Item Details

An Analysis of the Distinction Between Deep and Shallow Expert Systems

Peter D. Karp and David C. Wilkins
Format
Book; Online; EBook
Published
Urbana, Illinois : University of Illinois at Urbana-Champaign, Department of Computer Science, 1989.
Language
English
Series
UIUCDCS-R-89
Report
Summary
This paper analyzes the relationship between the techniques used to build expert systems and the behaviors they exhibit to show that there is not sufficient evidence to link the behavioral shortcomings of first-generation expert systems to the shallow methods of representation and inference they employ. There is only evidence that the shortcomings are a consequence of a general lack of knowledge. Moreover, the paper shows that the first-generation of expert systems employ both shallow methods and most of the so-called deep methods. Lastly, we show that deeper methods augment but do not replace shallow reasoning methods; most expert systems should possess both."
Abstract: "The first generation of expert systems (e.g., MYCIN, DENDRAL, R1) is often characterized as only using shallow methods of representation and inference, such as the use of production rules to encode empirical knowledge. First-generation expert systems are often dismissed on the grounds that shallow methods have inherent and fatal shortcomings which prevent them from achieving problem-solving behaviors that expert systems should possess. Examples of such desirable behaviors include graceful performance degradation, the handling of novel problems, and the ability of the expert system to detect its problem-solving limits.
Description
40 leaves : ill. ; 28 cm.
Mode of access: Internet.
Notes
  • "August 1989."
  • "UIUCDCS-R-89-1536"
  • "To appear in International Journal of Expert Systems, 1989."
  • Includes bibliographical references.
Series Statement
UIUCDCS-R-89 ; no. 1536
Report ; no. 1536
Logo for Copyright Not EvaluatedCopyright Not Evaluated
Technical Details

  • LEADER 03497cam a2200517Ia 4500
    001 100697383
    003 MiAaHDL
    005 20150526000000.0
    006 m d
    007 cr bn ---auaua
    008 901109s1989 ilua bt 000 0 eng d
    035
      
      
    a| sdr-uiuc7219346
    035
      
      
    a| (OCoLC)22654483
    040
      
      
    a| PMC b| eng c| PMC d| UWW d| OCLCQ d| UIU
    049
      
      
    a| UIUU
    050
      
    4
    a| QA76.76.E95 b| K37 1989
    082
    0
    4
    a| 510.78
    088
      
      
    a| UILU-ENG-89-1755
    100
    1
      
    a| Karp, Peter D.
    245
    1
    3
    a| An analysis of the distinction between deep and shallow expert systems / c| Peter D. Karp and David C. Wilkins.
    260
      
      
    a| Urbana, Illinois : b| University of Illinois at Urbana-Champaign, Department of Computer Science, c| 1989.
    300
      
      
    a| 40 leaves : b| ill. ; c| 28 cm.
    490
    0
      
    a| UIUCDCS-R-89 ; v| no. 1536
    490
    0
      
    a| Report ; v| no. 1536
    500
      
      
    a| "August 1989."
    500
      
      
    a| "UIUCDCS-R-89-1536"
    500
      
      
    a| "To appear in International Journal of Expert Systems, 1989."
    504
      
      
    a| Includes bibliographical references.
    520
      
      
    a| This paper analyzes the relationship between the techniques used to build expert systems and the behaviors they exhibit to show that there is not sufficient evidence to link the behavioral shortcomings of first-generation expert systems to the shallow methods of representation and inference they employ. There is only evidence that the shortcomings are a consequence of a general lack of knowledge. Moreover, the paper shows that the first-generation of expert systems employ both shallow methods and most of the so-called deep methods. Lastly, we show that deeper methods augment but do not replace shallow reasoning methods; most expert systems should possess both."
    520
      
      
    a| Abstract: "The first generation of expert systems (e.g., MYCIN, DENDRAL, R1) is often characterized as only using shallow methods of representation and inference, such as the use of production rules to encode empirical knowledge. First-generation expert systems are often dismissed on the grounds that shallow methods have inherent and fatal shortcomings which prevent them from achieving problem-solving behaviors that expert systems should possess. Examples of such desirable behaviors include graceful performance degradation, the handling of novel problems, and the ability of the expert system to detect its problem-solving limits.
    536
      
      
    a| Supported in part by DARPA. c| N00039-83-C-0136
    536
      
      
    a| Supported in part by the NIH. c| RR-00785
    536
      
      
    a| Supported in part by the NSF. c| MCS83-10236
    536
      
      
    a| Supported in part by the ONR. c| N00014-88K0124
    538
      
      
    a| Mode of access: Internet.
    650
      
    0
    a| Expert systems (Computer science)
    700
    1
      
    a| Wilkins, David C. e| author.
    710
    2
      
    a| University of Illinois at Urbana-Champaign. b| Department of Computer Science.
    974
      
      
    8| ia.analysisofdistin1536karp b| UIU c| IUIUC d| 20180315 s| ia u| uiuo.ark:/13960/t3st95h26 z| Report (University of Illinois at Urbana-Champaign. Dept. of Computer Science) no.1536 y| 1989 r| cc-by-nc-sa-3.0 q| con

Access online

Google Preview