Ontology alignment in geographical hard-soft information fusion systems

Proceedings of the 2011 IEEE National Aerospace and Electronics Conference (NAECON)

Dezert

Valin

2011

In this paper, we explore the use of the DezertSmarandache Theory (DSmT) for seismic and acoustic sensor fusion. The seismic/acoustic data is noisy which leads to classification errors and conflicts in declarations. DSmT affords the redistribution of masses when there is a conflict. The goal of this paper is to present an application and comparison on DSmT with other classifier methods to include the support vector machine(SVM) and DempsterShafer (DS) methods. The work is based on two key references (1) Marco Duarte with the initial SVM classifier application of the seismic and acoustic sensor data and (2) Arnaud Martin in Vol. 3 with the Proportional Conflict Redistribution Rule 5/6 (PCR5/PCR6) developments. By using the developments of Duarte and Martin, we were able to explore the various aspects of DSmT in an unattended ground sensor scenario. Using the receiver operator curve (ROC), we compare the methods for individual classification as well as a measure of overall classification using the area under the curve (AUC). Conclusions of the work show that the DSmT results with a maximum forced choice are comparable to the SVM.

Section: Discussionmentioning

confidence: 99%

DSmT applied to seismic and acoustic sensor fusion

Proceedings of the 2011 IEEE National Aerospace and Electronics Conference (NAECON)

Dezert

Valin

2011

“…An alternative is given via the λ divergence, (10) which can be interpreted as the expected information gain about X from discovering which probability distribution X is drawn from, P or Q, if they currently have probabilities (11) where M is the average of the two distributions, M = (½) (P + Q) (12) D JS can also be interpreted as the capacity of a noisy information channel with two inputs giving the output distributions p and q. The Jensen-Shannon divergence is the square of a metric that is equivalent to the Hellinger metric, and the Jensen-Shannon divergence is also equal to one-half the so-called Jeffrey's divergence [68 ,69].…”

Section: Symmetrized Divergencementioning

confidence: 99%

“…In order to support information fusion and decision making, it is necessary to qualify and quantify such uncertainty. The goal is to quality the uncertainty associated with alignment of two ontological information fusion systems [12] for decision making [13]. Information theory has long been regarded as a method of assessing vocabulary content, however, there is need to also evaluate the uncertainty in meaning for the successful merging of databases.…”

Section: 1mentioning

confidence: 99%

Ontology alignment using relative entropy for semantic uncertainty analysis

Proceedings of the IEEE 2010 National Aerospace &Amp; Electronics Conference

Dorion

Valin

et al. 2010

The development and use of many diverse ontologies to support the representational needs of different sources and different contexts is common and necessary. However, the increased sharing of databases implementing heterogeneous ontologies pose the problem of ontological alignment. Ontology alignment typically consists of manual operations from users with different experiences and understandings and limited reporting is conducted in the quality of mappings. To assist the International Organization for Standards (ISO) in standards development for information and data quality assessment, we propose an approach using relative entropy for semantic uncertainty analysis. Information theory has widely been adopted and provides uncertainty assessment for quality of service (QOS) analysis. Quality of information (QOI) or Information Quality (IQ) definitions for semantic assessment can be used to bridge the gap between ontology (semantic) uncertainty alignment and information theory (symbolic) analysis.Pragmatically aiding users of the shared ontologies requires assessments of the cognitive mental models, recognition of semantic classifications, and action over timeliness, throughput, confidence, and accuracy of the translations. In this paper, we explore issues of ontology uncertainty alignment utilizing the elements of information theory (KL divergence or relative entropy). A maritime domain situational awareness example with ship semantic labels is shown to demonstrate ontology alignment uncertainty assessment for data quality standards to assist users for pragmatic surveillance.

“…Situation and Knowledge Representations [16,17] HLIF system design [18,19] HLIF for decision support [20] HLIF evaluation [21,22] Three common themes throughout the papers include: (A) Information fusion designs support situational awareness. Advanced techniques in design (e.g.…”

Section: Fusion10 Panel Papers Overviewmentioning

confidence: 99%

High Level Information Fusion developments, issues, and grand challenges: Fusion 2010 panel discussion

2010 13th International Conference on Information Fusion

Llinas

Lambert

et al. 2010

Self Cite

-The goal of the High-Level Information Fusion (HLIF) Panel Discussion is to present contemporary HLIF advances and developments to determine unsolved grand challenges and issues. The discussion will address the issues between low-level (signal processing and object state estimation and characterization) and high-level information fusion (control, situational understanding, and relationships to the environment). Specific areas of interest include modeling (situations, environments), representations (semantic, knowledge, and complex), systems design (scenario-based, user-based, distributedagent) The HLIF panel discussion's goal is to highlight the unsolved problems and concerns to motivate the information fusion community towards systems-level solutions. The panelists' expert perspectives are based on three areas: (1) previous panel discussions and summaries, (2) an integrated list of HLIF challenges, and (3) companion papers presented at the Fusion2010 conference (note we switch to Fusion10 to refer to the conference). Previous Related Panel DiscussionsPanel discussions provide a valuable resource to the community to overview the current techniques and provide areas of concern for future research. Previous Fusion Conference panel discussion papers related to HLIF include knowledge representation (Fusion05) [7], resource management coordination with situation and threat assessment (Fusion06) [8, 9, 10], agent-based design (Fusion07) [11], and HLIF challenges (Fusion08) [12]. Three panel discussions were conducted at Fusion09 without papers: Many of the authors of this Fusion10 HLIF panel coordinated on previous publications, but continual refinement of HLIF contemporary are desired. The panel discussion follows from a day-long event special session. There are most likely other papers at Fusion10 that are related that would validate good questions from the audience to the panelists. Many of the participants to the special session would be encouraged to voice their opinions and questions to the moderated panel.