Roger T. Hartley

The MDR algorithm and its application to the generation of explanations for novel events

Abstract This paper presents an algorithm for reasoning about novel events. Termed Model Generative Reasoning (MGR), we replace deductive reasoning with an abductive procedure based on the generation of alternative, intensional domain descriptions (models) to cover problem assumptions, which are then evaluated against domain facts as alternative explanations for queried events. The algorithm is principally illustrated using a problem from process control.

A uniform representation for time and space and their mutual constraints

Abstract Much recent work in reasoning systems has concentrated on the role of time in planning, action modeling, and tasks in domains where time is important. On the other hand, there are systems that concentrate on spatial reasoning, especially where manipulation or managing of the environment is important, as in robot route planning. The integration of the two themes is a goal which, if possible, would allow the interaction of space and time to be explored. A problem-solving system would then be able to reason about the times at which actions might occur, in the light of spatial constraints, or vice versa, to reason about the places in which actions take place, and the temporal constraints involved. This paper shows a way to integrate the representation of both time and space in a framework that allows uniform reasoning across both dimensions. An ontology for objects, events, states, and processes is provided using conceptual graphs for representation, and a syntactic extension to Sowas conceptual graph formalism (see Sowas article, this volume) is presented to support the effort.

REASONING WITH GRAPH OPERATIONS

Abstract Problem solving is an analog to scientific method, wherein abduction and deduction operate in a cyclic fashion to generate and refine a series of hypotheses that purport to explain the observed data. Model generative reasoning implements this cycle through a family of operations on representations based on conceptual graphs. Specialize , the operator that implements abduction, generates alternative hypotheses. Fragment removes potential incoherences from hypotheses, while preserving coherence with the observations. This is seen as a form of deduction with the aim of allowing more hypotheses to be generated in the next cycle.

Temporal, spatial, and constraint handling in the Conceptual Programming environment, CP

The goal of logic synthesis is to obtain high-quality designs from specifications. Current approaches to logic synthesis often trade off design quality for technology independence. In this paper, we present a model of logic synthesis that uses technology-specific design rules and extends rule-based search to functional decomposition and technology mapping. While this model improves design quality by taking advantage of the target technology, it is not robust to technology changes. To improve robustness, we augment the model with two learning components: one for acquiring rules that make use of physical cells in a technology library, and another for acquiring rules that make use of appropriate design styles. These components are related to work in the learning of macro-operators and explanation-based learning.

Quality of OCR for degraded text images

Commercial OCR packages work best with high-quality scanned images. They often produce poor results when the image is degraded, either because the original itself was poor quality, or because of excessive photocopying. The ability to predict the word failure rate of OCR from a statistical analysis of the image can help in making decisions in the trade-off between the success rate of OCR and the cost of human correction of errors. This paper describes an investigation of OCR of degraded text images using a standard OCR engine (Adobe Capture). The documents were selected from those in the archive at Los Alamos National Laboratory. By introducing noise in a controlled manner into perfect documents, we show how the quality of OCR can be predicted from the nature of the noise. The preliminary results show that a simple noise model can give good prediction of the number of OCR errors.

Visual representation of procedural knowledge

Traditionally, knowledge representation (KR) languages have declarative semantics based on classical logic and have a concrete syntax that is textual. Conversely, programming languages (PL) have mainly procedural semantics and are represented in visual terms. Moreover, PLs can only represent procedures at a very low level. Our conceptual programming language (CPL) bridges this gap, while still retaining visual appeal. It has both the declarative semantics of a KR language and the procedural semantics of a PL. CPL is a visual language for expressing procedural knowledge explicitly as programs.

A Comparison of Different Conceptual Structures Projection Algorithms

Knowledge representation (KR) is used to store and retrieve meaningful data. This data is saved using dynamic data structures that are suitable for the style of KR being implemented. The KR allows the system to manipulate the knowledge in the data by using reasoning operations. The data structure, together with the contents of the transformed knowledge, is known as the knowledge base (KB). An algorithm and the associated data structures make up the reasoning operation, and the performance of this operation is dependent on the KB. In this paper, the basic reasoning operation for a query-answer system, projection, is explored using different theoretical algorithms. Within this discussion, the associated algorithms will be using different KBs for their Conceptual Graph (CG) knowledge representation. The basic projection algorithm defined using the CG representation is looking for a graph morphism of a query graph onto a graph from the KB. The overall running time for the projection operation is known to be a NP class problem; however, by modifying the algorithm, taking into account the associated KB, the actual time needed for discovering and creating the projection/s can be improved. In fact, a new projection algorithm will be defined that, given a typical query onto a carefully defined KB, presents a running time for the actual projection that only grows with the number of projections present.

Belief ascription and model generative reasoning: joining two paradigms to a robust parser of messages

This paper discusses the extension of ViewGen, a program for belief ascription, to the area of intensional object identification with applications to battle environments, and its combination in a overall system with MGR, a Model-Generative Reasoning system, and PREMO a semantics-based parser for robust parsing of noisy message data.ViewGen represents the beliefs of agents as explicit, partitioned proposition-sets known as environments. Environments are convenient, even essential, for addressing important pragmatic issues of reasoning. The paper concentrates on showing that the transfer of information in intensional object identification and belief ascription itself can both be seen as different manifestations of a single environment-amalgamation process. The entities we shall be concerned with will be ones, for example, the system itself believes to be separate entities while it is computing the beliefs and reasoning of a hostile agent that believes them to be the same entity (e.g. we believe enemy radar shows two of our ships to be the same ship, or vice-versa. The KAL disaster should bring the right kind of scenario to mind). The representational issue we address is how to represent that fictional single entity in the belief space of the other agent, and what content it should have given that it is an amalgamation of two real entities.A major feature of the paper is our work on embedding within the ViewGen belief-and-point-of-view system the knowledge representation system of our MGR reasoner, and then bringing together the multiple viewpoints offered by ViewGen with the multiple representations of MGR. The fusing of these techniques, we believe, offers a very strong system for extracting message gists from texts and reasoning about them.

e-MGR: an architecture for symbolic plasticity

Abstract The e-MGR architecture was designed for symbolic problem solving in task environments where data are noisy and problems are ill-defined. e-MGR is an operator-based system which integrates problem-solving ideas from symbolic artificial intelligence (AI) and adaptive systems research.

An object-based methodology for knowledge representation in SGML

An object-based methodology for knowledge representation and its Standard Generalized Markup Language (SGML) implementation is presented. The methodology includes class, perspective, domain and event constructs for representing knowledge within an object paradigm. The perspective construct allows for the representation of knowledge from multiple and varying viewpoints. The event construct allows actual use of knowledge to be represented. The SGML implementation of the methodology facilitates usability, structured, yet flexible knowledge design, and sharing and re-use of knowledge class libraries.