Browse Prior Art Database

An Experiment in Distributed Interpretation

IP.com Disclosure Number: IPCOM000128658D
Original Publication Date: 1979-Dec-31
Included in the Prior Art Database: 2005-Sep-16

Publishing Venue

Software Patent Institute

Related People

Victor R. Lesser: AUTHOR [+4]

Abstract

The range of application areas to which distributed processing has been applied effectively is limited. In order to extend this range, new models for organizing distributed systems must be developed.

This text was extracted from a PDF file.
This is the abbreviated version, containing approximately 3% of the total text.

Page 1 of 32

THIS DOCUMENT IS AN APPROXIMATE REPRESENTATION OF THE ORIGINAL.

An Experiment in Distributed Interpretation

Victor R. Lesser University of Massachusetts Lee D. Erman USC Information Sciences Institute

ARPA ORDER NO. 2223 ISIIRR-79 -7G May 197> UNIVERSITY OP SOUTHERN CALIFORNIA

1`NFORMATION SCIENCES INSTITUTE

46 76 flilnriralty Way/MarinadelReylCalifornia 90291 ( 213 ) 822-1 S I 1 This resbarch`was supported in,part.by Defense Advanced Research~Project# .4~ricy,Fontract F44620-73-C-0074 to Caiiieg -Melton c:. University, National Science Foundation grant MC$78-04212' to University of, MassachNSe~tts, and DARPA contract DAHC 1572-0-

0308 to University of Southern California Information Sciences Institute. V'kews=and conclusions contained in this document are the authors' and.shaild not be interpreted. as representIni-the offi~cia.1 opi,nion.orpolicy of'DARPA, the U.S. Government, or.any other person or agency connected with them.' > r,

ABSTRACT

The range of application areas to which distributed processing has been applied effectively is limited. In order to extend this range, new models for organizing distributed systems must be developed.

We present a new model, in which the distributed system is able to function effectively even though processing nodes have inconsistent and incomplete views of the data bases necessary for their computations. This model differs from conventional approaches in its emphasis on dealing with distribution-caused uncertainty and errors in control, data, and algorithm as an integral part of the network problem-solving process.

We show how this new model can be applied to the problem of distributed interpretation. Experimental results with an actual interpretation system support these ideas.

This report is being published simultaneously by USC/Information Sciences Institute (as RR-79- 76) and Carnegie-Mellon University (as CMU-CS-79-120).

CONTENTS

1. INTRODUCTION I

2. OVERVIEW OF HEARSAY-II: A System that Handles Uncertainty 3 2.1. The Model 3 2.2. The Architecture 4

3. ISSUES IN DISTRIBUTING HEARSAY-II 7

4. A NETWORK OF HEARSAY-II SYSTEMS 10

University of Southern California Page 1 Dec 31, 1979

Page 2 of 32

An Experiment in Distributed Interpretation

4.1. Intranode Considerations -- Selection and Focusing of KSs 11

4.2. Network Configurations 14

4.3. lnternode Communication -- Mechanism 16

4.4. Internode Communication -- Policies 16

4.4.1. The Basic Policy 20

4.4.2. Variants 20

5. THE EXPERIMENT 23 5.1. Simulating a Network 23 5.2. Selection of KSs and Areas-of- Interest 23 5.3. Communication Strategy ?6 5.4. Knowledge-Source Changes ?7

6. RESULTS 29 6.1. Network versus Centralized 30 6.2. Transmission Policies 37 6.3. Communication with Errors 37

7. CONCLUSIONS 39

ACKNOWLEDGMENTS 45

REFERENCES 45

1. INTRODUCTION'

An interpretation system accepts a set of signals from some environment and produces higher- level descriptions of objects and events in the environment. Speech and image understanding, medica...