File size: 72,606 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 | {
"paper_id": "P95-1017",
"header": {
"generated_with": "S2ORC 1.0.0",
"date_generated": "2023-01-19T08:33:51.897684Z"
},
"title": "Evaluating Automated and Manual Acquisition of Anaphora Resolution Strategies",
"authors": [
{
"first": "Chinatsu",
"middle": [],
"last": "Aone",
"suffix": "",
"affiliation": {
"laboratory": "",
"institution": "Systems Research and Applications Corporation (SRA)",
"location": {
"addrLine": "2000 15th Street North Arlington",
"postCode": "22201",
"region": "VA"
}
},
"email": ""
},
{
"first": "Scott",
"middle": [
"William"
],
"last": "Bennett",
"suffix": "",
"affiliation": {
"laboratory": "",
"institution": "Systems Research and Applications Corporation (SRA)",
"location": {
"addrLine": "2000 15th Street North Arlington",
"postCode": "22201",
"region": "VA"
}
},
"email": ""
}
],
"year": "",
"venue": null,
"identifiers": {},
"abstract": "We describe one approach to build an automatically trainable anaphora resolution system. In this approach, we use Japanese newspaper articles tagged with discourse information as training examples for a machine learning algorithm which employs the C4.5 decision tree algorithm by Quinlan (Quinlan, 1993). Then, we evaluate and compare the results of several variants of the machine learning-based approach with those of our existing anaphora resolution system which uses manually-designed knowledge sources. Finally, we compare our algorithms with existing theories of anaphora, in particular, Japanese zero pronouns.",
"pdf_parse": {
"paper_id": "P95-1017",
"_pdf_hash": "",
"abstract": [
{
"text": "We describe one approach to build an automatically trainable anaphora resolution system. In this approach, we use Japanese newspaper articles tagged with discourse information as training examples for a machine learning algorithm which employs the C4.5 decision tree algorithm by Quinlan (Quinlan, 1993). Then, we evaluate and compare the results of several variants of the machine learning-based approach with those of our existing anaphora resolution system which uses manually-designed knowledge sources. Finally, we compare our algorithms with existing theories of anaphora, in particular, Japanese zero pronouns.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Abstract",
"sec_num": null
}
],
"body_text": [
{
"text": "Anaphora resolution is an important but still difficult problem for various large-scale natural language processing (NLP) applications, such as information extraction and machine tr~slation. Thus far, no theories of anaphora have been tested on an empirical basis, and therefore there is no answer to the \"best\" anaphora resolution algorithm. I Moreover, an anaphora resolution system within an NLP system for real applications must handle:",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Introduction",
"sec_num": "1"
},
{
"text": "\u2022 degraded or missing input (no NLP system has complete lexicons, grammars, or semantic knowledge and outputs perfect results), and",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Introduction",
"sec_num": "1"
},
{
"text": "\u2022 different anaphoric phenomena in different domains, languages, and applications.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Introduction",
"sec_num": "1"
},
{
"text": "Thus, even if there exists a perfect theory, it might not work well with noisy input, or it would not cover all the anaphoric phenomena.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Introduction",
"sec_num": "1"
},
{
"text": "1Walker (Walker, 1989) compares Brennan, Friedman a~ad Pollard's centering approach (Brennan et al., 1987) with Hobbs' algorithm (Hohbs, 1976 ) on a theoretical basis.",
"cite_spans": [
{
"start": 8,
"end": 22,
"text": "(Walker, 1989)",
"ref_id": "BIBREF17"
},
{
"start": 84,
"end": 106,
"text": "(Brennan et al., 1987)",
"ref_id": "BIBREF5"
},
{
"start": 112,
"end": 141,
"text": "Hobbs' algorithm (Hohbs, 1976",
"ref_id": null
}
],
"ref_spans": [],
"eq_spans": [],
"section": "Introduction",
"sec_num": "1"
},
{
"text": "These requirements have motivated us to develop robust, extensible, and trainable anaphora resolution systems.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Introduction",
"sec_num": "1"
},
{
"text": "Previously (Aone and Mc-Kee, 1993) , we reported our data-driven multilingual anaphora resolution system, which is robust, exteusible, and manually trainable. It uses discourse knowledge sources (KS's) which are manually selected and ordered. (Henceforth, we call the system the Manually-Designed Resolver, or MDR.) We wanted to develop, however, truly automatically trainable systems, hoping to improve resolution performance and reduce the overhead of manually constructing and arranging such discourse data.",
"cite_spans": [
{
"start": 11,
"end": 34,
"text": "(Aone and Mc-Kee, 1993)",
"ref_id": null
}
],
"ref_spans": [],
"eq_spans": [],
"section": "Introduction",
"sec_num": "1"
},
{
"text": "In this paper, we first describe one approach we are taking to build an automatically trainable anaphora resolution system. In this approach, we tag corpora with discourse information, and use them as training examples for a machine learning algorithm. (Henceforth, we call the system the Machine Learning-based Resolver, or MLR.) Specifically, we have tagged Japanese newspaper articles about joint ventures and used the C4.5 decision tree algorithm by Quinlan (Quinlan, 1993) . Then, we evaluate and compare the results of the MLR with those produced by the MDR. Finally, we compare our algorithms with existing theories of anaphora, in particular, Japanese zero pronouns.",
"cite_spans": [
{
"start": 462,
"end": 477,
"text": "(Quinlan, 1993)",
"ref_id": "BIBREF13"
}
],
"ref_spans": [],
"eq_spans": [],
"section": "Introduction",
"sec_num": "1"
},
{
"text": "In this section, we first discuss corpora which we created for training and testing. Then, we describe the learning approach chosen, and discuss training features and training methods that we employed for our current experiments.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Applying a Machine Learning Technique to Anaphora Resolution",
"sec_num": "2"
},
{
"text": "In order to both train and evaluate an anaphora resolution system, we have been developing corpora which are tagged with discourse information. The tagging has been done using a GUI-based tool called the Discourse Tagging Tool (DTTool) according to \"The Discourse Tagging Guidelines\" we have developed. 2 The tool allows a user to link an anaphor with its antecedent and specify the type of the anaphor (e.g. pronouns, definite NP's, etc.). The tagged result can be written out to an SGMLmarked file, as shown in Figure 1 . For our experiments, we have used a discoursetagged corpus which consists of Japanese newspaper articles about joint ventures. The tool lets a user define types of anaphora as necessary. The anaphoric types used to tag this corpus are shown in Table 1. NAME anaphora are tagged when proper names are used anaphorically. For example, in Figure 1 , \"Yamaichi (ID=3)\" and \"Sony-Prudential (ID=5)\" referring back to \"Yamaichi Shouken (ID=4)\" (Yamaichi Securities) and \"Sony-Prudential Seimeihoken (ID=6)\" (Sony-Prudential Life Insurance) respectively are NAME anaphora. NAME anaphora in Japanese are different from those in English in that any combination of characters in an antecedent can be NAME anaphora as long as the character order is preserved (e.g. \"abe\" can be an anaphor of \"abcde\").",
"cite_spans": [],
"ref_spans": [
{
"start": 513,
"end": 521,
"text": "Figure 1",
"ref_id": null
},
{
"start": 768,
"end": 776,
"text": "Table 1.",
"ref_id": "TABREF0"
},
{
"start": 860,
"end": 868,
"text": "Figure 1",
"ref_id": null
}
],
"eq_spans": [],
"section": "Training and Test Corpora",
"sec_num": "2.1"
},
{
"text": "Japanese definite NPs (i.e. DNP anaphora) are those prefixed by \"dou\" (literally meaning \"the same\"), \"ryou\" (literally meaning \"the two\"), and deictic determiners like \"kono\"(this) and \"sono\" (that). For example, \"dou-sha\" is equivalent to \"the company\", and \"ryou-koku\" to \"the two countries\". The DNP anaphora with \"dou\" and \"ryou\" prefixes are characteristic of written, but not spoken, Japanese texts.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Training and Test Corpora",
"sec_num": "2.1"
},
{
"text": "Unlike English, Japanese has so-called zero pronouns, which are not explicit in the text. In these cases, the DTTool lets the user insert a \"Z\" marker just before the main predicate of the zero pronoun to indicate the existence of the anaphor. We made distinction between QZPRO and ZPRO when tagging zero pronouns. QZPRO (\"quasi-zero pronoun\") is chosen when a sentence has multiple clauses (subordinate or coordinate), and the zero pronouns in these clauses refer back to the subject of the initial clause in the same sentence, as shown in Figure 2 .",
"cite_spans": [],
"ref_spans": [
{
"start": 541,
"end": 549,
"text": "Figure 2",
"ref_id": null
}
],
"eq_spans": [],
"section": "Training and Test Corpora",
"sec_num": "2.1"
},
{
"text": "The anaphoric types are sub-divided according to more semantic criteria such as organizations, people, locations, etc. This is because the current application of our multilingual NLP system is information extraction , i.e. extracting from texts information about which organizations are forming joint ventures with whom. Thus, resolving certain anaphora (e.g. various ways to refer back to organizations) affects the task performance more than others, as we previously reported (Aone, 1994) . Our goal is to customize and evaluate anaphora resolution systems according to the types of anaphora when necessary.",
"cite_spans": [
{
"start": 478,
"end": 490,
"text": "(Aone, 1994)",
"ref_id": "BIBREF4"
}
],
"ref_spans": [],
"eq_spans": [],
"section": "Training and Test Corpora",
"sec_num": "2.1"
},
{
"text": "2Our work on the DTTool and tagged corpora was reported in a recent paper (Aone and Bennett, 1994) .",
"cite_spans": [
{
"start": 74,
"end": 98,
"text": "(Aone and Bennett, 1994)",
"ref_id": "BIBREF0"
}
],
"ref_spans": [],
"eq_spans": [],
"section": "Training and Test Corpora",
"sec_num": "2.1"
},
{
"text": "While several inductive learning approaches could have been taken for construction of the trainable anaphoric resolution system, we found it useful to be able to observe the resulting classifier in the form of a decision tree. The tree and the features used could most easily be compared to existing theories. Therefore, our initial approach has been to employ Quinlan's C4.5 algorithm at the heart of our classification approach. We discuss the features used for learning below and go on to discuss the training methods and how the resulting tree is used in our anaphora resolution algorithm.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Learning Method",
"sec_num": "2.2"
},
{
"text": "In our current machine learning experiments, we have taken an approach where we train a decision tree by feeding feature vectors for pairs of an anaphor and its possible antecedent. Currently we use 66 features, and they include lezical (e.g. category), syntactic (e.g. grammatical role), semantic (e.g. semantic class), and positional (e.g. distance between anaphor and antecedent) features. Those features can be either unary features (i.e. features of either an anaphor or an antecedent such as syntactic number values) or binary features (i.e. features concerning relations between the pairs such as the positional relation between an anaphor and an antecedent.) We started with the features used by the MDR, generalized them, and added new features. The features that we employed are common across domains and languages though the feature values may change in different domains or languages. Example of training features are shown in Table 2 .",
"cite_spans": [],
"ref_spans": [
{
"start": 939,
"end": 946,
"text": "Table 2",
"ref_id": "TABREF1"
}
],
"eq_spans": [],
"section": "Training Features",
"sec_num": "2.3"
},
{
"text": "The feature values are obtained automatically by processing a set of texts with our NLP system, which performs lexical, syntactic and semantic analysis and then creates discourse markers (Kamp, 1981) for each NP and S. 3 Since discourse markers store the output of lexical, syntactic and semantic processing, the feature vectors are automatically calculated from them. Because the system output is not always perfect (especially given the complex newspaper articles), however, there is some noise in feature values.",
"cite_spans": [
{
"start": 187,
"end": 199,
"text": "(Kamp, 1981)",
"ref_id": "BIBREF10"
}
],
"ref_spans": [],
"eq_spans": [],
"section": "Training Features",
"sec_num": "2.3"
},
{
"text": "We have employed different training methods using three parameters: anaphoric chains, anaphoric type identification, and confidence factors.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Training Methods",
"sec_num": "2.4"
},
{
"text": "The anaphoric chain parameter is used in selecting training examples. When this parameter is on, we select a set of positive training examples and a set of negative training examples for each anaphor in a text in the following way: <CORe: m=' I\"><COREF n~'4\">ttl--lEff-</mR~:<u.~J-m='s'>y-'--\u2022 ~')l,~Y:,,~,)t,\u00a2.@~l~ (~P,-'ll~l~:.~t, :\u00a24t. lr)~) <CORE]: m='O\" rcPE='~ RB:='i\"></COR~>III@b~. ~)q~'~6<COR~ ZD='2e rVPE='ZPm-t~-\" REFf'I\"></COREF>~Ii 3\"~. <CORe: ZD='~' WRf\"NANE--OR6\" RB:f'4\">ttI--</COE~<COREF ~\"8\">q~,l,~ltC)~e't-\"~.'3tt~ttll~:~'~'& </COR~<COR~ m='s\" WR='tt~E-O~ REFf\"#'>y-'---. ~')t,-~>-b,,v)l,</mR~{:~-, <COmF n)=\"\u00a2' WPE='Dm\" REF='8\"> C r~ 5, ~-7\" I, <,'CUT~ ~ <CORBF m='9\" WR='ZT4~O-O~ 8EEf'5\"> </OR~ ff -~ T <CO~ m=\" ~o\" TYR='~O-U~ RE~='5\"> Figure 1 : Text Tagged with Discourse Information using SGML Positive training examples are those anaphorantecedent pairs whose anaphor is directly linked to its antecedent in the tagged corpus and also whose anaphor is paired with one of the antecedents on the anaphoric chain, i.e. the transitive closure between the anaphor and the first mention of the antecedent. For example, if B refers to A and C refers to B, C-A is a positive training example as well as B-A and C-B.",
"cite_spans": [],
"ref_spans": [
{
"start": 757,
"end": 765,
"text": "Figure 1",
"ref_id": null
}
],
"eq_spans": [],
"section": "Training Methods",
"sec_num": "2.4"
},
{
"text": "Tags DNP DNP-F DNP-L DNP-ORG DNP-P DNP-T DNP-BOTH DNP-BOTH-ORG DNP-BOTH-L DNP-BOTH-P REFLEXIVE NAME NAME-F NAME-L NAME-ORG NAME-P DPRO LOCI TIMEI QZPRO QZPRO-ORG QZPRO-P ZPRO ZPRO-IMP ZPRO-ORG ZPRO-P",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Training Methods",
"sec_num": "2.4"
},
{
"text": "Negative training examples are chosen by pairing an anaphor with all the possible antecedents in a text except for those on the transitive closure described above. Thus, if there are possible antecedents in the text which are not in the C-B-A transitive closure, say D, C-D and B-D are negative training examples.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Training Methods",
"sec_num": "2.4"
},
{
"text": "When the anaphoric chain parameter is off, only those anaphor-antecedent pairs whose anaphora are directly linked to their antecedents in the corpus are considered as positive examples. Because of the way in which the corpus was tagged (according to our tagging guidelines), an anaphor is linked to the most recent antecedent, except for a zero pronoun, which is linked to its most recent overt antecedent. In other words, a zero pronoun is never linked to another zero pronoun.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Training Methods",
"sec_num": "2.4"
},
{
"text": "The anaphoric type identification parameter is utilized in training decision trees. With this parameter on, a decision tree is trained to answer \"no\" when a pair of an anaphor and a possible antecedent are not co-referential, or answer the anaphoric type when they are co-referential. If the parameter is off, a binary decision tree is trained to answer just \"yes\" or \"no\" and does not have to answer the types of anaphora.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Training Methods",
"sec_num": "2.4"
},
{
"text": "The confidence factor parameter (0-100) is used in pruning decision trees. With a higher confidence factor, less pruning of the tree is performed, and thus it tends to overfit the training examples. With a lower confidence factor, more pruning is performed, resulting in a smaller, more generalized tree. We used confidence factors of 25, 50, 75 and 100%. The anaphoric chain parameter described above was employed because an anaphor may have more than one \"correct\" antecedent, in which case there is no absolute answer as to whether one antecedent is better than the others. The decision tree approach we have taken may thus predict more than one antecedent to pair with a given anaphor. Currently, confidence values returned from the decision tree are employed when it is desired that a single antecedent be selected for a given anaphor. We are experimenting with techniques to break ties in confidence values from the tree. One approach is to use a particular bias, say, in preferring the antecedent closest to the anaphor among those with the highest confidence (as in the results reported here). Although use of the confidence values from the tree works well in practice, these values were only intended as a heuristic for pruning in Quinlan's C4.5. We have plans to use cross-validation across the training set as a method of determining error-rates by which to prefer one predicted antecedent over another.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Training Methods",
"sec_num": "2.4"
},
{
"text": "Another approach is to use a hybrid method where a preference-trained decision tree is brought in to supplement the decision process. Preference-trained trees, like that discussed in Connolly et al. (Connolly et al., 1994) , are trained by presenting the learning algorithm with examples of when one anaphorantecedent pair should be preferred over another. Despite the fact that such trees are learning preferences, they may not produce sufficient preferences to permit selection of a single best anaphor-antecedent combination (see the \"Related Work\" section below).",
"cite_spans": [
{
"start": 199,
"end": 222,
"text": "(Connolly et al., 1994)",
"ref_id": "BIBREF6"
}
],
"ref_spans": [],
"eq_spans": [],
"section": "Training Methods",
"sec_num": "2.4"
},
{
"text": "In this section, we first discuss how we configured and developed the MLRs and the MDR for testing. Next, we describe the scoring methods used, and then the testing results of the MLRs and the MDR. In this paper, we report the results of the four types of anaphora, namely NAME-ORG, QZPRO-ORG, DNP-ORG, and ZPRO-ORG, since they are the majority of the anaphora appearing in the texts and most important for the current domain (i.e. joint ventures) and application (i.e. information extraction).",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Testing",
"sec_num": "3"
},
{
"text": "To build MLRs, we first trained decision trees with 1971 anaphora 4 (of which 929 were NAME-ORG; 546 QZPRO-ORG; 87 DNP-ORG; 282 ZPRO-ORG) in 295 training texts. The six MLRs using decision trees with different parameter combinations are described in Table 3 .",
"cite_spans": [],
"ref_spans": [
{
"start": 250,
"end": 257,
"text": "Table 3",
"ref_id": "TABREF2"
}
],
"eq_spans": [],
"section": "Testing the MLRa",
"sec_num": "3.1"
},
{
"text": "Then, we trained decision trees in the MLR-2 configuration with varied numbers of training texts, namely 50, 100, 150,200 and 250 texts. This is done to find out the minimum number of training texts to achieve the optimal performance.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Testing the MLRa",
"sec_num": "3.1"
},
{
"text": "The same training texts used by the MLRs served as development data for the MDR. Because the NLP system is used for extracting information about joint ventures, the MDR was configured to handle only the crucial subset of anaphoric types for this experiment, namely all the name anaphora and zero pronouns and the definite NPs referring to organizations (i.e. DNP-ORG). The MDR applies different sets of generators, filters and orderers to resolve different anaphoric types . A generator generates a set of possible antecedent hypotheses for each anaphor, while a filter eliminates *In both training and testing, we did not include anaphora which refer to multiple discontinuous antecedents. unlikely hypotheses from the set. An orderer ranks hypotheses in a preference order if there is more than one hypothesis left in the set after applying all the applicable filters. Table 4 shows KS's employed for the four anaphoric types.",
"cite_spans": [],
"ref_spans": [
{
"start": 871,
"end": 878,
"text": "Table 4",
"ref_id": "TABREF4"
}
],
"eq_spans": [],
"section": "Testing the MDR",
"sec_num": "3.2"
},
{
"text": "We used recall and precision metrics, as shown in Table 5 , to evaluate the performance of anaphora resolution. It is important to use both measures because one can build a high recall-low precision system or a low recall-high precision system, neither of which may be appropriate in certain situations.",
"cite_spans": [],
"ref_spans": [
{
"start": 50,
"end": 57,
"text": "Table 5",
"ref_id": "TABREF3"
}
],
"eq_spans": [],
"section": "Scoring Method",
"sec_num": "3.3"
},
{
"text": "The NLP system sometimes fails to create discourse markers exactly corresponding to anaphora in texts due to failures of hxical or syntactic processing. In order to evaluate the performance of the anaphora resolution systems themselves, we only considered anaphora whose discourse markers were identified by the NLP system in our evaluation. Thus, the system performance evaluated against all the anaphora in texts could be different. ",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Scoring Method",
"sec_num": "3.3"
},
{
"text": "The testing was done using 1359 anaphora (of which 1271 were one of the four anaphoric types) in 200 blind test texts for both the MLRs and the MDR. It should be noted that both the training and testing texts are newspaper articles about joint ventures, and that each article always talks about more than one organization. Thus, finding antecedents of organizational anaphora is not straightforward. Table 6 shows the results of six different MLRs and the MDR for the four types of anaphora, while Table 7 shows the results of the MLR-2 with different sizes of training examples, 4 Evaluation 4.1 The MLRs vs. the MDR Using F-measures 5 as an indicator for overall performance, the MLRs with the chain parameters turned on and type identification turned off (i.e. MLR-1, 2, 3, and 4) performed the best. MLR-1, 2, 3, 4, and 5 all exceeded the MDR in overall performance based on F-measure.",
"cite_spans": [],
"ref_spans": [
{
"start": 400,
"end": 407,
"text": "Table 6",
"ref_id": "TABREF5"
},
{
"start": 498,
"end": 505,
"text": "Table 7",
"ref_id": "TABREF6"
}
],
"eq_spans": [],
"section": "Testing Results",
"sec_num": "3.4"
},
{
"text": "Both the MLRs and the MDR used the character subsequence, the proper noun category, and the semantic class feature values for NAME-ORG anaphora (in MLR-5, using anaphoric type identification). It is interesting to see that the MLR additionally uses the topicalization feature before testing the semantic class feature. This indicates that, information theoretically, if the topicalization feature is present, the semantic class feature is not needed for the classification. The performance of NAME-ORG is better than other anaphoric phenomena because the character subsequence feature has very high antecedent predictive power.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Testing Results",
"sec_num": "3.4"
},
{
"text": "Changing the three parameters in the MLRs caused changes in anaphora resolution performance. As Table 6 shows, using anaphoric chains without anaphoric type identification helped improve the MLRs. Our experiments with the confidence factor parameter indicates the trade off between recall and precision. With 100% confidence factor, which means no pruning of the tree, the tree overfits the examples, and leads to spurious uses of features such as the number of sentences between an anaphor and an antecedent near the leaves of the generated tree. This causes the system to attempt more anaphor resolutions albeit with lower precision. Conversely, too much pruning can also yield poorer results.",
"cite_spans": [],
"ref_spans": [
{
"start": 96,
"end": 103,
"text": "Table 6",
"ref_id": "TABREF5"
}
],
"eq_spans": [],
"section": "Evaluation of the MLIts",
"sec_num": "4.1.1"
},
{
"text": "MLR-5 illustrates that when anaphoric type identification is turned on the MLR's performance drops SF-measure is calculated by:",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Evaluation of the MLIts",
"sec_num": "4.1.1"
},
{
"text": "F= (~2+1.0) \u00d7 P x R #2 x P+R",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Evaluation of the MLIts",
"sec_num": "4.1.1"
},
{
"text": "where P is precision, R is recall, and /3 is the relative importance given to recall over precision. In this case, = 1.0.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Evaluation of the MLIts",
"sec_num": "4.1.1"
},
{
"text": "ZPRO-ORG but still exceeds that of the MDR. MLR-6 shows the effect of not training on anaphoric chains. It results in poorer performance than the MLR-1, 2, 3, 4, and 5 configurations and the MDR. One of the advantages of the MLRs is that due to the number of different anaphoric types present in the training data, they also learned classifiers for several additional anaphoric types beyond what the MDR could handle. While additional coding would have been required for each of these types in the MDR, the MLRs picked them up without additional work. The additional anaphoric types included DPRO, REFLEXIVE, and TIMEI (cf. Table 1). Another advantage is that, unlike the MDR, whose features are hand picked, the MLRs automatically select and use necessary features.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "QZPRO-ORG",
"sec_num": null
},
{
"text": "We suspect that the poorer performance of ZPRO-OR(; and DNP-ORG may be due to the following deficiency of the current MLR algorithms: Because anaphora resolution is performed in a \"batch mode\" for the MLRs, there is currently no way to percolate the information on an anaphor-antecedent link found by a system after each resolution. For example, if a zero pronoun (Z-2) refers to another zero pronoun (Z-l), which in turn refers to an overt NP, knowing which is the antecedent of Z-1 may be important for Z-2 to resolve its antecedent correctly. However, such information is not available to the MLRs when resolving Z-2.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Generators",
"sec_num": null
},
{
"text": "One advantage of the MDR is that a tagged training corpus is not required for hand-coding the resolution algorithms. Of course, such a tagged corpus is necessary to evaluate system performance quantitatively and is also useful to consult with during algorithm construction.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Evaluation of the MDR",
"sec_num": "4.1.2"
},
{
"text": "However, the MLR results seem to indicate the limitation of the MDR in the way it uses orderer KS's. Currently, the MDR uses an ordered list of multiple orderer KS's for each anaphoric type (cf . Table 4) , where the first applicable orderer KS in the list is used to pick the best antecedent when there is more than one possibility. Such selection ignores the fact that even anaphora of the same type may use different orderers (i.e. have different preferences), depending on the types of possible antecedents and on the context in which the particular anaphor was used in the text. Table 7 indicates that with even 50 training texts, the MLR achieves better performance than the MDR. Performance seems to reach a plateau at about 250 training examples with a F-measure of around 77.4.",
"cite_spans": [],
"ref_spans": [
{
"start": 194,
"end": 205,
"text": ". Table 4)",
"ref_id": "TABREF4"
},
{
"start": 585,
"end": 592,
"text": "Table 7",
"ref_id": "TABREF6"
}
],
"eq_spans": [],
"section": "Evaluation of the MDR",
"sec_num": "4.1.2"
},
{
"text": "Anaphora resolution systems for English texts based on various machine learning algorithms, including a decision tree algorithm, are reported in Connolly et al. (Connolly et al., 1994) . Our approach is different from theirs in that their decision tree identifies which of the two possible antecedents for a given anaphor is \"better\". The assumption seems to be that the closest antecedent is the \"correct\" antecedent. However, they note a problem with their decision tree in that it is not guaranteed to return consistent classifications given that the \"preference\" relationship between two possible antecedents is not transitive. Soderland and Lehnert's machine learning-based information extraction system (Soderland and Lehnert, 1994) is used specifically for filling particular templates from text input. Although a part of its task is to merge multiple referents when they corefer (i.e. anaphora resolution), it is hard to evaluate how their anaphora resolution capability compares with ours, since it is not a separate module. The only evaluation result provided is their extraction result. Our anaphora resolution system is modular, and can be used for other NLP-based applications such as machine translation. Soderland and Lehnert's approach relies on a large set of filled templates used for training. Domain-specific features from those templates are employed for the learning. Consequently, the learned classifiers are very domain-specific, and thus the approach relies on the availability of new filled template sets for porting to other domains. While some such template sets exist, such as those assembled for the Message Understanding Conferences, collecting such large amounts of training data for each new domain may be impractical.",
"cite_spans": [
{
"start": 145,
"end": 184,
"text": "Connolly et al. (Connolly et al., 1994)",
"ref_id": "BIBREF6"
},
{
"start": 709,
"end": 738,
"text": "(Soderland and Lehnert, 1994)",
"ref_id": null
}
],
"ref_spans": [],
"eq_spans": [],
"section": "Related Work",
"sec_num": "5"
},
{
"text": "Zero pronoun resolution for machine translation reported by Nakaiwa and Ikehara (Nakaiwa and Ikehara, 1992 ) used only semantic attributes of verbs in a restricted domain. The small test results (102 sentences from 29 articles) had high success rate of 93%. However, the input was only the first paragraphs of newspaper articles which contained relatively short sentences. Our anaphora resolution systems reported here have the advantages of domain-independence and full-text handling without the need for creating an extensive domain knowledge base.",
"cite_spans": [
{
"start": 60,
"end": 106,
"text": "Nakaiwa and Ikehara (Nakaiwa and Ikehara, 1992",
"ref_id": "BIBREF11"
}
],
"ref_spans": [],
"eq_spans": [],
"section": "Related Work",
"sec_num": "5"
},
{
"text": "Various theories of Japanese zero pronouns have been proposed by computational linguists, for example, Kameyama (Kameyama, 1988) and Walker et aL (Walker et al., 1994) . Although these theories are based on dialogue examples rather than texts, \"features\" used by these theories and those by the decision trees overlap interestingly. For example, Walker et ai. proposes the following ranking scheme to select antecedents of zero pronouns.",
"cite_spans": [
{
"start": 112,
"end": 128,
"text": "(Kameyama, 1988)",
"ref_id": "BIBREF9"
},
{
"start": 146,
"end": 167,
"text": "(Walker et al., 1994)",
"ref_id": "BIBREF16"
}
],
"ref_spans": [],
"eq_spans": [],
"section": "Related Work",
"sec_num": "5"
},
{
"text": "(GRAMMATICAL or ZERO) TOPIC > EMPATHY > SUBJECT > OBJECT2 > OBJECT > OTHERS",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Related Work",
"sec_num": "5"
},
{
"text": "In examining decision trees produced with anaphoric type identification turned on, the following features were used for QZPRO-ORG in this order: topicalization, distance between an anaphor and an antecedent, semantic class of an anaphor and an antecedent, and subject NP. We plan to analyze further the features which the decision tree has used for zero pronouns and compare them with these theories.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Related Work",
"sec_num": "5"
},
{
"text": "This paper compared our automated and manual acquisition of anaphora resolution strategies, and reported optimistic results for the former. We plan to continue to improve machine learning-based system performance by introducing other relevant features. For example, discourse structure information (Passonneau and Litman, 1993; Hearst, 1994) , if obtained reliably and automatically, will be another useful domain-independent feature. In addition, we will explore the possibility of combining machine learning results with manual encoding of discourse knowledge. This can be accomplished by allowing the user to interact with the produced classifters, tracing decisions back to particular examples and allowing users to edit features and to evaluate the efficacy of changes.",
"cite_spans": [
{
"start": 298,
"end": 327,
"text": "(Passonneau and Litman, 1993;",
"ref_id": "BIBREF12"
},
{
"start": 328,
"end": 341,
"text": "Hearst, 1994)",
"ref_id": "BIBREF7"
}
],
"ref_spans": [],
"eq_spans": [],
"section": "Summary and Future Work",
"sec_num": null
},
{
"text": "Existence of zero pronouns in sentences is detected by the syntax module, and discourse maxkers are created for them.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "",
"sec_num": null
}
],
"back_matter": [],
"bib_entries": {
"BIBREF0": {
"ref_id": "b0",
"title": "Discourse Tagging Tool and Discourse-tagged Multilingual Corpora",
"authors": [
{
"first": "Chinatsu",
"middle": [],
"last": "Aone",
"suffix": ""
},
{
"first": "Scott",
"middle": [
"W"
],
"last": "Bennett",
"suffix": ""
}
],
"year": 1994,
"venue": "Proceedings of International Workshop on Sharable Natural Language Resources (SNLR)",
"volume": "",
"issue": "",
"pages": "",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Chinatsu Aone and Scott W. Bennett. 1994. Dis- course Tagging Tool and Discourse-tagged Mul- tilingual Corpora. In Proceedings of Interna- tional Workshop on Sharable Natural Language Resources (SNLR).",
"links": null
},
"BIBREF2": {
"ref_id": "b2",
"title": "Language-Independent Anaphora Resolution System for Understanding Multilingual Texts",
"authors": [],
"year": null,
"venue": "Proceedings of 31st Annual Meeting of the ACL",
"volume": "",
"issue": "",
"pages": "",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Language-Independent Anaphora Resolution Sys- tem for Understanding Multilingual Texts. In Proceedings of 31st Annual Meeting of the ACL.",
"links": null
},
"BIBREF3": {
"ref_id": "b3",
"title": "SRA: Description of the SOLOMON System as Used for MUC-5",
"authors": [
{
"first": "Chinatsu",
"middle": [],
"last": "Aone",
"suffix": ""
},
{
"first": "Sharon",
"middle": [],
"last": "Flank",
"suffix": ""
},
{
"first": "Paul",
"middle": [],
"last": "Krause",
"suffix": ""
},
{
"first": "Doug",
"middle": [],
"last": "Mckee",
"suffix": ""
}
],
"year": 1993,
"venue": "Proceedings of Fourth Message Understanding Conference (MUC-5)",
"volume": "",
"issue": "",
"pages": "",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Chinatsu Aone, Sharon Flank, Paul Krause, and Doug McKee. 1993. SRA: Description of the SOLOMON System as Used for MUC-5. In Pro- ceedings of Fourth Message Understanding Con- ference (MUC-5).",
"links": null
},
"BIBREF4": {
"ref_id": "b4",
"title": "Customizing and Evaluating a Multilingual Discourse Module",
"authors": [
{
"first": "Chinatsu",
"middle": [],
"last": "Aone",
"suffix": ""
}
],
"year": 1994,
"venue": "Proceedings of the 15th International Conference on Computational Linguistics (COLING)",
"volume": "",
"issue": "",
"pages": "",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Chinatsu Aone. 1994. Customizing and Evaluating a Multilingual Discourse Module. In Proceedings of the 15th International Conference on Compu- tational Linguistics (COLING).",
"links": null
},
"BIBREF5": {
"ref_id": "b5",
"title": "A Centering Approach to Pronouns",
"authors": [
{
"first": "Susan",
"middle": [],
"last": "Brennan",
"suffix": ""
},
{
"first": "Marilyn",
"middle": [],
"last": "Friedman",
"suffix": ""
},
{
"first": "Carl",
"middle": [],
"last": "Pollard",
"suffix": ""
}
],
"year": 1987,
"venue": "Proceedings of 25th Annual Meeting of the ACL",
"volume": "",
"issue": "",
"pages": "",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Susan Brennan, Marilyn Friedman, and Carl Pol- lard. 1987. A Centering Approach to Pronouns. In Proceedings of 25th Annual Meeting of the ACL.",
"links": null
},
"BIBREF6": {
"ref_id": "b6",
"title": "A Machine Learning Approach to Anaphoric Reference",
"authors": [
{
"first": "Dennis",
"middle": [],
"last": "Connolly",
"suffix": ""
},
{
"first": "John",
"middle": [
"D"
],
"last": "Burger",
"suffix": ""
},
{
"first": "David",
"middle": [
"S"
],
"last": "Day",
"suffix": ""
}
],
"year": 1994,
"venue": "Proceedings of International Conference on New Methods in Language Processing (NEMLAP)",
"volume": "",
"issue": "",
"pages": "",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Dennis Connolly, John D. Burger, and David S. Day. 1994. A Machine Learning Approach to Anaphoric Reference. In Proceedings of Interna- tional Conference on New Methods in Language Processing (NEMLAP).",
"links": null
},
"BIBREF7": {
"ref_id": "b7",
"title": "Multi-Paragraph Segmentation of Expository Text",
"authors": [
{
"first": "Marti",
"middle": [
"A"
],
"last": "Hearst",
"suffix": ""
}
],
"year": 1994,
"venue": "Proceedings of 32nd Annual Meeting of the ACL",
"volume": "",
"issue": "",
"pages": "",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Marti A. Hearst. 1994. Multi-Paragraph Segmenta- tion of Expository Text. In Proceedings of 32nd Annual Meeting of the ACL.",
"links": null
},
"BIBREF8": {
"ref_id": "b8",
"title": "Pronoun Resolution",
"authors": [
{
"first": "R",
"middle": [],
"last": "Jerry",
"suffix": ""
},
{
"first": "",
"middle": [],
"last": "Hobbs",
"suffix": ""
}
],
"year": 1976,
"venue": "",
"volume": "",
"issue": "",
"pages": "",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Jerry R. Hobbs. 1976. Pronoun Resolution. Tech- nical Report 76-1, Department of Computer Sci- ence, City College, City University of New York.",
"links": null
},
"BIBREF9": {
"ref_id": "b9",
"title": "Japanese Zero Pronominal Binding, where Syntax and Discourse Meet",
"authors": [
{
"first": "Megumi",
"middle": [],
"last": "Kameyama",
"suffix": ""
}
],
"year": 1988,
"venue": "Papers from the Second International Worksho on Japanese Syntax",
"volume": "",
"issue": "",
"pages": "",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Megumi Kameyama. 1988. Japanese Zero Pronom- inal Binding, where Syntax and Discourse Meet. In Papers from the Second International Worksho on Japanese Syntax.",
"links": null
},
"BIBREF10": {
"ref_id": "b10",
"title": "A Theory of Truth and Semantic Representation",
"authors": [
{
"first": "Hans",
"middle": [],
"last": "Kamp",
"suffix": ""
}
],
"year": 1981,
"venue": "Formal Methods in the Study of Language. Mathematical Centre",
"volume": "",
"issue": "",
"pages": "",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Hans Kamp. 1981. A Theory of Truth and Semantic Representation. In J. Groenendijk et al., editors, Formal Methods in the Study of Language. Math- ematical Centre, Amsterdam.",
"links": null
},
"BIBREF11": {
"ref_id": "b11",
"title": "Zero Pronoun Resolution in a Japanese to English Machine Translation Systemby using Verbal Semantic Attribute",
"authors": [
{
"first": "Hiromi",
"middle": [],
"last": "Nakaiwa",
"suffix": ""
},
{
"first": "Satoru",
"middle": [],
"last": "Ikehara",
"suffix": ""
}
],
"year": 1992,
"venue": "Proceedings of the Fourth Conference on Applied Natural Language Processing",
"volume": "",
"issue": "",
"pages": "",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Hiromi Nakaiwa and Satoru Ikehara. 1992. Zero Pronoun Resolution in a Japanese to English Ma- chine Translation Systemby using Verbal Seman- tic Attribute. In Proceedings of the Fourth Con- ference on Applied Natural Language Processing.",
"links": null
},
"BIBREF12": {
"ref_id": "b12",
"title": "Intention-Based Segmentation: Human Reliability and Correlation with Linguistic Cues",
"authors": [
{
"first": "Rebecca",
"middle": [
"J"
],
"last": "Passonneau",
"suffix": ""
},
{
"first": "Diane",
"middle": [
"J"
],
"last": "Litman",
"suffix": ""
}
],
"year": 1993,
"venue": "Proceedings of 31st Annual Meeting of the ACL",
"volume": "",
"issue": "",
"pages": "",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Rebecca J. Passonneau and Diane J. Litman. 1993. Intention-Based Segmentation: Human Reliabil- ity and Correlation with Linguistic Cues. In Pro- ceedings of 31st Annual Meeting of the ACL.",
"links": null
},
"BIBREF13": {
"ref_id": "b13",
"title": "C~.5: Programs forMachine Learning",
"authors": [
{
"first": "J",
"middle": [],
"last": "",
"suffix": ""
},
{
"first": "Ross",
"middle": [],
"last": "Quinlan",
"suffix": ""
}
],
"year": 1993,
"venue": "",
"volume": "",
"issue": "",
"pages": "",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "J. Ross quinlan. 1993. C~.5: Programs forMachine Learning. Morgan Kaufmann Publishers.",
"links": null
},
"BIBREF15": {
"ref_id": "b15",
"title": "Corpus-driven Knowledge Acquisition for Discourse Analysis",
"authors": [],
"year": null,
"venue": "Proceedings of AAAI",
"volume": "",
"issue": "",
"pages": "",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Corpus-driven Knowledge Acquisition for Dis- course Analysis. In Proceedings of AAAI.",
"links": null
},
"BIBREF16": {
"ref_id": "b16",
"title": "Japanese Discourse and the Process of Centering",
"authors": [
{
"first": "Marilyn",
"middle": [],
"last": "Walker",
"suffix": ""
},
{
"first": "Masayo",
"middle": [],
"last": "Iida",
"suffix": ""
},
{
"first": "Sharon",
"middle": [],
"last": "Cote",
"suffix": ""
}
],
"year": 1994,
"venue": "Computational Linguistics",
"volume": "",
"issue": "2",
"pages": "",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Marilyn Walker, Masayo Iida, and Sharon Cote. 1994. Japanese Discourse and the Process of Cen- tering. Computational Linguistics, 20(2).",
"links": null
},
"BIBREF17": {
"ref_id": "b17",
"title": "Evaluating Discourse Processing Algorithms",
"authors": [
{
"first": "Marilyn",
"middle": [
"A"
],
"last": "Walker",
"suffix": ""
}
],
"year": 1989,
"venue": "Proceedings of 27th Annual Meeting of the ACL",
"volume": "",
"issue": "",
"pages": "",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Marilyn A. Walker. 1989. Evaluating Discourse Pro- cessing Algorithms. In Proceedings of 27th Annual Meeting of the ACL.",
"links": null
}
},
"ref_entries": {
"FIGREF0": {
"text": "Figure 2: QZPRO Example",
"type_str": "figure",
"uris": null,
"num": null
},
"TABREF0": {
"content": "<table><tr><td/><td>Meaning</td><td/><td/><td/></tr><tr><td/><td colspan=\"2\">Definite NP</td><td/><td/></tr><tr><td/><td colspan=\"2\">Definite NP</td><td colspan=\"2\">whose referent is a facility</td></tr><tr><td/><td colspan=\"2\">Definite NP</td><td colspan=\"2\">whose referent is a location</td></tr><tr><td/><td colspan=\"2\">Definite NP</td><td colspan=\"3\">whose referent is an organization</td></tr><tr><td/><td colspan=\"2\">Definite NP</td><td colspan=\"2\">whose referent is a person</td></tr><tr><td/><td colspan=\"2\">Definite NP</td><td colspan=\"2\">whose referent is time</td></tr><tr><td/><td colspan=\"4\">Definite NP whose referent is two entities</td></tr><tr><td/><td colspan=\"5\">Definite NP whose referent is two organization entities</td></tr><tr><td/><td colspan=\"5\">Definite NP whose referent is two location entities</td></tr><tr><td/><td colspan=\"5\">Definite NP whose referent is two person entities</td></tr><tr><td/><td colspan=\"4\">Reflexive expressions (e.$. \"jisha ~)</td></tr><tr><td/><td colspan=\"3\">Proper name</td><td/></tr><tr><td/><td colspan=\"3\">Proper name for facility</td><td/></tr><tr><td/><td colspan=\"3\">Proper name for location</td><td/></tr><tr><td/><td colspan=\"4\">Proper name for organization</td></tr><tr><td/><td colspan=\"3\">Proper name for person</td><td/></tr><tr><td/><td colspan=\"4\">Deictic pronoun (this, these)</td></tr><tr><td/><td colspan=\"4\">Locational indexical (here, there)</td></tr><tr><td/><td colspan=\"4\">Time indexical (now, then, later)</td></tr><tr><td/><td colspan=\"3\">Quasi-zero pronoun</td><td/></tr><tr><td/><td colspan=\"5\">Quasi-zero pronoun whose referent is an organization</td></tr><tr><td/><td colspan=\"5\">Quasi-zero pronoun whose referent is a person</td></tr><tr><td/><td colspan=\"3\">Zero pronoun</td><td/></tr><tr><td/><td colspan=\"4\">Zero pronoun in an impersonal construction</td></tr><tr><td/><td colspan=\"5\">Zero pronoun whose referent is an organization</td></tr><tr><td/><td colspan=\"4\">Zero pronoun whose referent is a person</td></tr><tr><td>JDEL</td><td colspan=\"2\">Dou-ellipsis</td><td/><td/></tr><tr><td>SONY-wa</td><td>RCA-to</td><td colspan=\"2\">teikeishi,</td><td>VCR-wo</td><td>QZPRO</td></tr><tr><td>Sony-subj</td><td colspan=\"4\">RCA-with joint venture VCR-obj</td><td>(it)</td></tr><tr><td>kaihatsusuru</td><td>to</td><td colspan=\"2\">QZPRO</td><td>happyoushita</td></tr><tr><td>develop</td><td>that</td><td colspan=\"2\">(it)</td><td>announced</td></tr><tr><td colspan=\"6\">\"(SONY) announced that SONY will form a joint venture with RCA</td></tr><tr><td colspan=\"3\">and (it) will develop VCR's.\"</td><td/><td/></tr></table>",
"type_str": "table",
"num": null,
"text": "Summary of Anaphoric Types",
"html": null
},
"TABREF1": {
"content": "<table><tr><td/><td colspan=\"2\">Unary feature Binaxy feature</td></tr><tr><td>Lexical</td><td>category</td><td>matching-category</td></tr><tr><td>Syntactic</td><td>topicalized</td><td>matching-topicalized</td></tr><tr><td>Semantic</td><td colspan=\"2\">semantic-class subsuming-semantic-class</td></tr><tr><td>Positional</td><td/><td>antecedent-precedes-anaphor</td></tr></table>",
"type_str": "table",
"num": null,
"text": "Examples of Training Features",
"html": null
},
"TABREF2": {
"content": "<table><tr><td/><td/><td>confidence factor</td></tr><tr><td>yes</td><td>no</td><td>lOO%</td></tr><tr><td>yes</td><td>no</td><td>75% '</td></tr><tr><td>yes</td><td>no</td><td>50% \"</td></tr><tr><td>yes</td><td>no</td><td>25%</td></tr><tr><td>yes</td><td>yes</td><td>75%</td></tr><tr><td>no</td><td>no</td><td>75%</td></tr></table>",
"type_str": "table",
"num": null,
"text": "Six Configurations of MLRs",
"html": null
},
"TABREF3": {
"content": "<table><tr><td/><td>Recall = Nc/I, Precision = Nc/Nn</td></tr><tr><td>I</td><td>Number of system-identified anaphora in input</td></tr><tr><td colspan=\"2\">N~ Number of correct resolutions</td></tr><tr><td colspan=\"2\">Nh Number of resolutions attempted</td></tr></table>",
"type_str": "table",
"num": null,
"text": "Recall and Precision Metrics for Evaluation",
"html": null
},
"TABREF4": {
"content": "<table><tr><td/><td>Filters</td><td>Orderers</td></tr><tr><td>current-text</td><td>syntactic-category-propn</td><td>reverse-recency</td></tr><tr><td/><td>nam~chax-subsequence</td><td/></tr><tr><td/><td>semantic-class-org</td><td/></tr><tr><td>current-text</td><td>semantic-dass-org</td><td>topica]ization</td></tr><tr><td/><td>semantic-amount-singular</td><td>subject-np</td></tr><tr><td/><td/><td>recency</td></tr><tr><td>current-paragraph</td><td>not-in-the-same-dc</td><td>topica]ization</td></tr><tr><td/><td>semantic-dass-from-pred</td><td>subject-np</td></tr><tr><td/><td/><td>category-np</td></tr><tr><td/><td/><td>recency</td></tr><tr><td>current-paragraph</td><td>not-in-the-same-dc</td><td>topicalization</td></tr><tr><td/><td>sere antic-dass-from-pred</td><td>subject-np</td></tr><tr><td/><td/><td>category-np</td></tr><tr><td/><td/><td>recency</td></tr><tr><td># exmpls</td><td/><td/></tr><tr><td>MLR-1</td><td/><td/></tr><tr><td>MLR-2</td><td/><td/></tr><tr><td>MLR-3</td><td/><td/></tr><tr><td>MLR-4</td><td/><td/></tr><tr><td>MLR-5</td><td/><td/></tr><tr><td>MLR-6</td><td/><td/></tr><tr><td>MDR</td><td/><td/></tr></table>",
"type_str": "table",
"num": null,
"text": "KS's used by the MDR",
"html": null
},
"TABREF5": {
"content": "<table><tr><td colspan=\"2\">NAME-ORG</td><td colspan=\"2\">DNP-ORG</td><td/><td/><td colspan=\"2\">ZPRO-ORG</td><td colspan=\"2\">Average</td><td>F-measure</td></tr><tr><td>631</td><td/><td>54</td><td/><td/><td>383</td><td/><td>203</td><td colspan=\"2\">1271</td><td>1271</td></tr><tr><td>R</td><td>P</td><td>R</td><td>P</td><td>R</td><td>P</td><td>R</td><td>P</td><td>R</td><td>P</td><td>F</td></tr><tr><td>84.79</td><td>92.24</td><td>44.44</td><td>50.00</td><td>65.62</td><td>80.25</td><td colspan=\"2\">4O.78 64.62</td><td colspan=\"2\">70.20 83.49</td><td>76.27</td></tr><tr><td>84.79</td><td>93.04</td><td>44.44</td><td>52.17</td><td>64.84</td><td>84.69</td><td>39.32</td><td>73.64</td><td colspan=\"2\">69.73 86.73</td><td>77.30</td></tr><tr><td>83.20</td><td>94.09</td><td>37.04</td><td>58.82</td><td>63.02</td><td>84.91</td><td>35.92</td><td>73.27</td><td colspan=\"2\">67.53 88.04</td><td>76.43</td></tr><tr><td>83.84</td><td>94.30</td><td>38.89</td><td>60.00</td><td>64.06</td><td>85.12</td><td>37.86</td><td>76.47</td><td>68.55</td><td>88.55</td><td>77.28</td></tr><tr><td>85.74</td><td>92.80</td><td>44.44</td><td>55.81</td><td>56.51</td><td>89.67</td><td>15.53</td><td>78.05</td><td>63.84</td><td>89.55</td><td>74.54</td></tr><tr><td>68.30</td><td>91.70</td><td>29.63</td><td>64.00</td><td>54.17</td><td>90.83</td><td colspan=\"2\">13.11 75.00</td><td>53.49</td><td>89.74</td><td>67.03</td></tr><tr><td>76.39</td><td>90.09</td><td>35.19</td><td>50.00</td><td>67.19</td><td>67.19</td><td colspan=\"2\">43.20 43.20</td><td>66.51</td><td>72.91</td><td>69.57</td></tr><tr><td>texts</td><td/><td/><td/><td/><td/><td/><td/><td/><td/><td/></tr><tr><td>50</td><td/><td/><td/><td/><td/><td/><td/><td/><td/><td/></tr><tr><td>I00</td><td/><td/><td/><td/><td/><td/><td/><td/><td/><td/></tr><tr><td>150</td><td/><td/><td/><td/><td/><td/><td/><td/><td/><td/></tr><tr><td>2OO</td><td/><td/><td/><td/><td/><td/><td/><td/><td/><td/></tr><tr><td>25O</td><td/><td/><td/><td/><td/><td/><td/><td/><td/><td/></tr><tr><td>295</td><td/><td/><td/><td/><td/><td/><td/><td/><td/><td/></tr><tr><td>MDR</td><td/><td/><td/><td/><td/><td/><td/><td/><td/><td/></tr></table>",
"type_str": "table",
"num": null,
"text": "Recall and Precision of the MLRs and the MDR",
"html": null
},
"TABREF6": {
"content": "<table><tr><td colspan=\"2\">NAME-ORG</td><td colspan=\"6\">DNP-ORG QZPRO-ORG ZPRO-ORG</td><td colspan=\"2\">Average</td><td>F-measure</td></tr><tr><td>R</td><td>P</td><td>R</td><td>P</td><td>R</td><td>P</td><td>R</td><td>P</td><td>R</td><td>P</td><td>F</td></tr><tr><td colspan=\"5\">81.30 91.94 35.19 48.72 59.38</td><td colspan=\"5\">76.77 29.13 56.07 64.31 81.92</td><td>72.06</td></tr><tr><td colspan=\"5\">82.09 92.01 38.89 53.85 63.02</td><td colspan=\"5\">85.82 28.64 62.77 65.88 85.89</td><td>74.57</td></tr><tr><td colspan=\"5\">82.57 91.89 48.15 60.47 55.73</td><td colspan=\"5\">85.60 20.39 70.00 62.98 87.28</td><td>73.17</td></tr><tr><td colspan=\"5\">83.99 91.70 46.30 60.98 63.02</td><td colspan=\"5\">82.88 36.41 65.22 68.39 84.99</td><td>75.79</td></tr><tr><td colspan=\"5\">84.79 93.21 44.44 53.33 65.10</td><td colspan=\"5\">83.89 40.78 73.04 70.04 86.53</td><td>77.42</td></tr><tr><td colspan=\"5\">84.79 93.04 44.44 52.17 64.84</td><td colspan=\"5\">84.69 39.32 73.64 69.73 86.73</td><td>77.30</td></tr><tr><td colspan=\"5\">76.39 90.09 35.19 50.00 67.19</td><td colspan=\"5\">67.19 43.20 43.20 66.51 72.91</td><td>69.57</td></tr></table>",
"type_str": "table",
"num": null,
"text": "MLR-2 Configuration with Varied Training Data Sizes",
"html": null
}
}
}
} |