James G. Schmolze

The KL-ONE family

Abstract The knowledge representation system KL-ONE has been one of the most influential and imitated knowledge representation systems in the Artificial Intelligence community. Begun at Bolt Beranek and Newman in 1978, KL-ONE pioneered the development of taxonomic representations that can automatically classify and assimilate new concepts based on a criterion of terminological subsumption. This theme generated considerab interest in both the formal community and a large community of potential users. The KL-ONE community has since expanded to include many systems at many institutions and in many different countries. This paper introduces the KL-ONE family and discusses some of the main themes explored by KL-ONE and its successors. We give an overview of current research, describe some of the systems that have been developed, and outline some future research directions.

Guaranteeing serializable results in synchronous parallel production systems

Abstract To speed up production systems, researchers have studied how to execute many rules simultaneously. Unfortunately, such systems can yield results that are impossible for a serial system to produce, leading to erroneous behaviors. We present a formal solution to the problem of guaranteeing serializable behavior in synchronous parallel production systems that execute many rules simultaneously. We also present a variety of algorithms that implement this solution. Our framework builds upon the work of Ishida and Stolfo [14] and improves upon their solution. The practical advantages of these strategies are demonstrated with measurements from an initial implementation.

Planning with POMDPs using a compact, logic-based representation

Partially observable Markov decision processes (POMDPs) provide a general framework for AI planning, but they lack the structure for representing real world planning problems in a convenient and efficient way. Representations built on logic allow for problems to be specified in a compact and transparent manner. Moreover, decision making algorithms can assume and exploit structure found in the state space, actions, observations, and success criteria, and can solve with relative efficiency problems with large state spaces. In recent years researchers have sought to combine the benefits of logic with the expressiveness of POMDPs. In this paper, we show how to build upon and extend the results in this fusing of logic and decision theory. In particular, we present a compact representation of POMDPs and a method to update beliefs after actions and observations. The key contribution is our compact representation of belief states and of the operations used to update them. We then use heuristic search to find optimal plans that maximize expected total reward given an initial belief state

Detecting redundancy among production rules using term rewrite semantics

We present a general method for detecting wide classes of redundant production rules (PRs) based on the term rewrite semantics. We present the semantic account, define rule execution over both ground memories and memory schemas, and define redundancy for the PRs. From those definitions, an algorithm is developed that detects wide classes of redundant rules, and which improves upon the previously published methods.

Rewrite Semantics for Production Rule Systems: Theory and Applications

Production rules are a fundamental computational paradigm in artificial intelligence, perhaps being best known as the basis for expert systems. However, to this point there has been little formal study of their properties as a method of deduction. In this paper we initiate such a study by presenting a rewrite semantics for production rule systems. Such a formalization is both interesting per se as a paradigm for deduction and also useful in providing a formal framework for analyzing properties of production rule systems. We show how to represent production rules as rewrite rules operating on collections of atoms, thereby allowing us to import techniques from equational reasoning (confluence checking and completion, critical pair criteria, termination orderings, etc.). An interesting feature of this representation is the use of symbolic constraints to represent the negation-as-failure interpretation of negative conditions in production rules. Practical applications of this theory provide for the development of a comprehensive environment for verifying and validating the correctness of production rule systems, as well as the development of convergent production rule systems for special applications such as parallel evaluation and real-time control.

Using Confluence to Control Parallel Production Systems

Publisher Summary This chapter describes how to test a class of production rule sets for confluence. The chapter focuses on future research to produce a tool that will in converting a non-confluent rule set into a confluent one. Production systems can be thought of as term rewriting systems if the focus is changed from rewriting terms to rewriting sets of literals. The chapter reviews term rewriting systems, including the notions of confluence and convergence. The chapter also explains the procedure to determine confluence in production rule sets. The chapter also explains how confluence testing can be applied to certain types of rule systems that have metacontrol schemes. Confluence testing has the potential of helping to verify rule sets. If there are multiple normal forms for each given starting state, then the task of verification is likely to be simplified by analyzing an equivalent, completed rule set that yields only one normal form per starting state.

Efficient Open World Reasoning for Planning

We consider the problem of reasoning and planning with incomplete knowledge and deterministic actions. We introduce a knowledge representation scheme called PSIPLAN that can effectively represent incompleteness of an agents knowledge while allowing for sound, complete and tractable entailment in domains where the set of all objects is either unknown or infinite. We present a procedure for state update resulting from taking an action in PSIPLAN that is correct, complete and has only polynomial complexity. State update is performed without considering the set of all possible worlds corresponding to the knowledge state. As a result, planning with PSIPLAN is done without direct manipulation of possible worlds. PSIPLAN representation underlies the PSIPOP planning algorithm that handles quantified goals with or without exceptions that no other domain independent planner has been shown to achieve. PSIPLAN has been implemented in Common Lisp and used in an application on planning in a collaborative interface.

Guaranteeing Serializability in Parallel Production Systems

To speed up production systems, researchers have studied how to execute many rules simultaneously. Unfortunately, such systems can yield results that are impossible for a serial system to produce, leading to erroneous behaviors. We present a formal solution to the problem of guaranteeing serializable behavior in parallel production systems that execute many rules simultaneously. We also present a variety of algorithms that implement this solution. Our framework builds upon the work of Ishida and Stolfo [12] and improves upon their solution. The practical advantages and limitations of these strategies is demonstrated with measurements from an implementation.

Practical reasoning about knowledge states for open world planning with sensing

We present a representation for reasoning and planning with an incomplete state description (open-world) called PSIPLAN-S. The presented formalism has several properties critical for application domains with a large degree of incompleteness in the state description, particularly, in domains with a large or unknown set of all objects. The formalism offers (1) considerably expressive state and goal description language, that includes limited universal quantification, (2) representation of sensing actions and knowledge goals, (3) a correct and complete state update procedure, and (4) complete reasoning within a substantial subset of the language. The approach is illustrated by examples from a working system.