File size: 103,864 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 | {
"paper_id": "D09-1037",
"header": {
"generated_with": "S2ORC 1.0.0",
"date_generated": "2023-01-19T16:39:40.337463Z"
},
"title": "A Bayesian Model of Syntax-Directed Tree to String Grammar Induction",
"authors": [
{
"first": "Trevor",
"middle": [],
"last": "Cohn",
"suffix": "",
"affiliation": {
"laboratory": "",
"institution": "University of Edinburgh",
"location": {
"addrLine": "10 Crichton Street",
"postCode": "EH8 9AB",
"settlement": "Edinburgh, Scotland",
"country": "United Kingdom"
}
},
"email": "tcohn@inf.ed.ac.uk"
},
{
"first": "Phil",
"middle": [],
"last": "Blunsom",
"suffix": "",
"affiliation": {
"laboratory": "",
"institution": "University of Edinburgh",
"location": {
"addrLine": "10 Crichton Street",
"postCode": "EH8 9AB",
"settlement": "Edinburgh, Scotland",
"country": "United Kingdom"
}
},
"email": "pblunsom@inf.ed.ac.uk"
}
],
"year": "",
"venue": null,
"identifiers": {},
"abstract": "Tree based translation models are a compelling means of integrating linguistic information into machine translation. Syntax can inform lexical selection and reordering choices and thereby improve translation quality. Research to date has focussed primarily on decoding with such models, but less on the difficult problem of inducing the bilingual grammar from data. We propose a generative Bayesian model of tree-to-string translation which induces grammars that are both smaller and produce better translations than the previous heuristic two-stage approach which employs a separate word alignment step.",
"pdf_parse": {
"paper_id": "D09-1037",
"_pdf_hash": "",
"abstract": [
{
"text": "Tree based translation models are a compelling means of integrating linguistic information into machine translation. Syntax can inform lexical selection and reordering choices and thereby improve translation quality. Research to date has focussed primarily on decoding with such models, but less on the difficult problem of inducing the bilingual grammar from data. We propose a generative Bayesian model of tree-to-string translation which induces grammars that are both smaller and produce better translations than the previous heuristic two-stage approach which employs a separate word alignment step.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Abstract",
"sec_num": null
}
],
"body_text": [
{
"text": "Many recent advances in statistical machine translation (SMT) are a result of the incorporation of syntactic knowledge into the translation process Zollmann and Venugopal, 2006) . This has been facilitated by the use of synchronous grammars to model translation as a generative process over pairs of strings in two languages. Such models are particularly attractive for translating between languages with divergent word orders, such as Chinese and English, where syntax-inspired translation rules can succinctly describe the requisite reordering operations. In contrast, standard phrase-based models (Koehn et al., 2003) assume a mostly monotone mapping between source and target, and therefore cannot adequately model these phenomena. Currently the most successful paradigm for the use of synchronous grammars in translation is that of stringto-tree transduction (Galley et al., 2004; Zollmann and Venugopal, 2006; Galley et al., 2006; . In this case a grammar is extracted from a parallel corpus, with strings on its source side and syntax trees on its target side, which is then used to translate novel sentences by performing inference over the space of target syntax trees licensed by the grammar.",
"cite_spans": [
{
"start": 148,
"end": 177,
"text": "Zollmann and Venugopal, 2006)",
"ref_id": "BIBREF25"
},
{
"start": 600,
"end": 620,
"text": "(Koehn et al., 2003)",
"ref_id": "BIBREF16"
},
{
"start": 864,
"end": 885,
"text": "(Galley et al., 2004;",
"ref_id": "BIBREF9"
},
{
"start": 886,
"end": 915,
"text": "Zollmann and Venugopal, 2006;",
"ref_id": "BIBREF25"
},
{
"start": 916,
"end": 936,
"text": "Galley et al., 2006;",
"ref_id": "BIBREF10"
}
],
"ref_spans": [],
"eq_spans": [],
"section": "Introduction",
"sec_num": "1"
},
{
"text": "To date grammar-based translation models have relied on heuristics to extract a grammar from a word-aligned parallel corpus. These heuristics are extensions of those developed for phrase-based models (Koehn et al., 2003) , and involve symmetrising two directional word alignments followed by a projection step which uses the alignments to find a mapping between source words and nodes in the target parse trees (Galley et al., 2004) . However, such approaches leave much to be desired. Word-alignments rarely factorise cleanly with parse trees (i.e., alignment points cross constituent structures), resulting in large and implausible translation rules which generalise poorly to unseen data (Fossum et al., 2008) . The principal reason for employing a grammar based formalism is to induce rules which capture long-range reorderings between source and target. However if the grammar itself is extracted using wordalignments induced with models that are unable to capture such reorderings, it is unlikely that the grammar will live up to expectations.",
"cite_spans": [
{
"start": 200,
"end": 220,
"text": "(Koehn et al., 2003)",
"ref_id": "BIBREF16"
},
{
"start": 411,
"end": 432,
"text": "(Galley et al., 2004)",
"ref_id": "BIBREF9"
},
{
"start": 691,
"end": 712,
"text": "(Fossum et al., 2008)",
"ref_id": "BIBREF8"
}
],
"ref_spans": [],
"eq_spans": [],
"section": "Introduction",
"sec_num": "1"
},
{
"text": "In this work we draw on recent advances in Bayesian modelling of grammar induction (Johnson et al., 2007; Cohn et al., 2009) to propose a non-parametric model of synchronous tree substitution grammar (STSG), continuing a recent trend in SMT to seek principled probabilistic formulations for heuristic translation models (Zhang et al., 2008; DeNero et al., 2008; Blunsom et al., 2009b; Blunsom et al., 2009a) . This model leverages a hierarchical Bayesian prior to induce a compact translation grammar directly from a parsed parallel corpus, unconstrained by word-alignments. We show that the induced grammars are more plausible and improve translation output.",
"cite_spans": [
{
"start": 83,
"end": 105,
"text": "(Johnson et al., 2007;",
"ref_id": "BIBREF13"
},
{
"start": 106,
"end": 124,
"text": "Cohn et al., 2009)",
"ref_id": "BIBREF4"
},
{
"start": 320,
"end": 340,
"text": "(Zhang et al., 2008;",
"ref_id": "BIBREF24"
},
{
"start": 341,
"end": 361,
"text": "DeNero et al., 2008;",
"ref_id": "BIBREF6"
},
{
"start": 362,
"end": 384,
"text": "Blunsom et al., 2009b;",
"ref_id": "BIBREF1"
},
{
"start": 385,
"end": 407,
"text": "Blunsom et al., 2009a)",
"ref_id": "BIBREF0"
}
],
"ref_spans": [],
"eq_spans": [],
"section": "Introduction",
"sec_num": "1"
},
{
"text": "This paper is structured as follows: In Section 2 we introduce the STSG formalism and describe current heuristic approaches to grammar induction. We define a principled Bayesian model of string-to-tree translation in Section 3, and describe an inference technique using Gibbs sampling in Section 4. In Section 5 we analyse an induced grammar on a corpus of Chinese\u2192English translation, comparing them with a heuristic grammar in terms of grammar size and translation quality.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Introduction",
"sec_num": "1"
},
{
"text": "Current tree-to-string translation models are a form of Synchronous Tree Substitution Grammar (STSG; Eisner (2003) ). Formally, a STSG is a 5-tuple, G = (T, T , N, S, R), where T and T are sets of terminal symbols in the target and source languages respectively, N is a set of nonterminal symbols, S \u2208 N is the distinguished root non-terminal and R is a set of productions (a.k.a. rules). Each production is a tuple comprising an elementary tree and a string, the former referring to a tree fragment of depth \u2265 1 where each internal node is labelled with a non-terminal and each leaf is labelled with either a terminal or a non-terminal. The string part of the rule describes the lexical component of the rule in the source language and includes a special variable for each frontier non-terminal in the elementary tree. These variables describe the reordering and form the recursion sites in the generative process of creating tree and string pairs with the grammar. For example, the rule",
"cite_spans": [
{
"start": 101,
"end": 114,
"text": "Eisner (2003)",
"ref_id": "BIBREF7"
}
],
"ref_spans": [],
"eq_spans": [],
"section": "Background",
"sec_num": "2"
},
{
"text": "EQUATION",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [
{
"start": 0,
"end": 8,
"text": "EQUATION",
"ref_id": "EQREF",
"raw_str": "(NP NP 1 (PP (IN of) NP 2 )), 2 \u7684 1",
"eq_num": "(1)"
}
],
"section": "Background",
"sec_num": "2"
},
{
"text": "rewrites a noun-phrase (NP) as a NP and prepositional phrase (PP) headed by 'of' in the target language. The rule generates the token '\u7684' in the source and reverses the order of the two child noun-phrases, indicated by the numbering of the variables in the string part. A derivation creates a (tree, string) pair by starting with the root non-terminal and an empty string, then choosing a rule to rewrite (substitute) the nonterminal and expand the string. This process repeats by rewriting all frontier non-terminals until there are none remaining. A Probabilistic STSG assigns a probability to each rule in the grammar. The probability of a derivation is the product of the probabilities of its component rules, and the probability of a (tree, string) pair is the sum of the probabilities over all its derivations.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Background",
"sec_num": "2"
},
{
"text": "Grammar based SMT models almost exclusively follow the same two-stage approach to grammar induction developed for phrase-based methods (Koehn et al., 2003) . In this approach they induce a finite-state grammar with phrase-pairs as rules by taking a sentence aligned parallel corpus and 1) predicting word alignments before 2) extracting transduction rules that are 'consistent' with the word aligned data. Although empirically effective, this two stage approach is less than ideal due to the disconnect between the wordbased models used for alignment and the phrasebased translation model. This is problematic as the word-based model cannot recognise phrase-based phenomena. Moreover, it raises the problem of identifying and weighting the rules from the word alignment.",
"cite_spans": [
{
"start": 135,
"end": 155,
"text": "(Koehn et al., 2003)",
"ref_id": "BIBREF16"
}
],
"ref_spans": [],
"eq_spans": [],
"section": "Heuristic Grammar Induction",
"sec_num": "2.1"
},
{
"text": "The same criticisms levied at the phrase-based models apply equally to the two-stage technique used for synchronous grammar induction (Galley et al., 2004; Zollmann and Venugopal, 2006; Galley et al., 2006; . Namely that the word alignment models typically do not use any syntax and therefore will not be able to model, e.g., consistent syntactic reordering effects, or the impact of the syntactic category on phrase translations. The identification and estimation of grammar rules from word aligned data is also nontrivial. Galley et al. (2004) describe an algorithm for inducing a string-to-tree grammar using a parallel corpus with syntax trees on target side. Their method projects the source strings onto nodes of the target tree using the word alignment, and then extracts the minimal transduction rules as well as rules composed of adjacent minimal units. The production weights are estimated either by heuristic counting (Koehn et al., 2003) or using the EM algorithm. Both estimation techniques are flawed. The heuristic method is inconsistent in the limit (Johnson, 2002) while EM is degenerate, placing disproportionate probability mass on the largest rules in order to describe the data with as few a rules as possible (DeNero et al., 2006) . With no limit on rule size this method will learn a single rule for every training instance, and therefore will not generalise to unseen sentences. These problems can be ameliorated by imposing limits on rule size or early stopping of EM training, however neither of these techniques addresses the underlying problems.",
"cite_spans": [
{
"start": 134,
"end": 155,
"text": "(Galley et al., 2004;",
"ref_id": "BIBREF9"
},
{
"start": 156,
"end": 185,
"text": "Zollmann and Venugopal, 2006;",
"ref_id": "BIBREF25"
},
{
"start": 186,
"end": 206,
"text": "Galley et al., 2006;",
"ref_id": "BIBREF10"
},
{
"start": 525,
"end": 545,
"text": "Galley et al. (2004)",
"ref_id": "BIBREF9"
},
{
"start": 929,
"end": 949,
"text": "(Koehn et al., 2003)",
"ref_id": "BIBREF16"
},
{
"start": 1066,
"end": 1081,
"text": "(Johnson, 2002)",
"ref_id": "BIBREF14"
},
{
"start": 1231,
"end": 1252,
"text": "(DeNero et al., 2006)",
"ref_id": "BIBREF5"
}
],
"ref_spans": [],
"eq_spans": [],
"section": "Heuristic Grammar Induction",
"sec_num": "2.1"
},
{
"text": "In contrast, our model is trained in a single step, i.e., the alignment model is the translation model. This allows syntax to directly inform the alignments. We infer a grammar without resorting to word alignment constraints or limits on rule size. The model uses a prior to bias towards a compact grammar with small rules, thus solving the degeneracy problem.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Heuristic Grammar Induction",
"sec_num": "2.1"
},
{
"text": "Our training data comprises parallel target trees and source strings and our aim is to induce a STSG that best describes this data. This is achieved by inferring a distribution over the derivations for each training instance, where the set of derivations collectively specify the grammar. In the following, we denote the source trees as t, target strings s, and derivations r which are sequences of grammar rules, r.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Model",
"sec_num": "3"
},
{
"text": "As described in section 2.1, previous methods for estimating a STSG have suffered from degeneracy. A principled way to correct for such degenerated behaviour is to use a prior over rules which biases towards small rules. This matches our intuition: we expect good translation rules to be small, with few internal nodes, frontier non-terminals and terminal strings. However, we recognise that on occasion larger rules will be necessary; we allow such rules when there is sufficient support in the data.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Model",
"sec_num": "3"
},
{
"text": "We model the grammar as a set of distributions, G c , over the productions for each non-terminal symbol, c. We adopt a non-parametric Bayesian approach by treating each G c as a random variable with a Dirichlet process (DP) prior,",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Model",
"sec_num": "3"
},
{
"text": "r|c \u223c G c G c |\u03b1 c , P 0 \u223c DP(\u03b1 c , P 0 (\u2022|c)) ,",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Model",
"sec_num": "3"
},
{
"text": "where P 0 (\u2022|c) (the base distribution) is a distribution over the infinite space of trees rooted with c, and \u03b1 c (the concentration parameter) controls the model's tendency towards either reusing existing rules or creating novel ones as each training instance is encountered (and consequently, the tendency to infer larger or smaller grammars). We discuss the base distribution in more detail below.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Model",
"sec_num": "3"
},
{
"text": "Rather than representing the distribution G c explicitly, we integrate over all possible values of G c . This leads to the following predictive distribution for the rule r i given the previously observed rules",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Model",
"sec_num": "3"
},
{
"text": "EQUATION",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [
{
"start": 0,
"end": 8,
"text": "EQUATION",
"ref_id": "EQREF",
"raw_str": "r \u2212i = r 1 . . . r i\u22121 , p(r i |r \u2212i , c, \u03b1 c , P 0 ) = n \u2212i r i + \u03b1 c P 0 (r i |c) n \u2212i c + \u03b1 c ,",
"eq_num": "(2)"
}
],
"section": "Model",
"sec_num": "3"
},
{
"text": "where n \u2212i r i is the number number of times r i has been used to rewrite c in r \u2212i , and",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Model",
"sec_num": "3"
},
{
"text": "n \u2212i c = r,R(r)=c n \u2212i",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Model",
"sec_num": "3"
},
{
"text": "r is the total count of rewriting c (here R(r) is the root non-terminal of r). The distribution is exchangeable, meaning that all permutations of the input sequence are assigned the same probability. This allows us to treat any item as being the last, which is fundamental for efficient Gibbs sampling. Henceforth we adopt the notation r \u2212 and n \u2212 to refer to the rules and counts for the whole data set excluding the current rules under consideration, irrespective of their location in the corpus.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Model",
"sec_num": "3"
},
{
"text": "The base distribution, P 0 , assigns a prior probability to an infinite number of rules, where each rule is an (elementary tree, source string) pair denoted r = (e, w). While there are a myriad of possible distributions, we developed a very simple one. We decompose the probability into two factors, P 0 (e, w|c) = P (e|c)P (w|e) ,",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Model",
"sec_num": "3"
},
{
"text": "the probability of the target elementary tree and the probability of the source string, where c = R(r). The tree probability, P (e|c) in (3), is modelled using generative process whereby the root category c is expanded into a sequence of child nonterminals, then each of these are either expanded or left as-is. This process continues until there are no unprocessed children. The number of child nodes for each expansion is drawn from a geometric prior with parameter p child , except in the case of pre-terminals where the number of children is always one. The binary expansion decisions are drawn from a Bernoulli prior with parameter p expand , and non-terminals and terminals are drawn uniformly from N and T respectively. For example, the source side of rule (1) was generated as follows: 1) the NP was rewritten as two children; 2) an NP; and 4) a PP; 5) the NP child was not expanded; 6) the PP child was expanded; 7) as an IN; and 8) a NP; 9) the IN was expanded to the terminal 'of'; and 10) the final NP was not expanded. Each of these steps is a draw from the relevant distribution, and the total probability is the product of the probabilities for each step. The second factor in (3) is P (w|e), the probability of the source string (a sequence of source terminals and variables). We assume that the elementary tree is generated first, and condition the string probability on l = F (e), its number of frontier nodes (i.e., variables). The string is then created by choosing a number of terminals from a geometric prior with parameter p term then drawing each terminal from a uniform distribution over T . Finally each of the l variables are inserted into the string one at a time using a uniform distribution over the possible placements. For the example rule in (1) the generative process corresponds to 1) deciding to create one terminal; 2) with value \u7684; 3) inserting the first variable after the terminal; and 4) inserting the second variable before the terminal. Again, the probability of the string is simply the product of the probabilities for each step.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Model",
"sec_num": "3"
},
{
"text": "Together both the factors in the base distribution penalise large trees with many nodes and long strings with many terminals and variables. P 0 decreases exponentially with rule size, thus discouraging the model from using larger rules; for this to occur the rules must significantly increase the likelihood.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Model",
"sec_num": "3"
},
{
"text": "To train our model we use Gibbs sampling (Geman and Geman, 1984) , a Markov chain Monte Carlo method (Gilks et al., 1996) in which variables are repeatedly sampled conditioned on the values of all other variables in the model. 1 After a period of burn-in, each sampler state (set of variable assignments) is a sample from the posterior distribution of the model. In our case, we wish to sample from the posterior over the grammar, P (r|t, s, \u03b1).",
"cite_spans": [
{
"start": 41,
"end": 64,
"text": "(Geman and Geman, 1984)",
"ref_id": "BIBREF11"
},
{
"start": 101,
"end": 121,
"text": "(Gilks et al., 1996)",
"ref_id": "BIBREF12"
}
],
"ref_spans": [],
"eq_spans": [],
"section": "Training",
"sec_num": "4"
},
{
"text": "To simplify matters we associate an alignment variable, a, with every internal node of the trees in the training set. This variable specifies the span of source tokens to which node is aligned. Alternatively, the node can be unaligned, which is encoded as an empty span. I.e. a \u2208 (J \u00d7 J) \u222a \u2205 where J is the set of target word indices. Spans are written [i, j): inclusive of i and exclusive of j. Each aligned node (a = \u2205) forms the root of a rule as well as being a frontier non-terminal of an ancestor rule, while unaligned nodes form part of an ancestor rule. 2 The set of valid alignments are constrained by the tree in a number of ways. Child nodes can be aligned only to subspans of their ancestor nodes' alignments and no two nodes' alignments can overlap. Finally, the root node of the tree must be aligned to the full 1 Previous approaches to bilingual grammar induction have used variational inference to optimise a bound on the data log-likelihood (Zhang et al., 2008; Blunsom et al., 2009b) . Both these approaches truncated the grammar a priori in order to permit tractable inference. In contrast our Gibbs sampler can perform inference over the full space of grammars. See also Blunsom et al. (2009a) where we present a Gibbs sampler for inducing SCFGs without truncation.",
"cite_spans": [
{
"start": 562,
"end": 563,
"text": "2",
"ref_id": null
},
{
"start": 958,
"end": 978,
"text": "(Zhang et al., 2008;",
"ref_id": "BIBREF24"
},
{
"start": 979,
"end": 1001,
"text": "Blunsom et al., 2009b)",
"ref_id": "BIBREF1"
},
{
"start": 1191,
"end": 1213,
"text": "Blunsom et al. (2009a)",
"ref_id": "BIBREF0"
}
],
"ref_spans": [],
"eq_spans": [],
"section": "Training",
"sec_num": "4"
},
{
"text": "2 The Gibbs sampler is an extension of our sampler for monolingual tree-substitution grammar (Cohn et al., 2009) , which used a binary substitution variable at each node to encode the segmentation of a training tree into elementary trees.",
"cite_spans": [
{
"start": 93,
"end": 112,
"text": "(Cohn et al., 2009)",
"ref_id": "BIBREF4"
}
],
"ref_spans": [],
"eq_spans": [],
"section": "Training",
"sec_num": "4"
},
{
"text": "(S (NP NP 1 PP 2 ) VP 3 . 4 ), 2 1 3 4 (NP DT 1 NN 2 ), 1 2 (DT Every), \u6bcf span of source words.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Training",
"sec_num": "4"
},
{
"text": "(NN corner), \u4e00 \u4e2a \u89d2\u843d (PP (IN of) NP 1 ), 1 \u7684 (NP (NNP Hong) (NNP Kong)), \u9999\u6e2f (VP (VBZ is) VP 1 ), 1 (VP VBN 1 PP 2 ), \u90fd 1 2 (VBN filled), \u5145 \u7740 (PP (IN with) (NP NN 1 )), 1 (NN fun), \u8da3 (. .), \u3002",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Training",
"sec_num": "4"
},
{
"text": "Collectively, the training trees and alignment variables specify the sequence of rules r, which in turn specify the grammar. Figure 1 shows an example derivation with alignment variables. The corresponding STSG rules are shown in Table 1 .",
"cite_spans": [],
"ref_spans": [
{
"start": 125,
"end": 133,
"text": "Figure 1",
"ref_id": "FIGREF0"
},
{
"start": 230,
"end": 237,
"text": "Table 1",
"ref_id": "TABREF0"
}
],
"eq_spans": [],
"section": "Training",
"sec_num": "4"
},
{
"text": "The Gibbs sampler works by sampling new values of the alignment variables, using two different Gibbs operators to make the updates. The first operator, EXPAND, takes a tree node, v, and samples a new alignment, a v , given the alignments of all other nodes in the same tree and all other trees in the corpus, denoted a \u2212 . The set of valid labels is constrained by the other alignments in the tree, specifically that of the node's closest aligned ancestor, a p , its closest aligned descendants, a d , and its aligned siblings, a s (the aligned descendants of a). The alignment variable may be empty, a v = \u2205, while non-empty values must obey the tree constraints. Specifically the span must be a subspan of its ancestor, a v \u2286 a p , subsume its descendants, a v \u2287 a d , and not overlap its siblings, j \u2208 a s , \u2200j \u2208 a v . Figure 2 shows an example with the range of valid values for corner/NN's alignment variable and the corresponding alignments that these encode. Each alignment in the set of valid outcomes defines a set of grammar rules. The non-aligned outcome results in a single rule r p rooted at ancestor node p. While the various aligned outcomes re- sult in a pair of rules, r p and r v , rooted at p and v respectively. In the example in Figure 2 , the topright outcome has a v = \u2205 and",
"cite_spans": [],
"ref_spans": [
{
"start": 822,
"end": 830,
"text": "Figure 2",
"ref_id": "FIGREF1"
},
{
"start": 1250,
"end": 1258,
"text": "Figure 2",
"ref_id": "FIGREF1"
}
],
"eq_spans": [],
"section": "Gibbs operators",
"sec_num": "4.1"
},
{
"text": "r p = (NP DT 1 (NN corner)), 1 \u4e00 \u4e2a \u89d2\u843d .",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Gibbs operators",
"sec_num": "4.1"
},
{
"text": "The bottom-right outcome, a v = [4, 5), describes the pair of rules:",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Gibbs operators",
"sec_num": "4.1"
},
{
"text": "r p = (NP DT 1 NN 2 ), 1 \u4e00 2 \u89d2\u843d and",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Gibbs operators",
"sec_num": "4.1"
},
{
"text": "r v = (NN corner), \u4e2a .",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Gibbs operators",
"sec_num": "4.1"
},
{
"text": "The set of valid options are then scored according to the probability of their rules as follows:",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Gibbs operators",
"sec_num": "4.1"
},
{
"text": "P (r p |r \u2212 ) = n \u2212 rp + \u03b1P 0 (r p |c p ) n \u2212 cp + \u03b1 (4) P (r p , r v |r \u2212 ) = P (r p |r \u2212 )P (r v |r \u2212 , r p ) = n \u2212 r p + \u03b1P 0 (r p |c p ) n \u2212 cp + \u03b1 \u00d7 (5) n \u2212 rv + \u03b4(r p , r v ) + \u03b1P 0 (r v |c v ) n \u2212 cv + \u03b4(c p , c v ) + \u03b1",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Gibbs operators",
"sec_num": "4.1"
},
{
"text": "where c p is the non-terminal at node p (similarly for c v ), n \u2212 denote counts of trees (e.g., n \u2212 rp ) or the sum over all trees expanding a nonterminal (e.g., n \u2212 cv ) in the conditioning context, r \u2212 , and \u03b4(\u2022, \u2022) is the Kronecker delta function, which returns 1 when its arguments are identical and 0 otherwise. For clarity, we have omitted some items from the conditioning context in (4) and (5), namely t, s and hyper-parameters \u03b1, p child , p expand , p term . The \u03b4 terms in the second factor of (5) account for the changes to n \u2212 that would occur after observing r p , which forms part of the conditioning context for r v . If the rules r p and r v are identical, then the count n \u2212 rv would increase by one, and if the rules expand the same root non-terminal, then n \u2212 cv would increase by one. Equation 4is evaluated once for the unaligned outcome, a v = \u2205, and (5) is evaluated for each valid alignment. The probabilities are normalised and an outcome sampled.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Gibbs operators",
"sec_num": "4.1"
},
{
"text": "The EXPAND operator is sufficient to move from one derivation to any other valid derivation, however it may take many steps to do so. These intermediate steps may require the sampler to pass through highly improbable regions of the state space, and consequently such moves are unlikely. The second operator, SWAP, is designed to help address this problem by increasing the mobility of the sampler, allowing it to mix more quickly. The operator considers pairs of nodes, v, w, in one tree and attempts to swap their alignment values. 3 This is illustrated in the example in Figure 3 . There are two options being compared: preserving the alignments (left) or swapping them (right). This can change three rules implied by the derivation: that rooted at the nodes' common aligned ancestor, p, and those rooted at v and w. For the example, the left option implies rules {r p = (NP DT 1 NN 2 ), 1 2 , r v = (DT Every), \u6bcf , r w = (NN corner), \u4e00 \u4e2a \u89d2\u843d } , and the right option implies rules",
"cite_spans": [],
"ref_spans": [
{
"start": 573,
"end": 581,
"text": "Figure 3",
"ref_id": null
}
],
"eq_spans": [],
"section": "Gibbs operators",
"sec_num": "4.1"
},
{
"text": "{r p = (NP DT 1 NN 2 ), 2 1 , r v = (DT Every), \u4e00 \u4e2a \u89d2\u843d , r w = (NN corner), \u6bcf } .",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Gibbs operators",
"sec_num": "4.1"
},
{
"text": "We simply evaluate the probability of both triples of rules under our model,",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Gibbs operators",
"sec_num": "4.1"
},
{
"text": "P (r p , r v , r w |r \u2212 ) = P (r p |r \u2212 )P (r v |r \u2212 , r p ) P (r w |r \u2212 , r p , r v )",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Gibbs operators",
"sec_num": "4.1"
},
{
"text": ", where the additional rules in the conditioning context signify their inclusion in the counts r \u2212 before applying (2) to evaluate the probability (much the same as in (5) where 3 We rarely need to consider the full quadratic space of node pairs, as the validity constraints mean that the only candidates for swapping are siblings (i.e., share the closest aligned ancestor) which do not have any aligned descendants. the \u03b4 functions encode the changes to the counts). An outcome is then sampled according to the normalised probabilities of the preserve vs. swap rules.",
"cite_spans": [
{
"start": 178,
"end": 179,
"text": "3",
"ref_id": null
}
],
"ref_spans": [],
"eq_spans": [],
"section": "Gibbs operators",
"sec_num": "4.1"
},
{
"text": "The Gibbs sampler makes use of both operators. The algorithm visits each (tree, string) pair in the training set in random order and applies the EX-PAND operator to every node in the tree. After the tree has been processed, the SWAP operator is applied to all candidate pairs of nodes. Visiting all sentence pairs in this way constitutes a single sample from the Gibbs sampler.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Gibbs operators",
"sec_num": "4.1"
},
{
"text": "We evaluate our non-parametric model of grammar induction on a subset of the NIST Chinese-English translation evaluation, representing a realistic SMT experiment with millions of words and long sentences. The Chinese-English training data consists of the FBIS corpus (LDC2003E14) and the first 100k sentence pairs from the Sinorama corpus (LDC2005E47). The Chinese text was segmented with a CRF-based Chinese segmenter optimized for MT (Chang et al., 2008) , and the English text was parsed using the Stanford parser (Klein and Manning, 2003) .",
"cite_spans": [
{
"start": 436,
"end": 456,
"text": "(Chang et al., 2008)",
"ref_id": "BIBREF2"
},
{
"start": 517,
"end": 542,
"text": "(Klein and Manning, 2003)",
"ref_id": "BIBREF15"
}
],
"ref_spans": [],
"eq_spans": [],
"section": "Experiments",
"sec_num": "5"
},
{
"text": "As a baseline we implemented the heuristic grammar extraction technique of Galley et al. (2004) (henceforth GHKM) . This method finds the minimum sized translation rules which are consistent with a word-aligned sentence pair, as described in section 2.1. The rules are then treated as events in a relative frequency estimate. 4 We used Giza++ Model 4 to obtain word alignments (Och and Ney, 2003) , using the grow-diag-final-and heuristic to symmetrise the two directional predictions (Koehn et al., 2003) .",
"cite_spans": [
{
"start": 75,
"end": 113,
"text": "Galley et al. (2004) (henceforth GHKM)",
"ref_id": null
},
{
"start": 326,
"end": 327,
"text": "4",
"ref_id": null
},
{
"start": 377,
"end": 396,
"text": "(Och and Ney, 2003)",
"ref_id": "BIBREF19"
},
{
"start": 485,
"end": 505,
"text": "(Koehn et al., 2003)",
"ref_id": "BIBREF16"
}
],
"ref_spans": [],
"eq_spans": [],
"section": "Experiments",
"sec_num": "5"
},
{
"text": "The model was sampled for 300 iterations to 'burn-in', where in each iteration we applied both sampling operators to all nodes (or node pairs) of all training instances. We initialised the sampler using the GHKM derivation of the training data (the baseline system). The final state of the sampler was used to extract the grammar. The hyperparameters were set by hand to \u03b1 = 10 6 , p child = 0.5, p expand = 0.5, and p term = 0.5. 5 Overall the model took on average 2,218s per full iteration of Gibbs sampling and 1 week in total to train, using a single core of a 2.3Ghz AMD Opteron machine.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Experiments",
"sec_num": "5"
},
{
"text": "The resulting grammar had 1.62M rules, almost identical to the GHKM grammar which had 1.63M. Despite their similarity in size the grammars were quite different, as illustrated in Figure 4 , which shows histograms over various measures of rule size for the two grammars. Under each measure the sampled grammar finds many more simple rules -shallower with fewer internal nodes, fewer variables and fewer terminals -than the GHKM method. This demonstrates that the prior is effective in shifting mass away from complex rules. To show how the rules themselves differ, Table 3 lists rules in the sampled grammar that are not in the GHKM grammar. Note that many of these rules are highly plausible, describing regular tree structures and lexicalisation. These rules have not been specified to the same extent in the GHKM grammar. For example the first rule incorporates the TOP symbol, while the GHKM grammar instead relies on the rule (TOP S 1 ), 1 to produce the same fragment. The model has learnt to distinguish between sentence-spanning and subsentential S constituents, which typically do not include final punctuation. The third and ninth (last) rule are particularly interesting. These rules encode reordering effects relating to noun phrases and subordinate prepositional phrases, in particular that Chinese prepositional modifiers precede the nouns they modify. Differences in word order such as these are quite common in Chinese-English corpora, so it is imperative that they are modelled accurately. The rules in the GHKM grammar that do not appear in the sampled grammar are shown in Table 4 . In contrast to the rules only present in the sampled grammar, these have much lower counts, i.e., are less probable. Each of these rules has been specified further by the Bayesian model. (TOP (S NP 1 VP 2 . 3 )), 1 2 3 (S (VP (TO to) VP 1 )), 1 (NP NP 1 (PP (IN of) NP 2 )), 2 1 (PP (IN in) NP 1 ), \u5728 1 (NP NP 1 (PP (IN of) NP 2 )), 1 2 (NP (DT the) NN 1 ), \u7684 1 (S (VP TO 1 VP 2 )), 1 2 (VP (VBZ is) NP 1 ), \u662f 1 (NP (NP (DT the) NN 1 ) (PP (IN of) NP 2 )), 2 1 Table 3 : Top ten rules in the sampled grammar that do not appear in the GHKM grammar. All the above rules are quite high probability, with counts between 37,118 and 7,275 from first to last.",
"cite_spans": [],
"ref_spans": [
{
"start": 179,
"end": 187,
"text": "Figure 4",
"ref_id": "FIGREF4"
},
{
"start": 564,
"end": 571,
"text": "Table 3",
"ref_id": null
},
{
"start": 1591,
"end": 1598,
"text": "Table 4",
"ref_id": null
},
{
"start": 2062,
"end": 2069,
"text": "Table 3",
"ref_id": null
}
],
"eq_spans": [],
"section": "Grammar Analysis",
"sec_num": "5.1"
},
{
"text": "(PP (IN at) (NP DT 1 (NNS levels))), 1 \u7d1a (NP NP 1 , 2 NP 3 (, ,) CC 4 NP 5 ), 1 2 3 4 5 (NP NP 1 , 2 NP 3 , 4 NP 5 (, ,) (CC and) NP 6 ), 1 2 3 4 5 , 6 (S S 1 (NP (PRP They)) VP 2 . 3 ), 1 2 3 (S PP 1 , 2 NP 3 VP 4 . 5 \" 6 ), 1 2 3 4 6 5 (S PP 1 , 2 NP 3 VP 4 . 5 ), 1 \u4e2d 2 3 4 5 (NP (NNP Foreign) (NNP Ministry) NN 1 (NNP Zhu) (NNP Bangzao)), \u5916\u4ea4\u90e8 1 \u6731\u90a6\u9020 (S S 1 S 2 ), 1 2 (S S 1 (NP (PRP We)) VP 2 . 3 ), 1 2 3 (NP (DT the) (NNS people) POS 1 ), \u4eba\u6c11 1 Table 4 : Top ten rules in the GHKM grammar that do not appear in the sampled grammar. These are quite low probability rules: their counts range from 1,137 to 103.",
"cite_spans": [],
"ref_spans": [
{
"start": 450,
"end": 457,
"text": "Table 4",
"ref_id": null
}
],
"eq_spans": [],
"section": "Grammar Analysis",
"sec_num": "5.1"
},
{
"text": "For example, every instance of the first rule had the same determiner and target translation, (PP (IN at) (NP (DT all) (NNS levels))), \u5404 \u7d1a , and therefore the model specified the determiner, resulting in a single rule. The model has correctly learnt that other translations for (DT all) are not appropriate in this context (e.g., \u90fd, \u6240 \u6709 or \u4e00 \u5207). In a number of the remaining rules the commas were lexicalised, or S rules were extended to include the TOP symbol.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Grammar Analysis",
"sec_num": "5.1"
},
{
"text": "To further illustrate the differences between the grammars, Table 5 shows the rules which include the possessive particle, \u7684, and at least one variable. In both grammars there are many fully lexicalised rules which translate the token to, e.g., a determiner or a preposition. The grammars differ on the complex rules which combine lexicalisation and frontier non-terminals. The GHKM rules are all very simple depth-1 SCFG rules, containing minimal information. In contrast, the sampled rules are more lexicalised, licensing the insertion of various English tokens and tree substructure. Note particularly the second and forth rule which succinctly describe the reordering of prepositional Sampled Grammar (NP (DT the) NN 1 ), \u7684 1 (NP (NP (DT the) NN 1 ) (PP (IN of) NP 2 )), 2 \u7684 1 (NP (DT the) NN 1 ), 1 \u7684 (NP (NP (DT the) JJ 1 NN 2 ) (PP (IN of) NP 3 )), 3 \u7684 1 2 (PP (IN of) NP 1 ), 1 \u7684 GHKM Grammar (NP JJ 1 NNS 2 ), 1 \u7684 2 (NP JJ 1 NN 2 ), 1 \u7684 2 (NP DT 1 JJ 2 NN 3 ), 1 2 \u7684 3 (NP PRP$ 1 NN 2 ), 1 \u7684 2 (NP NP 1 PP 2 ), 2 \u7684 1 Table 5 : Top five rules which include the possessive particle \u7684 and at least one variable.",
"cite_spans": [],
"ref_spans": [
{
"start": 60,
"end": 67,
"text": "Table 5",
"ref_id": null
},
{
"start": 1026,
"end": 1033,
"text": "Table 5",
"ref_id": null
}
],
"eq_spans": [],
"section": "Grammar Analysis",
"sec_num": "5.1"
},
{
"text": "phrases with an noun phrase.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Grammar Analysis",
"sec_num": "5.1"
},
{
"text": "In order to test the translation performance of the grammars induced by our model and the GHKM method 6 we report BLEU (Papineni et al., 2002) scores on sentences of up to twenty words in length from the MT03 NIST evaluation. We built a synchronous beam search decoder to find the maximum scoring derivation, based on the CYK+ chart parsing algorithm and the cubepruning method of Chiang (2007) . Parse edges for all constituents spanning a given chart cell were cube-pruned together using a beam of width 1000, and only edges from the top ten constituents in each cell were retained. No artificial glue-rules or rule span limits were employed. 7 The parameters of the translation system were trained to maximize BLEU on the MT02 test set (Och, 2003) . Decoding took roughly 10s per sentence for both grammars, using a 8-core 2.6Ghz Intel Xeon machine. Table 6 shows the BLEU scores for the baseline using the GHKM rule induction algorithm, and our non-parametric Bayesian grammar induction method. We see a small increase in generalisation performance from our model. Our previous anal- 6 Our decoder was unable to process unary rules (those which consume nothing in the source). Monolingual parsing with unary productions is fairly straightforward (Stolcke, 1995) , however in the transductive setting these rules can licence infinite insertions in the target string. This is further complicated by the language model integration. Therefore we composed each unary rule instance with its descendant rule(s) to create a non-unary rule.",
"cite_spans": [
{
"start": 119,
"end": 142,
"text": "(Papineni et al., 2002)",
"ref_id": "BIBREF21"
},
{
"start": 381,
"end": 394,
"text": "Chiang (2007)",
"ref_id": "BIBREF3"
},
{
"start": 645,
"end": 646,
"text": "7",
"ref_id": null
},
{
"start": 739,
"end": 750,
"text": "(Och, 2003)",
"ref_id": "BIBREF20"
},
{
"start": 1088,
"end": 1089,
"text": "6",
"ref_id": null
},
{
"start": 1250,
"end": 1265,
"text": "(Stolcke, 1995)",
"ref_id": "BIBREF22"
}
],
"ref_spans": [
{
"start": 853,
"end": 860,
"text": "Table 6",
"ref_id": null
}
],
"eq_spans": [],
"section": "Translation",
"sec_num": "5.2"
},
{
"text": "7 Our decoder lacks certain features shown to be beneficial to synchronous grammar decoding, in particular rule binarisation . As such the reported results for MT03 lag the state-of-the-art: the Moses phrase-based decoder (Koehn et al., 2007) achieves 26.8. We believe that improvements from a better decoder implementation would be orthogonal to the improvements presented here (and would allow us to relax the length restriction on the test set).",
"cite_spans": [
{
"start": 222,
"end": 242,
"text": "(Koehn et al., 2007)",
"ref_id": "BIBREF17"
}
],
"ref_spans": [],
"eq_spans": [],
"section": "Translation",
"sec_num": "5.2"
},
{
"text": "BLEU score GHKM 26.0 Our model 26.6 Table 6 : Translation results on the NIST test set MT03 for sentences of length \u2264 20.",
"cite_spans": [],
"ref_spans": [
{
"start": 36,
"end": 43,
"text": "Table 6",
"ref_id": null
}
],
"eq_spans": [],
"section": "Model",
"sec_num": null
},
{
"text": "ysis (Section 5.1) of the grammars produced by the two approaches showed our method produced better lexicalised rules than those induced by the GHKM algorithm. Galley et al. (2006) noted that the GHKM algorithm often over generalised and proposed combining minimal rules to form composed rules as a solution. Although composing rules was effective at improving BLEU scores, the result was a massive expansion in the size of the grammar. By learning the appropriate level of lexicalisation we believe that our inference algorithm is having a similar effect as composing rules (Galley et al., 2006) , however the resulting grammar remains compact, a significant advantage of our approach.",
"cite_spans": [
{
"start": 160,
"end": 180,
"text": "Galley et al. (2006)",
"ref_id": "BIBREF10"
},
{
"start": 575,
"end": 596,
"text": "(Galley et al., 2006)",
"ref_id": "BIBREF10"
}
],
"ref_spans": [],
"eq_spans": [],
"section": "Model",
"sec_num": null
},
{
"text": "In this paper we have presented a method for inducing a tree-to-string grammar which removes the need for various heuristics and constraints from models of word alignment. Instead the model is capable of directly inferring a grammar in one step, using the syntactic fragments that it has learnt to better align the source and target data. Using a prior which favours sparse distributions and simpler rules, we demonstrate that the model finds a more parsimonious grammar than the heuristic technique. Moreover, this grammar results in improved translations on a standard evaluation set.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Conclusion",
"sec_num": "6"
},
{
"text": "We expect that various extensions to the model would improve its performance. One avenue is to develop a more sophisticated prior over rules, e.g., one that recognises common types of rule via the shape of the tree and ordering pattern in the target. A second avenue is to develop better means of inference under the grammar, in order to ensure faster mixing and a means to escape from local optima. Finally, we wish to develop a method for decoding under the full Bayesian model, instead of the current beam search. With these extensions we expect that our model of grammar induction has the potential to greatly improve translation output.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Conclusion",
"sec_num": "6"
},
{
"text": "Our implementation of the GHKM algorithm attaches unaligned source words to the highest possible node in the source tree, rather than allowing all attachment points as in the original presentation(Galley et al., 2004). Allowing all attachments made no difference to translation performance, but did make the grammar considerably larger. We implemented only the minimal rule extraction, i.e., with no rule composition(Galley et al., 2006). Consequently there is no derivational ambiguity, obviating the need for expectation maximisation or similar for rule estimation.5 Note that although \u03b1 seems large, it still encourages sparse distributions as the P0 values are typically much smaller than its reciprocal, 10 \u22126 , especially if the rule is large. \u03b1P0 < 1 implies a sparse Dirichlet prior.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "",
"sec_num": null
}
],
"back_matter": [
{
"text": "The authors acknowledge the support of the EP-SRC (grants GR/T04557/01 and EP/D074959/1). This work has made use of the resources provided by the Edinburgh Compute and Data Facility (ECDF). The ECDF is partially supported by the eDIKT initiative.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Acknowledgements",
"sec_num": null
}
],
"bib_entries": {
"BIBREF0": {
"ref_id": "b0",
"title": "A Gibbs sampler for phrasal synchronous grammar induction",
"authors": [
{
"first": "P",
"middle": [],
"last": "Blunsom",
"suffix": ""
},
{
"first": "T",
"middle": [],
"last": "Cohn",
"suffix": ""
},
{
"first": "C",
"middle": [],
"last": "Dyer",
"suffix": ""
},
{
"first": "M",
"middle": [],
"last": "Osborne",
"suffix": ""
}
],
"year": 2009,
"venue": "Proceedings of the Joint conference of the 47th Annual Meeting of the Association for Computational Linguistics and the 4th International Joint Conference on Natural Language Processing of the Asian Federation of Natural Language Processing",
"volume": "",
"issue": "",
"pages": "",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "P. Blunsom, T. Cohn, C. Dyer, M. Osborne. 2009a. A Gibbs sampler for phrasal synchronous grammar induction. In Proceedings of the Joint conference of the 47th Annual Meeting of the Association for Computational Linguistics and the 4th International Joint Conference on Natural Language Processing of the Asian Federation of Natural Language Pro- cessing, Singapore. To appear.",
"links": null
},
"BIBREF1": {
"ref_id": "b1",
"title": "Bayesian synchronous grammar induction",
"authors": [
{
"first": "P",
"middle": [],
"last": "Blunsom",
"suffix": ""
},
{
"first": "T",
"middle": [],
"last": "Cohn",
"suffix": ""
},
{
"first": "M",
"middle": [],
"last": "Osborne",
"suffix": ""
}
],
"year": 2009,
"venue": "Advances in Neural Information Processing Systems",
"volume": "21",
"issue": "",
"pages": "161--168",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "P. Blunsom, T. Cohn, M. Osborne. 2009b. Bayesian synchronous grammar induction. In D. Koller, D. Schuurmans, Y. Bengio, L. Bottou, eds., Ad- vances in Neural Information Processing Systems 21, 161-168. MIT Press, Cambridge, MA.",
"links": null
},
"BIBREF2": {
"ref_id": "b2",
"title": "Optimizing Chinese word segmentation for machine translation performance",
"authors": [
{
"first": "P.-C",
"middle": [],
"last": "Chang",
"suffix": ""
},
{
"first": "M",
"middle": [],
"last": "Galley",
"suffix": ""
},
{
"first": "C",
"middle": [
"D"
],
"last": "Manning",
"suffix": ""
}
],
"year": 2008,
"venue": "Proceedings of the Third Workshop on Statistical Machine Translation",
"volume": "",
"issue": "",
"pages": "224--232",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "P.-C. Chang, M. Galley, C. D. Manning. 2008. Op- timizing Chinese word segmentation for machine translation performance. In Proceedings of the Third Workshop on Statistical Machine Translation, 224-232, Columbus, Ohio.",
"links": null
},
"BIBREF3": {
"ref_id": "b3",
"title": "Hierarchical phrase-based translation",
"authors": [
{
"first": "D",
"middle": [],
"last": "Chiang",
"suffix": ""
}
],
"year": 2007,
"venue": "Computational Linguistics",
"volume": "33",
"issue": "2",
"pages": "201--228",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "D. Chiang. 2007. Hierarchical phrase-based transla- tion. Computational Linguistics, 33(2):201-228.",
"links": null
},
"BIBREF4": {
"ref_id": "b4",
"title": "Inducing compact but accurate tree-substitution grammars",
"authors": [
{
"first": "T",
"middle": [],
"last": "Cohn",
"suffix": ""
},
{
"first": "S",
"middle": [],
"last": "Goldwater",
"suffix": ""
},
{
"first": "P",
"middle": [],
"last": "Blunsom",
"suffix": ""
}
],
"year": 2009,
"venue": "Proceedings of Human Language Technologies: The 2009 Annual Conference of the North American Chapter of the Association for Computational Linguistics",
"volume": "",
"issue": "",
"pages": "548--556",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "T. Cohn, S. Goldwater, P. Blunsom. 2009. Inducing compact but accurate tree-substitution grammars. In Proceedings of Human Language Technologies: The 2009 Annual Conference of the North American Chapter of the Association for Computational Lin- guistics, 548-556, Boulder, Colorado.",
"links": null
},
"BIBREF5": {
"ref_id": "b5",
"title": "Why generative phrase models underperform surface heuristics",
"authors": [
{
"first": "J",
"middle": [],
"last": "Denero",
"suffix": ""
},
{
"first": "D",
"middle": [],
"last": "Gillick",
"suffix": ""
},
{
"first": "J",
"middle": [],
"last": "Zhang",
"suffix": ""
},
{
"first": "D",
"middle": [],
"last": "Klein",
"suffix": ""
}
],
"year": 2006,
"venue": "Proceedings on the Workshop on Statistical Machine Translation",
"volume": "",
"issue": "",
"pages": "31--38",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "J. DeNero, D. Gillick, J. Zhang, D. Klein. 2006. Why generative phrase models underperform sur- face heuristics. In Proceedings on the Workshop on Statistical Machine Translation, 31-38, New York City, NY.",
"links": null
},
"BIBREF6": {
"ref_id": "b6",
"title": "Sampling alignment structure under a Bayesian translation model",
"authors": [
{
"first": "J",
"middle": [],
"last": "Denero",
"suffix": ""
},
{
"first": "A",
"middle": [],
"last": "Bouchard-C\u00f4t\u00e9",
"suffix": ""
},
{
"first": "D",
"middle": [],
"last": "Klein",
"suffix": ""
}
],
"year": 2008,
"venue": "Proceedings of the 2008 Conference on Empirical Methods in Natural Language Processing",
"volume": "",
"issue": "",
"pages": "314--323",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "J. DeNero, A. Bouchard-C\u00f4t\u00e9, D. Klein. 2008. Sam- pling alignment structure under a Bayesian transla- tion model. In Proceedings of the 2008 Conference on Empirical Methods in Natural Language Pro- cessing, 314-323, Honolulu, Hawaii.",
"links": null
},
"BIBREF7": {
"ref_id": "b7",
"title": "Learning non-isomorphic tree mappings for machine translation",
"authors": [
{
"first": "J",
"middle": [],
"last": "Eisner",
"suffix": ""
}
],
"year": 2003,
"venue": "The Companion Volume to the Proceedings of 41st Annual Meeting of the Association for Computational Linguistics",
"volume": "",
"issue": "",
"pages": "205--208",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "J. Eisner. 2003. Learning non-isomorphic tree map- pings for machine translation. In The Companion Volume to the Proceedings of 41st Annual Meeting of the Association for Computational Linguistics, 205- 208, Sapporo, Japan.",
"links": null
},
"BIBREF8": {
"ref_id": "b8",
"title": "Using syntax to improve word alignment precision for syntaxbased machine translation",
"authors": [
{
"first": "V",
"middle": [],
"last": "Fossum",
"suffix": ""
},
{
"first": "K",
"middle": [],
"last": "Knight",
"suffix": ""
},
{
"first": "S",
"middle": [],
"last": "Abney",
"suffix": ""
}
],
"year": 2008,
"venue": "Proceedings of the Third Workshop on Statistical Machine Translation",
"volume": "",
"issue": "",
"pages": "44--52",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "V. Fossum, K. Knight, S. Abney. 2008. Using syn- tax to improve word alignment precision for syntax- based machine translation. In Proceedings of the Third Workshop on Statistical Machine Translation, 44-52, Columbus, Ohio.",
"links": null
},
"BIBREF9": {
"ref_id": "b9",
"title": "What's in a translation rule",
"authors": [
{
"first": "M",
"middle": [],
"last": "Galley",
"suffix": ""
},
{
"first": "M",
"middle": [],
"last": "Hopkins",
"suffix": ""
},
{
"first": "K",
"middle": [],
"last": "Knight",
"suffix": ""
},
{
"first": "D",
"middle": [],
"last": "Marcu",
"suffix": ""
}
],
"year": 2004,
"venue": "Proceedings of the 2004 Human Language Technology Conference of the North American Chapter of the Association for Computational Linguistics",
"volume": "",
"issue": "",
"pages": "273--280",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "M. Galley, M. Hopkins, K. Knight, D. Marcu. 2004. What's in a translation rule? In Proceedings of the 2004 Human Language Technology Conference of the North American Chapter of the Association for Computational Linguistics, 273-280, Boston, MA.",
"links": null
},
"BIBREF10": {
"ref_id": "b10",
"title": "Scalable inference and training of context-rich syntactic translation models",
"authors": [
{
"first": "M",
"middle": [],
"last": "Galley",
"suffix": ""
},
{
"first": "J",
"middle": [],
"last": "Graehl",
"suffix": ""
},
{
"first": "K",
"middle": [],
"last": "Knight",
"suffix": ""
},
{
"first": "D",
"middle": [],
"last": "Marcu",
"suffix": ""
},
{
"first": "S",
"middle": [],
"last": "Deneefe",
"suffix": ""
},
{
"first": "W",
"middle": [],
"last": "Wang",
"suffix": ""
},
{
"first": "I",
"middle": [],
"last": "Thayer",
"suffix": ""
}
],
"year": 2006,
"venue": "Proceedings of the 21st International Conference on Computational Linguistics and 44th Annual Meeting of the Association for Computational Linguistics",
"volume": "",
"issue": "",
"pages": "961--968",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "M. Galley, J. Graehl, K. Knight, D. Marcu, S. DeNeefe, W. Wang, I. Thayer. 2006. Scalable inference and training of context-rich syntactic translation mod- els. In Proceedings of the 21st International Con- ference on Computational Linguistics and 44th An- nual Meeting of the Association for Computational Linguistics, 961-968, Sydney, Australia.",
"links": null
},
"BIBREF11": {
"ref_id": "b11",
"title": "Stochastic relaxation, Gibbs distributions and the Bayesian restoration of images",
"authors": [
{
"first": "S",
"middle": [],
"last": "Geman",
"suffix": ""
},
{
"first": "D",
"middle": [],
"last": "Geman",
"suffix": ""
}
],
"year": 1984,
"venue": "IEEE Transactions on Pattern Analysis and Machine Intelligence",
"volume": "6",
"issue": "",
"pages": "721--741",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "S. Geman, D. Geman. 1984. Stochastic relaxation, Gibbs distributions and the Bayesian restoration of images. IEEE Transactions on Pattern Analysis and Machine Intelligence, 6:721-741.",
"links": null
},
"BIBREF12": {
"ref_id": "b12",
"title": "Markov Chain Monte Carlo in Practice",
"authors": [
{
"first": "W",
"middle": [],
"last": "Gilks",
"suffix": ""
},
{
"first": "S",
"middle": [],
"last": "Richardson",
"suffix": ""
},
{
"first": "D",
"middle": [
"J"
],
"last": "Spiegelhalter",
"suffix": ""
}
],
"year": 1996,
"venue": "",
"volume": "",
"issue": "",
"pages": "",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "W. Gilks, S. Richardson, D. J. Spiegelhalter, eds. 1996. Markov Chain Monte Carlo in Practice. Chapman and Hall, Suffolk.",
"links": null
},
"BIBREF13": {
"ref_id": "b13",
"title": "Adaptor grammars: A framework for specifying compositional nonparametric Bayesian models",
"authors": [
{
"first": "M",
"middle": [],
"last": "Johnson",
"suffix": ""
},
{
"first": "T",
"middle": [
"L"
],
"last": "Griffiths",
"suffix": ""
},
{
"first": "S",
"middle": [],
"last": "Goldwater",
"suffix": ""
}
],
"year": 2007,
"venue": "Advances in Neural Information Processing Systems",
"volume": "19",
"issue": "",
"pages": "641--648",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "M. Johnson, T. L. Griffiths, S. Goldwater. 2007. Adaptor grammars: A framework for specifying compositional nonparametric Bayesian models. In B. Sch\u00f6lkopf, J. Platt, T. Hoffman, eds., Advances in Neural Information Processing Systems 19, 641- 648. MIT Press, Cambridge, MA.",
"links": null
},
"BIBREF14": {
"ref_id": "b14",
"title": "The DOP estimation method is biased and inconsistent",
"authors": [
{
"first": "M",
"middle": [],
"last": "Johnson",
"suffix": ""
}
],
"year": 2002,
"venue": "Computational Lingusitics",
"volume": "28",
"issue": "1",
"pages": "71--76",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "M. Johnson. 2002. The DOP estimation method is biased and inconsistent. Computational Lingusitics, 28(1):71-76.",
"links": null
},
"BIBREF15": {
"ref_id": "b15",
"title": "Fast exact inference with a factored model for natural language parsing",
"authors": [
{
"first": "D",
"middle": [],
"last": "Klein",
"suffix": ""
},
{
"first": "C",
"middle": [
"D"
],
"last": "Manning",
"suffix": ""
}
],
"year": 2003,
"venue": "Advances in Neural Information Processing Systems",
"volume": "15",
"issue": "",
"pages": "3--10",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "D. Klein, C. D. Manning. 2003. Fast exact inference with a factored model for natural language parsing. In In Advances in Neural Information Processing Systems 15, 3-10. MIT Press.",
"links": null
},
"BIBREF16": {
"ref_id": "b16",
"title": "Statistical phrase-based translation",
"authors": [
{
"first": "P",
"middle": [],
"last": "Koehn",
"suffix": ""
},
{
"first": "F",
"middle": [
"J"
],
"last": "Och",
"suffix": ""
},
{
"first": "D",
"middle": [],
"last": "Marcu",
"suffix": ""
}
],
"year": 2003,
"venue": "Proceedings of the 2003 Human Language Technology Conference of the North American Chapter of the Association for Computational Linguistics",
"volume": "",
"issue": "",
"pages": "48--54",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "P. Koehn, F. J. Och, D. Marcu. 2003. Statisti- cal phrase-based translation. In Proceedings of the 2003 Human Language Technology Conference of the North American Chapter of the Association for Computational Linguistics, 48-54, Edmonton, Canada.",
"links": null
},
"BIBREF17": {
"ref_id": "b17",
"title": "Moses: Open source toolkit for statistical machine translation",
"authors": [
{
"first": "P",
"middle": [],
"last": "Koehn",
"suffix": ""
},
{
"first": "H",
"middle": [],
"last": "Hoang",
"suffix": ""
},
{
"first": "A",
"middle": [],
"last": "Birch",
"suffix": ""
},
{
"first": "C",
"middle": [],
"last": "Callison-Burch",
"suffix": ""
},
{
"first": "M",
"middle": [],
"last": "Federico",
"suffix": ""
},
{
"first": "N",
"middle": [],
"last": "Bertoldi",
"suffix": ""
},
{
"first": "B",
"middle": [],
"last": "Cowan",
"suffix": ""
},
{
"first": "W",
"middle": [],
"last": "Shen",
"suffix": ""
},
{
"first": "C",
"middle": [],
"last": "Moran",
"suffix": ""
},
{
"first": "R",
"middle": [],
"last": "Zens",
"suffix": ""
},
{
"first": "C",
"middle": [],
"last": "Dyer",
"suffix": ""
},
{
"first": "O",
"middle": [],
"last": "Bojar",
"suffix": ""
},
{
"first": "A",
"middle": [],
"last": "Constantin",
"suffix": ""
},
{
"first": "E",
"middle": [],
"last": "Herbst",
"suffix": ""
}
],
"year": 2007,
"venue": "Proceedings of the 45th Annual Meeting of the Association for Computational Linguistics Companion",
"volume": "",
"issue": "",
"pages": "177--180",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "P. Koehn, H. Hoang, A. Birch, C. Callison-Burch, M. Federico, N. Bertoldi, B. Cowan, W. Shen, C. Moran, R. Zens, C. Dyer, O. Bojar, A. Con- stantin, E. Herbst. 2007. Moses: Open source toolkit for statistical machine translation. In Pro- ceedings of the 45th Annual Meeting of the Associa- tion for Computational Linguistics Companion Vol- ume Proceedings of the Demo and Poster Sessions, 177-180, Prague, Czech Republic.",
"links": null
},
"BIBREF18": {
"ref_id": "b18",
"title": "SPMT: Statistical machine translation with syntactified target language phrases",
"authors": [
{
"first": "D",
"middle": [],
"last": "Marcu",
"suffix": ""
},
{
"first": "W",
"middle": [],
"last": "Wang",
"suffix": ""
},
{
"first": "A",
"middle": [],
"last": "Echihabi",
"suffix": ""
},
{
"first": "K",
"middle": [],
"last": "Knight",
"suffix": ""
}
],
"year": 2006,
"venue": "Proceedings of the 2006 Conference on Empirical Methods in Natural Language Processing",
"volume": "",
"issue": "",
"pages": "44--52",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "D. Marcu, W. Wang, A. Echihabi, K. Knight. 2006. SPMT: Statistical machine translation with syntact- ified target language phrases. In Proceedings of the 2006 Conference on Empirical Methods in Natural Language Processing, 44-52, Sydney, Australia.",
"links": null
},
"BIBREF19": {
"ref_id": "b19",
"title": "A systematic comparison of various statistical alignment models",
"authors": [
{
"first": "F",
"middle": [
"J"
],
"last": "Och",
"suffix": ""
},
{
"first": "H",
"middle": [],
"last": "Ney",
"suffix": ""
}
],
"year": 2003,
"venue": "Computational Linguistics",
"volume": "29",
"issue": "1",
"pages": "19--51",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "F. J. Och, H. Ney. 2003. A systematic comparison of various statistical alignment models. Computational Linguistics, 29(1):19-51.",
"links": null
},
"BIBREF20": {
"ref_id": "b20",
"title": "Minimum error rate training in statistical machine translation",
"authors": [
{
"first": "F",
"middle": [
"J"
],
"last": "Och",
"suffix": ""
}
],
"year": 2003,
"venue": "Proceedings of the 41st Annual Meeting of the Association for Computational Linguistics",
"volume": "",
"issue": "",
"pages": "160--167",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "F. J. Och. 2003. Minimum error rate training in sta- tistical machine translation. In Proceedings of the 41st Annual Meeting of the Association for Compu- tational Linguistics, 160-167, Sapporo, Japan.",
"links": null
},
"BIBREF21": {
"ref_id": "b21",
"title": "BLEU: a method for automatic evaluation of machine translation",
"authors": [
{
"first": "K",
"middle": [],
"last": "Papineni",
"suffix": ""
},
{
"first": "S",
"middle": [],
"last": "Roukos",
"suffix": ""
},
{
"first": "T",
"middle": [],
"last": "Ward",
"suffix": ""
},
{
"first": "W.-J",
"middle": [],
"last": "Zhu",
"suffix": ""
}
],
"year": 2002,
"venue": "Proceedings of the 40th Annual Meeting of the Association for Computational Linguistics",
"volume": "",
"issue": "",
"pages": "311--318",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "K. Papineni, S. Roukos, T. Ward, W.-J. Zhu. 2002. BLEU: a method for automatic evaluation of ma- chine translation. In Proceedings of the 40th Annual Meeting of the Association for Computational Lin- guistics, 311-318, Philadelphia, PA.",
"links": null
},
"BIBREF22": {
"ref_id": "b22",
"title": "An efficient probabilistic contextfree parsing algorithm that computes prefix probabilities",
"authors": [
{
"first": "A",
"middle": [],
"last": "Stolcke",
"suffix": ""
}
],
"year": 1995,
"venue": "Computational Linguistics",
"volume": "21",
"issue": "2",
"pages": "",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "A. Stolcke. 1995. An efficient probabilistic context- free parsing algorithm that computes prefix proba- bilities. Computational Linguistics, 21(2).",
"links": null
},
"BIBREF23": {
"ref_id": "b23",
"title": "Synchronous binarization for machine translation",
"authors": [
{
"first": "H",
"middle": [],
"last": "Zhang",
"suffix": ""
},
{
"first": "L",
"middle": [],
"last": "Huang",
"suffix": ""
},
{
"first": "D",
"middle": [],
"last": "Gildea",
"suffix": ""
},
{
"first": "K",
"middle": [],
"last": "Knight",
"suffix": ""
}
],
"year": 2006,
"venue": "Proceedings of the 2006 Human Language Technology Conference of the North American Chapter of the Association for Computational Linguistics",
"volume": "",
"issue": "",
"pages": "256--263",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "H. Zhang, L. Huang, D. Gildea, K. Knight. 2006. Syn- chronous binarization for machine translation. In Proceedings of the 2006 Human Language Technol- ogy Conference of the North American Chapter of the Association for Computational Linguistics, 256- 263.",
"links": null
},
"BIBREF24": {
"ref_id": "b24",
"title": "Bayesian learning of non-compositional phrases with synchronous parsing",
"authors": [
{
"first": "H",
"middle": [],
"last": "Zhang",
"suffix": ""
},
{
"first": "C",
"middle": [],
"last": "Quirk",
"suffix": ""
},
{
"first": "R",
"middle": [
"C"
],
"last": "Moore",
"suffix": ""
},
{
"first": "D",
"middle": [],
"last": "Gildea",
"suffix": ""
}
],
"year": 2008,
"venue": "Proceedings of ACL-08: HLT, 97-105",
"volume": "",
"issue": "",
"pages": "",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "H. Zhang, C. Quirk, R. C. Moore, D. Gildea. 2008. Bayesian learning of non-compositional phrases with synchronous parsing. In Proceedings of ACL- 08: HLT, 97-105, Columbus, Ohio.",
"links": null
},
"BIBREF25": {
"ref_id": "b25",
"title": "Syntax augmented machine translation via chart parsing",
"authors": [
{
"first": "A",
"middle": [],
"last": "Zollmann",
"suffix": ""
},
{
"first": "A",
"middle": [],
"last": "Venugopal",
"suffix": ""
}
],
"year": 2006,
"venue": "Proceedings of the 2006 Human Language Technology Conference of the North American Chapter of the Association for Computational Linguistics",
"volume": "",
"issue": "",
"pages": "138--141",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "A. Zollmann, A. Venugopal. 2006. Syntax augmented machine translation via chart parsing. In Proceed- ings of the 2006 Human Language Technology Con- ference of the North American Chapter of the As- sociation for Computational Linguistics, 138-141, New York City, NY.",
"links": null
}
},
"ref_entries": {
"FIGREF0": {
"text": "Example derivation. Each node is annotated with their span in the target string (aligned nodes are shaded). The dotted edges show the implied alignments. Preterminals are displayed with their child terminal in the leaf nodes.",
"type_str": "figure",
"num": null,
"uris": null
},
"FIGREF1": {
"text": "Possible state updates for the (NN corner) node using the EXPAND operator.",
"type_str": "figure",
"num": null,
"uris": null
},
"FIGREF4": {
"text": "Histograms over rule statistics comparing the heuristic grammar (GHKM) and learnt grammar (Gibbs).",
"type_str": "figure",
"num": null,
"uris": null
},
"TABREF0": {
"text": "Grammar rules specified by the derivation in Figure 1. Each rule is shown as a tuple comprising a target elementary tree and a source string. Boxed numbers show the alignment between string variables and frontier nonterminals.",
"content": "<table/>",
"html": null,
"type_str": "table",
"num": null
},
"TABREF2": {
"text": "",
"content": "<table><tr><td>:</td><td>NIST Chinese-English corpora statistics</td></tr><tr><td colspan=\"2\">(LDC2003E14, LDC2005E47).</td></tr></table>",
"html": null,
"type_str": "table",
"num": null
}
}
}
} |