File size: 69,274 Bytes
6fa4bc9
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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
1001
1002
1003
1004
1005
1006
1007
1008
1009
1010
1011
1012
1013
1014
1015
1016
1017
1018
1019
1020
1021
1022
1023
1024
1025
1026
1027
1028
1029
1030
1031
1032
1033
1034
1035
1036
1037
1038
1039
1040
1041
1042
1043
1044
1045
1046
1047
1048
1049
1050
1051
1052
1053
1054
1055
1056
1057
1058
1059
1060
1061
1062
1063
1064
1065
1066
1067
1068
1069
1070
1071
1072
1073
1074
1075
1076
1077
1078
1079
1080
1081
1082
1083
1084
1085
1086
1087
1088
1089
1090
1091
1092
1093
1094
1095
1096
1097
1098
1099
1100
1101
1102
1103
1104
1105
1106
1107
1108
1109
1110
1111
1112
1113
1114
1115
1116
1117
1118
1119
1120
1121
1122
1123
1124
1125
1126
1127
1128
1129
1130
1131
1132
1133
1134
1135
1136
1137
1138
1139
1140
1141
1142
1143
1144
1145
1146
1147
1148
1149
1150
1151
1152
1153
1154
1155
1156
1157
1158
1159
1160
1161
1162
1163
1164
1165
1166
1167
1168
1169
1170
1171
1172
1173
1174
1175
1176
1177
1178
1179
1180
1181
1182
1183
1184
1185
1186
1187
1188
1189
1190
1191
1192
1193
1194
1195
1196
1197
1198
1199
1200
1201
1202
1203
1204
1205
1206
1207
1208
1209
1210
1211
1212
1213
1214
1215
1216
1217
1218
1219
1220
1221
{
    "paper_id": "P02-1045",
    "header": {
        "generated_with": "S2ORC 1.0.0",
        "date_generated": "2023-01-19T09:31:00.362463Z"
    },
    "title": "Applying Co-Training to Reference Resolution",
    "authors": [
        {
            "first": "Christoph",
            "middle": [],
            "last": "M\u00fcller",
            "suffix": "",
            "affiliation": {
                "laboratory": "European Media Laboratory",
                "institution": "",
                "location": {
                    "addrLine": "GmbH Villa Bosch Schlo\u00df-Wolfsbrunnenweg 33",
                    "postCode": "69118",
                    "settlement": "Heidelberg",
                    "country": "Germany"
                }
            },
            "email": "mueller@eml.villa-bosch.de"
        },
        {
            "first": "Stefan",
            "middle": [],
            "last": "Rapp",
            "suffix": "",
            "affiliation": {},
            "email": "rapp@sony.de"
        },
        {
            "first": "Michael",
            "middle": [],
            "last": "Strube",
            "suffix": "",
            "affiliation": {
                "laboratory": "European Media Laboratory GmbH Villa Bosch Schlo\u00df-Wolfsbrunnenweg 33",
                "institution": "",
                "location": {
                    "postCode": "69118",
                    "settlement": "Heidelberg",
                    "country": "Germany"
                }
            },
            "email": "strube@eml.villa-bosch.de"
        }
    ],
    "year": "",
    "venue": null,
    "identifiers": {},
    "abstract": "In this paper, we investigate the practical applicability of Co-Training for the task of building a classifier for reference resolution. We are concerned with the question if Co-Training can significantly reduce the amount of manual labeling work and still produce a classifier with an acceptable performance.",
    "pdf_parse": {
        "paper_id": "P02-1045",
        "_pdf_hash": "",
        "abstract": [
            {
                "text": "In this paper, we investigate the practical applicability of Co-Training for the task of building a classifier for reference resolution. We are concerned with the question if Co-Training can significantly reduce the amount of manual labeling work and still produce a classifier with an acceptable performance.",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Abstract",
                "sec_num": null
            }
        ],
        "body_text": [
            {
                "text": "A major obstacle for natural language processing systems which analyze natural language texts or utterances is the need to identify the entities referred to by means of referring expressions. Among referring expressions, pronouns and definite noun phrases (NPs) are the most prominent.",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Introduction",
                "sec_num": "1"
            },
            {
                "text": "Supervised machine learning algorithms were used for pronoun resolution with good results (Ge et al., 1998) , and for definite NPs with fairly good results (Aone and Bennett, 1995; McCarthy and Lehnert, 1995; Soon et al., 2001 ). However, the deficiency of supervised machine learning approaches is the need for an unknown amount of annotated training data for optimal performance.",
                "cite_spans": [
                    {
                        "start": 90,
                        "end": 107,
                        "text": "(Ge et al., 1998)",
                        "ref_id": "BIBREF6"
                    },
                    {
                        "start": 156,
                        "end": 180,
                        "text": "(Aone and Bennett, 1995;",
                        "ref_id": "BIBREF0"
                    },
                    {
                        "start": 181,
                        "end": 208,
                        "text": "McCarthy and Lehnert, 1995;",
                        "ref_id": "BIBREF7"
                    },
                    {
                        "start": 209,
                        "end": 226,
                        "text": "Soon et al., 2001",
                        "ref_id": null
                    }
                ],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Introduction",
                "sec_num": "1"
            },
            {
                "text": "So, researchers in NLP began to experiment with weakly supervised machine learning algorithms such as Co-Training (Blum and Mitchell, 1998) . Among others Co-Training was applied to document classification (Blum and Mitchell, 1998) , namedentity recognition (Collins and Singer, 1999) , noun phrase bracketing (Pierce and Cardie, 2001) , and statistical parsing (Sarkar, 2001) . In this paper we apply Co-Training to the problem of reference resolution in German texts from the tourism domain in order to provide answers to the following questions:",
                "cite_spans": [
                    {
                        "start": 114,
                        "end": 139,
                        "text": "(Blum and Mitchell, 1998)",
                        "ref_id": "BIBREF1"
                    },
                    {
                        "start": 206,
                        "end": 231,
                        "text": "(Blum and Mitchell, 1998)",
                        "ref_id": "BIBREF1"
                    },
                    {
                        "start": 258,
                        "end": 284,
                        "text": "(Collins and Singer, 1999)",
                        "ref_id": "BIBREF5"
                    },
                    {
                        "start": 310,
                        "end": 335,
                        "text": "(Pierce and Cardie, 2001)",
                        "ref_id": "BIBREF9"
                    },
                    {
                        "start": 362,
                        "end": 376,
                        "text": "(Sarkar, 2001)",
                        "ref_id": "BIBREF10"
                    }
                ],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Introduction",
                "sec_num": "1"
            },
            {
                "text": "Does Co-Training work at all for this task (when compared to conventional C4.5 decision tree learning)?",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Introduction",
                "sec_num": "1"
            },
            {
                "text": "How much labeled training data is required for achieving a reasonable performance? First, we discuss features that have been found to be relevant for the task of reference resolution, and describe the feature set that we are using (Section 2). Then we briefly introduce the Co-Training paradigm (Section 3), which is followed by a description of the corpus we use, the corpus annotation, and the way we prepared the data for using a binary classifier in the Co-Training algorithm (Section 4). In Section 5 we specify the experimental setup and report on the results.",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Introduction",
                "sec_num": "1"
            },
            {
                "text": "Driven by the necessity to provide robust systems for the MUC system evaluations, researchers began to look for those features which were particular important for the task of reference resolution. While most features for pronoun resolution have been described in the literature for decades, researchers only recently began to look for robust and cheap features, i.e., those which perform well over several domains and can be annotated (semi-) automatically. Also, the relative quantitative contribution of each of these features came into focus only after the advent of corpus-based and statistical methods. In the following, we describe a few earlier contributions with respect to the features used.",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Previous Work",
                "sec_num": "2.1"
            },
            {
                "text": "Decision tree algorithms were used for reference resolution by Aone and Bennett (1995, C4.5), McCarthy and Lehnert (1995, C4.5) and Soon et al. (2001, C5.0) . This approach requires the definition of a set of training features describing pairs of anaphors and their antecedents. Aone and Bennett (1995) , working on reference resolution in Japanese newspaper articles, use 66 features. They do not mention all of these explicitly but emphasize the features POS-tag, grammatical role, semantic class and distance. The set of semantic classes they use appears to be rather elaborated and highly domain-dependent. Aone and Bennett (1995) report that their best classifier achieved an F-measure of about 77% after training on 250 documents. They mention that it was important for the training data to contain transitive positives, i.e., all possible coreference relations within an anaphoric chain.",
                "cite_spans": [
                    {
                        "start": 63,
                        "end": 71,
                        "text": "Aone and",
                        "ref_id": "BIBREF0"
                    },
                    {
                        "start": 72,
                        "end": 106,
                        "text": "Bennett (1995, C4.5), McCarthy and",
                        "ref_id": null
                    },
                    {
                        "start": 107,
                        "end": 127,
                        "text": "Lehnert (1995, C4.5)",
                        "ref_id": null
                    },
                    {
                        "start": 132,
                        "end": 156,
                        "text": "Soon et al. (2001, C5.0)",
                        "ref_id": null
                    },
                    {
                        "start": 279,
                        "end": 302,
                        "text": "Aone and Bennett (1995)",
                        "ref_id": "BIBREF0"
                    },
                    {
                        "start": 611,
                        "end": 634,
                        "text": "Aone and Bennett (1995)",
                        "ref_id": "BIBREF0"
                    }
                ],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Previous Work",
                "sec_num": "2.1"
            },
            {
                "text": "McCarthy and Lehnert (1995) describe a reference resolution component which they evaluated on the MUC-5 English Joint Venture corpus. They distinguish between features which focus on individual noun phrases (e.g. Does noun phrase contain a name?) and features which focus on the anaphoric relation (e.g. Do both share a common NP?). It was criticized (Soon et al., 2001 ) that the features used by McCarthy and Lehnert (1995) are highly idiosyncratic and applicable only to one particular domain. McCarthy and Lehnert (1995) achieved results of about 86% F-measure (evaluated according to Vilain et al. (1995) ) on the MUC-5 data set. However, only a defined subset of all possible reference resolution cases was considered relevant in the MUC-5 task description, e.g., only entity references. For this case, the domain-dependent features may have been particularly important, making it difficult to compare the results of this approach to others working on less restricted domains. Soon et al. (2001) use twelve features (see Table 1). They show a part of their decision tree in which the weak string identity feature (i.e. identity after determiners have been removed) appears to be the most important one. They also report on the relative contribution of the features where -distance in sentences between anaphor and antecedent -antecedent is a pronoun? -anaphor is a pronoun? -weak string identity between anaphor and antecedent -anaphor is a definite noun phrase? -anaphor is a demonstrative pronoun? -number agreement between anaphor and antecedent -semantic class agreement between anaphor and antecedent -gender agreement between anaphor and antecedent -anaphor and antecedent are both proper names? -an alias feature (used for proper names and acronyms) -an appositive feature the three features weak string identity, alias (which maps named entities in order to resolve dates, person names, acronyms, etc.) and appositive seem to cover most of the cases (the other nine features contribute only 2.3% F-measure for MUC-6 texts and 1% F-measure for MUC-7 texts). Soon et al. 2001include all noun phrases returned by their NP identifier and report an F-measure of 62.6% for MUC-6 data and 60.4% for MUC-7 data. They only used pairs of anaphors and their closest antecedents as positive examples in training, but evaluated according to Vilain et al. (1995) . Cardie and Wagstaff (1999) describe an unsupervised clustering approach to noun phrase coreference resolution in which features are assigned to single noun phrases only. They use the features shown in Table 2 , all of which are obtained automatically without any manual tagging.",
                "cite_spans": [
                    {
                        "start": 351,
                        "end": 369,
                        "text": "(Soon et al., 2001",
                        "ref_id": null
                    },
                    {
                        "start": 398,
                        "end": 425,
                        "text": "McCarthy and Lehnert (1995)",
                        "ref_id": "BIBREF7"
                    },
                    {
                        "start": 497,
                        "end": 524,
                        "text": "McCarthy and Lehnert (1995)",
                        "ref_id": "BIBREF7"
                    },
                    {
                        "start": 589,
                        "end": 609,
                        "text": "Vilain et al. (1995)",
                        "ref_id": "BIBREF14"
                    },
                    {
                        "start": 983,
                        "end": 1001,
                        "text": "Soon et al. (2001)",
                        "ref_id": null
                    },
                    {
                        "start": 2342,
                        "end": 2362,
                        "text": "Vilain et al. (1995)",
                        "ref_id": "BIBREF14"
                    },
                    {
                        "start": 2365,
                        "end": 2391,
                        "text": "Cardie and Wagstaff (1999)",
                        "ref_id": "BIBREF3"
                    }
                ],
                "ref_spans": [
                    {
                        "start": 2566,
                        "end": 2573,
                        "text": "Table 2",
                        "ref_id": "TABREF1"
                    }
                ],
                "eq_spans": [],
                "section": "Previous Work",
                "sec_num": "2.1"
            },
            {
                "text": "-position (NPs are numbered sequentially) -pronoun type (nom., acc., possessive, ambiguous) -article (indefinite, definite, none) -appositive (yes, no) -number (singular, plural) -proper name (yes, no) -semantic class (based on WordNet: time, city, animal, human, object; based on a separate algorithm: number, money, company) -gender (masculine, feminine, either, neuter) -animacy (anim, inanim) The feature semantic class used by Cardie and Wagstaff (1999) seems to be a domain-dependent one which can only be used for the MUC domain and similar ones. Cardie and Wagstaff (1999) report a performance of 53,6% F-measure (evaluated according to Vilain et al. (1995) ).",
                "cite_spans": [
                    {
                        "start": 432,
                        "end": 458,
                        "text": "Cardie and Wagstaff (1999)",
                        "ref_id": "BIBREF3"
                    },
                    {
                        "start": 554,
                        "end": 580,
                        "text": "Cardie and Wagstaff (1999)",
                        "ref_id": "BIBREF3"
                    },
                    {
                        "start": 645,
                        "end": 665,
                        "text": "Vilain et al. (1995)",
                        "ref_id": "BIBREF14"
                    }
                ],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Previous Work",
                "sec_num": "2.1"
            },
            {
                "text": "We consider the features we use for our weakly supervised approach to be domain-independent. We distinguish between features assigned to noun phrases and features assigned to the potential coreference relation. They are listed in Table 3 together with their respective possible values. In the literature on reference resolution it is claimed that the antecedent's grammatical function and its realization are important. Hence we introduce the features ante gram func and ante npform. The identity in grammatical function of a potential anaphor and antecedent is captured in the feature syn par. Since in German the gender and the semantic class do not necessarily coincide (i.e. objects are not necessarily neuter as in English) we also provide a semanticclass feature which captures the difference between human, concrete, and abstract objects. This basically corresponds to the gender attribute in English. The feature wdist captures the distance in words between anaphor and antecedent, the feature ddist captures the distance in sentences, the feature mdist the number of markables (NPs) between anaphor and antecedent. Features like the string ident and substring match features were used by other researchers (Soon et al., 2001) , while the features ante med and ana med were used by Strube et al. (2002) in order to improve the performance for definite NPs. The minimum edit distance (MED) computes the similarity of strings by taking into account the minimum number of editing operations (substitutions s, insertions i, deletions d) needed to transform one string into the other (Wagner and Fischer, 1974) . The MED is computed from these editing operations and the length of the potential antecedent m or the length of the anaphor n.",
                "cite_spans": [
                    {
                        "start": 1215,
                        "end": 1234,
                        "text": "(Soon et al., 2001)",
                        "ref_id": null
                    },
                    {
                        "start": 1290,
                        "end": 1310,
                        "text": "Strube et al. (2002)",
                        "ref_id": "BIBREF13"
                    },
                    {
                        "start": 1587,
                        "end": 1613,
                        "text": "(Wagner and Fischer, 1974)",
                        "ref_id": "BIBREF15"
                    }
                ],
                "ref_spans": [
                    {
                        "start": 230,
                        "end": 237,
                        "text": "Table 3",
                        "ref_id": null
                    }
                ],
                "eq_spans": [],
                "section": "Our Features",
                "sec_num": "2.2"
            },
            {
                "text": "Co-Training (Blum and Mitchell, 1998 ) is a metalearning algorithm which exploits unlabeled in addition to labeled training data for classifier learning. A Co-Training classifier is complex in the sense that it consists of two simple classifiers (most often Naive Bayes, e.g. by Blum and Mitchell (1998) and Pierce and Cardie (2001) ). Initially, these classifiers are trained in the conventional way using a small set of size L of labeled training data. In this process, each of the two classifiers is trained on a different subset of features of the training data. These feature subsets are commonly referred to as different views that the classifiers have on the data, i.e., each classifier describes a given instance in terms of different features. The Co-Training algorithm is supposed to bootstrap by gradually extending the training data with self-labeled instances. It utilizes the two classifiers by letting them in turn label the p best positive and n best negative instances from a set of size P of unlabeled training data (referred to in the literature as the pool). Instances labeled by one classifier are then added to the other's training data, and vice versa. After each turn, both classifiers are re-trained on their augmented training sets, and the pool is refilled with",
                "cite_spans": [
                    {
                        "start": 12,
                        "end": 36,
                        "text": "(Blum and Mitchell, 1998",
                        "ref_id": "BIBREF1"
                    },
                    {
                        "start": 279,
                        "end": 303,
                        "text": "Blum and Mitchell (1998)",
                        "ref_id": "BIBREF1"
                    },
                    {
                        "start": 308,
                        "end": 332,
                        "text": "Pierce and Cardie (2001)",
                        "ref_id": "BIBREF9"
                    }
                ],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Co-Training",
                "sec_num": "3"
            },
            {
                "text": "\u00a1 \u00a3 \u00a2 \u00a5 \u00a4 \u00a7 \u00a6 \u00a9",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Co-Training",
                "sec_num": "3"
            },
            {
                "text": "unlabeled training instances drawn at random. This process is repeated either for a given number of iterations I or until all the unlabeled data has been labeled. In particular the definition of the two data views appears to be a crucial factor which can strongly influence the behaviour of Co-Training. A number of requirements for these views are mentioned in the literature, e.g., that they have to be disjoint or even conditionally independent (but cf. Nigam and Ghani (2000) ). Another important factor is the ratio between p and n, i.e., the number of positive and negative instances added in each iteration. These values are commonly chosen in such a way as to reflect the empirical class distribution of the respective instances.",
                "cite_spans": [
                    {
                        "start": 457,
                        "end": 479,
                        "text": "Nigam and Ghani (2000)",
                        "ref_id": "BIBREF8"
                    }
                ],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Co-Training",
                "sec_num": "3"
            },
            {
                "text": "Our corpus consists of 250 short German texts (total 36924 tokens, 9399 NPs, 2179 anaphoric NPs) about sights, historic events and persons in Heidelberg. The average length of the texts was 149 tokens. The texts were POS-tagged using TnT (Brants, 2000) . A basic identification of markables (i.e. NPs) was obtained by using the NP-Chunker Chunkie (Skut and Brants, 1998) . The POS-tagger was also used for assigning attributes to markables (e.g. the NP form). The automatic annotation was followed by a man- ",
                "cite_spans": [
                    {
                        "start": 238,
                        "end": 252,
                        "text": "(Brants, 2000)",
                        "ref_id": "BIBREF2"
                    },
                    {
                        "start": 347,
                        "end": 370,
                        "text": "(Skut and Brants, 1998)",
                        "ref_id": "BIBREF11"
                    }
                ],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Text Corpus",
                "sec_num": "4.1"
            },
            {
                "text": "! \" # % $ ' & ) ( 1 0 2 0 4 3 6 5 \u00a9 7 9 8 A @ C B D E B 6 F H G 5",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Text Corpus",
                "sec_num": "4.1"
            },
            {
                "text": "17. ana med minimum edit distance to antecedent: Table 3 : Our Features ual correction and annotation phase in which further tags were assigned to the markables. In this phase manual coreference annotation was performed as well. In our annotation, coreference is represented in terms of a member attribute on markables (i.e., noun phrases). Markables with the same value in this attribute are considered coreferring expressions. The annotation was performed by two students. The reliability of the annotations was checked using the kappa statistic (Carletta, 1996) .",
                "cite_spans": [
                    {
                        "start": 548,
                        "end": 564,
                        "text": "(Carletta, 1996)",
                        "ref_id": "BIBREF4"
                    }
                ],
                "ref_spans": [
                    {
                        "start": 49,
                        "end": 56,
                        "text": "Table 3",
                        "ref_id": null
                    }
                ],
                "eq_spans": [],
                "section": "Text Corpus",
                "sec_num": "4.1"
            },
            {
                "text": "\" # I $ P & ) ( % 0 Q 0 4 3 6 R S 7 9 8 A @ C B D E B 6 F H G R",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Text Corpus",
                "sec_num": "4.1"
            },
            {
                "text": "The problem of coreference resolution can easily be formulated in such a way as to be amenable to Co-Training. The most straightforward definition turns the task into a binary classification: Given a pair of potential anaphor and potential antecedent, classify as positive if the antecedent is in fact the closest antecedent, and as negative otherwise. Note that the restriction of this rule to the closest antecedent means that transitive antecedents (i.e. those occuring further upwards in the text as the direct antecedent) are treated as negative in the training data. We favour this definition because it strengthens the predictive power of the word distance between potential anaphor and potential antecedent (as expressed in the wdist feature).",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Coreference resolution as binary classification",
                "sec_num": "4.2"
            },
            {
                "text": "From our annotated corpus, we created one initial training and test data set. For each text, a list of noun phrases in document order was generated. This list was then processed from end to beginning, the phrase at the current position being considered as a potential anaphor. Beginning with the directly preceding position, each noun phrase which appeared before was combined with the potential anaphor and both entities were considered a potential antecedentanaphor pair. If applied to a text with \u00a6 noun phrases, this algorithm produces a total of",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Test and Training Data Generation",
                "sec_num": "4.3"
            },
            {
                "text": "T V U Q W E T Y X \u00e0 c b d",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Test and Training Data Generation",
                "sec_num": "4.3"
            },
            {
                "text": "noun phrase pairs. However, a number of filters can reasonably be applied at this point. An antecedentanaphor pair is discarded if the anaphor is an indefinite NP, if one entity is embedded into the other, e.g., if the potential anaphor is the head of the potential antecedent NP (or vice versa), if both entities have different values in their semantic class attributes 1 , if either entity has a value other than 3rd person singular or plural in its agreement feature, if both entities have different values in their agreement features 2 .",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Test and Training Data Generation",
                "sec_num": "4.3"
            },
            {
                "text": "For some texts, these heuristics reduced to up to 50% the potential antecedent-anaphor pairs, all of which would have been negative cases. We regard these cases as irrelevant because they do not contribute any knowledge for the classifier. After application of these filters, the remaining candidate pairs were labeled as follows:",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Test and Training Data Generation",
                "sec_num": "4.3"
            },
            {
                "text": "Pairs of anaphors and their direct (i.e. closest) antecedents were labeled P. This means that each anaphoric expression produced exactly one positive instance.",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Test and Training Data Generation",
                "sec_num": "4.3"
            },
            {
                "text": "Pairs of anaphors and their indirect (transitive) antecedents were labeled TP.",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Test and Training Data Generation",
                "sec_num": "4.3"
            },
            {
                "text": "Pairs of anaphors and those non-antecedents which occurred before the direct antecedent were labeled N. The number of negative instances that each expression produced thus depended on the number of non-antecedents occurring before the direct antecedent (if any).",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Test and Training Data Generation",
                "sec_num": "4.3"
            },
            {
                "text": "Pairs of anaphors and non-antecedents were labeled DN (distant N) if at least one true antecedent occurred in between.",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Test and Training Data Generation",
                "sec_num": "4.3"
            },
            {
                "text": "This produced 250 data sets with a total of 92750 instances of potential antecedent-anaphor pairs (2074 P, 70021 N, 6014 TP and 14641 DN) . From this set the last 50 texts were used as a test set. From this set, all instances with class DN and TP were removed, resulting in a test set of 11033 instances. Removing DNs and TPs was motivated by the fact that initial experimentation with C4.5 had indicated that a four way classification gives no advantage over a two way classification. In addition, this kind of test set approximates the decisions made by a simple resolution algorithm that cause in a real-world setting, information about a pronoun's semantic class obviously is not available prior to its resolution.",
                "cite_spans": [
                    {
                        "start": 98,
                        "end": 137,
                        "text": "(2074 P, 70021 N, 6014 TP and 14641 DN)",
                        "ref_id": null
                    }
                ],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Test and Training Data Generation",
                "sec_num": "4.3"
            },
            {
                "text": "2 This filter applies only if the anaphor is a pronoun. This restriction is necessary because German allows for cases where an antecedent is referred back to by a non-pronoun anaphor which has a different grammatical gender. looks for an antecedent from the current position upwards until it finds one or reaches the beginning. Hence, our results are only indirectly comparable with the ones obtained by an evaluation according to Vilain et al. (1995) . However, in this paper we only compare results of this direct binary antecedentanaphor pair decision.",
                "cite_spans": [
                    {
                        "start": 431,
                        "end": 451,
                        "text": "Vilain et al. (1995)",
                        "ref_id": "BIBREF14"
                    }
                ],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Test and Training Data Generation",
                "sec_num": "4.3"
            },
            {
                "text": "The remaining texts were split in two sets of 50 resp. 150 texts. From the first, our labeled training set was produced by removing all instances with class DN and TP. The second set was used as our unlabeled training set. From this set, no instances were removed because no knowledge whatsoever about the data can be assumed in a realistic setting.",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Test and Training Data Generation",
                "sec_num": "4.3"
            },
            {
                "text": "For our experiments we implemented the standard Co-Training algorithm (as described in Section 3) in Java using the Weka machine learning library 3 . In contrast to other Co-Training approaches, we did not use Naive Bayes as base classifiers, but J48 decision trees, which are a Weka re-implementation of C4.5. The use of decision tree classifiers was motivated by the observation that they appeared to perform better on the task at hand.",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Experiments and Results",
                "sec_num": "5"
            },
            {
                "text": "We conducted a number of experiments to investigate the question if Co-Training is beneficial for the task of training a classifier for coreference resolution. In previous work (Strube et al., 2002) we obtained quite different results for different types of anaphora, i.e. if we split the data according to the ana np feature into personal and possessive pronouns (PPER PPOS), proper names (NE), and definite NPs (def NP). Therefore we performed Co-Training experiments on subsets of our data defined by these NP forms, and on the whole data set.",
                "cite_spans": [
                    {
                        "start": 177,
                        "end": 198,
                        "text": "(Strube et al., 2002)",
                        "ref_id": "BIBREF13"
                    }
                ],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Experiments and Results",
                "sec_num": "5"
            },
            {
                "text": "We determined the features for the two different views with the following procedure: We trained classifiers on each feature separately and chose the best one, adding the feature which produced it as the first feature of view 1. We then trained classifiers on all remaining features separately, again choosing the best one and adding its feature as the first feature of view 2. In the next step, we enhanced the first classifier by combining it with all remaining features separately. The classifier with the best performance was then chosen and its new feature added as the second feature of view 1. We then enhanced the second classifier in the same way by selecting from the remaining features the one that most improved it, adding this feature as the second one of view 2. This process was repeated until no features were left or no significant improvement was achieved, resulting in the views shown in Table 4 (features marked na were not available for the respective class). This way we determined two views which performed reasonably well separately.",
                "cite_spans": [],
                "ref_spans": [
                    {
                        "start": 906,
                        "end": 913,
                        "text": "Table 4",
                        "ref_id": "TABREF3"
                    }
                ],
                "eq_spans": [],
                "section": "Experiments and Results",
                "sec_num": "5"
            },
            {
                "text": "PPER NE def NP all PPOS features 1 2 1 2 1 2 1 2 2. ante gram func X X X X 3. ante npform X X X X 4. ante agree X X X X 5. ante semanticc. X X X X 6. ana gram func X X X 7. ana npform na na X 8. ana agree X X X 9. ana semanticc. For Co-Training, we committed ourselves to fixed parameter settings in order to reduce the complexity of the experiments. Settings are given in the relevant subsections, where the following abbreviations are used: L=size of labeled training set, P/N=number of positive/negative instances added per iteration. All reported Co-Training results are averaged over 5 runs utilizing randomized sequences of unlabeled instances.",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Experiments and Results",
                "sec_num": "5"
            },
            {
                "text": "na X X na 10. wdist X X X X 11. ddist X X X X 12. mdist X X X X 13. syn par X X X 14. string ident X X X X 15. string match X X X X 16. ante med X X X X 17. ana med X X X X",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Experiments and Results",
                "sec_num": "5"
            },
            {
                "text": "We compare the results we obtained with Co-Training with the initial result before the Co-Training process started (zero iterations, both views combined; denoted as XX 0its in the plots). For this, we used a conventional C4.5 decision tree classifier (J48 implementation, default settings) on labeled training data sets of the same size used for the respective Co-Training experiment. We did this in order to verify the quality of the training data and for obtaining reference values for comparison with the Co-Training classifiers. \"20\" using 2:9 \"20_0its\" using 2:6 \"100\" using 2:9 \"100_0its\" using 2:6 \"200\" using 2:9 \"200_0its\" using 2:6",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Experiments and Results",
                "sec_num": "5"
            },
            {
                "text": "Figure 1: F for PPER PPOS over iterations, baselines PPER PPOS. In Figure 1 , three curves and three baselines are plotted: For 20 (L=20), 20 0its is the baseline, i.e. the initial result obtained by just combining the two initial classifiers. For 100, L=100, and for 200, L=200. The other settings were: P=1, N=1, Pool=10. As can be seen, the baselines slightly outperform the Co-Training curves (except for 100). \"200\" using 2:9 \"200_0its\" using 2:6 \"1000\" using 2:9 \"1000_0its\" using 2:6 \"2000\" using 2:9 \"2000_0its\" using 2:6",
                "cite_spans": [],
                "ref_spans": [
                    {
                        "start": 67,
                        "end": 75,
                        "text": "Figure 1",
                        "ref_id": null
                    }
                ],
                "eq_spans": [],
                "section": "Experiments and Results",
                "sec_num": "5"
            },
            {
                "text": "Figure 2: F for NE over iterations, baselines NE. Then we ran the Co-Training experiment with the NP form NE (i.e. proper names). Since the distribution of positive and negative examples in the labeled training data was quite different from the previous experiment, we used P=1, N=33, Pool=120. Since all results with Lf 200 were equally poor, we started with L=200, where the results were closer to ones of classifiers using the whole data set. The resulting Co-Training curve degrades substantially. However, with a training size of 1000 and 2000 the Co-Training curves are above their baselines. \"500\" using 2:9 \"500_0its\" using 2:6 \"1000\" using 2:9 \"1000_0its\" using 2:6 \"2000\" using 2:9 \"2000_0its\" using 2:6",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Experiments and Results",
                "sec_num": "5"
            },
            {
                "text": "Figure 3: F for def NP over iterations, baselines def NP. In the next experiment we tested the NP form def NP, a concept which can be expected to be far more difficult to learn than the previous two NP forms. Used settings were P=1, N=30, Pool=120. For Lf 500, F-measure was near 0. With L=500 the Co-Training curve is way below the baseline. However, with L=1000 and L=2000 Co-Training does show some improvement. \"200\" using 2:9 \"200_0its\" using 2:6 \"1000\" using 2:9 \"1000_0its\" using 2:6 \"2000\" using 2:9 \"2000_0its\" using 2:6 All. In the last experiment we trained our classifier on all NP forms, using P=1, N=33, Pool=120.",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Experiments and Results",
                "sec_num": "5"
            },
            {
                "text": "With L=200 the baseline clearly outperforms Co-Training. Co-Training with L=1000 initially rises above the baselines, but then decreases after about 15 to 20 iterations. With L=2000 the Co-Training curve approximates its baseline and then degenerates.",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Experiments and Results",
                "sec_num": "5"
            },
            {
                "text": "Supervised learning of reference resolution classifiers is expensive since it needs unknown amounts of annotated data for training. However, reference resolution algorithms based on these classifiers achieve reasonable performance of about 60 to 63% F-measure (Soon et al., 2001) . Unsupervised learning might be an alternative, since it does not need any annotation at all. However, the cost is the decrease in performance to about 53% F-measure on the same data (Cardie and Wagstaff, 1999) which may be unsuitable for a lot of tasks. In this paper we tried to pioneer a path between the unsupervised and the supervised paradigm by using the Co-Training meta-learning algorithm.",
                "cite_spans": [
                    {
                        "start": 260,
                        "end": 279,
                        "text": "(Soon et al., 2001)",
                        "ref_id": null
                    },
                    {
                        "start": 464,
                        "end": 491,
                        "text": "(Cardie and Wagstaff, 1999)",
                        "ref_id": "BIBREF3"
                    }
                ],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Conclusions",
                "sec_num": "6"
            },
            {
                "text": "The results, however, are mostly negative. Although we did not try every possible setting for the Co-Training algorithm, we did experiment with different feature views, Pool sizes and positive/negative increments, and we assume the settings we used are reasonable. It seems that Co-Training is useful in rather specialized constellations only. For the classes PPER PPOS, NE and All, our Co-Training experiments did not yield any benefits worth reporting. Only for def NP, we observed a considerable improvement from about 17% to about 25% F-measure using an initial training set of 1000 labeled instances, and from about 19% to about 28% F-measure using 2000 labeled training instances. In Strube et al. (2002) we report results from other experiments for definite noun phrase reference resolution. Although based on much more labeled training data, these experiments did not yield significantly better results. In this case, therefore, Co-Training seems to be able to save manual annotation work. On the other hand, the definition of the feature views is non-trivial for the task of training a reference resolution classifier, where no obvious or natural feature split suggests itself. In practical terms, therefore, this could outweigh the advantage of annota-tion work saved.",
                "cite_spans": [
                    {
                        "start": 690,
                        "end": 710,
                        "text": "Strube et al. (2002)",
                        "ref_id": "BIBREF13"
                    }
                ],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Conclusions",
                "sec_num": "6"
            },
            {
                "text": "Another finding of our work is that for personal and possessive pronouns, rather small numbers of labeled training data (about 100) seem to be sufficient for obtaining classifiers with a performance of about 80% F-measure. To our knowledge, this fact has not yet been reported in the literature.",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Conclusions",
                "sec_num": "6"
            },
            {
                "text": "While we restricted ourselves in this work to rather small sets of labeled training data, future work on Co-Training will include further experiments with larger data sets.",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Conclusions",
                "sec_num": "6"
            },
            {
                "text": "This filter applies only if none of the expressions is a pronoun. Otherwise, filtering on semantic class is not possible be-",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "",
                "sec_num": null
            },
            {
                "text": "http://www.cs.waikato.ac.nz/e ml/weka",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "",
                "sec_num": null
            }
        ],
        "back_matter": [
            {
                "text": "Acknowledgments. The work presented here has been partially funded by the German Ministry of Research and Technology as part of the EMBASSI project (01 IL 904 D/2, 01 IL 904 S 8), by Sony International (Europe) GmbH and by the Klaus Tschira Foundation. We would like to thank our annotators Anna Bj\u00f6rk Nikul\u00e1sd\u00f4ttir, Berenike Loos and Lutz Wind.",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "acknowledgement",
                "sec_num": null
            }
        ],
        "bib_entries": {
            "BIBREF0": {
                "ref_id": "b0",
                "title": "Evaluating automated and manual acquisition of anaphora resolution strategies",
                "authors": [
                    {
                        "first": "Chinatsu",
                        "middle": [],
                        "last": "Aone",
                        "suffix": ""
                    },
                    {
                        "first": "Scott",
                        "middle": [
                            "W"
                        ],
                        "last": "Bennett",
                        "suffix": ""
                    }
                ],
                "year": 1995,
                "venue": "Proceedings of the 33rd Annual Meeting of the Association for Computational Linguistics",
                "volume": "",
                "issue": "",
                "pages": "122--129",
                "other_ids": {},
                "num": null,
                "urls": [],
                "raw_text": "Chinatsu Aone and Scott W. Bennett. 1995. Evaluating automated and manual acquisition of anaphora reso- lution strategies. In Proceedings of the 33rd Annual Meeting of the Association for Computational Linguis- tics, Cambridge, Mass., 26-30 June 1995, pages 122- 129.",
                "links": null
            },
            "BIBREF1": {
                "ref_id": "b1",
                "title": "Combining labeled and unlabeled data with Co-Training",
                "authors": [
                    {
                        "first": "Avrim",
                        "middle": [],
                        "last": "Blum",
                        "suffix": ""
                    },
                    {
                        "first": "Tom",
                        "middle": [],
                        "last": "Mitchell",
                        "suffix": ""
                    }
                ],
                "year": 1998,
                "venue": "Proceedings of the 11th Annual Conference on Learning Theory",
                "volume": "",
                "issue": "",
                "pages": "92--100",
                "other_ids": {},
                "num": null,
                "urls": [],
                "raw_text": "Avrim Blum and Tom Mitchell. 1998. Combining la- beled and unlabeled data with Co-Training. In Pro- ceedings of the 11th Annual Conference on Learning Theory, Madison, Wisc., 24-26 July, 1998, pages 92- 100.",
                "links": null
            },
            "BIBREF2": {
                "ref_id": "b2",
                "title": "TnT -A statistical Part-of-Speech tagger",
                "authors": [
                    {
                        "first": "Thorsten",
                        "middle": [],
                        "last": "Brants",
                        "suffix": ""
                    }
                ],
                "year": 2000,
                "venue": "Proceedings of the 6th Conference on Applied Natural Language Processing",
                "volume": "",
                "issue": "",
                "pages": "224--231",
                "other_ids": {},
                "num": null,
                "urls": [],
                "raw_text": "Thorsten Brants. 2000. TnT -A statistical Part-of- Speech tagger. In Proceedings of the 6th Confer- ence on Applied Natural Language Processing, Seat- tle, Wash., 29 April -4 May 2000, pages 224-231.",
                "links": null
            },
            "BIBREF3": {
                "ref_id": "b3",
                "title": "Noun phrase coreference as clustering",
                "authors": [
                    {
                        "first": "Claire",
                        "middle": [],
                        "last": "Cardie",
                        "suffix": ""
                    },
                    {
                        "first": "Kiri",
                        "middle": [],
                        "last": "Wagstaff",
                        "suffix": ""
                    }
                ],
                "year": 1999,
                "venue": "Proceedings of the 1999 SIGDAT Conference on Empirical Methods in Natural Language Processing and Very Large Corpora",
                "volume": "",
                "issue": "",
                "pages": "82--89",
                "other_ids": {},
                "num": null,
                "urls": [],
                "raw_text": "Claire Cardie and Kiri Wagstaff. 1999. Noun phrase coreference as clustering. In Proceedings of the 1999 SIGDAT Conference on Empirical Methods in Natural Language Processing and Very Large Corpora, Col- lege Park, Md., 21-22 June 1999, pages 82-89.",
                "links": null
            },
            "BIBREF4": {
                "ref_id": "b4",
                "title": "Assessing agreement on classification tasks: The kappa statistic",
                "authors": [
                    {
                        "first": "Jean",
                        "middle": [],
                        "last": "Carletta",
                        "suffix": ""
                    }
                ],
                "year": 1996,
                "venue": "Computational Linguistics",
                "volume": "22",
                "issue": "2",
                "pages": "249--254",
                "other_ids": {},
                "num": null,
                "urls": [],
                "raw_text": "Jean Carletta. 1996. Assessing agreement on classifi- cation tasks: The kappa statistic. Computational Lin- guistics, 22(2):249-254.",
                "links": null
            },
            "BIBREF5": {
                "ref_id": "b5",
                "title": "Unsupervised models for named entity classification",
                "authors": [
                    {
                        "first": "Michael",
                        "middle": [],
                        "last": "Collins",
                        "suffix": ""
                    },
                    {
                        "first": "Yoram",
                        "middle": [],
                        "last": "Singer",
                        "suffix": ""
                    }
                ],
                "year": 1999,
                "venue": "Proceedings of the 1999 SIGDAT Conference on Empirical Methods in Natural Language Processing and Very Large Corpora",
                "volume": "",
                "issue": "",
                "pages": "100--110",
                "other_ids": {},
                "num": null,
                "urls": [],
                "raw_text": "Michael Collins and Yoram Singer. 1999. Unsupervised models for named entity classification. In Proceedings of the 1999 SIGDAT Conference on Empirical Meth- ods in Natural Language Processing and Very Large Corpora, College Park, Md., 21-22 June 1999, pages 100-110.",
                "links": null
            },
            "BIBREF6": {
                "ref_id": "b6",
                "title": "A statistical approach to anaphora resolution",
                "authors": [
                    {
                        "first": "Niyu",
                        "middle": [],
                        "last": "Ge",
                        "suffix": ""
                    },
                    {
                        "first": "John",
                        "middle": [],
                        "last": "Hale",
                        "suffix": ""
                    },
                    {
                        "first": "Eugene",
                        "middle": [],
                        "last": "Charniak",
                        "suffix": ""
                    }
                ],
                "year": 1998,
                "venue": "Proceedings of the Sixth Workshop on Very Large Corpora",
                "volume": "",
                "issue": "",
                "pages": "161--170",
                "other_ids": {},
                "num": null,
                "urls": [],
                "raw_text": "Niyu Ge, John Hale, and Eugene Charniak. 1998. A sta- tistical approach to anaphora resolution. In Proceed- ings of the Sixth Workshop on Very Large Corpora, Montr\u00e9al, Canada, pages 161-170.",
                "links": null
            },
            "BIBREF7": {
                "ref_id": "b7",
                "title": "Using decision trees for coreference resolution",
                "authors": [
                    {
                        "first": "F",
                        "middle": [],
                        "last": "Joseph",
                        "suffix": ""
                    },
                    {
                        "first": "Wendy",
                        "middle": [
                            "G"
                        ],
                        "last": "Mccarthy",
                        "suffix": ""
                    },
                    {
                        "first": "",
                        "middle": [],
                        "last": "Lehnert",
                        "suffix": ""
                    }
                ],
                "year": 1995,
                "venue": "Proceedings of the 14th International Joint Conference on Artificial Intelligence",
                "volume": "",
                "issue": "",
                "pages": "1050--1055",
                "other_ids": {},
                "num": null,
                "urls": [],
                "raw_text": "Joseph F. McCarthy and Wendy G. Lehnert. 1995. Us- ing decision trees for coreference resolution. In Pro- ceedings of the 14th International Joint Conference on Artificial Intelligence, Montr\u00e9al, Canada, 1995, pages 1050-1055.",
                "links": null
            },
            "BIBREF8": {
                "ref_id": "b8",
                "title": "Analyzing the effectiveness and applicability of Co-Training",
                "authors": [
                    {
                        "first": "Kamal",
                        "middle": [],
                        "last": "Nigam",
                        "suffix": ""
                    },
                    {
                        "first": "Rayid",
                        "middle": [],
                        "last": "Ghani",
                        "suffix": ""
                    }
                ],
                "year": 2000,
                "venue": "Proceedings of the 9th International Conference on Information and Knowledge Management",
                "volume": "",
                "issue": "",
                "pages": "86--93",
                "other_ids": {},
                "num": null,
                "urls": [],
                "raw_text": "Kamal Nigam and Rayid Ghani. 2000. Analyzing the ef- fectiveness and applicability of Co-Training. In Pro- ceedings of the 9th International Conference on Infor- mation and Knowledge Management, pages pp. 86-93.",
                "links": null
            },
            "BIBREF9": {
                "ref_id": "b9",
                "title": "Limitations of Co-Training for natural language learning from large datasets",
                "authors": [
                    {
                        "first": "David",
                        "middle": [],
                        "last": "Pierce",
                        "suffix": ""
                    },
                    {
                        "first": "Claire",
                        "middle": [],
                        "last": "Cardie",
                        "suffix": ""
                    }
                ],
                "year": 2001,
                "venue": "Proceedings of the 2001 Conference on Empirical Methods in Natural Language Processing",
                "volume": "",
                "issue": "",
                "pages": "1--9",
                "other_ids": {},
                "num": null,
                "urls": [],
                "raw_text": "David Pierce and Claire Cardie. 2001. Limitations of Co-Training for natural language learning from large datasets. In Proceedings of the 2001 Conference on Empirical Methods in Natural Language Processing, Pittsburgh, Pa., 3-4 June 2001, pages 1-9.",
                "links": null
            },
            "BIBREF10": {
                "ref_id": "b10",
                "title": "Applying Co-Training methods to statistical parsing",
                "authors": [
                    {
                        "first": "Anoop",
                        "middle": [],
                        "last": "Sarkar",
                        "suffix": ""
                    }
                ],
                "year": 2001,
                "venue": "Proceedings of the 2nd Conference of the North American Chapter of the Association for Computational Linguistics",
                "volume": "",
                "issue": "",
                "pages": "175--182",
                "other_ids": {},
                "num": null,
                "urls": [],
                "raw_text": "Anoop Sarkar. 2001. Applying Co-Training methods to statistical parsing. In Proceedings of the 2nd Con- ference of the North American Chapter of the Asso- ciation for Computational Linguistics, Pittsburgh, Pa., 2-7 June, 2001, pages 175-182.",
                "links": null
            },
            "BIBREF11": {
                "ref_id": "b11",
                "title": "A maximumentropy partial parser for unrestricted text",
                "authors": [
                    {
                        "first": "Wojciech",
                        "middle": [],
                        "last": "Skut",
                        "suffix": ""
                    },
                    {
                        "first": "Thorsten",
                        "middle": [],
                        "last": "Brants",
                        "suffix": ""
                    }
                ],
                "year": 1998,
                "venue": "6th Workshop on Very Large Corpora",
                "volume": "",
                "issue": "",
                "pages": "143--151",
                "other_ids": {},
                "num": null,
                "urls": [],
                "raw_text": "Wojciech Skut and Thorsten Brants. 1998. A maximum- entropy partial parser for unrestricted text. In 6th Workshop on Very Large Corpora, Montreal, Canada, pages 143-151.",
                "links": null
            },
            "BIBREF12": {
                "ref_id": "b12",
                "title": "A machine learning approach to coreference resolution of noun phrases",
                "authors": [],
                "year": 2001,
                "venue": "Computational Linguistics",
                "volume": "27",
                "issue": "4",
                "pages": "521--544",
                "other_ids": {},
                "num": null,
                "urls": [],
                "raw_text": "Wee Meng Soon, Hwee Tou Ng, and Daniel Chung Yong Lim. 2001. A machine learning approach to corefer- ence resolution of noun phrases. Computational Lin- guistics, 27(4):521-544.",
                "links": null
            },
            "BIBREF13": {
                "ref_id": "b13",
                "title": "The influence of minimum edit distance on reference resolution",
                "authors": [
                    {
                        "first": "Michael",
                        "middle": [],
                        "last": "Strube",
                        "suffix": ""
                    },
                    {
                        "first": "Stefan",
                        "middle": [],
                        "last": "Rapp",
                        "suffix": ""
                    },
                    {
                        "first": "Christoph",
                        "middle": [],
                        "last": "M\u00fcller",
                        "suffix": ""
                    }
                ],
                "year": 2002,
                "venue": "Proceedings of the 2002 Conference on Empirical Methods in Natural Language Processing",
                "volume": "",
                "issue": "",
                "pages": "",
                "other_ids": {},
                "num": null,
                "urls": [],
                "raw_text": "Michael Strube, Stefan Rapp, and Christoph M\u00fcller. 2002. The influence of minimum edit distance on ref- erence resolution. In Proceedings of the 2002 Confer- ence on Empirical Methods in Natural Language Pro- cessing, Philadelphia, Pa., 6-7 July 2002. To appear.",
                "links": null
            },
            "BIBREF14": {
                "ref_id": "b14",
                "title": "A modeltheoretic coreference scoring scheme",
                "authors": [
                    {
                        "first": "Marc",
                        "middle": [],
                        "last": "Vilain",
                        "suffix": ""
                    },
                    {
                        "first": "John",
                        "middle": [],
                        "last": "Burger",
                        "suffix": ""
                    },
                    {
                        "first": "John",
                        "middle": [],
                        "last": "Aberdeen",
                        "suffix": ""
                    }
                ],
                "year": 1995,
                "venue": "Proceedings fo the 6th Message Understanding Conference (MUC-6)",
                "volume": "",
                "issue": "",
                "pages": "45--52",
                "other_ids": {},
                "num": null,
                "urls": [],
                "raw_text": "Marc Vilain, John Burger, John Aberdeen, Dennis Con- nolly, and Lynette Hirschman. 1995. A model- theoretic coreference scoring scheme. In Proceedings fo the 6th Message Understanding Conference (MUC- 6), pages 45-52, San Mateo, Cal. Morgan Kaufmann.",
                "links": null
            },
            "BIBREF15": {
                "ref_id": "b15",
                "title": "The string-to-string correction problem",
                "authors": [
                    {
                        "first": "A",
                        "middle": [],
                        "last": "Robert",
                        "suffix": ""
                    },
                    {
                        "first": "Michael",
                        "middle": [
                            "J"
                        ],
                        "last": "Wagner",
                        "suffix": ""
                    },
                    {
                        "first": "",
                        "middle": [],
                        "last": "Fischer",
                        "suffix": ""
                    }
                ],
                "year": 1974,
                "venue": "Journal of the ACM",
                "volume": "21",
                "issue": "1",
                "pages": "168--173",
                "other_ids": {},
                "num": null,
                "urls": [],
                "raw_text": "Robert A. Wagner and Michael J. Fischer. 1974. The string-to-string correction problem. Journal of the ACM, 21(1):168-173.",
                "links": null
            }
        },
        "ref_entries": {
            "FIGREF4": {
                "uris": null,
                "num": null,
                "text": "F for All over iterations, baselines",
                "type_str": "figure"
            },
            "TABREF0": {
                "content": "<table/>",
                "num": null,
                "text": "Features used by Soon et al.",
                "type_str": "table",
                "html": null
            },
            "TABREF1": {
                "content": "<table/>",
                "num": null,
                "text": "Features used by Cardie and Wagstaff",
                "type_str": "table",
                "html": null
            },
            "TABREF3": {
                "content": "<table/>",
                "num": null,
                "text": "Views used for the experiments",
                "type_str": "table",
                "html": null
            }
        }
    }
}