Cordell Green

Computer-based consultations in clinical therapeutics: Explanation and rule acquisition capabilities of the MYCIN system☆

Abstract This report describes progress in the development of an interactive computer program, termed MYCIN, that uses the clinical decision criteria of experts to advise physicans who request advice regarding selection of appropriate antimicrobial therapy for hospital patients with bacterial infections. Since patients with infectious diseases often require therapy before complete information about the organism becomes available, infectious disease experts have identified clinical and historical criteria that aid in the early selection of antimicrobial therapy. MYCIN gives advice in this area by means of three subprograms: (1) A Consultation System that uses information provided by the physician, together with its own knowledge base, to choose an appropriate drug or combination of drugs; (2) An Explanation System that understands simple English questions and answers them in order to justify its decisions or instruct the user; and (3) A Rule Acquisition System that acquires decision criteria during interactions with an expert and codes them for use during future consultation sessions. A variety of human engineering capabilities have been included to heighten the programs acceptability to the physicians who will use it. Early experience indicates that a sample knowledge base of 200 decision criteria can be used by MYCIN to give appropriate advice for many patients with bacteremia. The system will be made available for evaluation in the clinical setting after its reliability has been shown to approach that of infectious disease experts.

An empirical study of list structure in Lisp

Static measurements of the list structure of five large Lisp programs are reported and analyzed in this paper. These measurements reveal substantial regularity, or predictability, among pointers to atoms and especially among pointers to lists. Pointers to atoms are found to obey, roughly, Zipfs law, which governs word frequencies in natural languages; pointers to lists usually point to a location physically nearby in memory. The use of such regularities in the space-efficient representation of list structure is discussed. Linearization of lists, whereby successive cdrs (or cars) are placed in consecutive memory locations whenever possible, greatly strengthens the observed regularity of list structure. It is shown that under some reasonable assumptions, the entropy or information content of a car-cdr pair in the programs measured is about 10 to 15 bits before linearization, and about 7 to 12 bits after.

The use of theorem-proving techniques in question-answering systems

For the purpose of this paper, a question-answering system is a computer program that has at least the following three characteristics: (1) The ability to accept statements of fact and store them in its memory (2) The ability to search stored information efficiently and to recognize items that are relevant to a particular query (3) The ability to respond appropriately to a question by identifying and presenting the answer if it is present in memory, and by deducing a reasonable logical response from relevant knowledge if the complete answer is not explictly available.

Inferring LISP programs from examples

A program is described which infers certain recursive LISP programs from single example input-output pairs Synthesized programs may recur in more than one argument, and may involve the synthesis of auxiliary functions An actual user session with the program, called EXAMPLE, is presented, and the operation of the program and its important heuristics are outlined.

Knowledge-based programming self applied

ABSTRACT A knowledge-based programming system can utilize a very-high-level self description to rewrite and improve itself. This paper presents a specification, in the very-high-level language V, of the rule compiler component of the CHI knowledge-based programming system. From this specification of part of itself, CHI produces an efficient program satisfying the specification. This represents a modest application of a machine intelligence system to a real programming problem, namely improving one of the programming environments tools — the rule compiler. The high-level description and the use of a programming knowledge base provide potential for system performance to improve with added knowledge.

An overview of automated reasoning and related fields

This article provides an overview of automated reasoning and of the various fields for which it is relevant. It takes the form of a collection of articles, each covering some field and each written by an expert in that field. A field is introduced, its elements reviewed, the current state of the art given, the basic problems discussed, and the various goals listed. Although individually the goals of each field present a wide spectrum, collectively the fields share the interest of automating the process known as reasoning.

A Constructive Approach to Correctness, Exemplified by a Generator for Certified Java Card Applets

We present a constructive approach to correctness and exemplify it by describing a generator for certified Java Card applets that we are building. A proof of full functional correctness is generated, along with the code, from the specification; the proof can be independently checked by a simple proof checker, so that the larger and more complex generator needs not be trusted. We argue that such an approach is a valuable alternative to post-hoc verification, in addressing the Program Verifier Grand Challenge.

Knowledge-based debugging: session summary

Robert Balzer began the session by presenting an overview of where debugging fits into knowledge-based programming systems. He distinguished between differences in paradigm (the traditional software engineering approach in which the source code is the first formal representation vs. the operational specification approach in which the specification is both formal and executable and from which the implementation is formally derived) and differences in technology (traditional debuggers vs. knowledge-based tools), pointing out that it is possible to use traditional debugging techniques with the operational specification paradigm as well as use knowledge-based tools with the traditional software engineering paradigm. Balzer presented the operational specification paradigm for software development. The basic idea is that an implementation is derived from a specification through transformations chosen by the programmer and applied by the system. Assuming each transformation is correct, the resulting implementation will accurately reflect the intentions of the specification.

Automatic synthesis—programs and plans

This session focuses on the continuing effort to develop automatic techniques for transforming a specification of “what” is to be done into a specification of “how” the “what” is to be done. One presentation deals with programming tasks, where the results of the transformation are executable computer programs. The other with tasks being carried out by people or organizations, where the results of the transformation are plans and schedules.