Stéphane Ferrari

TempoWordNet for sentence time tagging

In this paper, we propose to build a temporal ontology, which may contribute to the success of time-related applications. Temporal classifiers are learned from a set of time-sensitive synsets and then applied to the whole WordNet to give rise to TempoWordNet. So, each synset is augmented with its intrinsic temporal value. To evaluate TempoWordNet, we use a semantic vector space representation for sentence temporal classification, which shows that improvements may be achieved with the time-augmented knowledge base against a bag-of-ngrams representation.

Propagation Strategies for Building Temporal Ontologies

In this paper, we propose to build temporal ontologies from WordNet. The underlying idea is that each synset is augmented with its temporal connotation. For that purpose, temporal classifiers are iteratively learned from an initial set of time-sensitive synsets and different propagation strategies to give rise to different TempoWordNets.

Tag Thunder: Web Page Skimming in Non Visual Environment Using Concurrent Speech

Skimming and scanning are two strategies generally used for speed reading. Skimming allows a reader to get a first glance of a document; scanning is the process of searching for a specific piece of information in a document. While both techniques are available in visual reading mode, it is rather difficult to use them in non visual environments. In this paper, we introduce the concept of tag thunder, which provides speed reading non-visual techniques similar to skimming and scanning. A tag thunder is the oral transposition of the tag cloud concept. Tag cloud key terms are presented using typographic effects which reflect their relevance and number of occurrences. Within a tag thunder, the relevance of a given key term is translated into specific speech effects and its position on the page is reflected in the position of the corresponding sound on a 2D stereo space. All key terms of a tag thunder are output according to a concurrent speech strategy, which exploits the cocktail party effect. In this paper, we present our implementation of the tag thunder concept. The results of the evaluation campaign show that tag thunders present a viable non-visual alternative to visual speed reading strategies.

Using Textual Clues to Improve Metaphor Processing

In this paper, we propose a textual clue approach to help metaphor detection, in order to improve the semantic processing of this figure. The previous works in the domain studied the semantic regularities only, overlooking an obvious set of regularities. A corpus-based analysis shows the existence of surface regularities related to metaphors. These clues can be characterized by syntactic structures and lexical markers. We present an object oriented model for representing the textual clues that were found. This representation is designed to help the choice of a semantic processing, in terms of possible non-literal meanings. A prototype implementing this model is currently under development, within an incremental approach allowing step-by-step evaluations.

Understanding Temporal Query Intent

Understanding the temporal orientation of web search queries is an important issue for the success of information access systems. In this paper, we propose a multi-objective ensemble learning solution that (1) allows to accurately classify queries along their temporal intent and (2) identifies a set of performing solutions thus offering a wide range of possible applications. Experiments show that correct representation of the problem can lead to great classification improvements when compared to recent state-of-the-art solutions and baseline ensemble techniques.