Thomas Delavallade

A Risk Assessment System with Automatic Extraction of Event Types

In this article we describe the joint effort of experts in linguistics, information extraction and risk assessment to integrate EventSpotter, an automatic event extraction engine, into ADAC, an automated early warning system. By detecting as early as possible weak signals of emerging risks ADAC provides a dynamic synthetic picture of situations involving risk. The ADAC system calculates risk on the basis of fuzzy logic rules operated on a template graph whose leaves are event types. EventSpotter is based on a general purpose natural language dependency parser, XIP, enhanced with domain-specific lexical resources (Lexicon-Grammar). Its role is to automatically feed the leaves with input data.

Proposition of a semi-automatic possibilistic information scoring process

This paper proposes a semi-automatic three step information scoring process that starts from constructs representing structured pieces of information and a user query. It first identifies the constructs relevant to answer the user question, based on their similarity to the query. The relevant items are then individually scored, taking into account both the reliability of their source and the certainty the latter expresses through its choice of linguistic terms. Lastly, these individual scores are fused, modeling a corroboration process that takes into account information obsolescence and source relations. This procedure is performed in the framework of possibility theory, relying on the definition of the appropriate aggregation operators.

MONITORING EVENT FLOWS AND MODELLING SCENARIOS FOR CRISIS PREDICTION: APPLICATION TO ETHNIC CONFLICTS FORECASTING

We have developed an early warning prototype, based on a knowledge management approach, so as to carry out the online detection of crises. Experts, with the help of automatic tools, design an ontology describing domain-specific crisis eruption processes. Then a recognition engine performs model-based inference, in order to identify among the events feeding the system, typical sequences that might trigger a crisis. Crises are described in the ontology through the template technique which provides also mechanisms to assess the similarity between stored scenarios and event flows related to the monitored process. This technique takes into account imperfect knowledge and uncertainties: for instance, imprecise temporal constraints between events are represented by fuzzy sets.

On the Challenges of Using Social Media for Crisis Management

In crisis situations, the challenge of understanding the current situation is tightly linked to the ability to process the variety and the amount of information provided by the multiple sources. In particular, social media can provide additional insight on real-time events, providing that the information that they relay is accurately retrieved, evaluated, and fused. In this chapter, we describe various mechanisms and functions necessary for information fusion and understanding, starting from social media exploration and retrieval, then describing the fusion process and the associated management of information uncertainty, concluding with a description of the methodology and experiments we use to tackle the intrinsic big volume of data and processing required for social media information analysis.

Two methods for Internet buzz detection exploiting the citation graph

This paper addresses the task of detecting Internet buzzes, defined as amplification phenomena, i.e. the diffusion on a very large scale of an Internet content, massively taken up within a short period of time. It proposes two approaches based on the citation graph that represents hyperlinks relation between websites. The first method detects temporal abnormalities in the number of citations of an information source, identifying information sources that undergo a surge of their direct citations. The second method exploits higher level cues, based on the definition of the dynamic cumulative visibility of an article. It captures the notion of citation cascade that is central to the specific type of buzzes related to rumour. Both detection approaches are illustrated, respectively on real data extracted from the Web and on realistic simulated data. The experimental study shows the relevance of the proposed methods and highlights their differences.