File size: 98,406 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 1492 1493 1494 1495 1496 1497 1498 1499 1500 1501 1502 1503 1504 1505 1506 1507 1508 1509 1510 1511 1512 1513 1514 1515 1516 1517 1518 1519 1520 1521 1522 1523 1524 1525 1526 1527 1528 1529 1530 1531 1532 1533 1534 1535 1536 1537 1538 1539 1540 1541 1542 1543 1544 1545 1546 1547 1548 1549 1550 1551 1552 1553 1554 1555 1556 1557 1558 1559 1560 1561 1562 1563 1564 1565 1566 1567 1568 1569 1570 1571 1572 1573 1574 1575 1576 1577 1578 1579 1580 1581 1582 1583 1584 1585 1586 1587 1588 1589 1590 1591 1592 1593 1594 1595 1596 1597 1598 1599 1600 1601 1602 1603 1604 1605 1606 1607 1608 1609 1610 1611 1612 1613 1614 1615 1616 1617 1618 1619 1620 1621 1622 1623 1624 1625 1626 1627 1628 1629 1630 1631 1632 1633 1634 1635 1636 1637 1638 1639 1640 1641 1642 1643 1644 1645 1646 1647 1648 1649 1650 1651 1652 1653 1654 1655 1656 1657 1658 1659 1660 1661 1662 1663 1664 1665 1666 1667 1668 1669 1670 1671 1672 1673 1674 1675 1676 1677 1678 1679 1680 1681 1682 1683 1684 1685 1686 1687 1688 1689 1690 1691 1692 1693 1694 1695 1696 1697 1698 1699 1700 1701 1702 1703 1704 1705 1706 1707 1708 1709 1710 1711 1712 1713 1714 1715 1716 1717 1718 1719 1720 1721 1722 1723 1724 1725 1726 1727 1728 1729 1730 1731 1732 1733 1734 1735 1736 1737 1738 1739 1740 1741 1742 1743 1744 1745 1746 1747 1748 1749 1750 1751 1752 1753 1754 1755 1756 1757 1758 1759 1760 1761 1762 1763 1764 1765 1766 1767 1768 1769 1770 1771 1772 1773 1774 1775 1776 1777 1778 1779 1780 1781 1782 1783 1784 1785 1786 1787 1788 1789 1790 1791 1792 1793 1794 1795 1796 1797 1798 1799 1800 1801 1802 1803 1804 1805 1806 1807 1808 1809 1810 1811 1812 1813 1814 1815 1816 1817 1818 1819 1820 1821 1822 1823 1824 1825 1826 1827 1828 1829 1830 1831 1832 1833 1834 1835 1836 1837 1838 1839 1840 1841 1842 1843 1844 1845 1846 1847 1848 1849 1850 1851 1852 1853 1854 1855 1856 1857 1858 1859 1860 1861 1862 1863 1864 1865 1866 1867 1868 1869 1870 1871 1872 1873 1874 1875 1876 1877 1878 1879 1880 1881 1882 1883 1884 1885 1886 1887 1888 1889 1890 1891 1892 1893 1894 1895 1896 1897 1898 1899 1900 1901 1902 1903 1904 1905 1906 1907 1908 1909 1910 1911 1912 1913 1914 1915 1916 1917 1918 1919 1920 1921 1922 1923 1924 1925 1926 1927 1928 1929 1930 1931 1932 1933 1934 |
{
"paper_id": "R11-1026",
"header": {
"generated_with": "S2ORC 1.0.0",
"date_generated": "2023-01-19T15:04:41.294289Z"
},
"title": "Cross-Domain Dutch Coreference Resolution",
"authors": [
{
"first": "Orph\u00e9e",
"middle": [],
"last": "De Clercq",
"suffix": "",
"affiliation": {},
"email": "orphee.declercq@hogent.be"
},
{
"first": "V\u00e9ronique",
"middle": [],
"last": "Hoste",
"suffix": "",
"affiliation": {},
"email": "veronique.hoste@hogent.be"
},
{
"first": "Iris",
"middle": [],
"last": "Hendrickx",
"suffix": "",
"affiliation": {},
"email": ""
}
],
"year": "",
"venue": null,
"identifiers": {},
"abstract": "This article explores the portability of a coreference resolver across a variety of eight text genres. Besides newspaper text, we also include administrative texts, autocues, texts used for external communication, instructive texts, wikipedia texts, medical texts and unedited new media texts. Three sets of experiments were conducted. First, we investigated each text genre individually, and studied the effect of larger training set sizes and including genre-specific training material. Then, we explored the predictive power of each genre for the other genres conducting cross-domain experiments. In a final step, we investigated whether excluding genres with less predictive power increases overall performance. For all experiments we use an existing Dutch mention-pair resolver and report on our experimental results using four metrics: MUC, B-cubed, CEAF and BLANC. We show that resolving out-of-domain genres works best when enough training data is included. This effect is further intensified by including a small amount of genre-specific text. As far as the cross-domain performance is concerned we see that especially genres of a very specific nature tend to have less generalization power.",
"pdf_parse": {
"paper_id": "R11-1026",
"_pdf_hash": "",
"abstract": [
{
"text": "This article explores the portability of a coreference resolver across a variety of eight text genres. Besides newspaper text, we also include administrative texts, autocues, texts used for external communication, instructive texts, wikipedia texts, medical texts and unedited new media texts. Three sets of experiments were conducted. First, we investigated each text genre individually, and studied the effect of larger training set sizes and including genre-specific training material. Then, we explored the predictive power of each genre for the other genres conducting cross-domain experiments. In a final step, we investigated whether excluding genres with less predictive power increases overall performance. For all experiments we use an existing Dutch mention-pair resolver and report on our experimental results using four metrics: MUC, B-cubed, CEAF and BLANC. We show that resolving out-of-domain genres works best when enough training data is included. This effect is further intensified by including a small amount of genre-specific text. As far as the cross-domain performance is concerned we see that especially genres of a very specific nature tend to have less generalization power.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Abstract",
"sec_num": null
}
],
"body_text": [
{
"text": "Coreference resolution is the task of automatically recognizing which words or expressions refer to the same discourse entity in a particular text or dialogue. 1 In the last decade considerable efforts have been put in annotating corpora with coreferential relations. Not only a widespread language such as English (e.g. ACE-2 (Doddington et al., 2004) , ARRAU (Poesio and Artstein, 2008) , OntoNotes 3.0 (Weischedel et al., 2009) ), but also Czech (PDT 2.0 (Ku\u010dov\u00e1 and Haji\u010dov\u00e1, 2004) ), Catalan (AnCora-Ca ) and Italian (I-CAB (Magnini et al., 2006) ) 2 can now rely on substantial resources for coreference research.",
"cite_spans": [
{
"start": 160,
"end": 161,
"text": "1",
"ref_id": null
},
{
"start": 327,
"end": 352,
"text": "(Doddington et al., 2004)",
"ref_id": "BIBREF7"
},
{
"start": 361,
"end": 388,
"text": "(Poesio and Artstein, 2008)",
"ref_id": "BIBREF22"
},
{
"start": 405,
"end": 430,
"text": "(Weischedel et al., 2009)",
"ref_id": null
},
{
"start": 458,
"end": 485,
"text": "(Ku\u010dov\u00e1 and Haji\u010dov\u00e1, 2004)",
"ref_id": "BIBREF15"
},
{
"start": 529,
"end": 551,
"text": "(Magnini et al., 2006)",
"ref_id": "BIBREF18"
}
],
"ref_spans": [],
"eq_spans": [],
"section": "Introduction",
"sec_num": "1"
},
{
"text": "One of the challenges in many current NLP tasks is to test their portability across different domains and languages. This portability to other languages was the main objective of the SemEval 2010 Task on Coreference Resolution in Multiple Languages . The issue of domain portability was the focus of the ACL 2010 Workshop on Domain Adaptation for NLP (Daum\u00e9 III et al., 2010) .",
"cite_spans": [
{
"start": 351,
"end": 375,
"text": "(Daum\u00e9 III et al., 2010)",
"ref_id": "BIBREF6"
}
],
"ref_spans": [],
"eq_spans": [],
"section": "Introduction",
"sec_num": "1"
},
{
"text": "In this paper we investigate the performance of an existing mention-pair coreference resolver for Dutch (Hoste, 2005; Hendrickx et al., 2008b) across various text genres. More specifically we want to know whether training on out-of-domain data can be done without performance loss. The above-mentioned corpora designed for coreference resolution consist almost exclusively of text from the same genre, i.e. newspaper texts, and as a consequence resulting coreference resolvers are mostly trained on this particular genre. Moreover, when other genres are included, the acquired data are rather scarce: 25K of dialogues in AR-RAU (Poesio and Artstein, 2008) , 23K manuals in AnATar (Hammami et al., 2009) or 50K of annotated blogs in LiveMemories (Rodr\u00edguez et al., 2010) . Another related study is the work of Longo and Todirascu (2010) . They analyzed a French corpus (50K) consisting of 5 different text genres to develop genre-specific features; in their study they use genre-specific features such as average length of the coreferential chain and average distance separating several mentions of the same referent. An exception to this observation of small datasets is the new OntoNotes 4.0 corpus that is used for the CoNLL 2011 Shared Task on unrestricted coreference resolution, as the corpus contains approximately 1 million words from 5 different text genres. 3 We do see a growing interest in one specific different text genre, namely biomedical text in many NLP tasks, including coreference resolution (e.g Yang et al. (2004) , Gasperin and Briscoe (2008) , Ngan Nguyen and Tsujii (2008) ).",
"cite_spans": [
{
"start": 104,
"end": 117,
"text": "(Hoste, 2005;",
"ref_id": "BIBREF13"
},
{
"start": 118,
"end": 142,
"text": "Hendrickx et al., 2008b)",
"ref_id": "BIBREF12"
},
{
"start": 628,
"end": 655,
"text": "(Poesio and Artstein, 2008)",
"ref_id": "BIBREF22"
},
{
"start": 680,
"end": 702,
"text": "(Hammami et al., 2009)",
"ref_id": "BIBREF9"
},
{
"start": 732,
"end": 769,
"text": "LiveMemories (Rodr\u00edguez et al., 2010)",
"ref_id": null
},
{
"start": 809,
"end": 835,
"text": "Longo and Todirascu (2010)",
"ref_id": "BIBREF16"
},
{
"start": 1367,
"end": 1368,
"text": "3",
"ref_id": null
},
{
"start": 1516,
"end": 1534,
"text": "Yang et al. (2004)",
"ref_id": "BIBREF36"
},
{
"start": 1537,
"end": 1564,
"text": "Gasperin and Briscoe (2008)",
"ref_id": "BIBREF8"
},
{
"start": 1583,
"end": 1596,
"text": "Tsujii (2008)",
"ref_id": "BIBREF21"
}
],
"ref_spans": [],
"eq_spans": [],
"section": "Introduction",
"sec_num": "1"
},
{
"text": "The data for the experiments come from three Dutch corpus projects in which coreference was annotated: COREA (Hendrickx et al., 2008a) , DuOMAn (Hendrickx and Hoste, 2009) and SoNaR (Schuurman et al., 2010) 4 . Combining these three resources allows us to work with diverse data spread over different text genres. Another advantage is that all data was annotated following the same approach: first all NPs were pre-tagged based on syntactic dependency structures (Bouma and Kloostermans, 2007) and secondly the COREA guidelines were reused in each project. Though the emphasis in this study is on edited text, we also include unedited text, viz. blogs and news comments (Hendrickx and Hoste, 2009) . With this crossdomain portability study, we aim to see which genres perform better or worse and whether it is possible to determine a priori which training data to add to our resolver so as to obtain better results. The results are presented using four of the more frequently used evaluation metrics for coreference research, namely MUC (Vilain et al., 1995) , Bcubed (Bagga and Baldwin, 1998) , CEAF (Luo and Zitouni, 2005) and BLANC (Recasens and Hovy, 2011) .",
"cite_spans": [
{
"start": 109,
"end": 134,
"text": "(Hendrickx et al., 2008a)",
"ref_id": "BIBREF11"
},
{
"start": 144,
"end": 171,
"text": "(Hendrickx and Hoste, 2009)",
"ref_id": "BIBREF10"
},
{
"start": 182,
"end": 208,
"text": "(Schuurman et al., 2010) 4",
"ref_id": null
},
{
"start": 463,
"end": 493,
"text": "(Bouma and Kloostermans, 2007)",
"ref_id": "BIBREF1"
},
{
"start": 670,
"end": 697,
"text": "(Hendrickx and Hoste, 2009)",
"ref_id": "BIBREF10"
},
{
"start": 1037,
"end": 1058,
"text": "(Vilain et al., 1995)",
"ref_id": "BIBREF33"
},
{
"start": 1068,
"end": 1093,
"text": "(Bagga and Baldwin, 1998)",
"ref_id": "BIBREF0"
},
{
"start": 1101,
"end": 1124,
"text": "(Luo and Zitouni, 2005)",
"ref_id": "BIBREF17"
},
{
"start": 1135,
"end": 1160,
"text": "(Recasens and Hovy, 2011)",
"ref_id": "BIBREF25"
}
],
"ref_spans": [],
"eq_spans": [],
"section": "Introduction",
"sec_num": "1"
},
{
"text": "We show that adding more data to training proves mostly beneficial, especially when genrespecific information is included. Moreover, training a resolver on each genre separately allows us to classify each genre as having good or bad generalization power when applied to other genres. This led us to conduct experiments in which we train on all genres while progressively leaving out the worst-performing cross-domain genres as an attempt to boost overall performance. Although the results are sometimes better, performance does not rise nor drop dramatically. We show that inclusion of some genre-specific training material is necessary, especially when less generalizable genres are to be labeled. However, most effect is perceived by adding more data to training.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Introduction",
"sec_num": "1"
},
{
"text": "The remainder of this paper is organized as follows. In Section 2, we present the datasets and experimental setup of our system and briefly discuss the different evaluation metrics. In Section 3 the results are presented and analyzed, and we report on our experience with the different evaluation metrics. Section 4 concludes this paper by formulating some conclusions and prospects for future work.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Introduction",
"sec_num": "1"
},
{
"text": "In the present study, we aim to investigate the cross-genre portability of an existing mention-pair coreference resolver for Dutch. In order to do so, our system's performance was compared on eight datasets: administrative texts (ADM), autocues (AUTO), texts used for external communication (EXT), instructive texts (INST), journalistic texts (JOUR), medical texts (MED), wikipedia (WIKI), and unedited text (DUO). All data were manually annotated using the COREA guidelines . These guidelines allow for the annotation of four relations and special cases are flagged. The four annotated relations are identity (NPs referring to the same discourse entity), bound (expressing properties of general categories), bridge (as in part-whole, superset-subset relations) and predicative. The following special cases were flagged: negations and expressions of modality, time-dependency and identity of sense (as in the so-called paycheck pronouns (Karttunen, 1976) ). As annotation environment, the MMAX2 annotation software 5 was used.",
"cite_spans": [
{
"start": 937,
"end": 954,
"text": "(Karttunen, 1976)",
"ref_id": "BIBREF14"
}
],
"ref_spans": [],
"eq_spans": [],
"section": "Datasets and Experimental Setup",
"sec_num": "2"
},
{
"text": "To rule out data size as a possible explanation for performance shifts, datasets of equal size (about 30K) were randomly selected. The focus of the current experiments was on resolving identity and predicative relations. Table 1 gives some statistics about each dataset, such as the average sentence length and the number of coreferring NPs.",
"cite_spans": [],
"ref_spans": [
{
"start": 221,
"end": 228,
"text": "Table 1",
"ref_id": "TABREF1"
}
],
"eq_spans": [],
"section": "Datasets and Experimental Setup",
"sec_num": "2"
},
{
"text": "For all experiments we used an existing coreference resolver for Dutch, developed by Hoste (2005) and Hendrickx et al. (2008b) . The system follows a machine learning approach 6 based on the seminal work of Soon et al. (2001) and represents a mention-pair model. First, a classifier is trained to decide whether a pair of NPs is coreferential or not, after which coreference chains are built for the pairs of NPs that were classified as coreferential. All datasets were preprocessed in the same way. Tokenisation, lemmatisation, Part-of-Speech tagging and grammatical relations were based on the manually verified output of the Alpino parser (Bouma et al., 2001 ), i.e. gold standard dependency structures. For the DuOMAn data, however, no gold standard dependency trees were available. Named entity recognition was performed using MBT (Daelemans et al., 2003) , trained on the 2002 CoNNL shared task Dutch dataset (Tjong Kim Sang, 2002) and an additional gazetteer lookup. As features we employ string matching, distance between sentences and NPs, grammatical role and named entity overlap, synonym/hypernym lookup using Cornetto (a Dutch database combining Dutch Wordnet (Vossen, 1998) and the Referentie Bestand Nederlands (Martin and Ploeger, 1999) ) and local context. All instances were built between NP pairs going 20 sentences back in context. NPs that are not part of a coreferential chain (singletons) are included as negative examples. For more information we refer to Hoste (2005) and Hendrickx et al. (2008a) .",
"cite_spans": [
{
"start": 85,
"end": 97,
"text": "Hoste (2005)",
"ref_id": "BIBREF13"
},
{
"start": 102,
"end": 126,
"text": "Hendrickx et al. (2008b)",
"ref_id": "BIBREF12"
},
{
"start": 207,
"end": 225,
"text": "Soon et al. (2001)",
"ref_id": null
},
{
"start": 642,
"end": 661,
"text": "(Bouma et al., 2001",
"ref_id": "BIBREF2"
},
{
"start": 836,
"end": 860,
"text": "(Daelemans et al., 2003)",
"ref_id": "BIBREF4"
},
{
"start": 926,
"end": 937,
"text": "Sang, 2002)",
"ref_id": "BIBREF32"
},
{
"start": 1173,
"end": 1187,
"text": "(Vossen, 1998)",
"ref_id": null
},
{
"start": 1226,
"end": 1252,
"text": "(Martin and Ploeger, 1999)",
"ref_id": "BIBREF19"
},
{
"start": 1480,
"end": 1492,
"text": "Hoste (2005)",
"ref_id": "BIBREF13"
},
{
"start": 1497,
"end": 1521,
"text": "Hendrickx et al. (2008a)",
"ref_id": "BIBREF11"
}
],
"ref_spans": [],
"eq_spans": [],
"section": "Datasets and Experimental Setup",
"sec_num": "2"
},
{
"text": "Since the focus of this study is on genre, we decided not to train on different NP types (pronouns, common nouns and proper names) individually. 7 For all experiments we used Timbl version 6.3 (Daelemans et al., 2010 ) with default parameter settings.",
"cite_spans": [
{
"start": 145,
"end": 146,
"text": "7",
"ref_id": null
},
{
"start": 193,
"end": 216,
"text": "(Daelemans et al., 2010",
"ref_id": "BIBREF5"
}
],
"ref_spans": [],
"eq_spans": [],
"section": "Datasets and Experimental Setup",
"sec_num": "2"
},
{
"text": "Our experimental results are evaluated using the four scoring metrics as implemented in the scoring script from the coreference resolution task from the SemEval-2010 competition :",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Datasets and Experimental Setup",
"sec_num": "2"
},
{
"text": "\u2022 The MUC scoring software (Vilain et al., 1995) counts the number of links between the coreferential elements in the text, and looks how many links are shared or not between the gold standard coreferential chains and the system predictions. As MUC concentrates on links, elements that are not part of a coreferential chain, entities that are only mentioned once (singletons), are not taken into account in this scoring method.",
"cite_spans": [
{
"start": 27,
"end": 48,
"text": "(Vilain et al., 1995)",
"ref_id": "BIBREF33"
}
],
"ref_spans": [],
"eq_spans": [],
"section": "Datasets and Experimental Setup",
"sec_num": "2"
},
{
"text": "\u2022 The B-cubed measure (Bagga and Baldwin, 1998) does not consider mere links between elements, but takes into account the coreferential clusters of elements referring to the same entity. B-cubed computes for every individual element in the text the precision and recall by counting how many elements are in the true coreferential cluster and how many in the predicted coreferential cluster.",
"cite_spans": [
{
"start": 22,
"end": 47,
"text": "(Bagga and Baldwin, 1998)",
"ref_id": "BIBREF0"
}
],
"ref_spans": [],
"eq_spans": [],
"section": "Datasets and Experimental Setup",
"sec_num": "2"
},
{
"text": "\u2022 The CEAF measure (Luo and Zitouni, 2005) focuses on a one-to-one mapping of elements in the true and predicted coreferential clusters. Both B-cubed and CEAF measures are sensitive to the presence of many singletons, the larger the percentage of singletons, the higher these scores become (Recasens and Hovy, 2011) .",
"cite_spans": [
{
"start": 19,
"end": 42,
"text": "(Luo and Zitouni, 2005)",
"ref_id": "BIBREF17"
},
{
"start": 290,
"end": 315,
"text": "(Recasens and Hovy, 2011)",
"ref_id": "BIBREF25"
}
],
"ref_spans": [],
"eq_spans": [],
"section": "Datasets and Experimental Setup",
"sec_num": "2"
},
{
"text": "\u2022 Recently, the BLANC measure (Recasens and Hovy, 2011) was developed to overcome problems with the other scoring methods. This measure is a variant of the Rand Index (Rand, 1971 ) adapted for coreference resolution and it averages over a score for correctly detecting singletons, and a score for detecting the correct cluster for coreferential elements.",
"cite_spans": [
{
"start": 167,
"end": 178,
"text": "(Rand, 1971",
"ref_id": "BIBREF24"
}
],
"ref_spans": [],
"eq_spans": [],
"section": "Datasets and Experimental Setup",
"sec_num": "2"
},
{
"text": "An important remark to make here is that our system does not take into account chains of only one element. As a consequence, contrary to the SemEval-2010 competition, when we compute of these NP types based on the motivation that the impact of different information sources varies per NP type. In order to test cross-genre portability, we ran three sets of experiments (Table 2): 1. In the first set of experiments, we wanted to investigate whether adding more data is beneficial for the classifier. We trained the classifier on each genre individually and compared performance with different training set sizes. Three experiments were conducted: we first trained on each individual genre and tested on the relevant genre using ten-fold cross validation (each fold 27K vs. 3K). In a second experiment, the classifier was trained on all genres except one and tested on the one that was left out (210K vs. 30K). In a third experiment, we used all data, including genrespecific training material for training the classifier, in a ten-fold cross validation set-up (each fold 237K vs. 3K).",
"cite_spans": [],
"ref_spans": [
{
"start": 369,
"end": 379,
"text": "(Table 2):",
"ref_id": "TABREF3"
}
],
"eq_spans": [],
"section": "Datasets and Experimental Setup",
"sec_num": "2"
},
{
"text": "2. In a second set of experiments, we focused on the actual cross-domain portability. In order to test this, we each time trained on one genre and tested the performance of the classifier for each of the other genres.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Datasets and Experimental Setup",
"sec_num": "2"
},
{
"text": "3. Based on the results obtained in the second batch of experiments, we investigated whether some particular genres actually decrease performance when training on all data. In other words, does excluding outlier genres from training data increase performance? This was done by each time leaving out the worst-performing cross-domain genres and performing ten-fold cross validation.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Datasets and Experimental Setup",
"sec_num": "2"
},
{
"text": "The results of the first round of experiments are presented in Figure 1 . The dots marked as individual present the experiments in which each classifier was trained and tested on the same material. The scores for All-individual present experiments in which the classifiers are trained on a large and diverse training set of all different genres except the genre that is held out as a test set. The last experiments in the graph All+individual show the result when training on all genres including the heldout genre. Though the B-cubed and CEAF scores are lower than MUC, they present the same tendency: adding more and diverse training material improves performance, especially when genrespecific information is also included. 8 BLANC, however, seems to contradict the other metrics. Though the scores are higher, they reveal that larger training data proves only beneficial for three genres: INST, JOUR and MED. BLANC thus suggests that training only on in-domain material of some genre is the best approach. This brings us to the cross-genre experiments, where we each time train on one genre and test on all the other genres individually until all genres have been once used as training data. 9 In order to represent the results, we ranked the classifier performance on each genre, ranging from the genreclassifier which on average performs worst when being applied to the other genres to the one performing best. We performed this ranking for each of the four evaluation metrics. The final ranking is visualized in Table 3 . Although there are some differences between the metrics -we again observe that BLANC tends to differ more from the others -they all seem to agree that MED (medical text), DUO (unedited text) and INST (instructive text) constitute poor cross-genre training material. JOUR has been selected by MUC, B3 and CEAF as the best material for training on other genres. As we mentioned in Section 1 that most of the currently available datasets annotated with coreferential information consist of newspaper text, this result shows that this might indeed be a good choice.",
"cite_spans": [],
"ref_spans": [
{
"start": 63,
"end": 71,
"text": "Figure 1",
"ref_id": null
},
{
"start": 1519,
"end": 1526,
"text": "Table 3",
"ref_id": "TABREF5"
}
],
"eq_spans": [],
"section": "Results",
"sec_num": "3"
},
{
"text": "The four metrics confirmed that three genres had less generalization power, viz. MED, DUO and INST. In the third experiment, we aim to op- timize our selection of training data to get the best possible general performance. We hypothesize that leaving out those genres with less predictive power for other genres from the training material will increase overall performance. In this set of experiments we train on all data, including genrespecific information, and test on one genre while progressively leaving out those three genres. The results of this reversed learning curve for all metrics can be found in Table 4 . Whenever a score is printed in bold, it is the best score obtained for a particular genre.",
"cite_spans": [],
"ref_spans": [
{
"start": 610,
"end": 617,
"text": "Table 4",
"ref_id": "TABREF7"
}
],
"eq_spans": [],
"section": "Results",
"sec_num": "3"
},
{
"text": "It is difficult to compare the different metrics with each other. We observe that only the BLANC metric confirms our expectation that the results are almost always better when poor training material is excluded from training. The results as measured with the other 3 metrics, however, show that leaving out data is only beneficial for half of the datasets. Overall, these results do not strongly confirm our hypothesis. An important observation to make is that, for all metrics, the performance gains which are obtained by leaving out data are modest, the effect of removing data is very small. Based on these observations we conclude that to get good generalization performance it is more important to have a large training set than to put time and effort in the composition of this training set.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Results",
"sec_num": "3"
},
{
"text": "Three genres, viz. MED, DUO and INST, did not score high in the cross-domain experiments and were the first genres to be left out in the final experiments. An error analysis on this data imposed itself. Looking at the data itself we see that MED includes data of a scientific nature consisting of various entries in a medical encyclopedia. DUO contains mostly user-generated text as it consists of texts from blogs and newspaper articles together with a large set of reader comments. This type of data is rather different from the other genres as it is unedited, subjective, informal and more similar to spoken language than the other genres. INST contains various patient information leaflets and manuals in which exactly the same sentences are often repeated with only one word -mostly the name of the product -different. The above observations already hint at the low generalizability of these three genres.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Error Analysis",
"sec_num": "3.1"
},
{
"text": "Compared to the other genres, who on average contain 25% of coreferential NPs, we note that MED and INST contain a high number of coreferential NPs (respectively 33% and 37%) and DUO a rather low amount (viz. 18%). Looking at the data statistics given in Table 1 , we observe that MED slightly differs from the others: it consists of 213 smaller documents and the average sentence length is shorter, viz. 14.4 words. Moreover, looking at the subdivision of NPs we see that MED contains a large number of common nouns (89%) and only few pronouns (5%) and proper nouns (6%). In the other five datasets, this division ranges between 70-75% common nouns and 10-15% pronouns and proper nouns. When using MED as training data this results in a higher number of introduced errors between common nouns. Especially when no string matching features are found between two common nouns the resolver has a lot of difficulty into correctly classifying them. Of all genres we see that with MED pronouns and proper nouns are harder to recognize, which can be explained by their low coverage in the training data. Having a closer look at the DUO dataset, we see that the division between common, proper and pronouns is 64% -14% -22% -which is a high number of pronouns. Counterintuitively, this does not mean that resolving pronouns goes better when training on DUO. On the contrary,we see that although the resolution of pronouns rises slightly, more errors are introduced. Dutch pronouns also turned out to be difficult to resolve ac-cording to Hoste (2005) because of the inability to distinguish between anaphoric and pleonastic pronouns. The NP subdivision in INST is comparable to the five other genres, with a small preference for proper nouns. The high amount of reoccurring sentences in the data is also reflected in the features, the INST dataset scored best when performing in-domain experiments because of the many exact matches. Furthermore, as many technical NPs are not covered by WordNet (and these semantic features are crucial for most genres), important links between two NPs are missed. In sum, these three genres have very specific features that seem to make them less predictive for other genres.",
"cite_spans": [
{
"start": 1530,
"end": 1542,
"text": "Hoste (2005)",
"ref_id": "BIBREF13"
}
],
"ref_spans": [
{
"start": 255,
"end": 262,
"text": "Table 1",
"ref_id": "TABREF1"
}
],
"eq_spans": [],
"section": "Error Analysis",
"sec_num": "3.1"
},
{
"text": "In this paper we explored the portability of an existing coreference resolver for Dutch when applied to eight different text genres: administrative texts, autocues, texts used for external communication, instructive texts, journalistic texts, medical texts, wikipedia and unedited new media texts. By comparing the performance on three sets of experiments, we found that larger training set size improves performance, especially when genre-specific training material (10%) is included. We saw that excluding poor cross-genre training material does not always results in better scores neither can a drop in performance be perceived. This might imply that training on more data with higher predictive power is more important than training on various text genres. This is something we definitely wish to look into in closer detail in future work. Moreover, we would like to find out how much genre-specific training data is exactly needed to optimize performance. We discovered that especially genres containing very specific (e.g. scientific or unedited) data and having a different subdivision between pronouns, common and proper nous are less equipped for crossgenre experiments and thus have less generalization power.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Conclusion",
"sec_num": "4"
},
{
"text": "We also observe that the different evaluation metrics for coreference research in use today, (MUC, B-cubed, CEAF and BLANC) tend to contradict each other and as a consequence hamper interpretation. This is a well-known problem within the community for which no solution has been found yet. In order to allow for a better comparison with the SemEval-2010 competition we intend to have a closer look at the effect of also scoring singletons.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Conclusion",
"sec_num": "4"
},
{
"text": "In this article we only discuss nominal coreference, i.e. which coreferential relations exist between noun phrases (common and proper nouns, pronouns).",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "",
"sec_num": null
},
{
"text": "For a more complete overview we refer to(Recasens, 2010) and(Poesio et al., forthcoming)",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "",
"sec_num": null
},
{
"text": "Website from CoNLL 2011: http://conll.bbn.com 4 SoNaR is currently still under development.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "",
"sec_num": null
},
{
"text": "http://mmax2.net",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "",
"sec_num": null
},
{
"text": "For an extensive overview of the different machine learning approaches for coreference resolution, we refer to the surveys ofNg (2010) andPoesio et al. (forthcoming) 7Hoste (2005) built a separate learning module for each",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "",
"sec_num": null
},
{
"text": "Because of space constraints we only incorporated two graphs in this paper.9 Train on ADM = test on AUTO; train on ADM test on DUO;....",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "",
"sec_num": null
}
],
"back_matter": [
{
"text": "The work presented in this paper was made possible by the STEVIN programme of the Dutch Language Union within the framework of the SoNaR project under grant number STE07014 and the Portuguese Science Foundation, FCT (Funda\u00e7\u00e3o para a Ci\u00eancia e a Tecnologia). We would like to thank the anonymous reviewers for their helpful comments and valuable suggestions.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Acknowledgments",
"sec_num": null
}
],
"bib_entries": {
"BIBREF0": {
"ref_id": "b0",
"title": "Algorithms for scoring coreference chains",
"authors": [
{
"first": "Amit",
"middle": [],
"last": "Bagga",
"suffix": ""
},
{
"first": "Breck",
"middle": [],
"last": "Baldwin",
"suffix": ""
}
],
"year": 1998,
"venue": "Proceedings of the First International Conference on Language Resources and Evaluation Workshop on Linguistic Coreference",
"volume": "",
"issue": "",
"pages": "563--566",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Amit Bagga and Breck Baldwin. 1998. Algorithms for scoring coreference chains. In Proceedings of the First International Conference on Language Resources and Evaluation Workshop on Linguistic Coreference, pages 563-566.",
"links": null
},
"BIBREF1": {
"ref_id": "b1",
"title": "Mining Syntactically Annotated Corpora using XQuery",
"authors": [
{
"first": "Gosse",
"middle": [],
"last": "Bouma",
"suffix": ""
},
{
"first": "Geert",
"middle": [],
"last": "Kloostermans",
"suffix": ""
}
],
"year": 2007,
"venue": "Proceedings of the Linguistic Annotation Workshop (held in conjunction with ACL 2007)",
"volume": "",
"issue": "",
"pages": "17--24",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Gosse Bouma and Geert Kloostermans. 2007. Mining Syntactically Annotated Corpora using XQuery. In Proceedings of the Linguistic Annotation Workshop (held in conjunction with ACL 2007), pages 17-24, Prague, Czech Republic.",
"links": null
},
"BIBREF2": {
"ref_id": "b2",
"title": "Alpino: Wide coverage computational analysis of dutch",
"authors": [
{
"first": "Gosse",
"middle": [],
"last": "Bouma",
"suffix": ""
},
{
"first": "Robert",
"middle": [],
"last": "Gertjan Van Noord",
"suffix": ""
},
{
"first": "",
"middle": [],
"last": "Malouf",
"suffix": ""
}
],
"year": 2001,
"venue": "Computational Linguistics in the Netherlands 2000: selected papers from the twentieth CLIN meeting",
"volume": "",
"issue": "",
"pages": "",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Gosse Bouma, Gertjan van Noord, and Robert Malouf. 2001. Alpino: Wide coverage computational anal- ysis of dutch. In Computational Linguistics in the Netherlands 2000: selected papers from the twenti- eth CLIN meeting.",
"links": null
},
"BIBREF3": {
"ref_id": "b3",
"title": "The COREA-project, manual for the annotation of coreference in Dutch texts",
"authors": [
{
"first": "Gosse",
"middle": [],
"last": "Bouma",
"suffix": ""
},
{
"first": "Walter",
"middle": [],
"last": "Daelemans",
"suffix": ""
},
{
"first": "Iris",
"middle": [],
"last": "Hendrickx",
"suffix": ""
},
{
"first": "V\u00e9ronique",
"middle": [],
"last": "Hoste",
"suffix": ""
},
{
"first": "Anne-Marie",
"middle": [],
"last": "Mineur",
"suffix": ""
}
],
"year": 2007,
"venue": "",
"volume": "",
"issue": "",
"pages": "",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Gosse Bouma, Walter Daelemans, Iris Hendrickx, V\u00e9ronique Hoste, and Anne-Marie Mineur. 2007. The COREA-project, manual for the annotation of coreference in Dutch texts. Technical report, Uni- versity Groningen.",
"links": null
},
"BIBREF4": {
"ref_id": "b4",
"title": "MBT: Memory Based Tagger, version 2.0, Reference Guide",
"authors": [
{
"first": "Walter",
"middle": [],
"last": "Daelemans",
"suffix": ""
},
{
"first": "Jakub",
"middle": [],
"last": "Zavrel",
"suffix": ""
},
{
"first": "",
"middle": [],
"last": "Van Den",
"suffix": ""
},
{
"first": "Ko",
"middle": [],
"last": "Bosch",
"suffix": ""
},
{
"first": "",
"middle": [],
"last": "Van Der",
"suffix": ""
},
{
"first": "",
"middle": [],
"last": "Sloot",
"suffix": ""
}
],
"year": 2003,
"venue": "",
"volume": "",
"issue": "",
"pages": "",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Walter Daelemans, Jakub Zavrel, Antal van den Bosch, and Ko van der Sloot. 2003. MBT: Memory Based Tagger, version 2.0, Reference Guide. Technical Re- port ILK Research Group Technical Report Series no. 03-13, Tilburg University.",
"links": null
},
"BIBREF5": {
"ref_id": "b5",
"title": "TiMBL: Tilburg Memory Based Learner, version 6.3, Reference Guide",
"authors": [
{
"first": "Walter",
"middle": [],
"last": "Daelemans",
"suffix": ""
},
{
"first": "Jakub",
"middle": [],
"last": "Zavrel",
"suffix": ""
},
{
"first": "Ko",
"middle": [],
"last": "Van Der",
"suffix": ""
},
{
"first": "Antal",
"middle": [],
"last": "Sloot",
"suffix": ""
},
{
"first": "",
"middle": [],
"last": "Van Den",
"suffix": ""
},
{
"first": "",
"middle": [],
"last": "Bosch",
"suffix": ""
}
],
"year": 2010,
"venue": "Technical Report ILK Research Group Technical Report Series",
"volume": "",
"issue": "10",
"pages": "",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Walter Daelemans, Jakub Zavrel, Ko Van der Sloot, and Antal van den Bosch. 2010. TiMBL: Tilburg Mem- ory Based Learner, version 6.3, Reference Guide. Technical Report ILK Research Group Technical Report Series no. 10-01, Tilburg University.",
"links": null
},
"BIBREF6": {
"ref_id": "b6",
"title": "Proceedings of the 2010 Workshop on Domain Adaptation for Natural Language Processing. Association for Computational Linguistics",
"authors": [
{
"first": "Hal",
"middle": [],
"last": "Daum\u00e9",
"suffix": ""
},
{
"first": "Iii",
"middle": [],
"last": "",
"suffix": ""
},
{
"first": "Tejaswini",
"middle": [],
"last": "Deoskar",
"suffix": ""
},
{
"first": "David",
"middle": [],
"last": "Mcclosky",
"suffix": ""
}
],
"year": 2010,
"venue": "",
"volume": "",
"issue": "",
"pages": "",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Hal Daum\u00e9 III, Tejaswini Deoskar, David McClosky, Barbara Plank, and J\u00f6rg Tiedemann, editors. 2010. Proceedings of the 2010 Workshop on Domain Adaptation for Natural Language Processing. As- sociation for Computational Linguistics, Uppsala, Sweden, July.",
"links": null
},
"BIBREF7": {
"ref_id": "b7",
"title": "The Automatic Content Extraction (ACE) Program Tasks, Data, and Evaluation",
"authors": [
{
"first": "George",
"middle": [],
"last": "Doddington",
"suffix": ""
},
{
"first": "Alexis",
"middle": [],
"last": "Mitchell",
"suffix": ""
},
{
"first": "Mark",
"middle": [],
"last": "Przybocki",
"suffix": ""
},
{
"first": "Lance",
"middle": [],
"last": "Ramshaw",
"suffix": ""
},
{
"first": "Stephanie",
"middle": [],
"last": "Strassel",
"suffix": ""
},
{
"first": "Ralph",
"middle": [],
"last": "Weischedel",
"suffix": ""
}
],
"year": 2004,
"venue": "Proceedings of LREC 2004",
"volume": "",
"issue": "",
"pages": "837--840",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "George Doddington, Alexis Mitchell, Mark Przybocki, Lance Ramshaw, Stephanie Strassel, and Ralph Weischedel. 2004. The Automatic Content Extrac- tion (ACE) Program Tasks, Data, and Evaluation. In Proceedings of LREC 2004, pages 837-840, Lisbon, Portugal.",
"links": null
},
"BIBREF8": {
"ref_id": "b8",
"title": "Statistical anaphora resolution in biomedical texts",
"authors": [
{
"first": "Caroline",
"middle": [],
"last": "Gasperin",
"suffix": ""
},
{
"first": "Ted",
"middle": [],
"last": "Briscoe",
"suffix": ""
}
],
"year": 2008,
"venue": "Proceedings of the 22nd International Conference on Computational Linguistics",
"volume": "",
"issue": "",
"pages": "257--264",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Caroline Gasperin and Ted Briscoe. 2008. Statis- tical anaphora resolution in biomedical texts. In Proceedings of the 22nd International Conference on Computational Linguistics (Coling 2008), pages 257-264, Manchester, UK, August. Coling 2008 Or- ganizing Committee.",
"links": null
},
"BIBREF9": {
"ref_id": "b9",
"title": "Arabic anaphora resolution: Corpora annotation with coreferential links",
"authors": [
{
"first": "Souha",
"middle": [],
"last": "Hammami",
"suffix": ""
},
{
"first": "Lamia",
"middle": [],
"last": "Belguith",
"suffix": ""
},
{
"first": "Abdelmajid Ben",
"middle": [],
"last": "Hamadou",
"suffix": ""
}
],
"year": 2009,
"venue": "The International Arab Journal of Information Technology",
"volume": "6",
"issue": "5",
"pages": "481--489",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Souha Hammami, Lamia Belguith, and Abdelma- jid Ben Hamadou. 2009. Arabic anaphora reso- lution: Corpora annotation with coreferential links. The International Arab Journal of Information Tech- nology, 6(5):481-489.",
"links": null
},
"BIBREF10": {
"ref_id": "b10",
"title": "Coreference Resolution on Blogs and Commented News",
"authors": [
{
"first": "Iris",
"middle": [],
"last": "Hendrickx",
"suffix": ""
},
{
"first": "V\u00e9ronique",
"middle": [],
"last": "Hoste",
"suffix": ""
}
],
"year": 2009,
"venue": "Anaphora Processing and Applications",
"volume": "5847",
"issue": "",
"pages": "43--53",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Iris Hendrickx and V\u00e9ronique Hoste. 2009. Coref- erence Resolution on Blogs and Commented News. In Anaphora Processing and Applications, Lecture Notes in Artificial Intelligence, volume 5847, pages 43-53, Heidelberg.",
"links": null
},
"BIBREF11": {
"ref_id": "b11",
"title": "A coreference corpus and resolution system for Dutch",
"authors": [
{
"first": "Iris",
"middle": [],
"last": "Hendrickx",
"suffix": ""
},
{
"first": "Gosse",
"middle": [],
"last": "Bouma",
"suffix": ""
},
{
"first": "Frederik",
"middle": [],
"last": "Coppens",
"suffix": ""
},
{
"first": "Walter",
"middle": [],
"last": "Daelemans",
"suffix": ""
},
{
"first": "V\u00e9ronique",
"middle": [],
"last": "Hoste",
"suffix": ""
},
{
"first": "Geert",
"middle": [],
"last": "Kloosterman",
"suffix": ""
},
{
"first": "Anne-Marie",
"middle": [],
"last": "Mineur",
"suffix": ""
},
{
"first": "Joeri",
"middle": [],
"last": "Van Der",
"suffix": ""
},
{
"first": "Jean-Luc",
"middle": [],
"last": "Vloet",
"suffix": ""
},
{
"first": "",
"middle": [],
"last": "Verschelde",
"suffix": ""
}
],
"year": 2008,
"venue": "Proceedings of LREC 2008",
"volume": "",
"issue": "",
"pages": "144--149",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Iris Hendrickx, Gosse Bouma, Frederik Coppens, Wal- ter Daelemans, V\u00e9ronique Hoste, Geert Klooster- man, Anne-Marie. Mineur, Joeri Van Der Vloet, and Jean-Luc Verschelde. 2008a. A coreference corpus and resolution system for Dutch. In Proceedings of LREC 2008, pages 144-149, Marrakech, Morocco.",
"links": null
},
"BIBREF12": {
"ref_id": "b12",
"title": "Semantic and Syntactic features for Anaphora Resolution for Dutch",
"authors": [
{
"first": "Iris",
"middle": [],
"last": "Hendrickx",
"suffix": ""
},
{
"first": "V\u00e9ronique",
"middle": [],
"last": "Hoste",
"suffix": ""
},
{
"first": "Walter",
"middle": [],
"last": "Daelemans",
"suffix": ""
}
],
"year": 2008,
"venue": "Proceedings of the 9th International Conference on Intelligent Text Processing and Computational Linguistics",
"volume": "4919",
"issue": "",
"pages": "351--361",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Iris Hendrickx, V\u00e9ronique Hoste, and Walter Daele- mans. 2008b. Semantic and Syntactic features for Anaphora Resolution for Dutch. In Proceedings of the 9th International Conference on Intelligent Text Processing and Computational Linguistics, Lecture Notes in Computer Science, volume 4919, pages 351-361, Haifa, Israel.",
"links": null
},
"BIBREF13": {
"ref_id": "b13",
"title": "Optimization Issues in Machine Learning of Coreference Resolution",
"authors": [
{
"first": "V\u00e9ronique",
"middle": [],
"last": "Hoste",
"suffix": ""
}
],
"year": 2005,
"venue": "",
"volume": "",
"issue": "",
"pages": "",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "V\u00e9ronique Hoste. 2005. Optimization Issues in Ma- chine Learning of Coreference Resolution. Ph.D. thesis, Antwerp University.",
"links": null
},
"BIBREF14": {
"ref_id": "b14",
"title": "Discourse referents. Syntax and Semantics",
"authors": [
{
"first": "Lauri",
"middle": [],
"last": "Karttunen",
"suffix": ""
}
],
"year": 1976,
"venue": "",
"volume": "7",
"issue": "",
"pages": "",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Lauri Karttunen. 1976. Discourse referents. Syntax and Semantics, 7.",
"links": null
},
"BIBREF15": {
"ref_id": "b15",
"title": "Coreferential relations in the Prague Dependency Treebank",
"authors": [
{
"first": "Lucie",
"middle": [],
"last": "Ku\u010dov\u00e1",
"suffix": ""
},
{
"first": "Eva",
"middle": [],
"last": "Haji\u010dov\u00e1",
"suffix": ""
}
],
"year": 2004,
"venue": "Proceedings of DAARC 2004",
"volume": "",
"issue": "",
"pages": "97--102",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Lucie Ku\u010dov\u00e1 and Eva Haji\u010dov\u00e1. 2004. Coreferen- tial relations in the Prague Dependency Treebank. In Proceedings of DAARC 2004, pages 97-102, Azores, Portugal.",
"links": null
},
"BIBREF16": {
"ref_id": "b16",
"title": "Genrebased reference chains identification for french",
"authors": [
{
"first": "Laurence",
"middle": [],
"last": "Longo",
"suffix": ""
},
{
"first": "Amalia",
"middle": [],
"last": "Todirascu",
"suffix": ""
}
],
"year": 2010,
"venue": "vestigationes Linguisticae",
"volume": "21",
"issue": "",
"pages": "57--75",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Laurence Longo and Amalia Todirascu. 2010. Genre- based reference chains identification for french. In- vestigationes Linguisticae, 21:57-75.",
"links": null
},
"BIBREF17": {
"ref_id": "b17",
"title": "Multi-lingual coreference resolution with syntactic features",
"authors": [
{
"first": "Xiaoqiang",
"middle": [],
"last": "Luo",
"suffix": ""
},
{
"first": "Imed",
"middle": [],
"last": "Zitouni",
"suffix": ""
}
],
"year": 2005,
"venue": "Proceedings of Human Language Technology Conference and Conference on Empirical Methods in Natural Language Processing",
"volume": "",
"issue": "",
"pages": "660--667",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Xiaoqiang Luo and Imed Zitouni. 2005. Multi-lingual coreference resolution with syntactic features. In Proceedings of Human Language Technology Con- ference and Conference on Empirical Methods in Natural Language Processing, pages 660-667.",
"links": null
},
"BIBREF18": {
"ref_id": "b18",
"title": "I-CAB: the Italian Content Annotation Bank",
"authors": [
{
"first": "Bernardo",
"middle": [],
"last": "Magnini",
"suffix": ""
},
{
"first": "Emanuele",
"middle": [],
"last": "Pianta",
"suffix": ""
},
{
"first": "Christian",
"middle": [],
"last": "Girardi",
"suffix": ""
},
{
"first": "Matteo",
"middle": [],
"last": "Negri",
"suffix": ""
},
{
"first": "Lorenza",
"middle": [],
"last": "Romano",
"suffix": ""
},
{
"first": "Manuela",
"middle": [],
"last": "Speranza",
"suffix": ""
},
{
"first": "Valentina",
"middle": [],
"last": "Bartalesi-Lenzi",
"suffix": ""
},
{
"first": "Rachele",
"middle": [],
"last": "Sprugnoli",
"suffix": ""
}
],
"year": 2006,
"venue": "Proceedings of LREC 2006",
"volume": "",
"issue": "",
"pages": "963--968",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Bernardo Magnini, Emanuele Pianta, Christian Girardi, Matteo Negri, Lorenza Romano, Manuela Sper- anza, Valentina Bartalesi-Lenzi, and Rachele Sprug- noli. 2006. I-CAB: the Italian Content Annotation Bank. In Proceedings of LREC 2006, pages 963- 968, Genoa, Italy.",
"links": null
},
"BIBREF19": {
"ref_id": "b19",
"title": "Tweetalige woordenboeken voor het Nederlands: het beleid van de Commissie Lexicografische Vertaalvoorzieningen",
"authors": [
{
"first": "Willy",
"middle": [],
"last": "Martin",
"suffix": ""
},
{
"first": "Jeannette",
"middle": [],
"last": "Ploeger",
"suffix": ""
}
],
"year": 1999,
"venue": "Neerlandica Extra Muros",
"volume": "37",
"issue": "",
"pages": "22--32",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Willy Martin and Jeannette Ploeger. 1999. Tweetalige woordenboeken voor het Nederlands: het beleid van de Commissie Lexicografische Vertaalvoorzienin- gen. Neerlandica Extra Muros, 37:22-32.",
"links": null
},
"BIBREF20": {
"ref_id": "b20",
"title": "Supervised Noun Phrase Coreference Research: The First Fifteen Years",
"authors": [
{
"first": "Vincent",
"middle": [],
"last": "Ng",
"suffix": ""
}
],
"year": 2010,
"venue": "Proceedings of the 48th Annual Meeting of the Association for Computational Linguistics",
"volume": "",
"issue": "",
"pages": "1396--1411",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Vincent Ng. 2010. Supervised Noun Phrase Corefer- ence Research: The First Fifteen Years. In Proceed- ings of the 48th Annual Meeting of the Association for Computational Linguistics, pages 1396-1411.",
"links": null
},
"BIBREF21": {
"ref_id": "b21",
"title": "Challenges in pronoun resolution system for biomedical text",
"authors": [
{
"first": "Jin-Dong Kim",
"middle": [],
"last": "",
"suffix": ""
},
{
"first": "Ngan",
"middle": [],
"last": "Nguyen",
"suffix": ""
},
{
"first": "Junichi",
"middle": [],
"last": "Tsujii",
"suffix": ""
}
],
"year": 2008,
"venue": "Proceedings of the Sixth International Language Resources and Evaluation (LREC'08)",
"volume": "",
"issue": "",
"pages": "",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Jin-Dong Kim Ngan Nguyen and Junichi Tsujii. 2008. Challenges in pronoun resolution system for biomedical text. In Proceedings of the Sixth International Language Resources and Evaluation (LREC'08), Marrakech, Morocco, may. European Language Resources Association (ELRA).",
"links": null
},
"BIBREF22": {
"ref_id": "b22",
"title": "Anaphoric annotation in the ARRAU corpus",
"authors": [
{
"first": "Massimo",
"middle": [],
"last": "Poesio",
"suffix": ""
},
{
"first": "Ron",
"middle": [],
"last": "Artstein",
"suffix": ""
}
],
"year": 2008,
"venue": "Proceedings of LREC 2008",
"volume": "",
"issue": "",
"pages": "1170--1174",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Massimo Poesio and Ron Artstein. 2008. Anaphoric annotation in the ARRAU corpus. In Proceedings of LREC 2008, pages 1170-1174, Marrakech, Mo- rocco.",
"links": null
},
"BIBREF23": {
"ref_id": "b23",
"title": "Computational models of anaphora resolution: A survey. Linguistic Issues in Language Technology",
"authors": [
{
"first": "Massimo",
"middle": [],
"last": "Poesio",
"suffix": ""
},
{
"first": "Simone",
"middle": [
"Paolo"
],
"last": "Ponzetto",
"suffix": ""
},
{
"first": "Yannick",
"middle": [],
"last": "Versley",
"suffix": ""
},
{
"first": "",
"middle": [],
"last": "Forthcoming",
"suffix": ""
}
],
"year": null,
"venue": "",
"volume": "",
"issue": "",
"pages": "",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Massimo Poesio, Simone Paolo Ponzetto, and Yannick Versley. forthcoming. Computational models of anaphora resolution: A survey. Linguistic Issues in Language Technology.",
"links": null
},
"BIBREF24": {
"ref_id": "b24",
"title": "Objective criteria for the evaluation of clustering methods",
"authors": [
{
"first": "William",
"middle": [
"M"
],
"last": "Rand",
"suffix": ""
}
],
"year": 1971,
"venue": "Journal of the American Statistical Association",
"volume": "66",
"issue": "336",
"pages": "846--850",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "William M. Rand. 1971. Objective criteria for the eval- uation of clustering methods. Journal of the Ameri- can Statistical Association, 66(336):846-850.",
"links": null
},
"BIBREF25": {
"ref_id": "b25",
"title": "Blanc: Implementing the rand index for coreference evaluation",
"authors": [
{
"first": "Marta",
"middle": [],
"last": "Recasens",
"suffix": ""
},
{
"first": "Eduard",
"middle": [],
"last": "Hovy",
"suffix": ""
}
],
"year": 2011,
"venue": "Natural Language Engineering",
"volume": "",
"issue": "",
"pages": "",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Marta Recasens and Eduard Hovy. 2011. Blanc: Im- plementing the rand index for coreference evalua- tion. Natural Language Engineering,.",
"links": null
},
"BIBREF26": {
"ref_id": "b26",
"title": "AnCora-CO: Coreferentially annotated corpora for Spanish and Catalan",
"authors": [
{
"first": "Marta",
"middle": [],
"last": "Recasens",
"suffix": ""
},
{
"first": "M",
"middle": [],
"last": "Ant\u00f2nia Mart\u00ed",
"suffix": ""
}
],
"year": 2010,
"venue": "Language Resources and Evaluation",
"volume": "44",
"issue": "4",
"pages": "315--345",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Marta Recasens and M. Ant\u00f2nia Mart\u00ed. 2010. AnCora- CO: Coreferentially annotated corpora for Spanish and Catalan. Language Resources and Evaluation, 44(4):315-345.",
"links": null
},
"BIBREF27": {
"ref_id": "b27",
"title": "SemEval-2010 Task 1: Coreference resolution in multiple languages",
"authors": [
{
"first": "Marta",
"middle": [],
"last": "Recasens",
"suffix": ""
},
{
"first": "Llu\u00edz",
"middle": [],
"last": "M\u00e1rquez",
"suffix": ""
},
{
"first": "Emili",
"middle": [],
"last": "Sapena",
"suffix": ""
},
{
"first": "M",
"middle": [
"Ant\u00f2nia"
],
"last": "Mart\u00ed",
"suffix": ""
},
{
"first": "Mariona",
"middle": [],
"last": "Taule\u00e9",
"suffix": ""
},
{
"first": "V\u00e9ronique",
"middle": [],
"last": "Hoste",
"suffix": ""
},
{
"first": "Massimo",
"middle": [],
"last": "Poesio",
"suffix": ""
},
{
"first": "Yannick",
"middle": [],
"last": "Versley",
"suffix": ""
}
],
"year": 2010,
"venue": "Proceedings of the 5th International Workshop on Semantic Evaluations (SemEval-2010)",
"volume": "",
"issue": "",
"pages": "1--8",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Marta Recasens, Llu\u00edz M\u00e1rquez, Emili Sapena, M. Ant\u00f2nia Mart\u00ed, Mariona Taule\u00e9, V\u00e9ronique Hoste, Massimo Poesio, and Yannick Versley. 2010. SemEval-2010 Task 1: Coreference resolution in multiple languages. In Proceedings of the 5th International Workshop on Semantic Evaluations (SemEval-2010), pages 1-8, Uppsala, Sweden.",
"links": null
},
"BIBREF28": {
"ref_id": "b28",
"title": "Coreference: Theory, Annotation, Resolution and Evaluation",
"authors": [
{
"first": "Marta",
"middle": [],
"last": "Recasens",
"suffix": ""
}
],
"year": 2010,
"venue": "",
"volume": "",
"issue": "",
"pages": "",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Marta Recasens. 2010. Coreference: Theory, An- notation, Resolution and Evaluation. Ph.D. thesis, Department of Linguistics, University of Barcelona, Barcelona, Spain, September.",
"links": null
},
"BIBREF29": {
"ref_id": "b29",
"title": "Anaphoric annotation of Wikipedia and blogs in the Live Memories Corpus",
"authors": [
{
"first": "Franceska",
"middle": [],
"last": "Kepa Joseba Rodr\u00edguez",
"suffix": ""
},
{
"first": "Yannick",
"middle": [],
"last": "Delogu",
"suffix": ""
},
{
"first": "Egon",
"middle": [],
"last": "Versley",
"suffix": ""
},
{
"first": "Massimo",
"middle": [],
"last": "Stemle",
"suffix": ""
},
{
"first": "",
"middle": [],
"last": "Poesio",
"suffix": ""
}
],
"year": 2010,
"venue": "Proceedings of LREC 2010",
"volume": "",
"issue": "",
"pages": "157--163",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Kepa Joseba Rodr\u00edguez, Franceska Delogu, Yannick Versley, Egon Stemle, and Massimo Poesio. 2010. Anaphoric annotation of Wikipedia and blogs in the Live Memories Corpus. In Proceedings of LREC 2010, pages 157-163, Valletta, Malta.",
"links": null
},
"BIBREF30": {
"ref_id": "b30",
"title": "Interacting Semantic Layers of Annotation in SoNaR, a Reference Corpus of Contemporary Written Dutch",
"authors": [
{
"first": "Ineke",
"middle": [],
"last": "Schuurman",
"suffix": ""
},
{
"first": "V\u00e9ronique",
"middle": [],
"last": "Hoste",
"suffix": ""
},
{
"first": "Paola",
"middle": [],
"last": "Monachesi",
"suffix": ""
}
],
"year": 2010,
"venue": "Proceedings of LREC 2010",
"volume": "",
"issue": "",
"pages": "2471--2477",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Ineke Schuurman, V\u00e9ronique Hoste, and Paola Monachesi. 2010. Interacting Semantic Layers of Annotation in SoNaR, a Reference Corpus of Con- temporary Written Dutch. In Proceedings of LREC 2010, pages 2471-2477, Valletta, Malta.",
"links": null
},
"BIBREF31": {
"ref_id": "b31",
"title": "A Machine Learning Approach to Coreference Resolution of Noun Phrases",
"authors": [],
"year": 2001,
"venue": "Computational Linguistics",
"volume": "27",
"issue": "4",
"pages": "521--544",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Wee Meng Soon, Hwee Tou Ng, and Daniel Chung Yong Lim. 2001. A Machine Learning Ap- proach to Coreference Resolution of Noun Phrases. Computational Linguistics, 27(4):521-544.",
"links": null
},
"BIBREF32": {
"ref_id": "b32",
"title": "Introduction to the CoNLL-2002 Shared Task: Language-Independent Named Entity Recognition",
"authors": [
{
"first": "Erik",
"middle": [
"F"
],
"last": "",
"suffix": ""
},
{
"first": "Tjong Kim",
"middle": [],
"last": "Sang",
"suffix": ""
}
],
"year": 2002,
"venue": "Proceedings of the 6th Conference on Natural Language Learning",
"volume": "",
"issue": "",
"pages": "155--158",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Erik F. Tjong Kim Sang. 2002. Introduction to the CoNLL-2002 Shared Task: Language-Independent Named Entity Recognition. In Proceedings of the 6th Conference on Natural Language Learning, pages 155-158, Taipei, Taiwan.",
"links": null
},
"BIBREF33": {
"ref_id": "b33",
"title": "A Model-Theoretic Coreference Scoring Scheme",
"authors": [
{
"first": "Marc",
"middle": [],
"last": "Vilain",
"suffix": ""
},
{
"first": "John",
"middle": [],
"last": "Burger",
"suffix": ""
},
{
"first": "John",
"middle": [],
"last": "Aberdeen",
"suffix": ""
}
],
"year": 1995,
"venue": "Proceedings of the Sixth Message Understanding Conference (MUC-6)",
"volume": "",
"issue": "",
"pages": "45--52",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Marc Vilain, John Burger, John Aberdeen, Dennis Con- nolly, and Lynette Hirschman. 1995. A Model- Theoretic Coreference Scoring Scheme. In Pro- ceedings of the Sixth Message Understanding Con- ference (MUC-6), pages 45-52.",
"links": null
},
"BIBREF34": {
"ref_id": "b34",
"title": "EuroWordNet: a multilingual database with lexical semantic networks",
"authors": [],
"year": 1998,
"venue": "",
"volume": "",
"issue": "",
"pages": "",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Piek Vossen, editor. 1998. EuroWordNet: a mul- tilingual database with lexical semantic networks. Kluwer Academic Publishers, Norwell, MA, USA.",
"links": null
},
"BIBREF36": {
"ref_id": "b36",
"title": "An np-cluster based approach to coreference resolution",
"authors": [
{
"first": "Xiaofeng",
"middle": [],
"last": "Yang",
"suffix": ""
},
{
"first": "Jian",
"middle": [],
"last": "Su",
"suffix": ""
},
{
"first": "Guodong",
"middle": [],
"last": "Zhou",
"suffix": ""
},
{
"first": "Chew Lim",
"middle": [],
"last": "Tan",
"suffix": ""
}
],
"year": 2004,
"venue": "Proceedings of Coling",
"volume": "",
"issue": "",
"pages": "226--232",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Xiaofeng Yang, Jian Su, GuoDong Zhou, and Chew Lim Tan. 2004. An np-cluster based ap- proach to coreference resolution. In Proceedings of Coling 2004, pages 226-232, Geneva, Switzerland, Aug 23-Aug 27.",
"links": null
}
},
"ref_entries": {
"TABREF1": {
"text": "",
"type_str": "table",
"content": "<table/>",
"num": null,
"html": null
},
"TABREF3": {
"text": "Three sets of experiments our scoring metrics, a singleton that is erroneously classified as part of a coreference chain is counted as an error. When it is correctly classified as a singleton, however, this is not represented in the scores.",
"type_str": "table",
"content": "<table/>",
"num": null,
"html": null
},
"TABREF5": {
"text": "",
"type_str": "table",
"content": "<table><tr><td>: Comparison of the worst (top) to</td></tr><tr><td>best-performing (bottom) cross-domain genres per</td></tr><tr><td>metric.</td></tr></table>",
"num": null,
"html": null
},
"TABREF6": {
"text": "TrainTestADM AUTO DUO EXT INST JOUR MED WIKI MUC ALL 37.10 34.61 43.61 42.09 44.81 43.63 35.57 54.48 1MinMED 37.26 34.41 43.56 42.01 44.61 44.03",
"type_str": "table",
"content": "<table><tr><td>P P</td><td>P</td><td>P P</td><td>P</td><td>P P P</td></tr><tr><td/><td/><td/><td/><td/><td>54.07</td></tr><tr><td colspan=\"4\">2MinDUO</td><td>37.39 34.85</td><td>42.29 44.51 44.56 35.44 54.35</td></tr><tr><td colspan=\"4\">3MinINST</td><td colspan=\"2\">37.06 34.00 31.02 41.81</td><td>44.46 34.72 54.21</td></tr><tr><td colspan=\"3\">B-cubed</td><td/><td/></tr><tr><td colspan=\"2\">ALL</td><td/><td/><td colspan=\"2\">27.83 29.77 31.45 30.64 31.66 31.23 26.08 30.84</td></tr><tr><td colspan=\"4\">1MinMED</td><td colspan=\"2\">27.74 29.64 31.68 30.18 31.66 31.34</td><td>30.46</td></tr><tr><td colspan=\"4\">2MinDUO</td><td>28.02 29.46</td><td>30.11 31.26 31.81 25.99 30.58</td></tr><tr><td colspan=\"4\">3MinINST</td><td colspan=\"2\">27.87 29.54 31.02 30.01</td><td>31.61 25.18 30.64</td></tr><tr><td colspan=\"3\">CEAF</td><td/><td/></tr><tr><td colspan=\"2\">ALL</td><td/><td/><td colspan=\"2\">29.48 30.61 29.79 31.36 28.42 31.42 29.49 26.31</td></tr><tr><td colspan=\"4\">1MinMED</td><td colspan=\"2\">29.11 30.33 29.96 30.26 28.47 30.86</td><td>26.40</td></tr><tr><td colspan=\"4\">2MinDUO</td><td>29.73 29.51</td><td>30.09 28.12 31.62 29.33 25.99</td></tr><tr><td colspan=\"4\">3MinINST</td><td colspan=\"2\">29.58 30.48 22.97 29.16</td><td>30.93 28.20 25.14</td></tr><tr><td colspan=\"3\">BLANC</td><td/><td/></tr><tr><td colspan=\"2\">ALL</td><td/><td/><td colspan=\"2\">48.10 51.11 52.87 48.29 50.21 49.74 49.01 55.73</td></tr><tr><td colspan=\"4\">1MinMED</td><td colspan=\"2\">48.49 51.37 54.70 48.51 50.72 49.55</td><td>56.66</td></tr><tr><td colspan=\"4\">2MinDUO</td><td>48.73 51.49</td><td>48.73 51.01 50.37 48.15 56.11</td></tr><tr><td colspan=\"4\">3MinINST</td><td colspan=\"2\">49.71 51.59 54.16 50.88</td><td>49.61 48.49 56.17</td></tr></table>",
"num": null,
"html": null
},
"TABREF7": {
"text": "Results of the third set of experiments for all metrics and in comparison with training on all data.",
"type_str": "table",
"content": "<table/>",
"num": null,
"html": null
}
}
}
} |