Archive for the ‘DFKI’ Category

Workshop on Seamless Integration of Semantic Technologies in Computer-supported Office Work

January 20, 2011

SISTCOW homepage

Held in conjunction with 15th International Conference on Knowledge-Based and Intelligent Information & Engineering Systems (KES 2011), 12, 13 & 14 September 2011, Kaiserslautern, Germany.

Call for papers

Integration of semantic technologies in Computer-supported Office Work (COW) is a hot research topic in knowledge management and semantic web communities. The aim of the research in this field is to increase the performance of knowledge work by:

  • Providing knowledge workers with tools for effective information organization and quick information access
  • Intelligent information search
  • Context-aware information assistance

During the last decade, dozens of research projects worldwide were targeted at developing methodologies, technologies and tools to solve problems of knowledge workers. However, after years, semantic technologies are still not a part of daily work performed by office workers outside computer laboratories. Reasons are manifold: low user acceptance, huge modeling effort needed to bootstrap solutions. The goal of the workshop is to gather researchers who are working on integrating semantic technologies in habitual office environment to exchange information and knowledge about problems preventing successful implementation of promising semantic web ideas in real office environments and to discuss about possible solutions of these problems.

Topics of interest include, but are not limited to:

  1. Semantic annotations in COW
    • Tagging documents and e-mails using semantic data structures, e.g. Linked Data
    • Using RDFa in HTML Mails
    • Annotating e-mail attachments
    • Enriching document content with semantic mashups
    • Semantic e-mail markups in e-mail headers
    • Semantic annotation in Instant Messaging
  2. Efficient search and navigation in semantically annotated collections of documents and e-mails
    • Semantic search in COW
    • Extending Microsoft search to support semantic search
  3. Semi-automatic knowledge extraction from user interaction logs in office environment
    • Enriching semantic data structures by user interaction data
    • Supporting identified interaction habits based on semantic data structures
  4. Semi-automatic document classification
    • E-mail/IM classification using background knowledge bases
    • Ontology-based classification
    • Classifying parts of bigger documents
  5. Architectures supporting semantic technologies in computer-aided office work
    • Architectures for storing semantic document metadata (server-based vs. embedded)
    • Semantic indexing (documents, e-mail, IM)
    • Storing metadata related to parts of larger documents
    • Semantic data formats and vocabularies supporting e-mails and document annotations: RDF, eRDF, Linked Data
    • Semantic data structures to annotate user-system interaction in office environments
    • Workflow-supportive architectures for digital collaborative work
  6. Extending Office applications to utilize semantic technologies
    • Developing Plugins for integrating semantic technologies in MSOffice documents
    • Extending MSOffice SmartTags functionality to use semantic structures
    • PlugIns leveraging application interaction based on semantic technologies
    • Application integration based on semantic technologies
    • Semantic Communication Support Systems

Important Dates

  • Manuscript due: 30 April 2011
  • Notification of acceptance: 15 May 2011
  • Final paper due: 1 June 2011
  • Registration deadline for presenting authors: 1 June 2011

Programm committee

Oleg Rostanin, HomePage
German Research Center for Arrtificial Intelligence (DFKI GmbH),
Trippstadterstr. 122 67633, Kaiserslautern, Germany

Simon Scerri
Digital Enterprise Research Institute (DERI),
National University of Ireland, Galway

Benedikt Schmidt
SAP Research,
Bleichstrasse 8 64283, Darmstadt, Germany

Enhanced by Zemanta

Powerful Search Engine for Your Windows Desktop

July 26, 2010

Since Windows Vista Windows Desktop Search(WDS) is an integrated part of the Windows OS.
In earlier Windows-Versions the user had an option to install the search engine.

WDS indexes specified drives and offline network folders using IFilters (adaptors for different data types) and makes them available for finding files both manually or programmically.
Depending on the file type, WDS indexes the whole textual content of the file or just attributes (title, author, categories, file type … see Windows Properties).

One can install specific IFilters (e.g. for PDFs PDFIFilter for 64bit).
IFilter for MSOffice comes along with the MSOffice software packet.
Unfortunatelly, IFilter for OpenOffice is commercial.

Support for software developers
WDS provides an API for using it or extending its features.
Searching for the documentation in the internet one has to be very carefull as there is a lot of documents describing previous versions of the WDS API that have minimal but crucial differences.

Developers have choice to use the Advanced Query Syntax ADS or or SQL Syntax to query the search engine.

WDS is the first search engine that I knew that supports normalized result relevance scale (from 0 to 1000).
Using SQLsyntax one can combine several subqueries using the OR-operator in the WHERE clause and weight the results of each of them separately.
The engine will combine the results according to these weights and sort it accordingly.

From the experience, AQS queries are translated to SQL automatically and the quality of this translation is not that good for complex queries.
Therefore the recommendation is to use SQL-queries profgrammatically that is very powerful. AQS is a really good thing for manual queries.

One has to notice explicitely that AQS-queries are platform language-dependent, i.e. AQS keywords are different for German and English Windows OS.
An alternative would be to use so called canonical queries.

FULLTEXT search
SQL Syntax of WDS contains support for FREETEXT and CONTAINS predicates that are very useful functions for the text search

Categories, structured search
Windows files can be assigned categories (comma-separated list of strings). MSOutlook items can be assigned categories either. Categories attribute can be searched by WDS (e.g. in ADS-Syntax: TaskNavigator System.Category:VOF will find files, Outlook or One-Note items catagorized with “VOF” and containing “TaskNavigator” in their attributes or text).

Unfortunately, categories are not hierarchical neither they are connected to each other as it is done in the Semantic Desktop.
Therefor for Microsoft it is the long way to go. However WDS with it’s “shortened” semantic OS descriptiuon can be very useful in many cases.

Performance
Form the experience, SQL queries constructed by developers is executed very quickly (interacrtive responce time). For AQS-Translated queries, there is no guarantee, that they are correct or efficient (I experienced that WindowsIndexer crashed as it was trying to process such a query). Also the memory consumption for such query is much higher.

Using AQS in MSOutlook

AQSsyntax can be used in MSOutlook for spontaneous search (von:”Max Musterman” kategorien:ADiWa).