X1
–
Multimodal alignment corpora: statistical modelling and information management
Philipp Cimiano, Hannes Rieser, David Schlangen
The X1 project focuses on the text technological modelling of
the input/output relation of alignment processes, the management and maintenance of
the resulting representations by means of an alignment corpus management system (ACMS) and
on the interchange of these representations between research projects inside and
outside Bielefeld. Within the project, an XML-based representation schema is developed
which is sensitive to the specific complexity of alignment data due to multiple
representation layers, informational uncertainty, the impact of alignment histories and the
dynamics of the evolvement of alignment processes. In order to model this specific
complexity, the project includes elementary statistical analyses of alignment data based on
a representation scheme of such analyses. That is, it integrates the standardized,
XML-based annotation with its further processing in the framework of statistical
modelling. In this sense, the project integrates computational linguistic and
text-technological modelling on the basis of a uniform framework in order to guarantee
informational sustainability of data representation models of alignment processes and
their instances.