The optimized compilation of Constraint Logic Programming (CLP) languages can give rise to impressive performance improvements, even more impressive than the ones obtainable for the compilation of Prolog. On the other hand, the global analysis techniques needed to derive the necessary information can be significantly more complicated than in the case of Prolog. The original contribution of the present work is the integration of approximate inference techniques, well known in the field of artificial intelligence (AI), with an appropriate framework for the definition of nonstandard semantics of CLP. This integration turns out to be particularly appropriate for the considered case of the abstract interpretation of CLP programs over numeric domains. One notable advantage of this approach is that it allows to close the often existing gap between the formalization of data-flow analysis in terms of abstract interpretation and the possibility of efficient implementations. Towards this aim we i...

A general problem in qualitative physics is determining the consequences of assumptions about the behavior of a system. If the space of behaviors is represented by an envisionment, many such consequences can be represented by pruning states from the envisionment. This paper provides a formal logic of occurrence which justifies the algorithms involved and provides a language for relating specific histories to envisionments. The concepts and axioms are general enough to be applicable to any system of qualitative physics. We further propose the concept of transverse quantities as a general solution to qualitative versions of Zeno's paradox. The utility of these ideas is illustrated by a rational reconstruction of the pruning algorithms used in FROB, a working AI program. December, 1 Introduction A goal of qualitative physics is to predict the behavior of physical systems. One technique, envisioning, generates all possible behaviors of a system, relative to a particular set of backgroun...

Constraint logic programming (CLP) is a generalization of the pure logic programming paradigm, having similar model-theoretic, fixpoint and operational semantics [9]. Since the basic operational step in program execution is a test for solvability of constraints in a given algebraic structure, CLP has in addition an algebraic semantics. CLP is then a general paradigm which may be instantiated on various semantic domains, thus achieving a good expressive power. One relevant feature is the distinction between testing for solvability and computing a solution of a given constraint formula. In the logic programming case, this corresponds to the unification process, which tests for solvability by computing a solution (a set of equations in solved form or most general unifier ). In CLP, the computation of a solution of a constraint is left to a constraint solver, which does not affect the semantic definition of the language. This allows different computational domains, e.g. real arithmetic, to...

This paper describes how to use an assumption-based truth maintenance system (ATMS) to significantly speed up qualitative reasoning. Specifically, we introduce three organizing abstractions for ATMS-based problem solvers (many-worlds databases, justify/interpret cycles, and closed-world tables). We illustrate their utility by describing the Qualitative Process Engine (qPE), an implementation of Qualitative Process theory that is roughly 95 times-faster and signficantly simpler than the previous implementation. After analyzing gPE's performance, we draw some general conclusions about the advantages and disadvantages of assumption-based truth maintenance systems. Program:ENGINEERING

We introduce new diagrammatic notations for probabilistic independence networks (including Bayes nets and graphical models). These notations include new node and link types that allow for natural representation of a wide range of probabilistic data models including complex hierarchical models. The diagrammatic notations also support models defined on variable numbers of complex objects and relationships. Node types include random variable nodes, index nodes, constraint nodes, and an object supernode. Link types include conditional dependency, indexing and index limitation, variable value limitation, and gating a dependency between nodes or objects by an arbitrary graph. Examples are shown for clustering problems, information retrieval, unknown graph structures in biological regulation, and other scientific domains. The diagrams may be taken as a shorthand notation for a more detailed syntactic representation by an algebraic expression for factored probability distributions, which in turn may be specified by stochastic parameterized grammar or graph grammar models. We illustrate these ideas with previously described applications and potential new ones. 1. Extending graph notation for dependency networks

The optimized compilation of Constraint Logic Programming (CLP) languages can give rise to impressive performance improvements, even more impressive than the ones obtainable for the compilation of Prolog. On the other hand, the global analysis techniques needed to derive the necessary information can be significantly more complicated than in the case of Prolog. The original contribution of the present work is the integration of approximate inference techniques, well known in the field of artificial intelligence (AI), with an appropriate framework for the definition of nonstandard semantics of CLP. This integration turns out to be particularly appropriate for the considered case of the abstract interpretation of CLP programs over numeric domains. One notable advantage of this approach is that it allows to close the often existing gap between the formalization of data-flow analysis in terms of abstract interpretation and the possibility of efficient implementations. Towards this aim we i...