Deductive Spreadsheet System - XcelLog
XcelLog is a user-centered deductive spreadsheet (DSS) system for empowering end-users to develop intelligent applications using the spreadsheet metaphor. It is implemented as an add-in to the highly popular Microsoft® Excel® spreadsheet system. Thus, XcelLog users can continue taking advantage of the benefits of traditional Excel® but with the added power of deduction provided by XcelLog.
Ever tried to use Excel® to find all employees of all branches in a complex hierarchical organization? A shortest path between distant cities given distances between adjacent cities?
The XcelLog computing model for deductive spreadsheets supports:
- Table cells that contain multiple values
- Statements that a cell must contain all the elements of another cell
- Recursive definitions through cyclic cell references, i.e., when a cell expression references another cell that references it.
Deductive Spreadsheet specifications are automatically translated to logic programs in a straightforward way, thus providing direct access to the deductive capabilities of the XSB Logic Programming System.
Untrained users can immediately begin to use the simple deductive capabilities of XcelLog and gradually learn to use its full power. XcelLog enables end-users to focus on the problem at hand and employ deduction-based problem solving methods using simple and intuitive gestures via the graphical spreadsheet user interface. XcelLog puts the power of deduction into the hands of end-users, but without the need to program and without the constraints of canned application packages.
Focused Crawlers
XSB is in the process of developing of a new generation of XRover® intelligent Web agents - Focused Crawlers. Focused Crawlers automatically determine relevant information on Web pages using ontologies that encode knowledge of domains of interest. This knowledge is applicable across multiple Web sites and may be reused for consequent data classification and attribute extraction. Its use eliminates the need for building site-specific navigation maps, and makes Web agents highly resilient to structural site changes.
Focused Crawler technology is based on automated techniques for partitioning the content of a Web page into segments consisting of semantically related items. Focused Crawlers traverse only the links within relevant partitions to harvest the data of interest to the user. Semantics-driven Web agents such as Focused Crawlers hold great promise as highly scaleable, low maintenance data acquisition technology.