6533b82afe1ef96bd128ccf3

RESEARCH PRODUCT

Reasoning with Vague Spatial Information from Upper Mesopotamia (2000BC)

Helbert ArenasChristophe CruzBenjamin Harbelot

subject

[ INFO.INFO-IR ] Computer Science [cs]/Information Retrieval [cs.IR]media_common.quotation_subjectReasonning02 engineering and technology[INFO.INFO-AI]Computer Science [cs]/Artificial Intelligence [cs.AI]Description logic0202 electrical engineering electronic engineering information engineeringMesopotamia ;[ INFO.INFO-AI ] Computer Science [cs]/Artificial Intelligence [cs.AI]Function (engineering)Spatial analysisGeneral Environmental ScienceMathematicsmedia_commondescription logicsInformation retrievalPoint (typography)Ontologybusiness.industryProbabilistic logic[INFO.INFO-LO]Computer Science [cs]/Logic in Computer Science [cs.LO]020207 software engineeringVaguenessspatial uncertainty[INFO.INFO-IR]Computer Science [cs]/Information Retrieval [cs.IR]Upper MesopotamiaOntologyGeneral Earth and Planetary Sciences[ INFO.INFO-LO ] Computer Science [cs]/Logic in Computer Science [cs.LO]020201 artificial intelligence & image processingArtificial intelligencebusinessConstruct (philosophy)

description

International audience; Concepts such as near, far, south of, etc., are by its own nature vague. However, they are quite common in human language. In the case of historical records, these concepts are often the only source of information regarding the position of ancient places whose exact location has been lost. In our research, we use digitized written records from Upper Mesopotamia (2000BC) from the HIGEOMES project. Our goal is to provide better understanding of the location of places, based on the analysis of spatial statements. In our approach, we analyse cardinal statements between places with known location. Using this information we construct a probabilistic function representing the vagueness of cardinal statements. For each place with unknown position, we combine multiple vague statements using our probabilistic function in order to limit the possible search area. At this point, our results look promising. In the near future, and thanks to new data, such as proximity statements as provided by the HIGEOMES project, we expect to improve our results.

https://doi.org/10.1016/j.proenv.2015.07.104