CBKB : Constraint Based Knowledge Brokers
The CBKB Prototype
CBKB is an implementation of the Knowledge Brokers model for intelligent information retrieval from the World Wide Web. It involves three kinds of agents:
- Users who input queries and read answers through a GUI;
- Wrappers capable to interrogate different Information Sources, which can provide answers to elementary queries (essentially various public bibliography databases available on the Web, but also an opera database in Budapest);
- Brokers, which can manage complex queries and mediate between the user and the different Wrappers.
The core of the system is given by the brokers, which provide various important services such as intelligent caching, filtering and knowledge combination. The current implementation uses the ForumTalk prototype. Brokers are implemented as ForumTalk coordinators programmed in LO. The GUI (programmed in Java) and the Wrappers (programmed in Python), are external processes accessed through ForumTalk plugs.
The Constraint Based Knowledge Brokers prototype includes language processing support such as stemming, translation, and language guessing. These extensions are developed and integrated in cooperation with the CA-group at XRCE. Xerox has recently released askOnce, a commercial product based on the Constraint Based Knowledge Brokers.
More like this
- PyStruct Extension for Typed CRF Graphs
- 21st International Conference on Knowledge-Based and Intelligent Information & Engineering Systems; Marseille, France
- Transkribus Python Toolkit
- 1st International Workshop on Open Services and Tools for Document Analysis (ICDAR-OST) at 14th IAPR ICDAR, Kyoto, Japan
- What to do when the access to the source data is constraint?