In recent years several high quality Sturm-Liouville software packages have been written (e.g., SLEDGE, SLEIGN, the NAG routine SL02FM). All excel at providing eigenfunction estimates at a nite set of points but only the NAG routine provides interpolation capability, and this by a fairly crude method (especially when compared with the sophistication of its eigenvalue solver). We describe an acc...

Austin Tate, Je Dalton and John Levine Arti cial Intelligence Applications Institute The University of Edinburgh 80 South Bridge, Edinburgh EH1 1HN, UK fa.tate, j.dalton, [email protected] http://www.aiai.ed.ac.uk/~oplan Abstract In this paper, we present a Web-based demonstration of a Course of Action (COA) comparison matrix being used as an interface to an O-Plan plan server to explore multi...

Training neural networks involves solving large-scale non-convex optimization problems. This task has long been believed to be extremely difficult, with fear of local minima and other obstacles motivating a variety of schemes to improve optimization, such as unsupervised pretraining. However, modern neural networks are able to achieve negligible training error on complex tasks, using only direc...

In this paper, we present a Web-based demonstration of a Course of Action (COA) comparison matrix being used as an interface to an O-Plan plan server to explore multiple qualitatively different plan options. The scenario used for this demonstration is concerned with crisis operations on the island of Pacifica. The COA comparison matrix allows the user to explore and evaluate several different p...

The objective in this paper is to present a framework for a system that describes objects in a qualitative fashion. A subset of spatial prepositions is chosen and an appropriate quantiication is applied to each of them that capture their inherent qualitative properties. The quantiications use such object attributes as area, centers, and elonga-tion properties. The familiar zeroth, rst, and seco...

Current methods for generating qualitatively different plans are either based on simple randomization of planning decisions and so cannot guarantee meaningful differences among generated plans, or require extensive user involvement to drive the system into different sections of the overall plan space. This paper presents a cost-effective method for automatically generating qualitatively differe...