-
Notifications
You must be signed in to change notification settings - Fork 0
/
Copy pathswartout-gil-sss96.ps
executable file
·2844 lines (2842 loc) · 56 KB
/
swartout-gil-sss96.ps
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
552
553
554
555
556
557
558
559
560
561
562
563
564
565
566
567
568
569
570
571
572
573
574
575
576
577
578
579
580
581
582
583
584
585
586
587
588
589
590
591
592
593
594
595
596
597
598
599
600
601
602
603
604
605
606
607
608
609
610
611
612
613
614
615
616
617
618
619
620
621
622
623
624
625
626
627
628
629
630
631
632
633
634
635
636
637
638
639
640
641
642
643
644
645
646
647
648
649
650
651
652
653
654
655
656
657
658
659
660
661
662
663
664
665
666
667
668
669
670
671
672
673
674
675
676
677
678
679
680
681
682
683
684
685
686
687
688
689
690
691
692
693
694
695
696
697
698
699
700
701
702
703
704
705
706
707
708
709
710
711
712
713
714
715
716
717
718
719
720
721
722
723
724
725
726
727
728
729
730
731
732
733
734
735
736
737
738
739
740
741
742
743
744
745
746
747
748
749
750
751
752
753
754
755
756
757
758
759
760
761
762
763
764
765
766
767
768
769
770
771
772
773
774
775
776
777
778
779
780
781
782
783
784
785
786
787
788
789
790
791
792
793
794
795
796
797
798
799
800
801
802
803
804
805
806
807
808
809
810
811
812
813
814
815
816
817
818
819
820
821
822
823
824
825
826
827
828
829
830
831
832
833
834
835
836
837
838
839
840
841
842
843
844
845
846
847
848
849
850
851
852
853
854
855
856
857
858
859
860
861
862
863
864
865
866
867
868
869
870
871
872
873
874
875
876
877
878
879
880
881
882
883
884
885
886
887
888
889
890
891
892
893
894
895
896
897
898
899
900
901
902
903
904
905
906
907
908
909
910
911
912
913
914
915
916
917
918
919
920
921
922
923
924
925
926
927
928
929
930
931
932
933
934
935
936
937
938
939
940
941
942
943
944
945
946
947
948
949
950
951
952
953
954
955
956
957
958
959
960
961
962
963
964
965
966
967
968
969
970
971
972
973
974
975
976
977
978
979
980
981
982
983
984
985
986
987
988
989
990
991
992
993
994
995
996
997
998
999
1000
%!PS-Adobe-3.0
%%Title: (Microsoft Word - Final-sss96.\311)
%%Creator: (Microsoft Word: PSPrinter 8.1.1)
%%CreationDate: (5:19 PM Tuesday, July 16, 1996)
%%For: (gil)
%%Pages: 6
%%DocumentFonts: Times-Bold Times-Roman Times-Italic Helvetica-Bold Helvetica
%%DocumentNeededFonts: Times-Bold Times-Roman Times-Italic Helvetica-Bold Helvetica
%%DocumentSuppliedFonts:
%%DocumentData: Clean7Bit
%%PageOrder: Ascend
%%Orientation: Portrait
%ADO_PaperArea: -12 -12 780 600
%ADO_ImageableArea: 0 0 768 588
%%EndComments
/md 150 dict def md begin
/currentpacking where {pop /sc_oldpacking currentpacking def true setpacking}if
%%BeginFile: adobe_psp_basic
%%Copyright: Copyright 1990-1993 Adobe Systems Incorporated. All Rights Reserved.
/bd{bind def}bind def
/xdf{exch def}bd
/xs{exch store}bd
/ld{load def}bd
/Z{0 def}bd
/T/true
/F/false
/:L/lineto
/lw/setlinewidth
/:M/moveto
/rl/rlineto
/rm/rmoveto
/:C/curveto
/:T/translate
/:K/closepath
/:mf/makefont
/gS/gsave
/gR/grestore
/np/newpath
14{ld}repeat
/$m matrix def
/av 81 def
/por true def
/normland false def
/psb-nosave{}bd
/pse-nosave{}bd
/us Z
/psb{/us save store}bd
/pse{us restore}bd
/level2
/languagelevel where
{
pop languagelevel 2 ge
}{
false
}ifelse
def
/featurecleanup
{
stopped
cleartomark
countdictstack exch sub dup 0 gt
{
{end}repeat
}{
pop
}ifelse
}bd
/noload Z
/startnoload
{
{/noload save store}if
}bd
/endnoload
{
{noload restore}if
}bd
level2 startnoload
/setjob
{
statusdict/jobname 3 -1 roll put
}bd
/setcopies
{
userdict/#copies 3 -1 roll put
}bd
level2 endnoload level2 not startnoload
/setjob
{
1 dict begin/JobName xdf currentdict end setuserparams
}bd
/setcopies
{
1 dict begin/NumCopies xdf currentdict end setpagedevice
}bd
level2 not endnoload
/pm Z
/mT Z
/sD Z
/realshowpage Z
/initializepage
{
/pm save store mT concat
}bd
/endp
{
pm restore showpage
}def
/$c/DeviceRGB def
/rectclip where
{
pop/rC/rectclip ld
}{
/rC
{
np 4 2 roll
:M
1 index 0 rl
0 exch rl
neg 0 rl
:K
clip np
}bd
}ifelse
/rectfill where
{
pop/rF/rectfill ld
}{
/rF
{
gS
np
4 2 roll
:M
1 index 0 rl
0 exch rl
neg 0 rl
fill
gR
}bd
}ifelse
/rectstroke where
{
pop/rS/rectstroke ld
}{
/rS
{
gS
np
4 2 roll
:M
1 index 0 rl
0 exch rl
neg 0 rl
:K
stroke
gR
}bd
}ifelse
%%EndFile
%%BeginFile: adobe_psp_colorspace_level1
%%Copyright: Copyright 1991-1993 Adobe Systems Incorporated. All Rights Reserved.
/G/setgray ld
/:F/setrgbcolor ld
%%EndFile
%%BeginFile: adobe_psp_uniform_graphics
%%Copyright: Copyright 1990-1993 Adobe Systems Incorporated. All Rights Reserved.
/@a
{
np :M 0 rl :L 0 exch rl 0 rl :L fill
}bd
/@b
{
np :M 0 rl 0 exch rl :L 0 rl 0 exch rl fill
}bd
/arct where
{
pop
}{
/arct
{
arcto pop pop pop pop
}bd
}ifelse
/x1 Z
/x2 Z
/y1 Z
/y2 Z
/rad Z
/@q
{
/rad xs
/y2 xs
/x2 xs
/y1 xs
/x1 xs
np
x2 x1 add 2 div y1 :M
x2 y1 x2 y2 rad arct
x2 y2 x1 y2 rad arct
x1 y2 x1 y1 rad arct
x1 y1 x2 y1 rad arct
fill
}bd
/@s
{
/rad xs
/y2 xs
/x2 xs
/y1 xs
/x1 xs
np
x2 x1 add 2 div y1 :M
x2 y1 x2 y2 rad arct
x2 y2 x1 y2 rad arct
x1 y2 x1 y1 rad arct
x1 y1 x2 y1 rad arct
:K
stroke
}bd
/@i
{
np 0 360 arc fill
}bd
/@j
{
gS
np
:T
scale
0 0 .5 0 360 arc
fill
gR
}bd
/@e
{
np
0 360 arc
:K
stroke
}bd
/@f
{
np
$m currentmatrix
pop
:T
scale
0 0 .5 0 360 arc
:K
$m setmatrix
stroke
}bd
/@k
{
gS
np
:T
0 0 :M
0 0 5 2 roll
arc fill
gR
}bd
/@l
{
gS
np
:T
0 0 :M
scale
0 0 .5 5 -2 roll arc
fill
gR
}bd
/@m
{
np
arc
stroke
}bd
/@n
{
np
$m currentmatrix
pop
:T
scale
0 0 .5 5 -2 roll arc
$m setmatrix
stroke
}bd
%%EndFile
%%BeginFile: adobe_psp_basic_text
%%Copyright: Copyright 1990-1993 Adobe Systems Incorporated. All Rights Reserved.
/S/show ld
/A{
0.0 exch ashow
}bd
/R{
0.0 exch 32 exch widthshow
}bd
/W{
0.0 3 1 roll widthshow
}bd
/J{
0.0 32 4 2 roll 0.0 exch awidthshow
}bd
/V{
0.0 4 1 roll 0.0 exch awidthshow
}bd
/fcflg true def
/fc{
fcflg{
vmstatus exch sub 50000 lt{
(%%[ Warning: Running out of memory ]%%\r)print flush/fcflg false store
}if pop
}if
}bd
/$f[1 0 0 -1 0 0]def
/:ff{$f :mf}bd
/MacEncoding StandardEncoding 256 array copy def
MacEncoding 39/quotesingle put
MacEncoding 96/grave put
/Adieresis/Aring/Ccedilla/Eacute/Ntilde/Odieresis/Udieresis/aacute
/agrave/acircumflex/adieresis/atilde/aring/ccedilla/eacute/egrave
/ecircumflex/edieresis/iacute/igrave/icircumflex/idieresis/ntilde/oacute
/ograve/ocircumflex/odieresis/otilde/uacute/ugrave/ucircumflex/udieresis
/dagger/degree/cent/sterling/section/bullet/paragraph/germandbls
/registered/copyright/trademark/acute/dieresis/notequal/AE/Oslash
/infinity/plusminus/lessequal/greaterequal/yen/mu/partialdiff/summation
/product/pi/integral/ordfeminine/ordmasculine/Omega/ae/oslash
/questiondown/exclamdown/logicalnot/radical/florin/approxequal/Delta/guillemotleft
/guillemotright/ellipsis/space/Agrave/Atilde/Otilde/OE/oe
/endash/emdash/quotedblleft/quotedblright/quoteleft/quoteright/divide/lozenge
/ydieresis/Ydieresis/fraction/currency/guilsinglleft/guilsinglright/fi/fl
/daggerdbl/periodcentered/quotesinglbase/quotedblbase/perthousand
/Acircumflex/Ecircumflex/Aacute/Edieresis/Egrave/Iacute/Icircumflex/Idieresis/Igrave
/Oacute/Ocircumflex/apple/Ograve/Uacute/Ucircumflex/Ugrave/dotlessi/circumflex/tilde
/macron/breve/dotaccent/ring/cedilla/hungarumlaut/ogonek/caron
MacEncoding 128 128 getinterval astore pop
level2 startnoload
/copyfontdict
{
findfont dup length dict
begin
{
1 index/FID ne{def}{pop pop}ifelse
}forall
}bd
level2 endnoload level2 not startnoload
/copyfontdict
{
findfont dup length dict
copy
begin
}bd
level2 not endnoload
md/fontname known not{
/fontname/customfont def
}if
/Encoding Z
/:mre
{
copyfontdict
/Encoding MacEncoding def
fontname currentdict
end
definefont :ff def
}bd
/:bsr
{
copyfontdict
/Encoding Encoding 256 array copy def
Encoding dup
}bd
/pd{put dup}bd
/:esr
{
pop pop
fontname currentdict
end
definefont :ff def
}bd
/scf
{
scalefont def
}bd
/scf-non
{
$m scale :mf setfont
}bd
/ps Z
/fz{/ps xs}bd
/sf/setfont ld
/cF/currentfont ld
/mbf
{
/makeblendedfont where
{
pop
makeblendedfont
/ABlend exch definefont
}{
pop
}ifelse
def
}def
%%EndFile
/currentpacking where {pop sc_oldpacking setpacking}if
end % md
%%EndProlog
%%BeginSetup
md begin
countdictstack
[
{%stopped
%%BeginFeature: *ManualFeed False
1 dict
dup /Policies 2 dict dup /PageSize 2 put dup /MediaType 0 put put
setpagedevice
1 dict
dup /ManualFeed false put
setpagedevice
%%EndFeature
}featurecleanup
countdictstack
[
{%stopped
%%BeginFeature: *PageSize Letter
1 dict
dup /Policies 2 dict dup /PageSize 2 put dup /MediaType 0 put put
setpagedevice
2 dict
dup /PageSize [612 792] put
dup /ImagingBBox null put
setpagedevice
%%EndFeature
}featurecleanup
(gil)setjob
/pT[1 0 0 -1 12 780]def/mT[1 0 0 -1 12 780]def
/sD 16 dict def
300 level2{1 dict dup/WaitTimeout 4 -1 roll put setuserparams}{statusdict/waittimeout 3 -1 roll put}ifelse
%%IncludeFont: Times-Bold
%%IncludeFont: Times-Roman
%%IncludeFont: Times-Italic
%%IncludeFont: Helvetica-Bold
%%IncludeFont: Helvetica
/f0_1/Times-Bold :mre
/f0_16 f0_1 16 scf
/f0_12 f0_1 12 scf
/f0_10 f0_1 10 scf
/f1_1/Times-Roman :mre
/f1_12 f1_1 12 scf
/f1_10 f1_1 10 scf
/f2_1/Times-Italic :mre
/f2_10 f2_1 10 scf
/f3_1 f1_1 1.087 scf
/f3_12 f3_1 12 scf
/f4_1/Helvetica-Bold :mre
/f4_128 f4_1 128 scf
/f4_112 f4_1 112 scf
/f4_96 f4_1 96 scf
/f5_1/Helvetica :mre
/f5_112 f5_1 112 scf
/f5_96 f5_1 96 scf
/Courier findfont[10 0 0 -10 0 0]:mf setfont
%%EndSetup
%%Page: 1 1
%%BeginPageSetup
initializepage
(gil; page: 1 of 6)setjob
%%EndPageSetup
gS 0 0 588 768 rC
153 76 :M
f0_16 sf
-.009(Flexible Knowledge Acquisition Through)A
144 100 :M
-.009(Explicit Representation of Knowledge Roles)A
260 131 :M
f0_12 sf
-.027(Bill Swartout)A
263 146 :M
-.034(Yolanda Gil)A
185 171 :M
31.819 0 rm
f3_12 sf
(USC/Information Sciences Institute)S
228 183 :M
21.157 0 rm
(4676 Admiralty Way)S
213 195 :M
20.359 0 rm
(Marina del Rey, CA 90292)S
205 207 :M
18.939 0 rm
275 243 :M
f0_10 sf
-.029(Abstract)A
52 254 :M
f1_10 sf
( )S
55 254 :M
1.115 .111(A system that acquires knowledge from a user should be able to reflect upon the knowledge that it has\321at each)J
52 265 :M
.184 .018(moment\321and understand what kinds of new knowledge it needs to learn. For the past two decades, research in the area)J
52 276 :M
.809 .081(of knowledge acquisition has been moving towards systems that have access to richer representations of knowledge)J
52 287 :M
.288 .029(about their task. This paper reviews some well-known knowledge acquisition tools representative of this trend. It also)J
52 298 :M
1.002 .1(describes our recent work in EXPECT, a system with explicit representations of knowledge about the task and the)J
52 309 :M
.135 .014(domain that supports knowledge acquisition for a wider range of tasks and applications than its predecessors. We hope)J
52 320 :M
1.07 .107(our observations will be useful to researchers in user interfaces and in machine learning concerned with acquiring)J
52 331 :M
-.031(information from users.)A
261 363 :M
f0_12 sf
-.029(Introduction)A
42 381 :M
f1_10 sf
.525 .052(The acquisition of knowledge about a task can be viewed as a process of incorporating new knowledge into some existing)J
42 392 :M
.536 .054(knowledge structure [Rosenbloom 1988]. The existing knowledge can guide and constrain the search for new knowledge,)J
42 403 :M
1.148 .115(and the process of integrating the new knowledge with the old may identify additional opportunities for learning. An)J
42 414 :M
.146 .015(acquisition system that takes this view needs to represent and understand the knowledge about the task as well as the process)J
42 425 :M
-.012(of finding and integrating new knowledge.)A
42 441 :M
.937 .094(A general trend in research on knowledge acquisition has been to make the knowledge structures that guide acquisition)J
42 452 :M
.035 .003(increasingly explicit. In early acquisition tools many of the requirements that needed to be satisfied when adding a new piece)J
42 463 :M
.146 .015(of knowledge were not stated. The result was that they could not provide very precise guidance for acquisition. Later, more)J
42 474 :M
.138 .014(of these requirements were made explicit, but they were embedded within the acquisition tools themselves. This allowed the)J
42 485 :M
.376 .038(tools to provide more specific guidance, but the requirements that were embedded in the tools could not be changed, which)J
42 496 :M
.298 .03(meant that some important aspects of the knowledge-based system being built could not be changed. More recent work has)J
42 507 :M
.747 .075(focused on making these requirements explicit and represented outside of the tool itself. A result of this has been more)J
42 518 :M
-.003(flexible acquisition tools that allow users to make a greater variety of changes to the knowledge-based system being built.)A
42 534 :M
.245 .024(To make this more concrete, we will begin by briefly reviewing three well-known acquisition systems: TEIREISIAS [Davis)J
42 545 :M
1.508 .151(1976], SALT [Marcus 1988], PROT\203G\203-II [Musen and Tu 1993]. We will then describe EXPECT, the acquisition)J
42 556 :M
.688 .069(framework we have been developing [Swartout and Gil 1995, Gil 1994, Gil and Paris 1994]. Each of these four systems)J
42 567 :M
.706 .071(represents a point along the trend we outlined above. We conclude with a discussion about how these systems make the)J
42 578 :M
.49 .049(knowledge structures they use for acquisition more or less explicit, and summarize the implications of these differences in)J
42 589 :M
-.01(terms of the kinds of knowledge acquisition they can support.)A
226 626 :M
f0_12 sf
-.044(Symbol-Level Approaches)A
42 644 :M
f1_10 sf
.091 .009(TEIREISIAS [Davis 1976] was one of the earliest knowledge acquisition systems. It was designed to help a user correct and)J
42 655 :M
.519 .052(extend MYCIN\325s knowledge base [Buchanan and Shortliffe 1984] for diagnosing infections. If MYCIN either incorrectly)J
42 666 :M
.418 .042(concluded that a disease was present, or missed a correct diagnosis, TEIREISIAS would walk the user through the trace of)J
42 677 :M
.094 .009(rule firings to determine where the error arose. If an incorrect diagnosis was concluded, TEIREISIAS would display the rule)J
42 688 :M
.48 .048(that led to the conclusion, and ask if any of the conditions on the rule needed to be changed. If so, the user was given the)J
endp
%%Page: 2 2
%%BeginPageSetup
initializepage
(gil; page: 2 of 6)setjob
%%EndPageSetup
gS 0 0 588 768 rC
42 69 :M
f1_10 sf
-.001(opportunity to make the change. If the rule was correct, but fired because some if its conditions were incorrectly asserted, the)A
42 80 :M
.911 .091(process would recurse and the user could look at the rules that made those assertions to determine if they were correct.)J
42 91 :M
.213 .021(Similarly, if a correct diagnosis was missed, TEIREISIAS would display the rule that could have caused that diagnosis to be)J
42 102 :M
.078 .008(reached, and would walk the user through the process of determining why those rules did not fire and correcting the problem,)J
42 113 :M
.899 .09(either by changing the conditions on the rules or by adding additional rules. When a new rule was added, TEIREISIAS)J
42 124 :M
.077 .008(provided guidance based on rule models derived from the existing rule base through statistical conceptual clustering. If rules)J
42 135 :M
1.207 .121(that concluded about a particular parameter )J
f2_10 sf
.45 .045(x )J
239 135 :M
f1_10 sf
1.053 .105(frequently mentioned some other parameter )J
427 135 :M
f2_10 sf
.965 .096(y )J
436 135 :M
f1_10 sf
1.143 .114( in their antecedents, then)J
42 146 :M
.133 .013(TEIREISIAS would point out a possible error if the user left the parameter )J
f2_10 sf
.057 .006(y )J
354 146 :M
f1_10 sf
.087 .009(out in a rule concluding about )J
f2_10 sf
.034 .003(x. )J
490 146 :M
f1_10 sf
-.027(This is useful,)A
42 157 :M
-.006(but a long way from actually understanding the role that the rule plays in problem solving.)A
42 173 :M
.436 .044(TEIREISIAS understood MYCIN at the symbol level [Newell 1982]. It understood rule patterns and why a particular rule)J
42 184 :M
.291 .029(fired or did not fire, but it did not have a global view of the overall algorithm that MYCIN was following. Indeed, work on)J
42 195 :M
.09 .009(TEIREISIAS pre-dated Clancey\325s analysis of MYCIN that showed that it was following the general problem solving strategy)J
42 206 :M
.316 .032(that he identified as heuristic classification [Clancey 1985]. TEIREISIAS did not capture the distinctions between rules for)J
42 217 :M
2.019 .202(data abstraction, heuristic match, and solution refinement that charaterize a heuristic classification system. Since)J
42 228 :M
.031 .003(TEIREISIAS did not understand the roles that the rules played in a system, it could not provide much help in guiding the user)J
42 239 :M
-.025(concerning the content of those rules.)A
226 276 :M
f0_12 sf
-.013(Role Limiting Approaches)A
42 294 :M
f1_10 sf
.563 .056(The next generation of knowledge acquistion tools represents what is called a role-limiting approach. It was based on the)J
42 305 :M
1.025 .102(observation that the kind of problem solving method that a system uses determines the kind of domain information the)J
42 316 :M
.669 .067(system needs [McDermott 1988]. Put another way, the role that a particular kind of knowledge plays in problem solving)J
42 327 :M
.208 .021(strongly constrains how that knowledge should be expressed \321 what is required for the system to function. A research goal)J
42 338 :M
.246 .025(during this stage was to try to understand a number of general methods used by knowledge based systems [Chandrasekaran,)J
42 349 :M
1.07 .107(1986], such as propose-and-revise and heuristic classification, and then to construct a knowledge acquisition tool for a)J
42 360 :M
.183 .018(particular method. Such tools could then be used to build knowledge based systems that used that particular problem solving)J
42 371 :M
.886 .089(approach. Examples of such tools include MORE [Eshelman 1988], SALT [Marcus 1988], and ROGET [Bennett 1985].)J
42 382 :M
.282 .028(Because these tools understood how knowledge would be used in problem solving, they could provide much more guidance)J
42 393 :M
-.004(in helping the user formulate the knowledge he was adding to a system correctly.)A
42 409 :M
.91 .091(SALT is a good example of a knowledge acquisition tool based on a role-limiting method. SALT was used to develop)J
42 420 :M
.617 .062(systems that use the propose-and-revise method to construct a solution to a problem. Examples included a configurer for)J
42 431 :M
(elevators and a flow-shop scheduler. In the propose-and-revise method, a system constructs an initial approximate solution to)S
42 442 :M
.152 .015(a problem which may have a number of aspects that still remain to be determined. It then proposes a design extension to fill)J
42 453 :M
.015 .002(in missing parts of the design, and looks for possible constraints that may be violated. If a constraint violation is detected, the)J
42 464 :M
.281 .028(system has knowledge of fixes that may be used to revise the solution and correct the problem. SALT captures this general)J
42 475 :M
.699 .07(algorithm, and understands that the knowledge that needs to be acquired to support this method is knowledge of ways of)J
42 486 :M
.193 .019(extending a design, the constraints that the design must satisfy, and ways of correcting constraint violations. In other words,)J
42 497 :M
1.222 .122(these are the roles that new knowledge plays within the propose-and-revise strategy. Because SALT was specifically)J
42 508 :M
.16 .016(designed for building propose-and-revise systems, it includes tools that detect and correct problems that can arise in building)J
42 519 :M
-.009(these systems, such as cycles in the fixes and constraints.)A
42 535 :M
.999 .1(Role-limiting approaches center knowledge base construction on filling the roles that knowledge plays in the particular)J
42 546 :M
1.209 .121(problem-solving method that they are designed for. However, the problem solving method is built into the knowledge)J
42 557 :M
.417 .042(acquisition tool itself. Thus, when one selects a tool, one also determines the problem-solving method that is employed by)J
42 568 :M
.604 .06(the application. One problem with this is that many large-scale systems are not homogeneous. That is, they do not use a)J
42 579 :M
.999 .1(single problem-solving method throughout\321some of the application can use one technique but different techniques are)J
42 590 :M
.44 .044(needed for other parts. As a result, an acquisition tool that only supports a single method has limited applicability [Musen)J
42 601 :M
-.166(1992].)A
201 638 :M
f0_12 sf
-.031(Composable Role-Limiting Methods)A
42 656 :M
f1_10 sf
.792 .079(Rather than embodying a single problem solving method, some knowledge acquisition environments contain a library of)J
42 667 :M
-.006(problem solving methods of smaller size. New applications are built by composing the overall problem-solving strategy from)A
endp
%%Page: 3 3
%%BeginPageSetup
initializepage
(gil; page: 3 of 6)setjob
%%EndPageSetup
gS 0 0 588 768 rC
42 389 :M
f1_10 sf
.331 .033(the smaller components of the library. Examples of systems that take this approach are SBF [Klinker et al. 1991], COMET)J
42 400 :M
-.018([Steels 1990], and PROT\203G\203-II [Musen and Tu 1993].)A
42 416 :M
.182 .018(PROT\203G\203-II guides acquisition based on fine-grained role-limiting methods. To build an application, a knowledge engineer)J
42 427 :M
.531 .053(configures the overall problem-~solving method using the components in the method library. At the same time, he or she)J
42 438 :M
.551 .055(builds a method ontology that contains the terms that are used by the method being configured. For example, \322constraint\323)J
42 449 :M
.87 .087(would be one of these terms and it would be identified as one of the kinds of knowledge needed by the component that)J
42 460 :M
1.581 .158(represents the revision stage of propose-and-revise. A domain ontology that contains domain-specific knowledge is)J
42 471 :M
1.184 .118(represented separately. The knowledge engineer then links the knowledge roles in the method ontology to the domain)J
42 482 :M
.648 .065(ontology, identifying how terms like \322constraint\323 map onto domain-specific terms. Once this mapping has been done, an)J
42 493 :M
1.114 .111(automatic interface builder uses the mapping to generate a knowledge acquisition tool that allows an end user to enter)J
42 504 :M
1.035 .103(domain-specific knowledge about the domain. Through this mapping the system understands just how domain-specific)J
42 515 :M
-.005(knowledge will actually be used, and the interface that is constructed thus requests the knowledge that will actually be needed)A
42 526 :M
-.049(for problem solving.)A
175 563 :M
f0_12 sf
-.021(Flexibility through Derived Interdependencies)A
42 581 :M
f1_10 sf
.017 .002(A remaining problem is that while acquisition systems allow a user to make changes to a system\325s factual knowledge, they do)J
42 592 :M
.332 .033(not allow the user to make changes to the problem solving methods that the system employs. Tools like SALT don\325t allow)J
42 603 :M
.321 .032(these changes because the problem solving method is implicitly encoded in the tool itself. PROT\203G\203-II doesn\325t allow such)J
42 614 :M
.295 .029(changes because its methods are pre-configured and the mapping between method and domain ontologies is fixed at system)J
42 625 :M
.474 .047(design time. Since PROT\203G\203-II\325s knowledge acquisition tool is derived from that mapping, it too is fixed at design time.)J
42 636 :M
.877 .088(The idea in EXPECT [Swartout and Gil 1995, Gil 1994, Gil and Paris 1994] is to derive the interdependencies between)J
42 647 :M
.107 .011(domain knowledge and problem solving methods automatically, and to be able to re-derive the dependencies as needed when)J
42 658 :M
.073 .007(changes are made to the problem solving knowledge. This approach allows a user to modify either the domain knowledge or)J
42 669 :M
.668 .067(the problem solving knowledge. We have used EXPECT to create systems in several domains. One of these domains is)J
.38 .56 .992 :F
74 66 453 297 rC
90 70 478 132 8 @q
.548 .956 .916 :F
86 66 473 128 8 @q
0 G
86.5 66.5 472.5 127.5 7.5 @s
gS
.086 .086 scale
1620.908 910.222 :M
1 G
f4_128 sf
0 G
-.004(Expectation-based Tools for Knowledge Acquisition)A
gR
1 G
244 85 314 115 3.5 @q
0 G
244.5 85.5 313.5 114.5 3 @s
gS
.086 .086 scale
2996.931 1131.943 :M
f5_96 sf
(expectation)S
3101.882 1236.969 :M
1 G
0 G
(builder)S
gR
1 G
353 84 419 114 3.5 @q
0 G
353.5 84.5 418.5 113.5 3 @s
gS
.086 .086 scale
4326.309 1120.273 :M
f5_96 sf
-.041(agenda)A
4302.986 1236.969 :M
1 G
0 G
-.121(manager)A
gR
1 G
147 85 210 115 3.5 @q
0 G
147.5 85.5 209.5 114.5 3 @s
gS
.086 .086 scale
2005.728 1131.943 :M
f5_96 sf
-.063(KB)A
1877.455 1236.969 :M
.38 .56 .992 :F
0 G
-.034(modifiers)A
gR
.38 .56 .992 :F
409 167 459 286 4.5 @q
1 G
405 163 455 282 4.5 @q
0 G
405.5 163.5 454.5 281.5 4 @s
gS
.086 .086 scale
4886.047 2287.225 :M
f5_112 sf
-.129(User)A
4746.112 2415.59 :M
-.078(Interaction)A
4769.435 2555.624 :M
(Manager/)S
4769.435 2683.989 :M
.38 .56 .992 :F
0 G
-.092(Explainer)A
gR
.38 .56 .992 :F
79 152 75 133 rF
1 G
75 148 75 133 rF
0 G
75.5 148.5 74 132 rS
gS
.086 .086 scale
909.574 1680.41 :M
1 G
f4_96 sf
0 G
-.028(Knowledge Bases)A
gR
1 G
79 152 67 38 rF
0 G
79.5 152.5 66 37 rS
gS
.086 .086 scale
1037.848 1902.131 :M
f5_96 sf
(Terminology)S
1224.427 2018.826 :M
-.051(and)A
1084.493 2135.521 :M
1 G
0 G
-.051(Domain K)A
gR
1 G
79 206 62 54 rF
0 G
79.5 206.5 61 53 rS
gS
.086 .086 scale
1096.154 2567.293 :M
f5_96 sf
-.055(Problem)A
1119.476 2683.989 :M
-.124(Solving)A
1037.848 2800.684 :M
(Knowledge)S
1037.848 2905.709 :M
.38 .56 .992 :F
0 G
(\(strategies\))S
gR
.38 .56 .992 :F
145 300 216 346 5.5 @q
.888 .744 1 :F
141 296 211 342 5.5 @q
0 G
141.5 296.5 210.5 341.5 5 @s
gS
.086 .086 scale
1772.504 3617.55 :M
f4_112 sf
-.053(Automatic)A
1842.471 3745.915 :M
(Method)S
1737.52 3885.949 :M
-.057(Instantiator)A
gR
-1 -1 116 318 1 1 115 280 @b
74 66 56 297 rC
116 317 -1 1 138 316 1 116 316 @a
gR
gS 74 66 453 297 rC
124 314 -1 1 138 317 1 124 313 @a
-1 -1 125 321 1 1 124 313 @b
-1 -1 125 321 1 1 137 317 @b
np 137 317 :M
124 313 :L
124 321 :L
137 317 :L
eofill
.38 .56 .992 :F
227 165 50 50 rF
1 G
223 161 50 50 rF
0 G
223.5 161.5 49 49 rS
gS
.086 .086 scale
2670.417 2053.835 :M
f5_112 sf
(Domain-)S
2693.739 2193.869 :M
-.05(specific)A
2775.368 2322.234 :M
.38 .56 .992 :F
0 G
-.055(KBS)A
gR
.38 .56 .992 :F
227 226 50 55 rF
1 G
223 222 50 54 rF
0 G
223.5 222.5 49 53 rS
gS
.086 .086 scale
2705.401 2870.701 :M
f5_112 sf
-.099(Design)A
2705.401 2999.066 :M
-.06(History)A
gR
74 193 453 170 rC
-1 -1 169 297 1 1 222 187 @b
gR
gS 74 66 453 297 rC
-1 -1 215 197 1 1 222 187 @b
214 197 -1 1 221 200 1 214 196 @a
-1 -1 221 201 1 1 222 187 @b
np 223 187 :M
214 196 :L
220 200 :L
223 187 :L
eofill
74 260 145 103 rC
-1 -1 201 296 1 1 221 254 @b
gR
gS 74 66 453 297 rC
-1 -1 214 264 1 1 222 254 @b
213 264 -1 1 221 266 1 213 263 @a
-1 -1 221 267 1 1 222 254 @b
np 223 254 :M
213 263 :L
220 267 :L
223 254 :L
eofill
.38 .56 .992 :F
314 202 380 231 3.5 @q
.992 .752 .896 :F
310 198 376 227 3.5 @q
0 G
310.5 198.5 375.5 226.5 3 @s
gS
.086 .086 scale
3731.587 2497.276 :M
.38 .56 .992 :F
f5_112 sf
0 G
-.046(Interpreter)A
gR
.38 .56 .992 :F
322 148 50 34 rF
1 G
318 144 50 34 rF
0 G
318.5 144.5 49 33 rS
gS
.086 .086 scale
3743.248 1832.114 :M
f5_112 sf
-.09(Execution)A
3848.199 1972.148 :M
-.051(Trace)A
gR
74 183 453 180 rC
-1 -1 343 199 1 1 342 177 @b