2008 - 2010 Social Media Miner / SMM funded by IBB Berlin
The project deals with trend detection and topic mining of social media resources.
The
blogrelated webpage contains more information and a screencast.
2008 Urban Sync / EU COST ACTION Sonic Interaction Design
The project collected realworld multimodal data about urban well-being.
The data is available to interested researchers via the
blog.
2006 - 2007 - Amplog
The project aims at developing a visual configuration tool to aggregate
weblogs. The manual usage of blog search engines, the subsequent
scanning of postings according to structural preferences (short
postings, audio included, etc.) and the aggreagtion via standard RSS
readers will be obsolete by using the AMPLOG configuration. Furthermore
the resulting topic-related metablogs may be converted to portable
formats such as podcasts.
2005 - 2007 - Smartweb
The project
Smartweb
is a BMBF-funded joint project of several academic and industrial
partners. It aims at natural-language based mobile access to multimedia
assets. Our part consists of the development of a multimedia ontology
and (semi)automatic annotation tools. Shankar Vembu is working on a
protege plugin to make these goals come true.
October 2004 - March 2005 - MP3 Konzert Archiv
The project
mp3 konzert archiv
was a research study that was held together with students at the
University of Kaiserslautern in 2005. The goal of the project was to do
inspirational brainstorm and to realise an exhibition on the condition
of music and digital audio of today. AI and beauty of design were key
issues.
June 2003 - May 2004 - MPeer
This project offered me the
possibility to develop a prototype for a mobile music recommendation
system. The academic outcome is documented in my Ph.D Thesis about
Artificial Listening Systems and the Ecological Evaluation of Multimodal Music Similarity Models. The online demonstration is still available so
enjoy!
Without my excellent co-workers Tim Pohle, Oliver Hummel and Shankar
Vembu it would not exist. All three of them are meanwhile working on
their Ph.D theses around the topic of music retrieval and the semantic
web.
June 2002 - May 2003 - EHunter
We did some web crawling of
personal websites and subsequent shallow linguistics in order to
extract relevant information in this project. The main algorithms could
be re-used beyond the headhunting application in order to extract
cultural metadata about artists.
November 2001 - May 2002 - Profinder
After being for a while
in an industrial context I started again research at DFKI within this
project. The combination of subsymbolic text retrieval with top-down
ontologies was the main purpose. The Profinder application was used to
match job candidates and their skill profiles to the existing document
database in a corporation. (The domain-independent text retrieval
components were later re-used to compute similarities of song lyrics).
January 1997 - June 1998 - Virtual Office
This was the last project of
my first phase
at DFKI. BMBF-funding to build the next generation document analysis
system for scanned documents. I was involved at proposal phase and
project leader. The prototype was successful and gained so much
interest from industry that we started the first spin-off of DFKI,
first named Smartlab, later on
Insiders Technologies
which established itself as a 40 people IT company. (Between 1992-1996
I did research on OpticalMusicRecognition (OMR) and
OnlineHandwritingRecognition.)