Claire Laudy

High-level fusion based on conceptual graphs

Most of studies in the field of information fusion focus on the production of high-level information from low-level data. The challenge is then to fuse this high-level information to produce a global and coherent information. Another approach consists in interpreting data as high-level information and fuse it at once. Our approach relies on the use of conceptual graphs model. The model is widely used for knowledge representation. We propose to go further and use it for information fusion. Conceptual graphs model contains aggregation operators such as join and maximal join. This paper is dedicated to the extension of the maximal join operator in order to manage heterogeneous information fusion. After describing the suitability of maximal join for high-level information fusion, we present the extension that we propose. The extension relies on relaxing the equality constraint on observations and on using fusion strategies. A case study illustrates our proposition.

Introducing semantic knowledge in high-level fusion

The necessity of taking into account high-level information has recently been reported by the information fusion community. We previously presented an approach for high-level symbolic fusion that relies on the use of the conceptual graphs model. Domain knowledge is a major point of the fusion process. The use of conceptual graphs for knowledge representation eases the process of expressing domain knowledge and fusion heuristics. We use conceptual graphs in order to model a situation of interest. Information items extracted from different sources are fused, using the model of situation in order to orient the fusion by delivering a context. This article details a step that was introduced prior to fusion, that aims at testing the compatibility of observations. It explains the way the domain knowledge is introduced inside the compatibility test, similarly as it was previously introduced in the fusion process. The proposed approach is illustrated within the context of a decision support system for geopolitical crisis management.

Fusion of Claude Bernard's Experiments for Scientific Discovery Reasoning

We are interested in using a fusion process to complete information prior to the reasoning process about scientific discoveries. In particular, using fusion to complete the set of experiments used as the source of information of the process that led Claude Bernard to his discovery about the effects of curare. Our reconstruction of the discovery process is based on his experiments as they are illustrated in his notebooks. Our main problem is the lack of some important information in his notebooks containing descriptions of his set of experiments. In order to fill in the gaps in his set of experiments, we propose to use fusion between experiments. Prior to fusion, we must ensure that the experiments are compatible according to some similarity measures and depending on the objectives of the fusion. The paper presents our domain-independent approach for similarity checking and fusion, including similarity and fusion strategies.

Adjustable Fusion to Support Cyber Security Operators

Cyber security operators use Security Information and Event Management systems to process and summarize the huge amount of heterogeneous logs and alerts. However, these systems do not give to the operator a concise view of the attack status or context, a mandatory feature to understand and remediate properly a threat. Moreover, the number of alerts to analyze for a single information system is high, and thus requires to be split into several levels of responsibility distributed among several operators. This layered security monitoring implies a decision problem as well as an automation problem tackled in this paper with the support of an attack graph-based feature. An attack graph is a risk assessment model that accurately describes, in a concise way, the threats on an information system. In this article, we describe how an attack graph can be used for pattern searching and fusion algorithms, in order to add context to the alerts. We also present recommendations for designing future interactive application based on adjustable fusion and a risk assessment model, for cyber security monitoring.