Link Discovery System
KOJAK is an integrated suite of link discovery tools that perform tasks such as:
- group detection (KOJAK Group Finder)
- anomaly detection (KOJAK UNICORN)
- pattern matching (KOJAK Pattern Finder)
- relationship and graph simplification (KOJAK SimpleRel)
KOJAK generally operates on data represented by labeled graphs where nodes
represent typed entities such as persons, organizations, events, etc. and links
represent different kinds of relationships between these entities (such graphs
are also referred to as semantic graphs or attributed relational
graphs). Graphs might be represented explicitly, or implicitly as views
over relational data (e.g., stored in an RDBMS). Optionally, graphs can be
enhanced with background ontologies and logic-based inferencing supported by
the PowerLoom KR&R system.
The software components outlined above are prototypes at different levels of
maturity. At this time, only the KOJAK Group Finder software is being
released. KOJAK UNICORN and SimpleRel are actively being developed and will
become available next.
KOJAK-Related Publications
- H. Chalupsky. Using KOJAK Link
Discovery Tools to Solve the Cell Phone Calls Mini Challenge.
Proceedings of the IEEE Symposium on Visual Analytics Science and
Technology (IEEE VAST 2008). DVD only, to appear. [pdf] (this is a summary of our
KOJAK submission to the VAST 2008 Cell Phone Calls Mini Challenge)
- H. Chalupsky.
KOJAK submission to the VAST 2008 Cell Phone Calls Mini
Challenge. 2008
- S. Lin and H. Chalupsky.
Discovering and Explaining Abnormal Nodes in Semantic Graphs.
IEEE Transactions on Knowledge and Data Engineering, 20(8):
pages 1039-1052, 2008. [draft pre-print]
- S. Lin.
Modeling, searching, and explaining abnormal instances in multi-relational
networks. Ph.D. Thesis, Department of Computer Science, University of Southern
California. Los Angeles, CA. 2006
- J. Adibi, T. Barrett, S. Bhatt, H. Chalupsky, J. Chame and
M. Hall. Processing-In-Memory Technology
for Knowledge Discovery Algorithms. In Proceedings of the Second International
Workshop on Data Management on New Hardware (DaMoN 2006). June, 2006 [pdf]
- J. Adibi and H. Chalupsky. Scalable Group Detection via a Mutual
Information Model. In Proceedings of the First International
Conference on Intelligence Analysis (IA-2005).
[pdf]
- J. Adibi, H. Chalupsky, E. Melz and A. Valente.
The KOJAK Group Finder: Connecting the Dots via Integrated
Knowledge-Based and Statistical Reasoning. In Proceedings of
the Sixteenth Innovative Applications of Artificial Intelligence
Conference (IAAI-04), 2004. [pdf]
- S. Lin and H. Chalupsky. Using Unsupervised
Link Discovery Methods to Find Interesting Facts and Connections in a
Bibliography Dataset. SIGKDD
Explorations, 5(2): pages 173-178, December 2003. This entry
made 2nd place in the Open Task of the 2003 KDD Cup.
[pdf]
- S. Lin and H. Chalupsky. Unsupervised
Link Discovery in Multi-relational Data via Rarity Analysis. In
Proceedings of the Third IEEE International Conference on Data Mining (ICDM
'03). 2003.
[pdf]
Documentation
The KOJAK Group Finder manual is available in the following formats:
For other formats such as the Emacs info format look in the
sources/KOJAK/doc
directory of the KOJAK release.
Questions and Comments
KOJAK development and maintenance has concluded and the software is not
available for download anymore. For any remaining questions or comments
please send mail to hans @
isi . edu.
Last modified:
June 15, 2018