar to a state-of-the-art unsupervised algorithm, operating over an offline corpus, but with the benefit of learning rules for different domains with no additional effort.