Real-time domains present a new and challenging environment for the application of knowledge-based problem-solving techniques. However, a substantial amount of research is still needed to solve many difficult problems before real-time expert systems can enhance current monitoring and control systems. In this article, we examine how the real-time problem domain is significantly different from those domains which have traditionally been solved by expert systems. We conduct a survey on the current state of the art in applying knowledge-based systems to real-time problems and describe the key issues that are pertinent in a real-time domain. The survey is divided into three areas: applications, tools, and theoretic issues. From the results of the survey, we identify a set of real-time research issues that have yet to be solved and point out limitations of current tools for real-time problems. Finally, we propose a set of requirements that a real-time knowledge-based system must satisfy.

This is Part 1 of a two-part article. Part 2 covers tort liability and computers as expert witnesses. It will appear in the Summer 1988 issue of AI Magazine. Technological developments that remove ever-increasing numbers of cognitive tasks from human control will alter the assumptions on which current legal rules are based. These rules will have a growing impact on AI researchers and entrepreneurs as their work reaches a growing audience of beneficiaries. In order to accommodate the needs of practitioners and their recipients, courts and lawmakers will be forced to reevaluate principles whose foundations were developed well before the implications of advanced technology could have been predicted. This article attempts to identify areas of law in which the need for accommodation will be greatest and provide some insight into the process and the direction of change.

This book is a collection of research papers covering a range of topics in expert systems and knowledge engineering.

We propose an approach for meeting real-time constraints in AI systems that views (1) time as a resource that should be considered when making control decisions, (2) plans as ways of expressing control decisions, and (3) approximate processing as a way of satisfying time constraints that cannot be achieved through normal processing. In this approach, a real-time problem solver estimates the time required to generate solutions and their quality. This estimate permits the system to anticipate whether the current objectives will be met in time. The system can then take corrective actions and form lower-quality solutions within the time constraints. These actions can involve modifying existing plans or forming radically different plans that utilize only rough data characteristics and approximate knowledge to achieve a desired speedup. A decision about how to change processing should be situation dependent, based on the current state of processing and the domain-dependent solution criteria. We present preliminary experiments that show how approximate processing helps a vehicle-monitoring problem solver meet deadlines and outline a framework for flexibly meeting real-time constraints.

In COLING-88, pp. 71–76

In Non- Monotonic Reasoning: 2nd International Workshop Proceedings, pp. 74–99.

Proceedings of the National Conference on Artificial Intelligence, 245-250

In Proceedings of the Fifth International Conference on Machine Learning, pages 54-64, Ann Arbor, Michigan.

Proc. AAAI-88, Vol. 2, pp. 607–611

In Mora, T. (Ed.), Applied Algebra, Algebraic Algorithms and Error-Correcting Codes, pp. 99–110. Springer-Verlag.