File size: 75,859 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 1222 1223 1224 1225 1226 1227 1228 1229 1230 1231 1232 1233 1234 1235 1236 1237 1238 1239 1240 1241 1242 1243 1244 1245 1246 1247 1248 1249 1250 1251 1252 1253 1254 1255 1256 1257 1258 1259 1260 1261 1262 1263 1264 1265 1266 1267 1268 1269 1270 1271 1272 1273 1274 1275 1276 1277 1278 1279 1280 1281 1282 1283 1284 1285 1286 1287 1288 1289 1290 1291 1292 1293 1294 1295 1296 1297 1298 1299 1300 1301 1302 1303 1304 1305 1306 1307 1308 1309 1310 1311 1312 1313 1314 1315 1316 1317 1318 1319 1320 1321 1322 1323 1324 1325 1326 1327 1328 1329 1330 1331 1332 1333 1334 1335 1336 1337 1338 1339 1340 1341 1342 1343 1344 1345 1346 1347 1348 1349 1350 1351 1352 1353 1354 1355 1356 1357 1358 1359 1360 1361 1362 1363 1364 1365 1366 1367 1368 1369 1370 1371 1372 1373 1374 1375 1376 1377 1378 1379 1380 1381 1382 1383 1384 1385 1386 1387 1388 1389 1390 1391 1392 1393 1394 1395 1396 1397 1398 1399 1400 1401 1402 1403 1404 1405 1406 1407 1408 1409 1410 1411 1412 1413 1414 1415 1416 1417 1418 1419 1420 1421 1422 1423 1424 1425 1426 1427 1428 1429 1430 1431 1432 1433 1434 1435 1436 1437 1438 1439 1440 1441 1442 1443 1444 1445 1446 1447 1448 1449 1450 1451 1452 1453 1454 1455 1456 1457 1458 1459 1460 1461 1462 1463 1464 1465 1466 1467 1468 1469 1470 1471 1472 1473 1474 1475 1476 1477 1478 1479 1480 1481 1482 1483 1484 1485 1486 1487 1488 1489 1490 1491 | {
"paper_id": "P95-1007",
"header": {
"generated_with": "S2ORC 1.0.0",
"date_generated": "2023-01-19T08:33:49.501516Z"
},
"title": "Corpus Statistics Meet the Noun Compound: Some Empirical Results",
"authors": [
{
"first": "Mark",
"middle": [],
"last": "Lauer",
"suffix": "",
"affiliation": {
"laboratory": "",
"institution": "Microsoft Institute",
"location": {
"addrLine": "65 Epping Road",
"postCode": "2113",
"settlement": "North Ryde",
"region": "NSW",
"country": "Australia"
}
},
"email": ""
}
],
"year": "",
"venue": null,
"identifiers": {},
"abstract": "A variety of statistical methods for noun compound anMysis are implemented and compared. The results support two main conclusions. First, the use of conceptual association not only enables a broad coverage, but also improves the accuracy. Second, an analysis model based on dependency grammar is substantially more accurate than one based on deepest constituents, even though the latter is more prevalent in the literature.",
"pdf_parse": {
"paper_id": "P95-1007",
"_pdf_hash": "",
"abstract": [
{
"text": "A variety of statistical methods for noun compound anMysis are implemented and compared. The results support two main conclusions. First, the use of conceptual association not only enables a broad coverage, but also improves the accuracy. Second, an analysis model based on dependency grammar is substantially more accurate than one based on deepest constituents, even though the latter is more prevalent in the literature.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Abstract",
"sec_num": null
}
],
"body_text": [
{
"text": "1 Background",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "",
"sec_num": null
},
{
"text": "If parsing is taken to be the first step in taming the natural language understanding task, then broad coverage NLP remains a jungle inhabited by wild beasts. For instance, parsing noun compounds appears to require detailed world knowledge that is unavailable outside a limited domain (Sparek Jones, 1983 ). Yet, far from being an obscure, endangered species, the noun compound is flourishing in modern language. It has already made five appearances in this paragraph and at least one diachronic study shows a veritable population explosion (Leonard, 1984) . While substantial work on noun compounds exists in both linguistics (e.g. Levi, 1978; Ryder, 1994) and computational linguistics (Finin, 1980; McDonald, 1982; Isabelle, 1984) , techniques suitable for broad coverage parsing remain unavailable. This paper explores the application of corpus statistics (Charniak, 1993) to noun compound parsing (other computational problems are addressed in Arens el al, 1987; Vanderwende, 1993 and Sproat, 1994) .",
"cite_spans": [
{
"start": 285,
"end": 304,
"text": "(Sparek Jones, 1983",
"ref_id": null
},
{
"start": 541,
"end": 556,
"text": "(Leonard, 1984)",
"ref_id": "BIBREF11"
},
{
"start": 633,
"end": 644,
"text": "Levi, 1978;",
"ref_id": "BIBREF12"
},
{
"start": 645,
"end": 657,
"text": "Ryder, 1994)",
"ref_id": "BIBREF19"
},
{
"start": 688,
"end": 701,
"text": "(Finin, 1980;",
"ref_id": "BIBREF4"
},
{
"start": 702,
"end": 717,
"text": "McDonald, 1982;",
"ref_id": "BIBREF15"
},
{
"start": 718,
"end": 733,
"text": "Isabelle, 1984)",
"ref_id": "BIBREF6"
},
{
"start": 860,
"end": 876,
"text": "(Charniak, 1993)",
"ref_id": "BIBREF3"
},
{
"start": 949,
"end": 967,
"text": "Arens el al, 1987;",
"ref_id": null
},
{
"start": 968,
"end": 989,
"text": "Vanderwende, 1993 and",
"ref_id": "BIBREF23"
},
{
"start": 990,
"end": 1003,
"text": "Sproat, 1994)",
"ref_id": "BIBREF22"
}
],
"ref_spans": [],
"eq_spans": [],
"section": "Compound Nouns",
"sec_num": "1.1"
},
{
"text": "The task is illustrated in example 1: Example 1 The parses assigned to these two compounds differ, even though the sequence of parts of speech are identical. The problem is analogous to the prepositional phrase attachment task explored in Hindle and Rooth (1993) . The approach they propose involves computing lexical associations from a corpus and using these to select the correct parse. A similar architecture may be applied to noun compounds.",
"cite_spans": [
{
"start": 239,
"end": 262,
"text": "Hindle and Rooth (1993)",
"ref_id": "BIBREF5"
}
],
"ref_spans": [],
"eq_spans": [],
"section": "Compound Nouns",
"sec_num": "1.1"
},
{
"text": "In the experiments below the accuracy of such a system is measured. Comparisons are made across five dimensions:",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Compound Nouns",
"sec_num": "1.1"
},
{
"text": "\u2022 Each of two analysis models are applied: adjacency and dependency.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Compound Nouns",
"sec_num": "1.1"
},
{
"text": "\u2022 Each of a range of training schemes are employed.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Compound Nouns",
"sec_num": "1.1"
},
{
"text": "\u2022 Results are computed with and without tuning factors suggested in the literature.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Compound Nouns",
"sec_num": "1.1"
},
{
"text": "\u2022 Each of two parameterisations are used: associations between words and associations between concepts.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Compound Nouns",
"sec_num": "1.1"
},
{
"text": "\u2022 Results are collected with and without machine tagging of the corpus.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Compound Nouns",
"sec_num": "1.1"
},
{
"text": "While Hindle and Rooth (1993) use a partial parser to acquire training data, such machinery appears unnecessary for noun compounds. Brent (1993) has proposed the use of simple word patterns for the acquisition of verb subcategorisation information. An analogous approach to compounds is used in Lauer (1994) and constitutes one scheme evaluated below. While such patterns produce false training examples, the resulting noise often only introduces minor distortions.",
"cite_spans": [
{
"start": 6,
"end": 29,
"text": "Hindle and Rooth (1993)",
"ref_id": "BIBREF5"
},
{
"start": 132,
"end": 144,
"text": "Brent (1993)",
"ref_id": "BIBREF1"
},
{
"start": 295,
"end": 307,
"text": "Lauer (1994)",
"ref_id": "BIBREF9"
}
],
"ref_spans": [],
"eq_spans": [],
"section": "Training Schemes",
"sec_num": "1.2"
},
{
"text": "A more liberal alternative is the use of a cooccurrence window. Yarowsky (1992) uses a fixed 100 word window to collect information used for sense disambiguation. Similarly, Smadja (1993) uses a six content word window to extract significant collocations. A range of windowed training schemes are employed below. Importantly, the use of a window provides a natural means of trading off the amount of data against its quality. When data sparseness undermines the system accuracy, a wider window may admit a sufficient volume of extra accurate data to outweigh the additional noise.",
"cite_spans": [
{
"start": 174,
"end": 187,
"text": "Smadja (1993)",
"ref_id": "BIBREF20"
}
],
"ref_spans": [],
"eq_spans": [],
"section": "Training Schemes",
"sec_num": "1.2"
},
{
"text": "There are at least four existing corpus-based algorithms proposed for syntactically analysing noun compounds. Only two of these have been subjected to evaluation, and in each case, no comparison to any of the other three was performed. In fact all authors appear unaware of the other three proposals. I will therefore briefly describe these algorithms.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Noun Compound Analysis",
"sec_num": "1.3"
},
{
"text": "Three of the algorithms use what I will call the ADJACENCY MODEL, an analysis procedure that goes back to Marcus (1980, p253) . Therein, the procedure is stated in terms of calls to an oracle which can determine if a noun compound is acceptable. It is reproduced here for reference:",
"cite_spans": [
{
"start": 106,
"end": 125,
"text": "Marcus (1980, p253)",
"ref_id": null
}
],
"ref_spans": [],
"eq_spans": [],
"section": "Noun Compound Analysis",
"sec_num": "1.3"
},
{
"text": "Given three nouns nl, n2 and nz: Only more recently has it been suggested that corpus statistics might provide the oracle, and this idea is the basis of the three algorithms which use the adjacency model. The simplest of these is reported in Pustejovsky et al (1993) . Given a three word compound, a search is conducted elsewhere in the corpus for each of the two possible subcomponents. Whichever is found is then chosen as the more closely bracketed pair. For example, when backup compiler disk is encountered, the analysis will be:",
"cite_spans": [
{
"start": 260,
"end": 266,
"text": "(1993)",
"ref_id": null
}
],
"ref_spans": [],
"eq_spans": [],
"section": "Noun Compound Analysis",
"sec_num": "1.3"
},
{
"text": "\u2022 If",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Noun Compound Analysis",
"sec_num": "1.3"
},
{
"text": "Example 2 (a) [backupN [compilerN diskN]]",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Noun Compound Analysis",
"sec_num": "1.3"
},
{
"text": "when compiler disk appears elsewhere",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Noun Compound Analysis",
"sec_num": "1.3"
},
{
"text": "(b) [[backupN compilerN] diskN]",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Noun Compound Analysis",
"sec_num": "1.3"
},
{
"text": "when backup compiler appears elsewhere Since this is proposed merely as a rough heuristic, it is not stated what the outcome is to be if neither or both subcomponents appear. Nor is there any evaluation of the algorithm.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Noun Compound Analysis",
"sec_num": "1.3"
},
{
"text": "The proposal of Liberman and Sproat (1992) is more sophisticated and allows for the frequency of the words in the compound. Their proposal involves comparing the mutual information between the two pairs of adjacent words and bracketing together whichever pair exhibits the highest. Again, there is no evaluation of the method other than a demonstration that four examples work correctly.",
"cite_spans": [
{
"start": 16,
"end": 42,
"text": "Liberman and Sproat (1992)",
"ref_id": "BIBREF13"
}
],
"ref_spans": [],
"eq_spans": [],
"section": "Noun Compound Analysis",
"sec_num": "1.3"
},
{
"text": "The third proposal based on the adjacency model appears in Resnik (1993) and is rather more complex again. The SELECTIONAL ASSOCIATION between a predicate and a word is defined based on the contribution of the word to the conditional entropy of the predicate. The association between each pair of words in the compound is then computed by taking the maximum selectional association from all possible ways of regarding the pair as predicate and argument. Whilst this association metric is complicated, the decision procedure still follows the outline devised by Marcus (1980) above. Resnik (1993) used unambiguous noun compounds from the parsed Wall Stree~ Journal (WSJ) corpus to estimate the association ~alues and analysed a test set of around 160 compounds. After some tuning, the accuracy was about 73%, as compared with a baseline of 64% achieved by always bracketing the first two nouns together.",
"cite_spans": [
{
"start": 59,
"end": 72,
"text": "Resnik (1993)",
"ref_id": "BIBREF17"
},
{
"start": 561,
"end": 574,
"text": "Marcus (1980)",
"ref_id": "BIBREF14"
},
{
"start": 582,
"end": 595,
"text": "Resnik (1993)",
"ref_id": "BIBREF17"
}
],
"ref_spans": [],
"eq_spans": [],
"section": "Noun Compound Analysis",
"sec_num": "1.3"
},
{
"text": "The fourth algorithm, first described in Lauer (1994) , differs in one striking manner from the other three. It uses what I will call the DEPENDENCY MO-DEL. This model utilises the following procedure when given three nouns at, n2 and n3:",
"cite_spans": [
{
"start": 41,
"end": 53,
"text": "Lauer (1994)",
"ref_id": "BIBREF9"
}
],
"ref_spans": [],
"eq_spans": [],
"section": "Noun Compound Analysis",
"sec_num": "1.3"
},
{
"text": "\u2022 \u2022 otherwise, build In1 rig.] first. Figure 1 shows a graphical comparison of the two analysis models.",
"cite_spans": [],
"ref_spans": [
{
"start": 38,
"end": 46,
"text": "Figure 1",
"ref_id": null
}
],
"eq_spans": [],
"section": "Noun Compound Analysis",
"sec_num": "1.3"
},
{
"text": "In Lauer (1994) , the degree of acceptability is again provided by statistical measures over a corpus. The metric used is a mutual information-like measure based on probabilities of modification relationships. This is derived from the idea that parse trees capture the structure of semantic relationships within a noun compound. 1",
"cite_spans": [
{
"start": 3,
"end": 15,
"text": "Lauer (1994)",
"ref_id": "BIBREF9"
}
],
"ref_spans": [],
"eq_spans": [],
"section": "Noun Compound Analysis",
"sec_num": "1.3"
},
{
"text": "The dependency model attempts to choose a parse which makes the resulting relationships as acceptable as possible. For example, when backup compiler disk is encountered, the analysis will be:",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Noun Compound Analysis",
"sec_num": "1.3"
},
{
"text": "Example 3 (a) [backupN [compilerN diskN]] when backup disk is more acceptable (b) [[backupN compilerN] diskN]",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Noun Compound Analysis",
"sec_num": "1.3"
},
{
"text": "when backup compiler is more acceptable I claim that the dependency model makes more intuitive sense for the following reason. Consider the compound calcium ion exchange, which is typically left-branching (that is, the first two words are bracketed together). There does not seem to be any reason why calcium ion should be any more frequent than ion exchange. Both are plausible compounds and regardless of the bracketing, ions are the object of an exchange. Instead, the correct parse depends on whether calcium characterises the ions or mediates the exchange.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Noun Compound Analysis",
"sec_num": "1.3"
},
{
"text": "Another significant difference between the models is the predictions they make about the proportion 1Lauer and Dras (1994) give a formal construction motivating the algorithm given in Lauer (1994) .",
"cite_spans": [
{
"start": 184,
"end": 196,
"text": "Lauer (1994)",
"ref_id": "BIBREF9"
}
],
"ref_spans": [],
"eq_spans": [],
"section": "Noun Compound Analysis",
"sec_num": "1.3"
},
{
"text": "L N2 t R Adjacency N3 t Prefer left-branching ig L is more acceptable than R L N1 N2 N3 t t R Dependency",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Noun Compound Analysis",
"sec_num": "1.3"
},
{
"text": "Figure 1: Two analysis models and the associations they compare of left and right-branching compounds. Lauer and Dras (1994) show that under a dependency model, left-branching compounds should occur twice as often as right-branching compounds (that is twothirds of the time). In the test set used here and in that of Resnik (1993) , the proportion of leftbranching compounds is 67% and 64% respectively. In contrast, the adjacency model appears to predict a proportion of 50%.",
"cite_spans": [
{
"start": 103,
"end": 124,
"text": "Lauer and Dras (1994)",
"ref_id": "BIBREF10"
},
{
"start": 317,
"end": 330,
"text": "Resnik (1993)",
"ref_id": "BIBREF17"
}
],
"ref_spans": [],
"eq_spans": [],
"section": "Noun Compound Analysis",
"sec_num": "1.3"
},
{
"text": "The dependency model has also been proposed by Kobayasi et al (1994) for analysing Japanese noun compounds, apparently independently. Using a corpus to acquire associations, they bracket sequences of Kanji with lengths four to six (roughly equivalent to two or three words). A simple calculation shows that using their own preprocessing hueristics to guess a bracketing provides a higher accuracy on their test set than their statistical model does. This renders their experiment inconclusive.",
"cite_spans": [
{
"start": 47,
"end": 68,
"text": "Kobayasi et al (1994)",
"ref_id": "BIBREF8"
}
],
"ref_spans": [],
"eq_spans": [],
"section": "Noun Compound Analysis",
"sec_num": "1.3"
},
{
"text": "A test set of syntactically ambiguous noun compounds was extracted from our 8 million word Grolier's encyclopedia corpus in the following way. 2 Because the corpus is not tagged or parsed, a somewhat conservative strategy of looking for unambiguous sequences of nouns was used. To distinguish nouns from other words, the University of Pennsylvania morphological analyser (described in Karp et al, 1992) was used to generate the set of words that can only be used as nouns (I shall henceforth call this set AZ). All consecutive sequences of these words were extracted, and the three word sequences used to form the test set. For reasons made clear below, only sequences consisting entirely of words from Roget's thesaurus were retained, giving a total of 308 test triples. 3 These triples were manually analysed using as context the entire article in which they appeared. In 2We would like to thank Grolier's for permission to use this material for research purposes.",
"cite_spans": [
{
"start": 385,
"end": 402,
"text": "Karp et al, 1992)",
"ref_id": "BIBREF7"
}
],
"ref_spans": [],
"eq_spans": [],
"section": "Extracting a Test Set",
"sec_num": "2.1"
},
{
"text": "3The 1911 version of Roget's used is available on-line and is in the public domain. some cases, the sequence was not a noun compound (nouns can appear adjacent to one another across various constituent boundaries) and was marked as an error. Other compounds exhibited what Hindie and Rooth (1993) have termed SEMANTIC INDE-TERMINACY where the two possible bracketings cannot be distinguished in the context. The remaining compounds were assigned either a left-branching or right-branching analysis. Table 1 shows the number of each kind and an example of each.",
"cite_spans": [],
"ref_spans": [
{
"start": 499,
"end": 506,
"text": "Table 1",
"ref_id": null
}
],
"eq_spans": [],
"section": "Extracting a Test Set",
"sec_num": "2.1"
},
{
"text": "Accuracy figures in all the results reported below were computed using only those 244 compounds which received a parse.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Extracting a Test Set",
"sec_num": "2.1"
},
{
"text": "One problem with applying lexical association to noun compounds is the enormous number of parameters required, one for every possible pair of nouns. Not only does this require a vast amount of memory space, it creates a severe data sparseness problem since we require at least some data about each parameter. Resnik and Hearst (1993) coined the term CONCEPTUAL ASSOCIATION to refer to association values computed between groups of words. By assuming that all words within a group behave similarly, the parameter space can be built in terms of the groups rather than in terms of the words.",
"cite_spans": [
{
"start": 309,
"end": 333,
"text": "Resnik and Hearst (1993)",
"ref_id": "BIBREF18"
}
],
"ref_spans": [],
"eq_spans": [],
"section": "Conceptual Association",
"sec_num": "2.2"
},
{
"text": "In this study, conceptual association is used with groups consisting of all categories from the 1911 version of Roget's thesaurus. 4 Given two thesaurus categories tl and t~, there is a parameter which represents the degree of acceptability of the structure [nine] where nl is a noun appearing in tl and n2 appears in t2. By the assumption that words within a group behave similarly, this is constant given the two categories. Following Lauer and Dras (1994) we can formally write this parameter as Pr(tl ~ t2) where the event tl ~ t2 denotes the modification of a noun in t2 by a noun in tl.",
"cite_spans": [
{
"start": 437,
"end": 458,
"text": "Lauer and Dras (1994)",
"ref_id": "BIBREF10"
}
],
"ref_spans": [],
"eq_spans": [],
"section": "Conceptual Association",
"sec_num": "2.2"
},
{
"text": "To ensure that the test set is disjoint from the training data, all occurrences of the test noun compounds have been removed from the training corpus. Two types of training scheme are explored in this study, both unsupervised. The first employs a pattern that follows Pustejovsky (1993) in counting the occurrences of subcomponents. A training instance is any sequence of four words WlW2W3W 4 where wl, w4 ~ .h/and w2, w3 E A/'. Let county(n1, n2) be the number of times a sequence wlnln2w4 occurs in the training corpus with wl, w4 ~ At'.",
"cite_spans": [
{
"start": 268,
"end": 286,
"text": "Pustejovsky (1993)",
"ref_id": "BIBREF16"
}
],
"ref_spans": [],
"eq_spans": [],
"section": "Training",
"sec_num": "2.3"
},
{
"text": "The second type uses a window to collect training instances by observing how often a pair of nouns cooccur within some fixed number of words. In this study, a variety of window sizes are used. For n > 2, let countn(nl, n2) be the number of times a sequence nlwl...wins occurs in the training corpus where i < n -2. Note that windowed counts are asymmetric. In the case of a window two words wide, this yields the mutual information metric proposed by Liberman and Sproat (1992) .",
"cite_spans": [
{
"start": 451,
"end": 477,
"text": "Liberman and Sproat (1992)",
"ref_id": "BIBREF13"
}
],
"ref_spans": [],
"eq_spans": [],
"section": "Training",
"sec_num": "2.3"
},
{
"text": "Using each of these different training schemes to arrive at appropriate counts it is then possible to estimate the parameters. Since these are expressed in terms of categories rather than words, it is necessary to combine the counts of words to arrive at estimates. In all cases the estimates used are: Here ambig(w) is the number of categories in which w appears. It has the effect of dividing the evidence from a training instance across all possible categories for the words. The normaliser ensures that all parameters for a head noun sum to unity.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Training",
"sec_num": "2.3"
},
{
"text": "Given the high level descriptions in section 1.3 it remains only to formalise the decision process used to analyse a noun compound. Each test compound presents a set of possible analyses and the goal is to choose which analysis is most likely. For three word compounds it suffices to compute the ratio of two probabilities, that of a left-branching analysis and that of a right-branching one. If this ratio is greater than unity, then the left-branching analysis is chosen. When it is less than unity, a right-branching analysis is chosen. ~ If the ratio is exactly unity, the analyser guesses left-branching, although this is fairly rare for conceptual association as shown by the experimental results below.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Analysing the Test Set",
"sec_num": "2.4"
},
{
"text": "For the adjacency model, when the given compound is WlW2W3, we can estimate this ratio as: In both cases, we sum over all possible categories for the words in the compound. Because the dependency model equations have two factors, they are affected more severely by data sparseness. If the probability estimate for Pr(t2 ~ t3) is zero for all possible categories t2 and t3 then both the numerator and the denominator will be zero. This will conceal any preference given by the parameters involving Q. In such cases, we observe that the test instance itself provides the information that the event t2 --~ t3 can occur and we recalculate the ratio using Pr(t2 ---* t3) = k for all possible categories t2,t a where k is any non-zero constant. However, no correction is made to the probability estimates for Pr(tl --~ t2) and Pr(Q --* t3) for unseen cases, thus putting the dependency model on an equal footing with the adjacency model above. The equations presented above for the dependency model differ from those developed in Lauer and Dras (1994) in one way. There, an additional weighting factor (of 2.0) is used to favour a left-branching analysis. This arises because their construction is based on the dependency model which predicts that left-branching analyses should occur twice as often. Also, the work reported in Lauer and Dras (1994) uses simplistic estimates of the probability of a word given its thesaurus category. The equations above assume these probabilities are uniformly constant. Section 3.2 below shows the result of making these two additions to the method. sit either probability estimate is zero, the other analysis is chosen. If both are zero the analysis is made as if the ratio were exactly unity. ",
"cite_spans": [
{
"start": 1024,
"end": 1045,
"text": "Lauer and Dras (1994)",
"ref_id": "BIBREF10"
},
{
"start": 1322,
"end": 1343,
"text": "Lauer and Dras (1994)",
"ref_id": "BIBREF10"
}
],
"ref_spans": [],
"eq_spans": [],
"section": "Analysing the Test Set",
"sec_num": "2.4"
},
{
"text": "Eight different training schemes have been used to estimate the parameters and each set of estimates used to analyse the test set under both the adjacency and the dependency model. The schemes used are:",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Dependency meets Adjacency",
"sec_num": "3.1"
},
{
"text": "\u2022 the pattern given in section 2.3; and",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Dependency meets Adjacency",
"sec_num": "3.1"
},
{
"text": "\u2022 windowed training schemes with window widths of 2, 3, 4, 5, 10, 50 and 100 words.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Dependency meets Adjacency",
"sec_num": "3.1"
},
{
"text": "The accuracy on the test set for all these experiments is shown in figure 2. As can be seen, the dependency model is more accurate than the adjacency model. This is true across the whole spectrum of training schemes. The proportion of cases in which the procedure was forced to guess, either because no data supported either analysis or because both were equally supported, is quite low. For the pattern and two-word window training schemes, the guess rate is less than 4% for both models. In the three-word window training scheme, the guess rates are less than 1%. For all larger windows, neither model is ever forced to guess.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Dependency meets Adjacency",
"sec_num": "3.1"
},
{
"text": "In the case of the pattern training scheme, the difference between 68.9% for adjacency and 77.5% for dependency is statistically significant at the 5% level (p = 0.0316), demonstrating the superiority of the dependency model, at least for the compounds within Grolier's encyclopedia.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Dependency meets Adjacency",
"sec_num": "3.1"
},
{
"text": "In no case do any of the windowed training schemes outperform the pattern scheme. It seems that additional instances admitted by the windowed schemes are too noisy to make an improvement.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Dependency meets Adjacency",
"sec_num": "3.1"
},
{
"text": "Initial results from applying these methods to the EMA corpus have been obtained by Wilco ter Stal (1995) , and support the conclusion that the dependency model is superior to the adjacency model. Lauer and Dras (1994) suggest two improvements to the method used above. These are:",
"cite_spans": [
{
"start": 84,
"end": 105,
"text": "Wilco ter Stal (1995)",
"ref_id": "BIBREF24"
},
{
"start": 197,
"end": 218,
"text": "Lauer and Dras (1994)",
"ref_id": "BIBREF10"
}
],
"ref_spans": [],
"eq_spans": [],
"section": "Dependency meets Adjacency",
"sec_num": "3.1"
},
{
"text": "\u2022 a factor favouring left-branching which arises from the formal dependency construction; and",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Tuning",
"sec_num": "3.2"
},
{
"text": "\u2022 factors allowing for naive estimates of the variation in the probability of categories.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Tuning",
"sec_num": "3.2"
},
{
"text": "While these changes are motivated by the dependency model, I have also applied them to the adjacency model for comparison. To implement them, equations 1 and 2 must be modified to incorporate 1 in each term of the sum and the a factor of entire ratio must be multiplied by two. Five training schemes have been applied with these extensions. The accuracy results are shown in figure 3. For comparison, the untuned accuracy figures are shown with dotted lines. A marked improvement is observed for the adjacency model, while the dependency model is only slightly improved.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Tuning",
"sec_num": "3.2"
},
{
"text": "To determine the difference made by conceptual association, the pattern training scheme has been retrained using lexical counts for both the dependency and adjacency model, but only for the words in the test set. If the same system were to be applied across all of Af (a total of 90,000 nouns), then around 8.1 billion parameters would be required. Left-branching is favoured by a factor of two as described in the previous section, but no estimates for the category probabilities are used (these being meaningless for the lexical association method).",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Lexical Association",
"sec_num": "3.3"
},
{
"text": "Accuracy and guess rates are shown in figure 4. Conceptual association outperforms lexical association, presumably because of its ability to generalise.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Lexical Association",
"sec_num": "3.3"
},
{
"text": "One problem with the training methods given in section 2.3 is the restriction of training data to nouns in .Af. Many nouns, especially common ones, have verbal or adiectival usages that preclude them from being in .Af. Yet when they occur as nouns, they still provide useful training information that the current system ignores. To test whether using tagged Training scheme (integers denote window widths) Figure 5 : Accuracy using a tagged corpus for various training schemes data would make a difference, the freely available Brill tagger (Brill, 1993) was applied to the corpus. Since no manually tagged training data is available for our corpus, the tagger's default rules were used (these rules were produced by Brill by training on the Brown corpus). This results in rather poor tagging accuracy, so it is quite possible that a manually tagged corpus would produce better results.",
"cite_spans": [
{
"start": 541,
"end": 554,
"text": "(Brill, 1993)",
"ref_id": "BIBREF2"
}
],
"ref_spans": [
{
"start": 406,
"end": 414,
"text": "Figure 5",
"ref_id": null
}
],
"eq_spans": [],
"section": "Using a Tagger",
"sec_num": "3.4"
},
{
"text": "Three training schemes have been used and the tuned analysis procedures applied to the test set. Figure 5 shows the resulting accuracy, with accuracy values from figure 3 displayed with dotted lines. If anything, admitting additional training data based on the tagger introduces more noise, reducing the accuracy. However, for the pattern training scheme an improvement was made to the dependency model, producing the highest overall accuracy of 81%.",
"cite_spans": [],
"ref_spans": [
{
"start": 97,
"end": 105,
"text": "Figure 5",
"ref_id": null
}
],
"eq_spans": [],
"section": "Using a Tagger",
"sec_num": "3.4"
},
{
"text": "The experiments above demonstrate a number of important points. The most general of these is that even quite crude corpus statistics can provide information about the syntax of compound nouns. At the very least, this information can be applied in broad coverage parsing to assist in the control of search. I have also shown that with a corpus of moderate size it is possible to get reasonable results without using a tagger or parser by employing a customised training pattern. While using windowed co-occurrence did not help here, it is possible that under more data sparse conditions better performance could be achieved by this method.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Conclusion",
"sec_num": "4"
},
{
"text": "The significance of the use of conceptual association deserves some mention. I have argued that without it a broad coverage system would be impossible. This is in contrast to previous work on conceptual association where it resulted in little improvement on a task which could already be performed. In this study, not only has the technique proved its worth by supporting generality, but through generalisation of training information it outperforms the equivalent lexical association approach given the same information.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Conclusion",
"sec_num": "4"
},
{
"text": "Amongst all the comparisons performed in these experiments one stands out as exhibiting the greatest contrast. In all experiments the dependency model provides a substantial advantage over the adjacency model, even though the latter is more prevalent in proposals within the literature. This result is in accordance with the informal reasoning given in section 1.3. The model also has the further commendation that it predicts correctly the observed proportion of left-branching compounds found in two independently extracted test sets.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Conclusion",
"sec_num": "4"
},
{
"text": "In all, the most accurate technique achieved an accuracy of 81% as compared to the 67% achieved by guessing left-branching. Given the high frequency of occurrence of noun compounds in many texts, this suggests tha; the use of these techniques in probabilistic parsers will result in higher performance in broad coverage natural language processing.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Conclusion",
"sec_num": "4"
},
{
"text": "nowledged from the Microsoft Institute and the Australian Government.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Conclusion",
"sec_num": "4"
}
],
"back_matter": [],
"bib_entries": {
"BIBREF0": {
"ref_id": "b0",
"title": "Phrasal Analysis of Long Noun Sequences",
"authors": [
{
"first": "Y",
"middle": [],
"last": "Arens",
"suffix": ""
},
{
"first": "J",
"middle": [],
"last": "Granacki",
"suffix": ""
},
{
"first": "A",
"middle": [],
"last": "Parker",
"suffix": ""
}
],
"year": 1987,
"venue": "Proceedings of the 25th Annual Meeting of the Association for Computational Linguistics",
"volume": "",
"issue": "",
"pages": "59--64",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Arens, Y., Granacki, J. and Parker, A. 1987. Phra- sal Analysis of Long Noun Sequences. In Procee- dings of the 25th Annual Meeting of the Associa- tion for Computational Linguistics, Stanford, CA. pp59-64.",
"links": null
},
"BIBREF1": {
"ref_id": "b1",
"title": "From Grammar to Lexicon: Unsupervised Learning of Lexical Syntax",
"authors": [
{
"first": "Michael",
"middle": [],
"last": "Brent",
"suffix": ""
}
],
"year": 1993,
"venue": "Special Issue on Using Large Corpora II",
"volume": "19",
"issue": "",
"pages": "243--62",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Brent, Michael. 1993. From Grammar to Lexi- con: Unsupervised Learning of Lexical Syntax. In Computational Linguistics, Vol 19(2), Special Is- sue on Using Large Corpora II, pp243-62.",
"links": null
},
"BIBREF2": {
"ref_id": "b2",
"title": "A Corpus-based Approach to Language Learning",
"authors": [
{
"first": "Eric",
"middle": [],
"last": "Brill",
"suffix": ""
}
],
"year": 1993,
"venue": "",
"volume": "",
"issue": "",
"pages": "",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Brill, Eric. 1993. A Corpus-based Approach to Lan- guage Learning. PhD Thesis, University of Penn- sylvania, Philadelphia, PA..",
"links": null
},
"BIBREF3": {
"ref_id": "b3",
"title": "Statistical Language Learning",
"authors": [
{
"first": "Eugene",
"middle": [],
"last": "Charniak",
"suffix": ""
}
],
"year": 1993,
"venue": "",
"volume": "",
"issue": "",
"pages": "",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Charniak, Eugene. 1993. Statistical Language Lear- ning. MIT Press, Cambridge, MA.",
"links": null
},
"BIBREF4": {
"ref_id": "b4",
"title": "The Semantic Interpretation of Compound Nominals",
"authors": [
{
"first": "Tim",
"middle": [],
"last": "Finin",
"suffix": ""
}
],
"year": 1980,
"venue": "",
"volume": "",
"issue": "",
"pages": "",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Finin, Tim. 1980. The Semantic Interpretation of Compound Nominals. PhD Thesis, Co-ordinated Science Laboratory, University of Illinois, Urbana, IL.",
"links": null
},
"BIBREF5": {
"ref_id": "b5",
"title": "Structural Ambiguity and Lexical Relations",
"authors": [
{
"first": "D",
"middle": [],
"last": "Hindle",
"suffix": ""
},
{
"first": "M",
"middle": [],
"last": "Rooth",
"suffix": ""
}
],
"year": 1993,
"venue": "Special Issue on Using Large Corpora I",
"volume": "19",
"issue": "",
"pages": "3--20",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Hindle, D. and Rooth, M. 1993. Structural Am- biguity and Lexical Relations. In Computational Linguistics Vol. 19(1), Special Issue on Using Large Corpora I, ppl03-20.",
"links": null
},
"BIBREF6": {
"ref_id": "b6",
"title": "Another Look At Nominal Compounds",
"authors": [
{
"first": "Pierre",
"middle": [],
"last": "Isabelle",
"suffix": ""
}
],
"year": 1984,
"venue": "Proceedings of COLING-84",
"volume": "",
"issue": "",
"pages": "509--525",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Isabelle, Pierre. 1984. Another Look At Nominal Compounds. In Proceedings of COLING-84, Stan- ford, CA. pp509-16.",
"links": null
},
"BIBREF7": {
"ref_id": "b7",
"title": "A Freely Available Wide Coverage Morphological Analyzer for English",
"authors": [
{
"first": "D",
"middle": [],
"last": "Karp",
"suffix": ""
},
{
"first": "Y",
"middle": [],
"last": "Schabes",
"suffix": ""
},
{
"first": "M",
"middle": [],
"last": "Zaidel",
"suffix": ""
},
{
"first": "D",
"middle": [],
"last": "Egedi",
"suffix": ""
}
],
"year": 1992,
"venue": "Proceedings of COLING-92",
"volume": "",
"issue": "",
"pages": "950--954",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Karp, D., Schabes, Y., Zaidel, M. and Egedi, D. 1992. A Freely Available Wide Coverage Mor- phological Analyzer for English. In Proceedings of COLING-92, Nantes, France, pp950-4.",
"links": null
},
"BIBREF8": {
"ref_id": "b8",
"title": "Analysis of Japanese Compound Nouns using Collocational Information",
"authors": [
{
"first": "Y",
"middle": [],
"last": "Kobayasi",
"suffix": ""
},
{
"first": "T",
"middle": [],
"last": "Tokunaga",
"suffix": ""
},
{
"first": "H",
"middle": [],
"last": "Tanaka",
"suffix": ""
}
],
"year": 1994,
"venue": "Proceedings of COLING-94",
"volume": "",
"issue": "",
"pages": "865--874",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Kobayasi, Y., Tokunaga, T. and Tanaka, H. 1994. Analysis of Japanese Compound Nouns using Collocational Information. In Proceedings of COLING-94, Kyoto, Japan, pp865-9.",
"links": null
},
"BIBREF9": {
"ref_id": "b9",
"title": "Conceptual Association for Compound Noun Analysis",
"authors": [
{
"first": "Mark",
"middle": [],
"last": "Lauer",
"suffix": ""
}
],
"year": 1994,
"venue": "Proceedings of the 32nd Annual Meeting of the Association for Computational Linguistics, Student Session, Las Cruces, NM",
"volume": "",
"issue": "",
"pages": "337--346",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Lauer, Mark. 1994. Conceptual Association for Compound Noun Analysis. In Proceedings of the 32nd Annual Meeting of the Association for Com- putational Linguistics, Student Session, Las Cru- ces, NM. pp337-9.",
"links": null
},
"BIBREF10": {
"ref_id": "b10",
"title": "A Probabilistic Model of Compound Nouns",
"authors": [
{
"first": "M",
"middle": [],
"last": "Lauer",
"suffix": ""
},
{
"first": "M",
"middle": [],
"last": "Dras",
"suffix": ""
}
],
"year": 1994,
"venue": "Proceedings of the 7th Australian Joint Conference on Artificial Intelligence",
"volume": "",
"issue": "",
"pages": "474--81",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Lauer, M. and Dras, M. 1994. A Probabilistic Mo- del of Compound Nouns. In Proceedings of the 7th Australian Joint Conference on Artificial Intelli- gence, Armidale, NSW, Australia. World Scienti- fic Press, pp474-81.",
"links": null
},
"BIBREF11": {
"ref_id": "b11",
"title": "The Interpretation of English Noun Sequences on the Computer",
"authors": [
{
"first": "Rosemary",
"middle": [],
"last": "Leonard",
"suffix": ""
}
],
"year": 1984,
"venue": "",
"volume": "",
"issue": "",
"pages": "",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Leonard, Rosemary. 1984. The Interpretation of English Noun Sequences on the Computer. North- Holland, Amsterdam.",
"links": null
},
"BIBREF12": {
"ref_id": "b12",
"title": "The Syntax and Semantics of Complex Nominals",
"authors": [
{
"first": "Judith",
"middle": [],
"last": "Levi",
"suffix": ""
}
],
"year": 1978,
"venue": "",
"volume": "",
"issue": "",
"pages": "",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Levi, Judith. 1978. The Syntax and Semantics of Complex Nominals. Academic Press, New York.",
"links": null
},
"BIBREF13": {
"ref_id": "b13",
"title": "The Stress and Structure of Modified Noun Phrases in English",
"authors": [
{
"first": "M",
"middle": [],
"last": "Liberman",
"suffix": ""
},
{
"first": "R",
"middle": [],
"last": "Sproat",
"suffix": ""
}
],
"year": 1992,
"venue": "Lexical Matters CSLI Lecture Notes No. 24",
"volume": "",
"issue": "",
"pages": "31--81",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Liberman, M. and Sproat, R. 1992. The Stress and Structure of Modified Noun Phrases in English. In Sag, I. and Szabolcsi, A., editors, Lexical Mat- ters CSLI Lecture Notes No. 24. University of Chicago Press, ppl31-81.",
"links": null
},
"BIBREF14": {
"ref_id": "b14",
"title": "A Theory of Syntactic Recognition for Natural Language",
"authors": [
{
"first": "Mit~",
"middle": [],
"last": "Marcus",
"suffix": ""
}
],
"year": 1980,
"venue": "",
"volume": "",
"issue": "",
"pages": "",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Marcus, Mit~.hell. 1980. A Theory of Syntactic Re- cognition for Natural Language. MIT Press, Cam- bridge, MA.",
"links": null
},
"BIBREF15": {
"ref_id": "b15",
"title": "Understanding Noun Compounds",
"authors": [
{
"first": "David",
"middle": [
"B"
],
"last": "Mcdonald",
"suffix": ""
}
],
"year": 1982,
"venue": "Carnegie-Mellon University",
"volume": "",
"issue": "",
"pages": "",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "McDonald, David B. 1982. Understanding Noun Compounds. PhD Thesis, Carnegie-Mellon Uni- versity, Pittsburgh, PA.",
"links": null
},
"BIBREF16": {
"ref_id": "b16",
"title": "Lexical Semantic Techniques for Corpus Analysis",
"authors": [
{
"first": "J",
"middle": [],
"last": "Pustejovsky",
"suffix": ""
},
{
"first": "S",
"middle": [],
"last": "Bergler",
"suffix": ""
},
{
"first": "P",
"middle": [],
"last": "Anick",
"suffix": ""
}
],
"year": 1993,
"venue": "Special Issue on Using Large Corpora II",
"volume": "19",
"issue": "",
"pages": "331--58",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Pustejovsky, J., Bergler, S. and Anick, P. 1993. Le- xical Semantic Techniques for Corpus Analysis. In Computational Linguistics Vol 19(2), Special Is- sue on Using Large Corpora II, pp331-58.",
"links": null
},
"BIBREF17": {
"ref_id": "b17",
"title": "Selection and Information: A Class.Based Approach to Lexical Relationships",
"authors": [
{
"first": "Philip",
"middle": [],
"last": "Resnik",
"suffix": ""
}
],
"year": 1993,
"venue": "",
"volume": "",
"issue": "",
"pages": "",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Resnik, Philip. 1993. Selection and Informa- tion: A Class.Based Approach to Lexical Relati- onships. PhD dissertation, University of Pennsyl- vania, Philadelphia, PA.",
"links": null
},
"BIBREF18": {
"ref_id": "b18",
"title": "Structural Ambiguity and Conceptual Relations",
"authors": [
{
"first": "P",
"middle": [],
"last": "Resnik",
"suffix": ""
},
{
"first": "M",
"middle": [],
"last": "Hearst",
"suffix": ""
}
],
"year": 1993,
"venue": "Proceedings of the Workshop on Very Large Corpora: Academic and Industrial Perspectives",
"volume": "",
"issue": "",
"pages": "58--64",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Resnik, P. and Hearst, M. 1993. Structural Ambi- guity and Conceptual Relations. In Proceedings of the Workshop on Very Large Corpora: Academic and Industrial Perspectives, June 22, Ohio State University, pp58-64.",
"links": null
},
"BIBREF19": {
"ref_id": "b19",
"title": "Ordered Chaos: The Interpretation of English Noun-Noun Compounds",
"authors": [
{
"first": "Mary",
"middle": [
"Ellen"
],
"last": "Ryder",
"suffix": ""
}
],
"year": 1994,
"venue": "Linguistics",
"volume": "123",
"issue": "",
"pages": "",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Ryder, Mary Ellen. 1994. Ordered Chaos: The In- terpretation of English Noun-Noun Compounds. University of California Press Publications in Lin- guistics, Vol 123.",
"links": null
},
"BIBREF20": {
"ref_id": "b20",
"title": "Retrieving Collocations from Text: Xtract",
"authors": [
{
"first": "Frank",
"middle": [],
"last": "Smadja",
"suffix": ""
}
],
"year": 1993,
"venue": "Special Issue on Using Large Corpora I",
"volume": "19",
"issue": "",
"pages": "143--177",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Smadja, Frank. 1993. Retrieving Collocations from Text: Xtract. In Computational Linguistics, Vol 19(1), Special Issue on Using Large Corpora I, pp143-177.",
"links": null
},
"BIBREF21": {
"ref_id": "b21",
"title": "Compound Noun Interpretation Problems",
"authors": [
{
"first": "Sparck",
"middle": [],
"last": "Jones",
"suffix": ""
},
{
"first": "Karen",
"middle": [],
"last": "",
"suffix": ""
}
],
"year": 1983,
"venue": "Computer Speech Processing",
"volume": "",
"issue": "",
"pages": "363--81",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Sparck Jones, Karen. 1983. Compound Noun Interpretation Problems. In Fallside, F. and Woods, W.A., editors, Computer Speech Proces- sing. Prentice-Hall, NJ. pp363-81.",
"links": null
},
"BIBREF22": {
"ref_id": "b22",
"title": "English noun-phrase accent prediction for text-to-speech",
"authors": [
{
"first": "Richard",
"middle": [],
"last": "Sproat",
"suffix": ""
}
],
"year": 1994,
"venue": "In Computer Speech and Language",
"volume": "8",
"issue": "",
"pages": "79--94",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Sproat, Richard. 1994. English noun-phrase accent prediction for text-to-speech. In Computer Speech and Language, Vol 8, pp79-94.",
"links": null
},
"BIBREF23": {
"ref_id": "b23",
"title": "SENS: The System for Evaluating Noun Sequences",
"authors": [
{
"first": "Lucy",
"middle": [],
"last": "Vanderwende",
"suffix": ""
}
],
"year": 1993,
"venue": "Natural Language Processing",
"volume": "",
"issue": "",
"pages": "161--73",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Vanderwende, Lucy. 1993. SENS: The System for Evaluating Noun Sequences. In Jensen, K., Hei- dorn, G. and Richardson, S., editors, Natural Lan- guage Processing: The PLNLP Approach. Kluwer Academic, pp161-73.",
"links": null
},
"BIBREF24": {
"ref_id": "b24",
"title": "Syntactic Disambiguation of Nominal Compounds Using Lexical and Conceptual Association. Memorandum UT-KBS-95-002",
"authors": [
{
"first": "",
"middle": [],
"last": "Ter Stal",
"suffix": ""
},
{
"first": "",
"middle": [],
"last": "Wilco",
"suffix": ""
}
],
"year": 1992,
"venue": "Proceedings of COLING-92",
"volume": "",
"issue": "",
"pages": "454--60",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "ter Stal, Wilco. 1995. Syntactic Disambiguation of Nominal Compounds Using Lexical and Concep- tual Association. Memorandum UT-KBS-95-002, University of Twente, Enschede, Netherlands. Yarowsky, David. 1992. Word-Sense Disambigua- tion Using Statistical Models of Roget's Catego- ries Trained on Large Corpora. In Proceedings of COLING-92, Nantes, France, pp454-60.",
"links": null
}
},
"ref_entries": {
"FIGREF0": {
"num": null,
"type_str": "figure",
"uris": null,
"text": "(a) [womanN [aidN workerN]] (b) [[hydrogenN ionN] exchangeN]"
},
"FIGREF1": {
"num": null,
"type_str": "figure",
"uris": null,
"text": "\u2022 ~,~j\u00a2 ambig(wl)ambig(w~) w2Et2"
},
"FIGREF2": {
"num": null,
"type_str": "figure",
"uris": null,
"text": "Ra4i : ~-~t,~cats(..~ Pr(tl ---* t2) (1) ~-'~t,ecats(-b Pr(t2 ---* t3) For the dependency model, the ratio is: Rdep = ~-~,,ec~ts(~,) Pr(Q ---* t~) Pr(t~ ---* ta) (2) )-~t,ec~ts(~) Pr(~l ---* t3) Pr(t2 ~ ta)"
},
"FIGREF3": {
"num": null,
"type_str": "figure",
"uris": null,
"text": "integers denote window widths) Figure 2: Accuracy of dependency and adjacency model for various training schemes 3Results"
},
"FIGREF4": {
"num": null,
"type_str": "figure",
"uris": null,
"text": "Figure 4: Accuracy and Guess Rates of Lexical and Conceptual Association"
},
"TABREF2": {
"num": null,
"type_str": "table",
"html": null,
"content": "<table><tr><td>Type</td><td>Number</td><td>Proportion</td></tr><tr><td>Error</td><td>29</td><td>9%</td></tr><tr><td>Indeterminate</td><td>35</td><td>11%</td></tr><tr><td>Left-branching</td><td>163</td><td>53%</td></tr><tr><td>Right-branching</td><td>81</td><td>26%</td></tr><tr><td/><td/><td>Table 1: Test set distribution</td></tr><tr><td/><td/><td>4It contains 1043 categories.</td></tr></table>",
"text": "In monsoon regions rainfall does not ... Most advanced aircraft have precision navigation systems. ...escaped punishment by the Allied war crimes tribunals. Ronald Reagan, who won two landslide election victories, ..."
}
}
}
} |