File size: 98,417 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 | {
"paper_id": "P09-1041",
"header": {
"generated_with": "S2ORC 1.0.0",
"date_generated": "2023-01-19T08:54:32.103379Z"
},
"title": "Semi-supervised Learning of Dependency Parsers using Generalized Expectation Criteria",
"authors": [
{
"first": "Gregory",
"middle": [],
"last": "Druck",
"suffix": "",
"affiliation": {
"laboratory": "",
"institution": "University of Massachusetts Amherst",
"location": {
"postCode": "01003",
"region": "MA"
}
},
"email": "gdruck@cs.umass.edu"
},
{
"first": "Gideon",
"middle": [],
"last": "Mann",
"suffix": "",
"affiliation": {},
"email": "gideon.mann@gmail.com"
},
{
"first": "Andrew",
"middle": [],
"last": "Mccallum",
"suffix": "",
"affiliation": {
"laboratory": "",
"institution": "University of Massachusetts Amherst",
"location": {
"postCode": "01003",
"region": "MA"
}
},
"email": "mccallum@cs.umass.edu"
}
],
"year": "",
"venue": null,
"identifiers": {},
"abstract": "In this paper, we propose a novel method for semi-supervised learning of nonprojective log-linear dependency parsers using directly expressed linguistic prior knowledge (e.g. a noun's parent is often a verb). Model parameters are estimated using a generalized expectation (GE) objective function that penalizes the mismatch between model predictions and linguistic expectation constraints. In a comparison with two prominent \"unsupervised\" learning methods that require indirect biasing toward the correct syntactic structure, we show that GE can attain better accuracy with as few as 20 intuitive constraints. We also present positive experimental results on longer sentences in multiple languages.",
"pdf_parse": {
"paper_id": "P09-1041",
"_pdf_hash": "",
"abstract": [
{
"text": "In this paper, we propose a novel method for semi-supervised learning of nonprojective log-linear dependency parsers using directly expressed linguistic prior knowledge (e.g. a noun's parent is often a verb). Model parameters are estimated using a generalized expectation (GE) objective function that penalizes the mismatch between model predictions and linguistic expectation constraints. In a comparison with two prominent \"unsupervised\" learning methods that require indirect biasing toward the correct syntactic structure, we show that GE can attain better accuracy with as few as 20 intuitive constraints. We also present positive experimental results on longer sentences in multiple languages.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Abstract",
"sec_num": null
}
],
"body_text": [
{
"text": "Early approaches to parsing assumed a grammar provided by human experts (Quirk et al., 1985) . Later approaches avoided grammar writing by learning the grammar from sentences explicitly annotated with their syntactic structure (Black et al., 1992) . While such supervised approaches have yielded accurate parsers (Charniak, 2001) , the syntactic annotation of corpora such as the Penn Treebank is extremely costly, and consequently there are few treebanks of comparable size.",
"cite_spans": [
{
"start": 72,
"end": 92,
"text": "(Quirk et al., 1985)",
"ref_id": "BIBREF17"
},
{
"start": 227,
"end": 247,
"text": "(Black et al., 1992)",
"ref_id": "BIBREF0"
},
{
"start": 313,
"end": 329,
"text": "(Charniak, 2001)",
"ref_id": "BIBREF2"
}
],
"ref_spans": [],
"eq_spans": [],
"section": "Introduction",
"sec_num": "1"
},
{
"text": "As a result, there has been recent interest in unsupervised parsing. However, in order to attain reasonable accuracy, these methods have to be carefully biased towards the desired syntactic structure. This weak supervision has been encoded using priors and initializations (Klein and Manning, 2004; , specialized models (Klein and Manning, 2004; Seginer, 2007; Bod, 2006) , and implicit negative evidence . These indirect methods for leveraging prior knowledge can be cumbersome and unintuitive for a non-machine-learning expert.",
"cite_spans": [
{
"start": 273,
"end": 298,
"text": "(Klein and Manning, 2004;",
"ref_id": "BIBREF9"
},
{
"start": 320,
"end": 345,
"text": "(Klein and Manning, 2004;",
"ref_id": "BIBREF9"
},
{
"start": 346,
"end": 360,
"text": "Seginer, 2007;",
"ref_id": "BIBREF18"
},
{
"start": 361,
"end": 371,
"text": "Bod, 2006)",
"ref_id": "BIBREF1"
}
],
"ref_spans": [],
"eq_spans": [],
"section": "Introduction",
"sec_num": "1"
},
{
"text": "This paper proposes a method for directly guiding the learning of dependency parsers with naturally encoded linguistic insights. Generalized expectation (GE) Druck et al., 2008 ) is a recently proposed framework for incorporating prior knowledge into the learning of conditional random fields (CRFs) (Lafferty et al., 2001) . GE criteria express a preference on the value of a model expectation. For example, we know that \"in English, when a determiner is directly to the left of a noun, the noun is usually the parent of the determiner\". With GE we may add a term to the objective function that encourages a feature-rich CRF to match this expectation on unlabeled data, and in the process learn about related features. In this paper we use a non-projective dependency tree CRF .",
"cite_spans": [
{
"start": 158,
"end": 176,
"text": "Druck et al., 2008",
"ref_id": "BIBREF4"
},
{
"start": 300,
"end": 323,
"text": "(Lafferty et al., 2001)",
"ref_id": "BIBREF11"
}
],
"ref_spans": [],
"eq_spans": [],
"section": "Introduction",
"sec_num": "1"
},
{
"text": "While a complete exploration of linguistic prior knowledge for dependency parsing is beyond the scope of this paper, we provide several promising demonstrations of the proposed method. On the English WSJ10 data set, GE training outperforms two prominent unsupervised methods using only 20 constraints either elicited from a human or provided by an \"oracle\" simulating a human. We also present experiments on longer sentences in Dutch, Spanish, and Turkish in which we obtain accuracy comparable to supervised learning with tens to hundreds of complete parsed sentences.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Introduction",
"sec_num": "1"
},
{
"text": "This work is closely related to the prototypedriven grammar induction method of Haghighi and Klein (2006) , which uses prototype phrases to guide the EM algorithm in learning a PCFG. Direct comparison with this method is not possible because we are interested in dependency syntax rather than phrase structure syntax. However, the approach we advocate has several significant advantages. GE is more general than prototypedriven learning because GE constraints can be uncertain. Additionally prototype-driven grammar induction needs to be used in conjunction with other unsupervised methods (distributional similarity and CCM (Klein and Manning, 2004) ) to attain reasonable accuracy, and is only evaluated on length 10 or less sentences with no lexical information. In contrast, GE uses only the provided constraints and unparsed sentences, and is used to train a feature-rich discriminative model.",
"cite_spans": [
{
"start": 80,
"end": 105,
"text": "Haghighi and Klein (2006)",
"ref_id": "BIBREF7"
},
{
"start": 625,
"end": 650,
"text": "(Klein and Manning, 2004)",
"ref_id": "BIBREF9"
}
],
"ref_spans": [],
"eq_spans": [],
"section": "Related Work",
"sec_num": "2"
},
{
"text": "Conventional semi-supervised learning requires parsed sentences. Kate and Mooney (2007) and McClosky et al. (2006) both use modified forms of self-training to bootstrap parsers from limited labeled data. Wang et al. (2008) combine a structured loss on parsed sentences with a least squares loss on unlabeled sentences. Koo et al. (2008) use a large unlabeled corpus to estimate cluster features which help the parser generalize with fewer examples. Smith and Eisner (2007) apply entropy regularization to dependency parsing. The above methods can be applied to small seed corpora, but McDonald 1 has criticized such methods as working from an unrealistic premise, as a significant amount of the effort required to build a treebank comes in the first 100 sentences (both because of the time it takes to create an appropriate rubric and to train annotators).",
"cite_spans": [
{
"start": 65,
"end": 87,
"text": "Kate and Mooney (2007)",
"ref_id": "BIBREF8"
},
{
"start": 92,
"end": 114,
"text": "McClosky et al. (2006)",
"ref_id": "BIBREF14"
},
{
"start": 204,
"end": 222,
"text": "Wang et al. (2008)",
"ref_id": "BIBREF25"
},
{
"start": 319,
"end": 336,
"text": "Koo et al. (2008)",
"ref_id": "BIBREF10"
},
{
"start": 449,
"end": 472,
"text": "Smith and Eisner (2007)",
"ref_id": "BIBREF21"
}
],
"ref_spans": [],
"eq_spans": [],
"section": "Related Work",
"sec_num": "2"
},
{
"text": "There are also a number of methods for unsupervised learning of dependency parsers. Klein and Manning (2004) use a carefully initialized and structured generative model (DMV) in conjunction with the EM algorithm to get the first positive results on unsupervised dependency parsing. As empirical evidence of the sensitivity of DMV to initialization, (pg. 37) uses three different initializations, and only one, the method of Klein and Manning (2004) , gives accuracy higher than 31% on the WSJ10 corpus (see Section 5). This initialization encodes the prior knowledge that long distance attachments are unlikely. develop contrastive estimation (CE), in which the model is encouraged to move probability mass away from implicit negative examples defined using a carefully chosen neighborhood function. For instance, function DEL1ORTRANS1 provides accuracy of 57.6% on WSJ10 (see Section 5). Another neighborhood, DEL1ORTRANS2, provides accuracy of 51.2%. The remaining six neighborhood functions provide accuracy below 50%. This demonstrates that constructing an appropriate neighborhood function can be delicate and challenging. Smith and Eisner (2006) propose structural annealing (SA), in which a strong bias for local dependency attachments is enforced early in learning, and then gradually relaxed. This method is sensitive to the annealing schedule. (pg. 136) use 10 annealing schedules in conjunction with three initializers. The best performing combination attains accuracy of 66.7% on WSJ10, but the worst attains accuracy of 32.5%.",
"cite_spans": [
{
"start": 84,
"end": 108,
"text": "Klein and Manning (2004)",
"ref_id": "BIBREF9"
},
{
"start": 424,
"end": 448,
"text": "Klein and Manning (2004)",
"ref_id": "BIBREF9"
},
{
"start": 1128,
"end": 1151,
"text": "Smith and Eisner (2006)",
"ref_id": "BIBREF20"
}
],
"ref_spans": [],
"eq_spans": [],
"section": "Related Work",
"sec_num": "2"
},
{
"text": "Finally, Seginer (2007) and Bod (2006) approach unsupervised parsing by constructing novel syntactic models. The development and tuning of the above methods constitute the encoding of prior domain knowledge about the desired syntactic structure. In contrast, our framework provides a straightforward and explicit method for incorporating prior knowledge. Ganchev et al. (2009) propose a related method that uses posterior constrained EM to learn a projective target language parser using only a source language parser and word alignments.",
"cite_spans": [
{
"start": 9,
"end": 23,
"text": "Seginer (2007)",
"ref_id": "BIBREF18"
},
{
"start": 28,
"end": 38,
"text": "Bod (2006)",
"ref_id": "BIBREF1"
},
{
"start": 355,
"end": 376,
"text": "Ganchev et al. (2009)",
"ref_id": "BIBREF6"
}
],
"ref_spans": [],
"eq_spans": [],
"section": "Related Work",
"sec_num": "2"
},
{
"text": "Generalized expectation criteria (Mann and Mc-Callum, 2008; Druck et al., 2008) are terms in a parameter estimation objective function that express a preference on the value of a model expectation. Let x represent input variables (i.e. a sentence) and y represent output variables (i.e. a parse tree). A generalized expectation term G(\u03bb) is defined by a constraint function G(y, x) that returns a non-negative real value given input and output variables, an empirical distributionp(x) over input variables (i.e. unlabeled data), a model distribution p \u03bb (y|x), and a score function S:",
"cite_spans": [
{
"start": 33,
"end": 59,
"text": "(Mann and Mc-Callum, 2008;",
"ref_id": null
},
{
"start": 60,
"end": 79,
"text": "Druck et al., 2008)",
"ref_id": "BIBREF4"
}
],
"ref_spans": [],
"eq_spans": [],
"section": "Generalized Expectation Criteria",
"sec_num": "3"
},
{
"text": "G(\u03bb) = S(Ep (x) [E p \u03bb (y|x) [G(y, x)]]).",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Generalized Expectation Criteria",
"sec_num": "3"
},
{
"text": "In this paper, we use a score function that is the squared difference of the model expectation of G and some target expectationG:",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Generalized Expectation Criteria",
"sec_num": "3"
},
{
"text": "S sq = \u2212(G \u2212 Ep (x) [E p \u03bb (y|x) [G(y, x)]]) 2 (1)",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Generalized Expectation Criteria",
"sec_num": "3"
},
{
"text": "We can incorporate prior knowledge into the training of p \u03bb (y|x) by specifying the from of the constraint function G and the target expectationG. Importantly, G does not need to match a particular feature in the underlying model. The complete objective function 2 includes multiple GE terms and a prior on parameters 3 ",
"cite_spans": [
{
"start": 318,
"end": 319,
"text": "3",
"ref_id": null
}
],
"ref_spans": [],
"eq_spans": [],
"section": "Generalized Expectation Criteria",
"sec_num": "3"
},
{
"text": ", p(\u03bb) O(\u03bb; D) = p(\u03bb) + G G(\u03bb)",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Generalized Expectation Criteria",
"sec_num": "3"
},
{
"text": "GE has been applied to logistic regression models (Mann and McCallum, 2007; Druck et al., 2008) and linear chain CRFs . In the following sections we apply GE to non-projective CRF dependency parsing.",
"cite_spans": [
{
"start": 50,
"end": 75,
"text": "(Mann and McCallum, 2007;",
"ref_id": "BIBREF12"
},
{
"start": 76,
"end": 95,
"text": "Druck et al., 2008)",
"ref_id": "BIBREF4"
}
],
"ref_spans": [],
"eq_spans": [],
"section": "Generalized Expectation Criteria",
"sec_num": "3"
},
{
"text": "We first consider an arbitrarily structured conditional random field (Lafferty et al., 2001 ) p \u03bb (y|x). We describe the CRF for non-projective dependency parsing in Section 3.2. The probability of an output y conditioned on an input x is",
"cite_spans": [
{
"start": 69,
"end": 91,
"text": "(Lafferty et al., 2001",
"ref_id": "BIBREF11"
}
],
"ref_spans": [],
"eq_spans": [],
"section": "GE in General CRFs",
"sec_num": "3.1"
},
{
"text": "p \u03bb (y|x) = 1 Z x exp j \u03bb j F j (y, x) ,",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "GE in General CRFs",
"sec_num": "3.1"
},
{
"text": "where F j are feature functions over the cliques of the graphical model and Z(x) is a normalizing constant that ensures p \u03bb (y|x) sums to 1. We are interested in the expectation of constraint function G(x, y) under this model. We abbreviate this model expectation as:",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "GE in General CRFs",
"sec_num": "3.1"
},
{
"text": "G \u03bb = Ep (x) [E p \u03bb (y|x) [G(y, x)]]",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "GE in General CRFs",
"sec_num": "3.1"
},
{
"text": "It can be shown that partial derivative of G(\u03bb) using S sq 4 with respect to model parameter \u03bb j is",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "GE in General CRFs",
"sec_num": "3.1"
},
{
"text": "\u2202 \u2202\u03bb j G(\u03bb) = 2(G \u2212 G \u03bb ) (2) Ep (x) E p \u03bb (y|x) [G(y, x)F j (y, x)] \u2212E p \u03bb (y|x) [G(y, x)] E p \u03bb (y|x) [F j (y, x)] .",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "GE in General CRFs",
"sec_num": "3.1"
},
{
"text": "Equation 2 has an intuitive interpretation. The first term (on the first line) is the difference between the model and target expectations. The second term (the rest of the equation) is the predicted covariance between the constraint function G and the model feature function F j . Therefore, if the constraint is not satisfied, GE updates parameters for features that the model predicts are related to the constraint function. If there are constraint functions G for all model feature functions F j , and the target expectations G are estimated from labeled data, then the globally optimal parameter setting under the GE objective function is equivalent to the maximum likelihood solution. However, GE does not require such a one-to-one correspondence between constraint functions and model feature functions. This allows bootstrapping of feature-rich models with a small number of prior expectation constraints.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "GE in General CRFs",
"sec_num": "3.1"
},
{
"text": "We now define a CRF p \u03bb (y|x) for unlabeled, nonprojective 5 dependency parsing. The tree y is represented as a vector of the same length as the sentence, where y i is the index of the parent of word i. The probability of a tree y given sentence x is",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Non-Projective Dependency Tree CRFs",
"sec_num": "3.2"
},
{
"text": "p \u03bb (y|x) = 1 Z x exp n i=1 j \u03bb j f j (x i , x y i , x) ,",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Non-Projective Dependency Tree CRFs",
"sec_num": "3.2"
},
{
"text": "where f j are edge-factored feature functions that consider the child input (word, tag, or other feature), the parent input, and the rest of the sentence. This factorization implies that dependency decisions are independent conditioned on the input sentence x if y is a tree. Computing Z x and the edge expectations needed for partial derivatives requires summing over all possible trees for x. By relating the sum of the scores of all possible trees to counting the number of spanning trees in a graph, it can be shown that Z x is the determinant of the Kirchoff matrix K, which is constructed using the scores of possible edges. (McDonald and Satta, 2007; . Computing the determinant takes O(n 3 ) time, where n is the length of the sentence. To compute the marginal probability of a particular edge k \u2192 i (i.e. y i = k), the score of any edge k \u2192 i such that k = k is set to 0. The determinant of the resulting modified Kirchoff matrix K k\u2192i is then the sum of the scores of all trees that include the edge k \u2192 i. The marginal p(y i = k|x; \u03b8) can be computed by dividing this score by Z x (McDonald and Satta, 2007) . Computing all edge expectations with this algorithm takes O(n 5 ) time. describe a more efficient algorithm that can compute all edge expectations in O(n 3 ) time using the inverse of the Kirchoff matrix K \u22121 .",
"cite_spans": [
{
"start": 631,
"end": 657,
"text": "(McDonald and Satta, 2007;",
"ref_id": "BIBREF15"
},
{
"start": 1092,
"end": 1118,
"text": "(McDonald and Satta, 2007)",
"ref_id": "BIBREF15"
}
],
"ref_spans": [],
"eq_spans": [],
"section": "Non-Projective Dependency Tree CRFs",
"sec_num": "3.2"
},
{
"text": "While in general constraint functions G may consider multiple edges, in this paper we use edge-factored constraint functions. In this case",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "GE for Non-Projective Dependency Tree CRFs",
"sec_num": "3.3"
},
{
"text": "E p \u03bb (y|x) [G(y, x)]E p \u03bb (y|x) [F j (y, x)]",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "GE for Non-Projective Dependency Tree CRFs",
"sec_num": "3.3"
},
{
"text": ", the second term of the covariance in Equation 2, can be computed using the edge marginal distributions p \u03bb (y i |x).",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "GE for Non-Projective Dependency Tree CRFs",
"sec_num": "3.3"
},
{
"text": "The first term of the covariance",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "GE for Non-Projective Dependency Tree CRFs",
"sec_num": "3.3"
},
{
"text": "E p \u03bb (y|x) [G(y, x)F j (y, x)]",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "GE for Non-Projective Dependency Tree CRFs",
"sec_num": "3.3"
},
{
"text": "is more difficult to compute because it requires the marginal probability of two edges p \u03bb (y i , y i |x). It is important to note that the model p \u03bb is still edge-factored.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "GE for Non-Projective Dependency Tree CRFs",
"sec_num": "3.3"
},
{
"text": "The sum of the scores of all trees that contain edges k \u2192 i and k \u2192 i can be computed by setting the scores of edges j \u2192 i such that j = k and j \u2192 i such that j = k to 0, and computing the determinant of the resulting modified Kirchoff matrix K k\u2192i,k \u2192i . There are O(n 4 ) pairs of possible edges, and the determinant computation takes time O(n 3 ), so this naive algorithm takes O(n 7 ) time.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "GE for Non-Projective Dependency Tree CRFs",
"sec_num": "3.3"
},
{
"text": "An improved algorithm computes, for each possible edge k \u2192 i, a modified Kirchoff matrix K k\u2192i that requires the presence of that edge. Then, the method of can be used to compute the probability of every possible edge conditioned on the presence of",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "GE for Non-Projective Dependency Tree CRFs",
"sec_num": "3.3"
},
{
"text": "k \u2192 i, p \u03bb (y i = k |y i = k, x), using K \u22121",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "GE for Non-Projective Dependency Tree CRFs",
"sec_num": "3.3"
},
{
"text": "k\u2192i . Multiplying this probability by p \u03bb (y i =k|x) yields the desired two edge marginal. Because this algorithm pulls the O(n 3 ) matrix operation out of the inner loop over edges, the run time is reduced to O(n 5 ).",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "GE for Non-Projective Dependency Tree CRFs",
"sec_num": "3.3"
},
{
"text": "If it were possible to perform only one O(n 3 ) matrix operation per sentence, then the gradient computation would take only O(n 4 ) time, the time required to consider all pairs of edges. Unfortunately, there is no straightforward generalization of the method of to the two edge marginal problem. Specifically, Laplace expansion generalizes to second-order matrix minors, but it is not clear how to compute secondorder cofactors from the inverse Kirchoff matrix alone (c.f. ).",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "GE for Non-Projective Dependency Tree CRFs",
"sec_num": "3.3"
},
{
"text": "Consequently, we also propose an approximation that can be used to speed up GE training at the expense of a less accurate covariance computation. We consider different cases of the edges k \u2192 i, and k \u2192 i .",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "GE for Non-Projective Dependency Tree CRFs",
"sec_num": "3.3"
},
{
"text": "\u2022 p \u03bb (y i =k, y i =k |x)=0 when i=i and k =k (different parent for the same word), or when i=k and k=i (cycle), because these pairs of edges break the tree constraint.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "GE for Non-Projective Dependency Tree CRFs",
"sec_num": "3.3"
},
{
"text": "EQUATION",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [
{
"start": 0,
"end": 8,
"text": "EQUATION",
"ref_id": "EQREF",
"raw_str": "\u2022 p \u03bb (y i =k, y i =k |x)=p \u03bb (y i =k|x) when i= i , k=k . \u2022 p \u03bb (y i = k, y i = k |x) \u2248 p \u03bb (y i = k|x)p \u03bb (y i =",
"eq_num": "k"
}
],
"section": "GE for Non-Projective Dependency Tree CRFs",
"sec_num": "3.3"
},
{
"text": "|x) when i = i and i = k or i = k (different words, do not create a cycle). This approximation assumes that pairs of edges that do not fall into one of the above cases are conditionally independent given x. This is not true because there are partial trees in which k \u2192 i and k \u2192 i can appear separately, but not together (for example if i = k and the partial tree contains i \u2192 k).",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "GE for Non-Projective Dependency Tree CRFs",
"sec_num": "3.3"
},
{
"text": "Using this approximation, the covariance for one sentence is approximately equal to",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "GE for Non-Projective Dependency Tree CRFs",
"sec_num": "3.3"
},
{
"text": "n i E p \u03bb (y i |x) [f j (x i , x y i , x)g(x i , x y i , x)] \u2212 n i E p \u03bb (y i |x) [f j (x i , x y i , x)]E p \u03bb (y i |x) [g(x i , x y i , x)] \u2212 n i,k p \u03bb (y i =k|x)p \u03bb (y k =i|x)f j (x i , x k , x)g(x k , x i , x).",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "GE for Non-Projective Dependency Tree CRFs",
"sec_num": "3.3"
},
{
"text": "Intuitively, the first and second terms compute a covariance over possible parents for a single word, and the third term accounts for cycles. Computing the above takes O(n 3 ) time, the time required to compute single edge marginals. In this paper, we use the O(n 5 ) exact method, though we find that the accuracy attained by approximate training is usually within 5% of the exact method. If G is not edge-factored, then we need to compute a marginal over three or more edges, making exact training intractable. An appealing alternative to a similar approximation to the above would use loopy belief propagation to efficiently approximate the marginals (Smith and Eisner, 2008) .",
"cite_spans": [
{
"start": 654,
"end": 678,
"text": "(Smith and Eisner, 2008)",
"ref_id": "BIBREF22"
}
],
"ref_spans": [],
"eq_spans": [],
"section": "GE for Non-Projective Dependency Tree CRFs",
"sec_num": "3.3"
},
{
"text": "In this paper g is binary and normalized by its total count in the corpus. The expectation of g is then the probability that it indicates a true edge.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "GE for Non-Projective Dependency Tree CRFs",
"sec_num": "3.3"
},
{
"text": "Training parsers using GE with the aid of linguists is an exciting direction for future work. In this paper, we use constraints derived from several basic types of linguistic knowledge.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Linguistic Prior Knowledge",
"sec_num": "4"
},
{
"text": "One simple form of linguistic knowledge is the set of possible parent tags for a given child tag. This type of constraint was used in the development of a rule-based dependency parser (Debusmann et al., 2004) . Additional information can be obtained from small grammar fragments. Haghighi and Klein (2006) provide a list of prototype phrase structure rules that can be augmented with dependencies and used to define constraints involving parent and child tags, surrounding or interposing tags, direction, and distance. Finally there are well known hypotheses about the direction and distance of attachments that can be used to define constraints. use the fact that short attachments are more common to improve unsupervised parsing accuracy.",
"cite_spans": [
{
"start": 184,
"end": 208,
"text": "(Debusmann et al., 2004)",
"ref_id": "BIBREF3"
},
{
"start": 280,
"end": 305,
"text": "Haghighi and Klein (2006)",
"ref_id": "BIBREF7"
}
],
"ref_spans": [],
"eq_spans": [],
"section": "Linguistic Prior Knowledge",
"sec_num": "4"
},
{
"text": "For some experiments that follow we use \"oracle\" constraints that are estimated from labeled data. This involves choosing feature templates (motivated by the linguistic knowledge described above) and estimating target expectations. Oracle methods used in this paper consider three simple statistics of candidate constraint functions: count c(g), edge countc edge (g), and edge probabilit\u1ef9 p(edge|g). Let D be the labeled corpus.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "\"Oracle\" constraints",
"sec_num": "4.1"
},
{
"text": "c(g) = x\u2208D i j g(x i , x j , x) c edge (g) = (x,y)\u2208D i g(x i , x y i , x) p(edge|g) =c edge (g) c(g)",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "\"Oracle\" constraints",
"sec_num": "4.1"
},
{
"text": "Constraint functions are selected according to some combination of the above statistics. In some cases we additionally prune the candidate set by considering only certain templates. To compute the target expectation, we simply use bin(p(edge|g)), where bin returns the closest value in the set {0, 0.1, 0.25, 0.5, 0.75, 1}. This can be viewed as specifying that g is very indicative of edge, somewhat indicative of edge, etc.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "\"Oracle\" constraints",
"sec_num": "4.1"
},
{
"text": "In this section we compare GE training with methods for unsupervised parsing. We use the WSJ10 corpus (as processed by ), which is comprised of English sentences of ten words or fewer (after stripping punctuation) from the WSJ portion of the Penn Treebank. As in previous work sentences contain only part-of-speech tags. We compare GE and supervised training of an edge-factored CRF with unsupervised learning of a DMV model (Klein and Manning, 2004) using EM and contrastive estimation (CE) . We also report the accuracy of an attach-right baseline 6 . Finally, we report the accuracy of a constraint baseline that assigns a score to each possible edge that is the sum of the target expectations for all constraints on that edge. Possible edges without constraints receive a score of 0. These scores are used as input to the maximum spanning tree algorithm, which returns the best tree. Note that this is a strong baseline because it can handle uncertain constraints, and the tree constraint imposed by the MST algorithm helps information propagate across edges.",
"cite_spans": [
{
"start": 425,
"end": 450,
"text": "(Klein and Manning, 2004)",
"ref_id": "BIBREF9"
}
],
"ref_spans": [],
"eq_spans": [],
"section": "Experimental Comparison with Unsupervised Learning",
"sec_num": "5"
},
{
"text": "We note that there are considerable differences between the DMV and CRF models. The DMV model is more expressive than the CRF because it can model the arity of a head as well as sibling relationships. Because these features consider multiple edges, including them in the CRF model would make exact inference intractable (McDonald and Satta, 2007) . However, the CRF may consider the distance between head and child, whereas DMV does not model distance. The CRF also models non-projective trees, which when evaluating on English is likely a disadvantage.",
"cite_spans": [
{
"start": 320,
"end": 346,
"text": "(McDonald and Satta, 2007)",
"ref_id": "BIBREF15"
}
],
"ref_spans": [],
"eq_spans": [],
"section": "Experimental Comparison with Unsupervised Learning",
"sec_num": "5"
},
{
"text": "Consequently, we experiment with two sets of features for the CRF model. The first, restricted set includes features that consider the head and child tags of the dependency conjoined with the direction of the attachment, (parent-POS,child-POS,direction). With this feature set, the CRF model is less expressive than DMV. The second full set includes standard features for edgefactored dependency parsers (McDonald et al., 2005) , though still unlexicalized. The CRF cannot consider valency even with the full feature set, but this is balanced by the ability to use distance. Table 1 : 20 constraints that give 61.3% accuracy on WSJ10. Tags are grouped according to heads, and are in the order they appear in the sentence, with the arrow pointing from head to modifier.",
"cite_spans": [
{
"start": 404,
"end": 427,
"text": "(McDonald et al., 2005)",
"ref_id": "BIBREF16"
}
],
"ref_spans": [
{
"start": 575,
"end": 582,
"text": "Table 1",
"ref_id": null
}
],
"eq_spans": [],
"section": "Experimental Comparison with Unsupervised Learning",
"sec_num": "5"
},
{
"text": "We generate constraints in two ways. First, we use oracle constraints of the form (parent-POS,child-POS,direction) such thatc(g) \u2265 200. We choose constraints in descending order of p(edge|g). The first 20 constraints selected using this method are displayed in Table 1 .",
"cite_spans": [],
"ref_spans": [
{
"start": 261,
"end": 268,
"text": "Table 1",
"ref_id": null
}
],
"eq_spans": [],
"section": "Experimental Comparison with Unsupervised Learning",
"sec_num": "5"
},
{
"text": "Although the reader can verify that the constraints in Table 1 are reasonable, we additionally experiment with human-provided constraints. We use the prototype phrase-structure constraints provided by Haghighi and Klein (2006) , and with the aid of head-finding rules, extract 14 (parent-pos,child-pos,direction) constraints. 7 We then estimated target expectations for these constraints using our prior knowledge, without looking at the training data. We also created a second constraint set with an additional six constraints for tag pairs that were previously underrepresented.",
"cite_spans": [
{
"start": 201,
"end": 226,
"text": "Haghighi and Klein (2006)",
"ref_id": "BIBREF7"
},
{
"start": 326,
"end": 327,
"text": "7",
"ref_id": null
}
],
"ref_spans": [
{
"start": 55,
"end": 62,
"text": "Table 1",
"ref_id": null
}
],
"eq_spans": [],
"section": "Experimental Comparison with Unsupervised Learning",
"sec_num": "5"
},
{
"text": "We present results varying the number of constraints in Figures 1 and 2. Figure 1 compares supervised and GE training of the CRF model, as well as the feature constraint baseline. First we note that GE training using the full feature set substantially outperforms the restricted feature set, despite the fact that the same set of constraints is used for both experiments. This result demonstrates GE's ability to learn about related but nonconstrained features. GE training also outperforms the baseline 8 .",
"cite_spans": [],
"ref_spans": [
{
"start": 56,
"end": 81,
"text": "Figures 1 and 2. Figure 1",
"ref_id": "FIGREF1"
}
],
"eq_spans": [],
"section": "Results",
"sec_num": "5.1"
},
{
"text": "We compare GE training of the CRF model with unsupervised learning of the DMV model in Figure 2 9 . Despite the fact that the restricted CRF is less expressive than DMV, GE training of this model outperforms EM with 30 constraints and CE with 50 constraints. GE training of the full CRF outperforms EM with 10 constraints and CE with 20 constraints (those displayed in Table 1 ). GE training of the full CRF with the set of 14 constraints from (Haghighi and Klein, 2006) , gives accuracy of 53.8%, which is above the interpolated oracle constraints curve (43.5% accuracy with 10 constraints, 61.3% accuracy with 20 constraints). With the 6 additional constraints, we obtain accuracy of 57.7% and match CE. Recall that CE, EM, and the DMV model incorporate prior knowledge indirectly, and that the reported results are heavily-tuned ideal cases (see Section 2). In contrast, GE provides a method to directly encode intuitive linguistic insights.",
"cite_spans": [
{
"start": 444,
"end": 470,
"text": "(Haghighi and Klein, 2006)",
"ref_id": "BIBREF7"
}
],
"ref_spans": [
{
"start": 87,
"end": 97,
"text": "Figure 2 9",
"ref_id": "FIGREF3"
},
{
"start": 369,
"end": 376,
"text": "Table 1",
"ref_id": null
}
],
"eq_spans": [],
"section": "Results",
"sec_num": "5.1"
},
{
"text": "Finally, note that structural annealing (Smith and Eisner, 2006) provides 66.7% accuracy on WSJ10 when choosing the best performing annealing schedule . As noted in Section 2 other annealing schedules provide accuracy as low as 32.5%. GE training of the full CRF attains accuracy of 67.0% with 30 constraints.",
"cite_spans": [
{
"start": 40,
"end": 64,
"text": "(Smith and Eisner, 2006)",
"ref_id": "BIBREF20"
}
],
"ref_spans": [],
"eq_spans": [],
"section": "Results",
"sec_num": "5.1"
},
{
"text": "Unsupervised parsing methods are typically evaluated on short sentences, as in Section 5. In this section we show that GE can be used to train parsers for longer sentences that provide comparable accuracy to supervised training with tens to hundreds of parsed sentences. We use the standard train/test splits of the Spanish, Dutch, and Turkish data from the 2006 CoNLL Shared Task. We also use standard edge-factored feature templates (McDonald et al., 2005) 10 . We experiment with versions of the dat- 9 Klein and Manning (2004) report 43.2% accuracy for DMV with EM on WSJ10. When jointly modeling constituency and dependencies, Klein and Manning (2004) report accuracy of 47.5%. Seginer (2007) and Bod (2006) propose unsupervised phrase structure parsing methods that give better unlabeled F-scores than DMV with EM, but they do not report directed dependency accuracy.",
"cite_spans": [
{
"start": 435,
"end": 458,
"text": "(McDonald et al., 2005)",
"ref_id": "BIBREF16"
},
{
"start": 504,
"end": 505,
"text": "9",
"ref_id": null
},
{
"start": 632,
"end": 656,
"text": "Klein and Manning (2004)",
"ref_id": "BIBREF9"
},
{
"start": 683,
"end": 697,
"text": "Seginer (2007)",
"ref_id": "BIBREF18"
},
{
"start": 702,
"end": 712,
"text": "Bod (2006)",
"ref_id": "BIBREF1"
}
],
"ref_spans": [],
"eq_spans": [],
"section": "Experimental Comparison with Supervised Training on Long Sentences",
"sec_num": "6"
},
{
"text": "10 Typical feature processing uses only supported features, or those features that occur on at least one true edge in the training data. Because we assume that the data is unlabeled, we instead use features on all possible edges. This generates tens of millions features, so we prune those features that occur fewer than 10 total times, as in (Smith and Eisner, 2007) . sets in which we remove sentences that are longer than 20 words and 60 words.",
"cite_spans": [
{
"start": 343,
"end": 367,
"text": "(Smith and Eisner, 2007)",
"ref_id": "BIBREF21"
}
],
"ref_spans": [],
"eq_spans": [],
"section": "Experimental Comparison with Supervised Training on Long Sentences",
"sec_num": "6"
},
{
"text": "For these experiments, we use an oracle constraint selection method motivated by the linguistic prior knowledge described in Section 4. The first set of constraints specify the most frequent head tag, attachment direction, and distance combinations for each child tag. Specifically, we select oracle constraints of the type (parent-CPOS,child-CPOS,direction,distance) 11 . We add constraints for every g such that c edge (g) > 100 for max length 60 data sets, and c edge (g) > 10 times for max length 20 data sets.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Experimental Comparison with Supervised Training on Long Sentences",
"sec_num": "6"
},
{
"text": "In some cases, the possible parent constraints described above will not be enough to provide high accuracy, because they do not consider other tags in the sentence (McDonald et al., 2005) . Consequently, we experiment with adding an additional 25 sequence constraints (for what are often called \"between\" and \"surrounding\" features). The oracle feature selection method aims to choose such constraints that help to reduce uncertainty in the possible parents constraint set. Consequently, we consider sequence features g s with p(edge|g s = 1) \u2265 0.75, and whose corresponding (parent-CPOS,child-CPOS,direction,distance) constraint g, has edge probabilityp(edge|g) \u2264 0.25. Among these candidates, we sort b\u1ef9 c(g s = 1), and select the top 25.",
"cite_spans": [
{
"start": 164,
"end": 187,
"text": "(McDonald et al., 2005)",
"ref_id": "BIBREF16"
}
],
"ref_spans": [],
"eq_spans": [],
"section": "Experimental Comparison with Supervised Training on Long Sentences",
"sec_num": "6"
},
{
"text": "We compare with the constraint baseline described in Section 5. Additionally, we report the number of parsed sentences required for supervised CRF training (averaged over 5 random splits) to match the accuracy of GE training using the possible parents + sequence constraint set.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Experimental Comparison with Supervised Training on Long Sentences",
"sec_num": "6"
},
{
"text": "The results are provided in Table 2 . We first observe that GE always beats the baseline, especially on parent decisions for which there are no constraints (not reported in Table 2 , but for example 53.8% vs. 20.5% on Turkish 20). Second, we note that accuracy is always improved by adding sequence constraints. Importantly, we observe that GE gives comparable performance to supervised training with tens or hundreds of parsed sentences. These parsed sentences provide a tremendous amount of information to the model, as for example in 20 Spanish length \u2264 60 sentences, a total of 1,630,466 features are observed, 330,856 of them unique. In contrast, the constraint-based methods are provided at most a few hundred constraints. When comparing the human costs of parsing sentences and specifying constraints, remember that parsing sentences requires the development of detailed annotation guidelines, which can be extremely time-consuming (see also the discussion is Section 2).",
"cite_spans": [],
"ref_spans": [
{
"start": 28,
"end": 35,
"text": "Table 2",
"ref_id": "TABREF3"
},
{
"start": 173,
"end": 180,
"text": "Table 2",
"ref_id": "TABREF3"
}
],
"eq_spans": [],
"section": "Experimental Comparison with Supervised Training on Long Sentences",
"sec_num": "6"
},
{
"text": "Finally, we experiment with iteratively adding constraints. We sort constraints with c(g) > 50 byp(edge|g), and ensure that 50% are (parent-CPOS,child-CPOS,direction,distance) constraints and 50% are sequence constraints. For lack of space, we only show the results for Spanish 60. In Figure 3 , we see that GE beats the baseline more soundly than above, and that Table 3 : Error analysis for GE training with possible parent + sequence constraints on Spanish 60 data. On the left, the predicted and true distribution over parent coarse part-of-speech tags. In the middle, the predicted and true distributions over attachment directions and distances. On the right, common features on false positive edges. adding constraints continues to increase accuracy.",
"cite_spans": [],
"ref_spans": [
{
"start": 285,
"end": 293,
"text": "Figure 3",
"ref_id": "FIGREF4"
},
{
"start": 364,
"end": 371,
"text": "Table 3",
"ref_id": null
}
],
"eq_spans": [],
"section": "Experimental Comparison with Supervised Training on Long Sentences",
"sec_num": "6"
},
{
"text": "In this section, we analyze the errors of the model learned with the possible parent + sequence constraints on the Spanish 60 data. In Table 3 , we present four types of analysis. First, we present the predicted and true distributions over coarsegrained parent part of speech tags. We can see that verb is being predicted as a parent tag more often then it should be, while most other tags are predicted less often than they should be. Next, we show the predicted and true distributions over attachment direction and distance. From this we see that the model is often incorrectly predicting left attachments, and is predicting too many short attachments. Finally, we show the most common parent-child tag with direction and distance fea-tures that occur on false positive edges. From this table, we see that many errors concern the attachments of punctuation. The second line indicates a prepositional phrase attachment ambiguity.",
"cite_spans": [],
"ref_spans": [
{
"start": 135,
"end": 142,
"text": "Table 3",
"ref_id": null
}
],
"eq_spans": [],
"section": "Error Analysis",
"sec_num": "7"
},
{
"text": "This analysis could also be performed by a linguist by looking at predicted trees for selected sentences. Once errors are identified, GE constraints could be added to address these problems.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Error Analysis",
"sec_num": "7"
},
{
"text": "In this paper, we developed a novel method for the semi-supervised learning of a non-projective CRF dependency parser that directly uses linguistic prior knowledge as a training signal. It is our hope that this method will permit more effective leveraging of linguistic insight and resources and enable the construction of parsers in languages and domains where treebanks are not available.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Conclusions",
"sec_num": "8"
},
{
"text": "In general, the objective function could also include the likelihood of available labeled data, but throughout this paper we assume we have no parsed sentences.3 Throughout this paper we use a Gaussian prior on parameters with \u03c3 2 = 10.4 In previous work, S was the KL-divergence from the target expectation. The partial derivative of the KL divergence score function includes the same covariance term as above but substitutes a different multiplicative term:G/G \u03bb .",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "",
"sec_num": null
},
{
"text": "Note that we could instead define a CRF for projective dependency parse trees and use a variant of the inside outside algorithm for inference. We choose non-projective because it is the more general case.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "",
"sec_num": null
},
{
"text": "The reported accuracies with the DMV model and the attach-right baseline are taken from.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "",
"sec_num": null
},
{
"text": "Because the CFG rules in(Haghighi and Klein, 2006) are \"flattened\" and in some cases do not generate appropriate dependency constraints, we only used a subset.8 The baseline eventually matches the accuracy of the restricted CRF but this is understandable because GE's ability to bootstrap is greatly reduced with the restricted feature set.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "",
"sec_num": null
},
{
"text": "For these experiments we use coarse-grained part-ofspeech tags in constraints.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "",
"sec_num": null
}
],
"back_matter": [
{
"text": "We thank Ryan McDonald, Keith Hall, John Hale, Xiaoyun Wu, and David Smith for helpful discussions. This work was completed in part while Gregory Druck was an intern at Google. This work was supported in part by the Center for Intelligent Information Retrieval, The Central Intelligence Agency, the National Security Agency and National Science Foundation under NSF grant #IIS-0326249, and by the Defense Advanced Research Projects Agency (DARPA) under Contract No. FA8750-07-D-0185/0004. Any opinions, findings and conclusions or recommendations expressed in this material are the author's and do not necessarily reflect those of the sponsor.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Acknowledgments",
"sec_num": null
}
],
"bib_entries": {
"BIBREF0": {
"ref_id": "b0",
"title": "Development and evaluation of a broad-coverage probabilistic grammar of english language computer manuals",
"authors": [
{
"first": "E",
"middle": [],
"last": "Black",
"suffix": ""
},
{
"first": "J",
"middle": [],
"last": "Lafferty",
"suffix": ""
},
{
"first": "S",
"middle": [],
"last": "Roukos",
"suffix": ""
}
],
"year": 1992,
"venue": "ACL",
"volume": "",
"issue": "",
"pages": "185--192",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "E. Black, J. Lafferty, and S. Roukos. 1992. Development and evaluation of a broad-coverage probabilistic grammar of english language computer manuals. In ACL, pages 185- 192.",
"links": null
},
"BIBREF1": {
"ref_id": "b1",
"title": "An all-subtrees approach to unsupervised parsing",
"authors": [
{
"first": "Rens",
"middle": [],
"last": "Bod",
"suffix": ""
}
],
"year": 2006,
"venue": "ACL",
"volume": "",
"issue": "",
"pages": "865--872",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Rens Bod. 2006. An all-subtrees approach to unsupervised parsing. In ACL, pages 865-872.",
"links": null
},
"BIBREF2": {
"ref_id": "b2",
"title": "Immediate-head parsing for language models",
"authors": [
{
"first": "E",
"middle": [],
"last": "Charniak",
"suffix": ""
}
],
"year": 2001,
"venue": "ACL",
"volume": "",
"issue": "",
"pages": "",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "E. Charniak. 2001. Immediate-head parsing for language models. In ACL.",
"links": null
},
"BIBREF3": {
"ref_id": "b3",
"title": "A relational syntaxsemantics interface based on dependency grammar",
"authors": [
{
"first": "R",
"middle": [],
"last": "Debusmann",
"suffix": ""
},
{
"first": "D",
"middle": [],
"last": "Duchier",
"suffix": ""
},
{
"first": "A",
"middle": [],
"last": "Koller",
"suffix": ""
},
{
"first": "M",
"middle": [],
"last": "Kuhlmann",
"suffix": ""
},
{
"first": "G",
"middle": [],
"last": "Smolka",
"suffix": ""
},
{
"first": "S",
"middle": [],
"last": "Thater",
"suffix": ""
}
],
"year": 2004,
"venue": "COLING",
"volume": "",
"issue": "",
"pages": "",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "R. Debusmann, D. Duchier, A. Koller, M. Kuhlmann, G. Smolka, and S. Thater. 2004. A relational syntax- semantics interface based on dependency grammar. In COLING.",
"links": null
},
"BIBREF4": {
"ref_id": "b4",
"title": "Learning from labeled features using generalized expectation criteria",
"authors": [
{
"first": "G",
"middle": [],
"last": "Druck",
"suffix": ""
},
{
"first": "G",
"middle": [
"S"
],
"last": "Mann",
"suffix": ""
},
{
"first": "A",
"middle": [],
"last": "Mccallum",
"suffix": ""
}
],
"year": 2008,
"venue": "SIGIR",
"volume": "",
"issue": "",
"pages": "",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "G. Druck, G. S. Mann, and A. McCallum. 2008. Learning from labeled features using generalized expectation crite- ria. In SIGIR.",
"links": null
},
"BIBREF5": {
"ref_id": "b5",
"title": "Parsing with soft and hard constraints on dependency length",
"authors": [
{
"first": "J",
"middle": [],
"last": "Eisner",
"suffix": ""
},
{
"first": "N",
"middle": [
"A"
],
"last": "Smith",
"suffix": ""
}
],
"year": 2005,
"venue": "IWPT",
"volume": "",
"issue": "",
"pages": "",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "J. Eisner and N.A. Smith. 2005. Parsing with soft and hard constraints on dependency length. In IWPT.",
"links": null
},
"BIBREF6": {
"ref_id": "b6",
"title": "Dependency grammar induction via bitext projection constraints",
"authors": [
{
"first": "Kuzman",
"middle": [],
"last": "Ganchev",
"suffix": ""
},
{
"first": "Jennifer",
"middle": [],
"last": "Gillenwater",
"suffix": ""
},
{
"first": "Ben",
"middle": [],
"last": "Taskar",
"suffix": ""
}
],
"year": 2009,
"venue": "ACL",
"volume": "",
"issue": "",
"pages": "",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Kuzman Ganchev, Jennifer Gillenwater, and Ben Taskar. 2009. Dependency grammar induction via bitext projec- tion constraints. In ACL.",
"links": null
},
"BIBREF7": {
"ref_id": "b7",
"title": "Prototype-driven grammar induction",
"authors": [
{
"first": "A",
"middle": [],
"last": "Haghighi",
"suffix": ""
},
{
"first": "D",
"middle": [],
"last": "Klein",
"suffix": ""
}
],
"year": 2006,
"venue": "COLING",
"volume": "",
"issue": "",
"pages": "",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "A. Haghighi and D. Klein. 2006. Prototype-driven grammar induction. In COLING.",
"links": null
},
"BIBREF8": {
"ref_id": "b8",
"title": "Semi-supervised learning for semantic parsing using support vector machines",
"authors": [
{
"first": "R",
"middle": [
"J"
],
"last": "Kate",
"suffix": ""
},
{
"first": "R",
"middle": [
"J"
],
"last": "Mooney",
"suffix": ""
}
],
"year": 2007,
"venue": "HLT-NAACL",
"volume": "",
"issue": "",
"pages": "",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "R. J. Kate and R. J. Mooney. 2007. Semi-supervised learning for semantic parsing using support vector machines. In HLT-NAACL (Short Papers).",
"links": null
},
"BIBREF9": {
"ref_id": "b9",
"title": "Corpus-based induction of syntactic structure: Models of dependency and constituency",
"authors": [
{
"first": "D",
"middle": [],
"last": "Klein",
"suffix": ""
},
{
"first": "C",
"middle": [],
"last": "Manning",
"suffix": ""
}
],
"year": 2004,
"venue": "ACL",
"volume": "",
"issue": "",
"pages": "",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "D. Klein and C. Manning. 2004. Corpus-based induction of syntactic structure: Models of dependency and con- stituency. In ACL.",
"links": null
},
"BIBREF10": {
"ref_id": "b10",
"title": "Simple semisupervised dependency parsing",
"authors": [
{
"first": "T",
"middle": [],
"last": "Koo",
"suffix": ""
},
{
"first": "X",
"middle": [],
"last": "Carreras",
"suffix": ""
},
{
"first": "M",
"middle": [],
"last": "Collins",
"suffix": ""
}
],
"year": 2008,
"venue": "ACL",
"volume": "",
"issue": "",
"pages": "",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "T. Koo, X. Carreras, and M. Collins. 2008. Simple semi- supervised dependency parsing. In ACL.",
"links": null
},
"BIBREF11": {
"ref_id": "b11",
"title": "Conditional random fields: Probabilistic models for segmenting and labeling sequence data",
"authors": [
{
"first": "J",
"middle": [],
"last": "Lafferty",
"suffix": ""
},
{
"first": "A",
"middle": [],
"last": "Mccallum",
"suffix": ""
},
{
"first": "F",
"middle": [],
"last": "Pereira",
"suffix": ""
}
],
"year": 2001,
"venue": "ICML",
"volume": "",
"issue": "",
"pages": "",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "J. Lafferty, A. McCallum, and F. Pereira. 2001. Conditional random fields: Probabilistic models for segmenting and labeling sequence data. In ICML.",
"links": null
},
"BIBREF12": {
"ref_id": "b12",
"title": "Simple, robust, scalable semi-supervised learning via expectation regularization",
"authors": [
{
"first": "G",
"middle": [],
"last": "Mann",
"suffix": ""
},
{
"first": "A",
"middle": [],
"last": "Mccallum",
"suffix": ""
}
],
"year": 2007,
"venue": "ICML",
"volume": "",
"issue": "",
"pages": "",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "G. Mann and A. McCallum. 2007. Simple, robust, scal- able semi-supervised learning via expectation regulariza- tion. In ICML.",
"links": null
},
"BIBREF13": {
"ref_id": "b13",
"title": "Generalized expectation criteria for semi-supervised learning of conditional random fields",
"authors": [
{
"first": "G",
"middle": [],
"last": "Mann",
"suffix": ""
},
{
"first": "A",
"middle": [],
"last": "Mccallum",
"suffix": ""
}
],
"year": 2008,
"venue": "ACL",
"volume": "",
"issue": "",
"pages": "",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "G. Mann and A. McCallum. 2008. Generalized expectation criteria for semi-supervised learning of conditional ran- dom fields. In ACL.",
"links": null
},
"BIBREF14": {
"ref_id": "b14",
"title": "Effective self-training for parsing",
"authors": [
{
"first": "D",
"middle": [],
"last": "Mcclosky",
"suffix": ""
},
{
"first": "E",
"middle": [],
"last": "Charniak",
"suffix": ""
},
{
"first": "M",
"middle": [],
"last": "Johnson",
"suffix": ""
}
],
"year": 2006,
"venue": "HLT-NAACL",
"volume": "",
"issue": "",
"pages": "",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "D. McClosky, E. Charniak, and M. Johnson. 2006. Effective self-training for parsing. In HLT-NAACL.",
"links": null
},
"BIBREF15": {
"ref_id": "b15",
"title": "On the complexity of non-projective data-driven dependency parsing",
"authors": [
{
"first": "Ryan",
"middle": [],
"last": "Mcdonald",
"suffix": ""
},
{
"first": "Giorgio",
"middle": [],
"last": "Satta",
"suffix": ""
}
],
"year": 2007,
"venue": "Proc. of IWPT",
"volume": "",
"issue": "",
"pages": "121--132",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Ryan McDonald and Giorgio Satta. 2007. On the complex- ity of non-projective data-driven dependency parsing. In Proc. of IWPT, pages 121-132.",
"links": null
},
"BIBREF16": {
"ref_id": "b16",
"title": "Online large-margin training of dependency parsers",
"authors": [
{
"first": "Ryan",
"middle": [],
"last": "Mcdonald",
"suffix": ""
},
{
"first": "Koby",
"middle": [],
"last": "Crammer",
"suffix": ""
},
{
"first": "Fernando",
"middle": [],
"last": "Pereira",
"suffix": ""
}
],
"year": 2005,
"venue": "ACL",
"volume": "",
"issue": "",
"pages": "91--98",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Ryan McDonald, Koby Crammer, and Fernando Pereira. 2005. Online large-margin training of dependency parsers. In ACL, pages 91-98.",
"links": null
},
"BIBREF17": {
"ref_id": "b17",
"title": "A Comprehensive Grammar of the English Language",
"authors": [
{
"first": "R",
"middle": [],
"last": "Quirk",
"suffix": ""
},
{
"first": "S",
"middle": [],
"last": "Greenbaum",
"suffix": ""
},
{
"first": "G",
"middle": [],
"last": "Leech",
"suffix": ""
},
{
"first": "J",
"middle": [],
"last": "Svartvik",
"suffix": ""
}
],
"year": 1985,
"venue": "",
"volume": "",
"issue": "",
"pages": "",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "R. Quirk, S. Greenbaum, G. Leech, and J. Svartvik. 1985. A Comprehensive Grammar of the English Language. Longman.",
"links": null
},
"BIBREF18": {
"ref_id": "b18",
"title": "Fast unsupervised incremental parsing",
"authors": [
{
"first": "Yoav",
"middle": [],
"last": "Seginer",
"suffix": ""
}
],
"year": 2007,
"venue": "ACL",
"volume": "",
"issue": "",
"pages": "384--391",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Yoav Seginer. 2007. Fast unsupervised incremental parsing. In ACL, pages 384-391, Prague, Czech Republic.",
"links": null
},
"BIBREF19": {
"ref_id": "b19",
"title": "Contrastive estimation: training log-linear models on unlabeled data",
"authors": [
{
"first": "A",
"middle": [],
"last": "Noah",
"suffix": ""
},
{
"first": "Jason",
"middle": [],
"last": "Smith",
"suffix": ""
},
{
"first": "",
"middle": [],
"last": "Eisner",
"suffix": ""
}
],
"year": 2005,
"venue": "ACL",
"volume": "",
"issue": "",
"pages": "354--362",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Noah A. Smith and Jason Eisner. 2005. Contrastive esti- mation: training log-linear models on unlabeled data. In ACL, pages 354-362.",
"links": null
},
"BIBREF20": {
"ref_id": "b20",
"title": "Annealing structural bias in multilingual weighted grammar induction",
"authors": [
{
"first": "A",
"middle": [],
"last": "Noah",
"suffix": ""
},
{
"first": "Jason",
"middle": [],
"last": "Smith",
"suffix": ""
},
{
"first": "",
"middle": [],
"last": "Eisner",
"suffix": ""
}
],
"year": 2006,
"venue": "COLING-ACL",
"volume": "",
"issue": "",
"pages": "569--576",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Noah A. Smith and Jason Eisner. 2006. Annealing struc- tural bias in multilingual weighted grammar induction. In COLING-ACL, pages 569-576.",
"links": null
},
"BIBREF21": {
"ref_id": "b21",
"title": "Bootstrapping feature-rich dependency parsers with entropic priors",
"authors": [
{
"first": "David",
"middle": [
"A"
],
"last": "Smith",
"suffix": ""
},
{
"first": "Jason",
"middle": [],
"last": "Eisner",
"suffix": ""
}
],
"year": 2007,
"venue": "EMNLP-CoNLL",
"volume": "",
"issue": "",
"pages": "667--677",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "David A. Smith and Jason Eisner. 2007. Bootstrapping feature-rich dependency parsers with entropic priors. In EMNLP-CoNLL, pages 667-677.",
"links": null
},
"BIBREF22": {
"ref_id": "b22",
"title": "Dependency parsing by belief propagation",
"authors": [
{
"first": "David",
"middle": [
"A"
],
"last": "Smith",
"suffix": ""
},
{
"first": "Jason",
"middle": [],
"last": "Eisner",
"suffix": ""
}
],
"year": 2008,
"venue": "EMNLP",
"volume": "",
"issue": "",
"pages": "",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "David A. Smith and Jason Eisner. 2008. Dependency parsing by belief propagation. In EMNLP.",
"links": null
},
"BIBREF23": {
"ref_id": "b23",
"title": "Probabilistic models of nonprojective dependency trees",
"authors": [
{
"first": "David",
"middle": [
"A"
],
"last": "Smith",
"suffix": ""
},
{
"first": "Noah",
"middle": [
"A"
],
"last": "Smith",
"suffix": ""
}
],
"year": 2007,
"venue": "EMNLP-CoNLL",
"volume": "",
"issue": "",
"pages": "132--140",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "David A. Smith and Noah A. Smith. 2007. Probabilistic models of nonprojective dependency trees. In EMNLP- CoNLL, pages 132-140.",
"links": null
},
"BIBREF24": {
"ref_id": "b24",
"title": "Novel Estimation Methods for Unsupervised Discovery of Latent Structure in Natural Language Text",
"authors": [
{
"first": "A",
"middle": [],
"last": "Noah",
"suffix": ""
},
{
"first": "",
"middle": [],
"last": "Smith",
"suffix": ""
}
],
"year": 2006,
"venue": "",
"volume": "",
"issue": "",
"pages": "",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Noah A. Smith. 2006. Novel Estimation Methods for Un- supervised Discovery of Latent Structure in Natural Lan- guage Text. Ph.D. thesis, Johns Hopkins University.",
"links": null
},
"BIBREF25": {
"ref_id": "b25",
"title": "Semi-supervised convex training for dependency parsing",
"authors": [
{
"first": "Qin Iris",
"middle": [],
"last": "Wang",
"suffix": ""
},
{
"first": "Dale",
"middle": [],
"last": "Schuurmans",
"suffix": ""
},
{
"first": "Dekang",
"middle": [],
"last": "Lin",
"suffix": ""
}
],
"year": 2008,
"venue": "ACL",
"volume": "",
"issue": "",
"pages": "532--540",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Qin Iris Wang, Dale Schuurmans, and Dekang Lin. 2008. Semi-supervised convex training for dependency parsing. In ACL, pages 532-540.",
"links": null
}
},
"ref_entries": {
"FIGREF1": {
"text": "Comparison of the constraint baseline and both GE and supervised training of the restricted and full CRF. Note that supervised training uses 5,301 parsed sentences. GE with human provided constraints closely matches the oracle results.",
"num": null,
"type_str": "figure",
"uris": null
},
"FIGREF3": {
"text": "Comparison of GE training of the restricted and full CRFs with unsupervised learning of DMV. GE training of the full CRF outperforms CE with just 20 constraints. GE also matches CE with 20 human provided constraints.",
"num": null,
"type_str": "figure",
"uris": null
},
"FIGREF4": {
"text": "Comparing GE training of a CRF and constraint baseline while increasing the number of oracle constraints.",
"num": null,
"type_str": "figure",
"uris": null
},
"TABREF1": {
"text": "POS 0.75 VBD \u2192 VBN 0.75 ROOT \u2192 MD 0.75 NNS \u2190 VBP 0.75 ROOT \u2192 VBD 1.00 PRP \u2190 VBP 0.75 ROOT \u2192 VBP 0.75 VBP \u2192 VBN 0.75",
"num": null,
"type_str": "table",
"content": "<table><tr><td>feature</td><td>ex.</td><td>feature</td><td>ex.</td></tr><tr><td>MD \u2192 VB</td><td colspan=\"3\">1.00 NNS \u2190 VBD 0.75</td></tr><tr><td>POS \u2190 NN</td><td colspan=\"3\">0.75 PRP \u2190 VBD 0.75</td></tr><tr><td>JJ \u2190 NNS</td><td>0.75</td><td>VBD \u2192 TO</td><td>1.00</td></tr><tr><td colspan=\"2\">NNP \u2190 ROOT \u2192 VBZ 0.75</td><td>PRP \u2190 VBZ</td><td>0.75</td></tr><tr><td>TO \u2192 VB</td><td>1.00</td><td>NN \u2190 VBZ</td><td>0.75</td></tr><tr><td>VBN \u2192 IN</td><td colspan=\"3\">0.75 VBZ \u2192 VBN 0.75</td></tr></table>",
"html": null
},
"TABREF3": {
"text": "Experiments on Dutch, Spanish, and Turkish with maximum sentence lengths of 20 and 60. Observe that GE outperforms the baseline, adding sequence constraints improves accuracy, and accuracy with GE training is comparable to supervised training with tens to hundreds of parsed sentences.",
"num": null,
"type_str": "table",
"content": "<table><tr><td>parent tag det. adv. conj. pron. verb adj. punc. noun prep.</td><td>true 0.005 0.018 0.012 0.011 0.355 0.067 0.031 0.276 0.181</td><td>predicted 0.005 0.013 0.001 0.009 0.405 0.075 0.013 0.272 0.165</td><td>direction right left distance 1 2 3 4 5 6-10 > 10</td><td>true 0.621 0.339 true 0.495 0.194 0.066 0.042 0.028 0.069 0.066</td><td>predicted 0.598 0.362 predicted 0.564 0.206 0.050 0.037 0.031 0.033 0.039</td><td>feature (distance) verb \u2192 punc. (>10) noun \u2192 prep. (1) adj. \u2192 prep. (1) verb \u2192 verb (6-10) verb \u2192 verb (>10) noun \u2190 punc. (1) verb \u2190 punc. (2) prep. \u2190 punc. (1) verb \u2192 punc. (4) verb \u2192 prep. (1)</td><td>false pos. occ. 1183 1139 855 756 569 512 509 476 427 422</td></tr></table>",
"html": null
}
}
}
} |