[Crm-sig] CAA modelling session

Øyvind Eide oyvind.eide at iln.uio.no
Wed Sep 12 14:49:19 EEST 2012

Dear Colleagues,

On behalf of the Computer Applications and Quantitative Methods in
Archaeology CAA 2013 Conference ‘Across Space and Time’, which will be
held at the University Club of Western Australia in Perth, Australia on
25-28 March 2013,  we would like to invite you to consider to attend 
session "S20. Modelling space and time as rooms of possibilities" that we 
will chair. The session abstract is included below.

For information about the Call for Papers please refer to the following web
pages: http://www.caa2013.org/drupal/papers (call for papers); 
http://www.caa2013.org/CAA2013sessions.pdf (list of sessions). 
The submission deadline is 10th October 2012.

Should you have any questions please do not hesitate to contact the
conference organiser, Dr Arianna Traviglia on arianna.traviglia at mq.edu.au or
Øyvind Eide on oyvind.eide at iln.uio.no.

We hope you will consider submitting a proposal for the session. The abstract is as follows:

S20. Modelling space and time as rooms of possibilities 
Chairs: Ø. Eide, J. Holmen, C.-E. Ore
Format: Paper presentation (LP)

During the last two decades, large collections of cultural heritage material have become available in digital format, including historical source texts, museum catalogues, and excavation reports. Although a number of advanced methods already exist for searching and integrating such resources, they are still under-utilised in storytelling and advanced reasoning. It is still the case that many scholarly users only access such resources through free text or simple metadata searches, if they use them at all.

On the other hand, there are several well defined standards and methods for data integration at various levels, including, e.g., CIDOC-CRM, Linked Data and RDF/OWL. But such methods are only applied to limited data sets. Although no uncontested history is known or even knowable based on available source materials, the sources can still be interpreted as representing a series of events happening in time and space, involving actors and objects. Even if these events do not give themselves to the researcher as a well ordered series steered by causality, they are still related.

We envisage systems which should be able to handle questions in the form of more or less developed hypotheses. This may imply that a tool can build up a scenario based on interpretation of available source material and also be used to scrutinise the plausibility of such scenarios, including likely connections between events in time and space. This will, of course, include speculations about causality.

For the session, we invite papers presenting ways to integrate historical events as they can be read from historical documents. Our experience leads us to believe that hybrid human-computer methods are better suited for solving such problems than fully algorithmic or fully manual approaches, but we are looking forward to being challenged by papers seeing this differently.

We are looking for well developed plans and prototypes for how such data integration can be done in practice, ideally based on real life examples. Rather than providing answers, we would like a focus on organising the information for the researchers so that they can use the systems for developing new questions; but again, we are happy to be challenged by proposed papers. Questions we would like to see addressed in paper proposals include, but are not restricted to:

- How to use various sources to verify co-references within one source material as well as across collections. Co-referring items could, e.g., refer to events, artefacts, places, or actors.

- If and how inferred information can be separated from «basic» information, at all levels from theory to the level of implementation.

- Can rules be established for how events and actors are distributed in time and space, e.g., based on maximum travel speed at a certain time and place?

- Can rules be established for how to use causality and other relationships between events to decide on and reduce uncertainty in their possible time spans and locations?

- Which tools are already available, and what would be needed in order to move forward in this area? Can existing tools be used in better ways?


Kind regards

Øyvind Eide
Jon Holmen
Christian-Emil Ore
Unit for Digital Documentation, University of Oslo.

More information about the Crm-sig mailing list