Polaris IP, LLC v. Google Inc. et al

Filing 273

Additional Attachments to Main Document: 272 MOTION Defendants' Joint Response Brief on Claim Construction.. (Attachments: # 1 Exhibit 1, # 2 Exhibit 2, # 3 Exhibit 3, # 4 Exhibit 4, # 5 Exhibit 5, part 1, # 6 Exhibit 5, part 2, # 7 Exhibit 5, part 3, # 8 Exhibit 5, part 4, # 9 Exhibit 6, # 10 Exhibit 7, # 11 Exhibit 8, part 1, # 12 Errata 8, part 2, # 13 Exhibit 9, # 14 Exhibit 10)(Perlson, David)

Polaris IP, LLC v. Google Inc. et al

Doc. 273 Att. 9

Dockets.Justia.com

ltac

A volume in the.. EEEE PRESS Selected Reprint`,Series prepared 'iun-de-r the sponsorsha q of the IEEE Computer Society.;

IEEE PRESS 1988 Editorial Board
Leonard Shaw ,,Editor in Chief

Deter Dorato , Editor, Selected Reprint Series
F. S. Barnes J. E. Brittain J. T. Cain Aileen Cavanagh

J. F. Hayes.
W. K. Jenkins

G. N. Saridis

A. E. Joel, Jr.
Shlomo Karni Albert Macovsld

C. B. Silio, Jr. M. I. Skolnik
G. S. Smith

S. H. Charap D. G. Childers H. W. Colborn
R. F. Cotellessa L. J. Greenstein

P. W. Smith
M. A. Soderstrand M. E. Van Valkenburg John Zaborsky

R_ G. Meyer
Seinosuke Narita W. E. Proebster

J. D. Ryder
W. R. Crone, Managing Editor Hans P. Leander, Technical Editor Laura J. Kelly, Administrative Editor Randi E. Scholnick, Associate Editor

Copyright © 1988 by THE INSTITUTE OF ELECTRICAL AND ELECTRONICS ENGINEERS, INC. 345 East 47th Street, New York, NY 10017-2394
All rights reserved.

PRINTED IN THE UNITED STATES OF AMERICA IEEE Order Number: PC0231-1
Library of Congress Cataloging -in-Publication Data Microcomputer-based expert systems. (IEEE Press selected reprint series) Includes indexes. L Expert Systems (Computer science) 2. Microcomputers - Programming. 1. Gupta , Amar. II. Prasad , Bandreddi E. (Bandreddi Eswara) III. Series. QA76.76.E95M53 1988 87-31059 006 .3'3

ISBN 0-87942-221-1

iv

Part I Introduction
HE area of microcomputer-based expert systems is one of T the fastest evolving components of the computer industry. Currently, the growth rate for this area , in terms of percentage increase per year, far exceeds the rate for microcomputers in general , as well as the rate of increase for the overall artificial intelligence industry . At the beginning of this decade , microcomputer-based expert systems were virtually nonexistent; by the end of the eighties , expert system tools will be in common use both in the office and at home. The rapid evolution of this area has been catalyzed by three factors. ( i) advances in expert systems technology; (ii) sdvances in microprocessor technology ; and (iii) proliferation of' microcomputers . The role of each of these factors is examined in the following.
ADVANCES IN I xPERT SYs" rEMs TECHNOLOGY

The honor of being known as the first expert system is usually given to the DENDRAL system developed at Stanford University in the early seventies [6l. Around the same time, tesearchers at the Massachusetts Institute of Technology were &iveloping the MACSYMA system [7]. These systems peritiitted inferences to be drawn in specific areas-the DENDRAL system focused on molecular structures and the M A:CSYMA system concentrated on algebraic areas. An important component of all expert systems is their nrp s of knowledge. This knowledge is frequently expressed t i It (wins of facts and rules. For example:

PACT: RUL :

PROFESSIONAL SOCIETIES PUBLISH EDUCATIONAL BOOKS. IF A BOOK HAS BEEN PUBLISHED BY A PROFESSIONAL SOCIETY, AND THE PROFESSIONAL SOCIETY IS IEEE, THEN THE BOOK HAS BEEN PUBLISHED IN NEW YORK.

s' °,fgjtemporary expert systems are organized to hold prob1rw 40(i aJn knowledge separate from problem -solution Irkivltloi,The former is termed the "knowledge base," 4ih,4 tho, Tatter is called the "inference engine." In both xt " 13 INI and DENDRAL, the focus was primarily on the ^^^°; rtlne . Such systems are usually classified as first{^aillelMl,) systems. Second -generation systems distinguish n ik;±,,,,r IIMt A owledge base and the inference engine. Most ; tt,li ^t,yallable expert systems fall,under this category. ^t+.ll ^,outlnues on the development of a third generation JV11i,i With more sophisticated knowledge- restructuring drawing capabilities than their predecessors [1]. v i i,3 4 different methods available today to structure a i lk? lijt4e , the rule-based approach and the frame-based ]l f) I , ,mo st commonly used. In a rifle -based representa# ?1& e; Is1 organized in the form of IF-THEN rules,

while a frame-based system uses a hierarchical organization based on frames. At the top level of the hierarchy, the object represents a "class " or a "parent"; below it, there are "slots" representing attributes that describe the main object. These slots in turn contain additional slots or frames. The inference engine is responsible for deciding which rules should be executed, using either a forward-chaining-approach or a backward-chaining approach, and when. In the forwardchaining approach, one starts with a given set of evidence and invokes all the production rules in a sequence ilntil all conclusions have been drawn. In the backward-chaining approach, one starts by making an assumption about the goal, and then works backward to see if any known fact or rule has been violated. If so, the assumption is incorrect. If not,. the assumption can indeed be true. Additional information may be needed to select one definite conclusion. Concepts in the areas of knowledge acquisition, knowledge representation, inference mechanisms, and heuristic programming have been studied and refined by researchers during the last three decades. In the mid-fifties, there was a utopian dream of using nascent computer technology to develop systems that would mimic human intelligence. During the next three decades, it became clear that since the computer functions very differently froin the human mind, 'thc latter is not the best paradigm for designing new generations of 'computers. The current thought is that while computer systems are far inferior to human beings in dealing with unforeseen problems, these systems can be infused with extensive expertise in narrow domains to .match, and even outperform, human experts. Examples include programs for playing chess and software for locating hidden materials. The harbingers of today's expert systems served two major functions. First, they helped to develop new techniques and methodologies for dealing with knowledge, as opposed to simple data. Second, the early systems helped to reduce expectations. Gradually, a consensus emerged in favor of dealing effectively with small areas of expertise rather than coming up with superficial solutions for larger scenarios. This emphasis on smaller domains and the advent of better tools and techniques led to increased popularity of expert systems. So far, we have focused on the field of expert systems. (This field is discussed in greater depth in [1].) However, even with all the advances in expert systems, the field of low-cost systems would have remained a distant reality had the microprocessor not been invented in the early seventies. We now turn our attention to advances in the field of microprocessors.
ADYANCEs IN MlcxorxocEssoR TECHNoLOGY

The term "microprocessor" was first used in 1972. However, the distinction of being the first "computer on a