IBM LanguageWare Miner for Multidimensional Socio-Semantic Networks
A unified API that helps in creating solutions for social computing, semantic search, and activity computing.
Date Posted: December 13, 2007
|
|
 |
|
|
Update: September 11, 2008
New version includes implementation of new, more generic, spreading activation functions and a text focus determiner algorithm based on those functions; updated versions of external dependencies.
What is IBM LanguageWare Miner for Multidimensional Socio-Semantic Networks?
The proliferation of Web 2.0 has brought about all kinds of digital artifacts: documents, people, concepts, vocabulary, tasks, activities, and more. However, there is no easy way to combine these assets in a way that provides meaningful user experience. IBM® LanguageWare Miner for Multidimensional Socio-Semantic Networks provides a unified API that helps in creating solutions for these types of multidimensional networks (people, documents, tasks, etc.) and provides an integrated platform for combining social computing, semantic processing, and activity-centered computing for enhanced user experience. Key features include the following:
- a unified and flexible API that allows applications to quickly adapt to changing needs and provides a high degree of personalization
- customization for clustering and fuzzy inference of information on the network through specialized techniques (activation spread on networks)
- semantic processing for both ontology and instance-based information. This might be used for automatic generation of meta-tags for use in semantic Web applications or in socio-semantic desktop applications such as NEPOMUK (Networked Environment for Personal Ontology-based Management of Unified Knowledge).
How does it work?
This technology contains an Eclipse application that provides a graphical interface to the basic features of the libraries. This distribution also comes with a small data set and some sample texts. The data set describes information about a fictional company, including management structures, products, locations, company terminology, etc. Using this application, you can build data into lexico-semantic dictionaries, explore the underlying multidimensional network, perform analysis on text, and perform some basic investigation into the additional information that can be inferred from the underlying network.
|
|
|
| | About the technology author(s):
All three engineers are working on a three-year integrated E.U. project called NEPOMUK (Networked Environment for Personal Ontology-based Management of Unified Knowledge), in which the IBM team is developing tools for semantic analysis of text. These tools allow for semi-automatic production of semantic meta-information needed for content management.
 Alexander Troussov, Ph.D., is chief scientist at the IBM Dublin Centre for Advanced Studies and chief scientist of the IBM LanguageWare group. He is a joint author of more than 30 peer-reviewed research publications and has five patents pending. As the LanguageWare group architect, Dr. Troussov developed new methods for the optimisation of finite state processing for morphological analysis, worked on computational models for compounding languages, and worked on other problems of subsententional text processing. As CAS chief scientist, he leads IBM's participation in the NEPOMUK project.
|
 John Judge, Ph.D., is a researcher at the IBM Dublin Software Lab, working as part of the LanguageWare team. His current duties include research and development for the NEPOMUK project. Dr. Judge is a joint author of six peer-reviewed research publications and has one patent pending.
|
 Mikhail Sogrin is a researcher and software engineer at the LanguageWare team in IBM Dublin Software Lab. He participated in, and took fifteenth place in, the World Finals of 2000 ACM International Collegiate Programming Contest. His current duties include research and development for the NEPOMUK project. Mr. Sogrin is a joint author of five peer-reviewed research publications and has one patent pending.
|
IBM is a trademark of IBM Corporation in the United States, other countries, or both.
Other company, product, or service names may be trademarks or service marks of others.
|
|
|
| |
