Evgeny Knutov


Sustainable Urban Delta Knowledge and Semantic Search

We introduce the SUD search system use-case for the flexible management
of a dynamic co-evolving document collection and knowledge structure
in a focused domain. The work reported here is in the context
of document and knowledge management activities in the domestic
heat and separation industry. Within the framework a
vast amount of unstructured information becomes available in the
form of different reports (primarily PDF) submitted by different
companies, and experts. There is a need to automate the processing
of these reports and to help domain experts to find and analyze
the most important information, and turn this information into a
knowledge base. In this note, we discuss how web standards and
tools are used in the design and give an overview
of the novel challenges we have faced and our solutions towards an
extensible and scalable evidence-based knowledge and document

The stream of documents and reports produced by organizations
is always increasing. Knowledge in these documents is increasing
as well, but is not easy to manage. The promise of the Semantic
Web is to make such knowledge more accessible as long as it
is published in a machine-readable format. Natural language
processing can help bridge the gap between documents and knowledge,
but existing platforms typically operate on static (snapshots
of) document sets.
In this presentation we report on the SUD infrastructure prepared as a pilot for the RVO which is both a knowledge base and a document manager. The knowledge base is developed semi-automatically, and in parallel with the evolving
document collection. The solution keeps the documents and
knowledge in sync by propagating changes in either direction thus
maintaining sustainability of the evidence documents (PDF documents)
and the respective knowledge representation (ontology).