Delphine Battistelli

Reasoning on aspectual-temporal information in French within conceptual graphs

This paper presents a modeling of time, aspect and verbal meanings in natural language processing within Simple Conceptual Graphs (SCG) by way of Semantico-Cognitive Schemes (SCS) and the aspectual-temporal theory. Our system translates a semantico-cognitive representation in terms of SCGs. The SCS allows us to build a representation of a text taking into account fine subtleties of natural language as the information about time and aspect. The Conceptual Graphs formalism provides a powerful inferential mechanism which makes it possible to reason from texts. Our work bears on French texts. A text is represented by two different structures both represented within the SCG model. The first structure models the semantico-cognitive representation while the second one is the temporal diagram representing the temporal constraints between the situations described in the text. Linking these structures entails an expansion of the original SCG model.

Semantics of calendar adverbials for information retrieval

Unlike most approaches in the field of temporal expressions annotation, we consider that temporal adverbials could be relevant units from the point of view of Information Retrieval. We present here the main principles of our semantic modeling approach to temporal adverbial units. It comprises two steps: functional modeling (using a small number of basic operators) and referential modeling (using calendar intervals). In order to establish relationships between calendar zones, our approach takes into account not only the calendar values involved in adverbial units but also the semantics of prepositional phrases involved in these units. Through a first experiment, we show how Information Retrieval systems could benefit from indexing calendar expressions in texts, defining relevance scores that combine keywords and temporal ranking models.

Symptom extraction issue

This work focuses on signs and symptoms recognition in biomedical texts abstracts. First, this specific task is described from a linguistic point of view. Then a methodology combining pattern mining and language processing is proposed. In the absence of an authoritative annotated corpus, our approach has the advantage of being weakly-supervised. Preliminary experimental results are discussed and reveal promising avenues.

Information Retrieval: Ranking Results According to Calendar Criteria

Our work deals with calendar information as it is expressed in natural language (NL), that is to say through textual units such as prepositional phrases or noun phrases (e.g. in the 90s, at the beginning of the XVth century). We call these textual units Calendar Expressions (CE). Our work aims at showing how Information Retrieval systems can benefit from dealing with CE. In this paper we describe our overall approach which consists in a formal analysis of CEs that leads to a semantic representation. We then detail an algorithm that uses this representation to filter and rank CEs embedded in texts, according to a query containing a CE. The algorithm is integrated in an experimental search engine (called CaSE). Our representation of calendar information as it is expressed in NL and the function which computes the proximity between the two CEs, one in the text and the other in the query, provides a mean to process a query without any overlapping.

Aspecto-Temporal Data and Lexical Representations in French within Simple Conceptual Graphs on the Basis of Semantico-Cognitive Schemes

This paper deals with the modeling of time, aspect and verbal meanings in natural language processing within Simple Conceptual Graphs (SCG) by way of Semantico-Cognitive Schemes (SCS) and the aspecto-temporal theory. The expression of a semantico-cognitive representation within SCGs is automatically tractable. The SCS allows us to build a representation of a text taking into account the information about time and aspect. It allows us to represent fine subtleties of natural language. On the other hand, the Conceptual Graphs formalism provides a powerful inferential mechanism which makes it possible to reason from texts. Our work bears on French texts. A text is represented by two different structures both represented within the SCG model. The first structure models the semantico-cognitive representation while the second one is the temporal diagram representing the temporal constraints between the situations described in the text. Linking these structures leads us to slightly extend the original SCG model.

Weakly-Supervised Symptom Recognition for Rare Diseases in Biomedical Text

In this paper, we tackle the issue of symptom recognition for rare diseases in biomedical texts. Symptoms typically have more complex and ambiguous structure than other biomedical named entities. Furthermore, existing resources are scarce and incomplete. Therefore, we propose a weakly-supervised framework based on a combination of two approaches: sequential pattern mining under constraints and sequence labeling. We use unannotated biomedical paper abstracts with dictionaries of rare diseases and symptoms to create our training data. Our experiments show that both approaches outperform simple projection of the dictionaries on text, and their combination is beneficial. We also introduce a novel pattern mining constraint based on semantic similarity between words inside patterns.