File size: 80,869 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 | {
"paper_id": "P96-1041",
"header": {
"generated_with": "S2ORC 1.0.0",
"date_generated": "2023-01-19T09:02:48.527387Z"
},
"title": "An Empirical Study of Smoothing Techniques for Language Modeling",
"authors": [
{
"first": "Stanley",
"middle": [
"F"
],
"last": "Chen",
"suffix": "",
"affiliation": {
"laboratory": "",
"institution": "Harvard University Aiken Computation Laboratory",
"location": {
"addrLine": "33 Oxford St",
"postCode": "02138",
"settlement": "Cambridge, sfc\u00a9eecs, harvard",
"region": "MA, edu"
}
},
"email": ""
},
{
"first": "Joshua",
"middle": [],
"last": "Goodman",
"suffix": "",
"affiliation": {
"laboratory": "Aiken Computation Laboratory",
"institution": "Harvard University",
"location": {
"addrLine": "33 Oxford St. Cambridge",
"postCode": "02138",
"region": "MA"
}
},
"email": ""
}
],
"year": "",
"venue": null,
"identifiers": {},
"abstract": "We present an extensive empirical comparison of several smoothing techniques in the domain of language modeling, including those described by Jelinek and Mercer (1980), Katz (1987), and Church and Gale (1991). We investigate for the first time how factors such as training data size, corpus (e.g., Brown versus Wall Street Journal), and n-gram order (bigram versus trigram) affect the relative performance of these methods, which we measure through the cross-entropy of test data. In addition, we introduce two novel smoothing techniques, one a variation of Jelinek-Mercer smoothing and one a very simple linear interpolation technique, both of which outperform existing methods.",
"pdf_parse": {
"paper_id": "P96-1041",
"_pdf_hash": "",
"abstract": [
{
"text": "We present an extensive empirical comparison of several smoothing techniques in the domain of language modeling, including those described by Jelinek and Mercer (1980), Katz (1987), and Church and Gale (1991). We investigate for the first time how factors such as training data size, corpus (e.g., Brown versus Wall Street Journal), and n-gram order (bigram versus trigram) affect the relative performance of these methods, which we measure through the cross-entropy of test data. In addition, we introduce two novel smoothing techniques, one a variation of Jelinek-Mercer smoothing and one a very simple linear interpolation technique, both of which outperform existing methods.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Abstract",
"sec_num": null
}
],
"body_text": [
{
"text": "Smoothing is a technique essential in the construction of n-gram language models, a staple in speech recognition (Bahl, Jelinek, and Mercer, 1983) as well as many other domains (Church, 1988; Brown et al., 1990 ; Kernighan, Church, and Gale, 1990) . A language model is a probability distribution over strings P(s) that attempts to reflect the frequency with which each string s occurs as a sentence in natural text. Language models are used in speech recognition to resolve acoustically ambiguous utterances.",
"cite_spans": [
{
"start": 113,
"end": 146,
"text": "(Bahl, Jelinek, and Mercer, 1983)",
"ref_id": "BIBREF0"
},
{
"start": 177,
"end": 191,
"text": "(Church, 1988;",
"ref_id": "BIBREF4"
},
{
"start": 192,
"end": 210,
"text": "Brown et al., 1990",
"ref_id": "BIBREF1"
},
{
"start": 213,
"end": 247,
"text": "Kernighan, Church, and Gale, 1990)",
"ref_id": "BIBREF16"
}
],
"ref_spans": [],
"eq_spans": [],
"section": "Introduction",
"sec_num": null
},
{
"text": "For example, if we have that P(it takes two) >> P(it takes too), then we know ceteris paribus to prefer the former transcription over the latter.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Introduction",
"sec_num": null
},
{
"text": "While smoothing is a central issue in language modeling, the literature lacks a definitive comparison between the many existing techniques. Previous studies (Nadas, 1984; Katz, 1987; Church and Gale, 1991; MacKay and Peto, 1995) only compare a small number of methods (typically two) on a single corpus and using a single training data size. As a result, it is currently difficult for a researcher to intelligently choose between smoothing schemes.",
"cite_spans": [
{
"start": 157,
"end": 170,
"text": "(Nadas, 1984;",
"ref_id": "BIBREF20"
},
{
"start": 171,
"end": 182,
"text": "Katz, 1987;",
"ref_id": "BIBREF15"
},
{
"start": 183,
"end": 205,
"text": "Church and Gale, 1991;",
"ref_id": "BIBREF5"
},
{
"start": 206,
"end": 228,
"text": "MacKay and Peto, 1995)",
"ref_id": "BIBREF18"
}
],
"ref_spans": [],
"eq_spans": [],
"section": "Introduction",
"sec_num": null
},
{
"text": "In this work, we carry out an extensive empirical comparison of the most widely used smoothing techniques, including those described by 3elinek and Mercer (1980) , Katz (1987) , and Church and Gale (1991) . We carry out experiments over many training data sizes on varied corpora using both bigram and trigram models. We demonstrate that the relative performance of techniques depends greatly on training data size and n-gram order. For example, for bigram models produced from large training sets Church-Gale smoothing has superior performance, while Katz smoothing performs best on bigram models produced from smaller data. For the methods with parameters that can be tuned to improve performance, we perform an automated search for optimal values and show that sub-optimal parameter selection can significantly decrease performance. To our knowledge, this is the first smoothing work that systematically investigates any of these issues.",
"cite_spans": [
{
"start": 136,
"end": 161,
"text": "3elinek and Mercer (1980)",
"ref_id": null
},
{
"start": 164,
"end": 175,
"text": "Katz (1987)",
"ref_id": "BIBREF15"
},
{
"start": 182,
"end": 204,
"text": "Church and Gale (1991)",
"ref_id": "BIBREF5"
}
],
"ref_spans": [],
"eq_spans": [],
"section": "Introduction",
"sec_num": null
},
{
"text": "In addition, we introduce two novel smoothing techniques: the first belonging to the class of smoothing models described by 3elinek and Mercer, the second a very simple linear interpolation method. While being relatively simple to implement, we show that these methods yield good performance in bigram models and superior performance in trigram models.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Introduction",
"sec_num": null
},
{
"text": "We take the performance of a method m to be its cross-entropy on test data",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Introduction",
"sec_num": null
},
{
"text": "1 IT IvT -log Pro(t,) i=1",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Introduction",
"sec_num": null
},
{
"text": "where Pm(ti) denotes the language model produced with method m and where the test data T is composed of sentences (tl,...,tzr) and contains a total of NT words. The entropy is inversely related to the average probability a model assigns to sentences in the test data, and it is generally assumed that lower entropy correlates with better performance in applications.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Introduction",
"sec_num": null
},
{
"text": "In n-gram language modeling, the probability of a string P(s) is expressed as the product of the probabilities of the words that compose the string, with each word probability conditional on the identity of the last n -1 words, i.e., ifs = wl-..wt we have l 1",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Smoothing n-gram Models",
"sec_num": "1.1"
},
{
"text": "P(s) = H P(wi[w{-1) ~ 1-~ P i-1 (1) i=1 i=1",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Smoothing n-gram Models",
"sec_num": "1.1"
},
{
"text": "where w i j denotes the words wi \u2022 \u2022. wj. Typically, n is taken to be two or three, corresponding to a bigram or trigram model, respectively. 1",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Smoothing n-gram Models",
"sec_num": "1.1"
},
{
"text": "Consider the case n = 2. To estimate the probabilities P(wilwi-,) in equation 1, one can acquire a large corpus of text, which we refer to as training data, and take",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Smoothing n-gram Models",
"sec_num": "1.1"
},
{
"text": "P(Wi-lWi) PML(Wil i-1) -- P(wi-1) c(wi-lWi)/Ns e(wi-1)/Ns c(wi_ w )",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Smoothing n-gram Models",
"sec_num": "1.1"
},
{
"text": "where c(c 0 denotes the number of times the string c~ occurs in the text and Ns denotes the total number of words. This is called the maximum likelihood (ML) estimate for P(wilwi_l).",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Smoothing n-gram Models",
"sec_num": "1.1"
},
{
"text": "While intuitive, the maximum likelihood estimate is a poor one when the amount of training data is small compared to the size of the model being built, as is generally the case in language modeling. For example, consider the situation where a pair of words, or bigram, say burnish the, doesn't occur in the training data. Then, we have PML(the Iburnish) = O, which is clearly inaccurate as this probability should be larger than zero. A zero bigram probability can lead to errors in speech recognition, as it disallows the bigram regardless of how informative the acoustic signal is. The term smoothing describes techniques for adjusting the maximum likelihood estimate to hopefully produce more accurate probabilities.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Smoothing n-gram Models",
"sec_num": "1.1"
},
{
"text": "As an example, one simple smoothing technique is to pretend each bigram occurs once more than it actually did (Lidstone, 1920; Johnson, 1932; Jeffreys, 1948) , yielding",
"cite_spans": [
{
"start": 110,
"end": 126,
"text": "(Lidstone, 1920;",
"ref_id": "BIBREF17"
},
{
"start": 127,
"end": 141,
"text": "Johnson, 1932;",
"ref_id": "BIBREF14"
},
{
"start": 142,
"end": 157,
"text": "Jeffreys, 1948)",
"ref_id": "BIBREF12"
}
],
"ref_spans": [],
"eq_spans": [],
"section": "Smoothing n-gram Models",
"sec_num": "1.1"
},
{
"text": "C(Wi-lWi) \"[-1 = + IVl",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Smoothing n-gram Models",
"sec_num": "1.1"
},
{
"text": "where V is the vocabulary, the set of all words being considered. This has the desirable quality of 1To make the term P(wdw [Z~,,+~) meaningful for i < n, one can pad the beginning of the string with a distinguished token. In this work, we assume there are n -1 such distinguished tokens preceding each sentence. preventing zero bigram probabilities. However, this scheme has the flaw of assigning the same probability to say, burnish the and burnish thou (assuming neither occurred in the training data), even though intuitively the former seems more likely because the word the is much more common than thou.",
"cite_spans": [
{
"start": 124,
"end": 132,
"text": "[Z~,,+~)",
"ref_id": null
}
],
"ref_spans": [],
"eq_spans": [],
"section": "Smoothing n-gram Models",
"sec_num": "1.1"
},
{
"text": "To address this, another smoothing technique is to interpolate the bigram model with a unigram model PML(Wi) = c(wi)/Ns, a model that reflects how often each word occurs in the training data. For example, we can take Pinto p( i J i-1) = APM (w pW _l) + (1getting the behavior that bigrams involving common words are assigned higher probabilities (Jelinek and Mercer, 1980) .",
"cite_spans": [
{
"start": 346,
"end": 372,
"text": "(Jelinek and Mercer, 1980)",
"ref_id": "BIBREF13"
}
],
"ref_spans": [],
"eq_spans": [],
"section": "Smoothing n-gram Models",
"sec_num": "1.1"
},
{
"text": "The simplest type of smoothing used in practice is additive smoothing (Lidstone, 1920; Johnson, 1932; aeffreys, 1948) , where we take i w i-1 e(wi_,,+l) + = + elVl",
"cite_spans": [
{
"start": 70,
"end": 86,
"text": "(Lidstone, 1920;",
"ref_id": "BIBREF17"
},
{
"start": 87,
"end": 101,
"text": "Johnson, 1932;",
"ref_id": "BIBREF14"
},
{
"start": 102,
"end": 117,
"text": "aeffreys, 1948)",
"ref_id": null
}
],
"ref_spans": [],
"eq_spans": [],
"section": "Previous Work",
"sec_num": "2"
},
{
"text": "( 2)and where Lidstone and Jeffreys advocate /i = 1. Gale and Church (1990; 1994) have argued that this method generally performs poorly. The Good-Turing estimate (Good, 1953) is central to many smoothing techniques. It is not used directly for n-gram smoothing because, like additive smoothing, it does not perform the interpolation of lower-and higher-order models essential for good performance. Good-Turing states that an n-gram that occurs r times should be treated as if it had occurred r* times, where r* = (r + 1)n~+l and where n~ is the number of n-grams that. occur exactly r times in the training data. Katz smoothing (1987) extends the intuitions of Good-Turing by adding the interpolation of higherorder models with lower-order models. It is perhaps the most widely used smoothing technique in speech recognition.",
"cite_spans": [
{
"start": 53,
"end": 75,
"text": "Gale and Church (1990;",
"ref_id": "BIBREF7"
},
{
"start": 76,
"end": 81,
"text": "1994)",
"ref_id": "BIBREF19"
},
{
"start": 163,
"end": 175,
"text": "(Good, 1953)",
"ref_id": "BIBREF11"
},
{
"start": 614,
"end": 635,
"text": "Katz smoothing (1987)",
"ref_id": null
}
],
"ref_spans": [],
"eq_spans": [],
"section": "Previous Work",
"sec_num": "2"
},
{
"text": "Church and Gale (1991) describe a smoothing method that combines the Good-Turing estimate with bucketing, the technique of partitioning a set, of n-grams into disjoint groups, where each group is characterized independently through a set of parameters. Like Katz, models are defined recursively in terms of lower-order models. Each n-gram is assigned to one of several buckets based on its frequency predicted from lower-order models. Each bucket is treated as a separate distribution and Good-Turing estimation is performed within each, giving corrected counts that are normalized to yield probabilities. ;.\u00b0*~a-:..",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Previous Work",
"sec_num": "2"
},
{
"text": "\u2022 % at ...,~;e.T\u00a2: \u00b0 . .. : The other smoothing technique besides Katz smoothing widely used in speech recognition is due to Jelinek and Mercer (1980) . They present a class of smoothing models that involve linear interpolation, e.g., Brown et al. (1992) ",
"cite_spans": [
{
"start": 125,
"end": 150,
"text": "Jelinek and Mercer (1980)",
"ref_id": "BIBREF13"
},
{
"start": 235,
"end": 254,
"text": "Brown et al. (1992)",
"ref_id": "BIBREF2"
}
],
"ref_spans": [],
"eq_spans": [],
"section": "\u2022 . \u00b0",
"sec_num": null
},
{
"text": "EQUATION",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [
{
"start": 0,
"end": 8,
"text": "EQUATION",
"ref_id": "EQREF",
"raw_str": "\u00b0 \u00b0 % o% **\u00b0 ~ - \u00b0 ~\u00b0~ \u00b0 o o \u00b0 \u00b0\u2022 \u00b0~ \u00b0 o* \u00b0 o o o , ,",
"eq_num": ",i , , ,i , , ,i , \" . ."
}
],
"section": "\u2022 . \u00b0",
"sec_num": null
},
{
"text": "take i--1 PML(Wi IWi-n+l) \"Iv ~Wi__ 1 i--1 i--n-]-I P~ /W i-1 , (1--)~to~-~ ) inte~pt i wi_n+2) (3) i--u-I-1",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "\u2022 . \u00b0",
"sec_num": null
},
{
"text": "That is, the maximum likelihood estimate is interpolated with the smoothed lower-order distribution, which is defined analogously. Training a distinct I ~-1 for each wi_,~+li-1 is not generally felicitous;",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "\u2022 . \u00b0",
"sec_num": null
},
{
"text": "Wi--n-{-1 Bahl, Jelinek, and Mercer (1983) ",
"cite_spans": [
{
"start": 10,
"end": 42,
"text": "Bahl, Jelinek, and Mercer (1983)",
"ref_id": "BIBREF0"
}
],
"ref_spans": [],
"eq_spans": [],
"section": "\u2022 . \u00b0",
"sec_num": null
},
{
"text": "suggest partition- i-1",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "\u2022 . \u00b0",
"sec_num": null
},
{
"text": "ing the 1~,~-~ into buckets according to c(wi_~+l),",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "\u2022 . \u00b0",
"sec_num": null
},
{
"text": "i--n-l-1",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "\u2022 . \u00b0",
"sec_num": null
},
{
"text": "where all )~w~-~ in the same bucket are constrained",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "\u2022 . \u00b0",
"sec_num": null
},
{
"text": "i--n-l-1",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "\u2022 . \u00b0",
"sec_num": null
},
{
"text": "to have the same value. To yield meaningful results, the data used to estimate the A~!-, need to be disjoint from the data Several smoothing techniques are motivated within a Bayesian framework, including work by Nadas (1984) and MacKay and Peto (1995) .",
"cite_spans": [
{
"start": 213,
"end": 225,
"text": "Nadas (1984)",
"ref_id": "BIBREF20"
},
{
"start": 230,
"end": 252,
"text": "MacKay and Peto (1995)",
"ref_id": "BIBREF18"
}
],
"ref_spans": [],
"eq_spans": [],
"section": "\u2022 . \u00b0",
"sec_num": null
},
{
"text": "Of the great many novel methods that we have tried, two techniques have performed especially well.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Novel Smoothing Techniques",
"sec_num": "3"
},
{
"text": "2When the same data is used to estimate both, setting all )~ ~-~ to one yields the optimal result. While larger i-1 c(wi_n+l) generally correspond to less sparse distributions, this quantity ignores the allocation of counts between words. For example, we would consider a distribution with ten counts distributed evenly among ten words to be much more sparse than a distribution with ten counts all on a single word. The average number of counts per word seems to more directly express the concept of sparseness, In Figure 1 , we graph the value of ~ assigned to each bucket under the original and new bucketing schemes on identical data. Notice that the new bucketing scheme results in a much tighter plot, indicating that it is better at grouping together distributions with similar behavior.",
"cite_spans": [],
"ref_spans": [
{
"start": 516,
"end": 524,
"text": "Figure 1",
"ref_id": "FIGREF1"
}
],
"eq_spans": [],
"section": "Novel Smoothing Techniques",
"sec_num": "3"
},
{
"text": "This technique combines two intuitions.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Method one-count",
"sec_num": "3.2"
},
{
"text": "First, MacKay and Peto (1995) argue that a reasonable form for a smoothed distribution is",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Method one-count",
"sec_num": "3.2"
},
{
"text": "\u2022 i-1 Pone(W i i-1 c(wL, +l) + Po,,e(wilw _ +9 IWi--nq-1) = i--1 c(wi_n+l) +",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Method one-count",
"sec_num": "3.2"
},
{
"text": "The parameter a can be thought of as the number of counts being added to the given distribution, where the new counts are distributed as in the lowerorder distribution. Secondly, the Good-Turing estimate can be interpreted as stating that the number of these extra counts should be proportional to the number of words with exactly one count in the given distribution. We have found that taking",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Method one-count",
"sec_num": "3.2"
},
{
"text": "EQUATION",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [
{
"start": 0,
"end": 8,
"text": "EQUATION",
"ref_id": "EQREF",
"raw_str": "i-1 O~ = \"y [nl(Wi_n+l) -~-~]",
"eq_num": "(4)"
}
],
"section": "Method one-count",
"sec_num": "3.2"
},
{
"text": "works well, where",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Method one-count",
"sec_num": "3.2"
},
{
"text": "i-i i",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Method one-count",
"sec_num": "3.2"
},
{
"text": "is the number of words with one count, and where/3 and 7 are constants.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Method one-count",
"sec_num": "3.2"
},
{
"text": "4.1",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Experimental Methodology",
"sec_num": "4"
},
{
"text": "We used the Penn treebauk and TIPSTER corpora distributed by the Linguistic Data Consortium. From the treebank, we extracted text from the tagged Brown corpus, yielding about one million words. From TIPSTER, we used the Associated Press (AP), Wall Street Journal (WSJ), and San Jose Mercury News (SJM) data, yielding 123, 84, and 43 million words respectively. We created two distinct vocabularies, one for the Brown corpus and one for the TIPSTER data. The former vocabulary contains all 53,850 words occurring in Brown; the latter vocabulary consists of the 65,173 words occurring at least 70 times in TIPSTER. For each experiment, we selected three segments of held-out data along with the segment of training data. One held-out segment was used as the test data for performance evaluation, and the other two were used as development test data for optimizing the parameters of each smoothing method. Each piece of held-out data was chosen to be roughly 50,000 words. This decision does not reflect practice very well, as when the training data size is less than 50,000 words it is not realistic to have so much development test data available. However, we made this decision to prevent us having to optimize the training versus held-out data tradeoff for each data size. In addition, the development test data is used to optimize typically very few parameters, so in practice small held-out sets are generally adequate, and perhaps can be avoided altogether with techniques such as deleted estimation.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Data",
"sec_num": null
},
{
"text": "In this section, we discuss the details of our implementations of various smoothing techniques. Due to space limitations, these descriptions are not comprehensive; a more complete discussion is presented in Chen (1996) . The titles of the following sections include the mnemonic we use to refer to the implementations in later sections. Unless otherwise specified, for those smoothing models defined recursively in terms of lower-order models, we end the recursion by taking the n = 0 distribution to be the uniform distribution Punif(wi) = l/IV[. For each method, we highlight the parameters (e.g., Am and 5 below) that can be tuned to optimize performance. Parameter values are determined through training on held-out data.",
"cite_spans": [
{
"start": 207,
"end": 218,
"text": "Chen (1996)",
"ref_id": "BIBREF3"
}
],
"ref_spans": [],
"eq_spans": [],
"section": "Smoothing Implementations",
"sec_num": "4.2"
},
{
"text": "Baseline Smoothing (interp-baseline) For our baseline smoothing method, we use an instance of Jelinek-Mercer smoothing where we constrain all A,~!-I to be equal to a single value A,~ for ,-n-hi each n, i.e., We consider two versions of additive smoothing. Referring to equation 2, we fix 5 = 1 in plus-one smoothing. In plus-delta, we consider any 6.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "4.2.1",
"sec_num": null
},
{
"text": "Katz Smoothing (katz) While the original paper (Katz, 1987) uses a single parameter k, we instead use a different k for each n > 1, k,~. We smooth the unigram distribution using additive smoothing with parameter 5.",
"cite_spans": [
{
"start": 47,
"end": 59,
"text": "(Katz, 1987)",
"ref_id": "BIBREF15"
}
],
"ref_spans": [],
"eq_spans": [],
"section": "4.2.3",
"sec_num": null
},
{
"text": "(church-gale)",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Church-Gale Smoothing",
"sec_num": "4.2.4"
},
{
"text": "To smooth the counts n~ needed for the Good-Turing estimate, we use the technique described by Gale and Sampson (1995) . We smooth the unigram distribution using Good-tiering without any bucketing.",
"cite_spans": [
{
"start": 95,
"end": 118,
"text": "Gale and Sampson (1995)",
"ref_id": "BIBREF10"
}
],
"ref_spans": [],
"eq_spans": [],
"section": "Church-Gale Smoothing",
"sec_num": "4.2.4"
},
{
"text": "Instead of the bucketing scheme described in the original paper, we use a scheme analogous to the one described by Bahl, Jelinek, and Mercer (1983) . We make the assumption that whether a bucket is large enough for accurate Good-Turing estimation depends on how many n-grams with non-zero counts occur in it. Thus, instead of partitioning the space of P(wi-JP(wi) values in some uniform way as was done by Church and Gale, we partition the space so that at least Cmi n non-zero n-grams fall in each bucket.",
"cite_spans": [
{
"start": 115,
"end": 147,
"text": "Bahl, Jelinek, and Mercer (1983)",
"ref_id": "BIBREF0"
}
],
"ref_spans": [],
"eq_spans": [],
"section": "Church-Gale Smoothing",
"sec_num": "4.2.4"
},
{
"text": "Finally, the original paper describes only bigram smoothing in detail; extending this method to trigram smoothing is ambiguous. In particular, it is unclear whether to bucket trigrams according to i-1 i--1 P(wi_JP . We chose the former; while the latter may yield better performance, our belief is that it is much more difficult to implement and that it requires a great deal more computation.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Church-Gale Smoothing",
"sec_num": "4.2.4"
},
{
"text": "(interp-held-out and interp-del-int)",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Jelinek-Mercer Smoothing",
"sec_num": "4.2.5"
},
{
"text": "We implemented two versions of Jelinek-Mercer smoothing differing only in what data is used to train the A's. We bucket the A ~-1 according to Wi--n-bl i-1 C(Wi_~+I) as suggested by Bahl et al. Similar to our Church-Gale implementation, we choose buckets to ensure that at least Cmi n words in the data used to train the A's fall in each bucket.",
"cite_spans": [
{
"start": 143,
"end": 151,
"text": "Wi--n-bl",
"ref_id": null
}
],
"ref_spans": [],
"eq_spans": [],
"section": "Jelinek-Mercer Smoothing",
"sec_num": "4.2.5"
},
{
"text": "In interp-held-out, the A's are trained using held-out interpolation on one of the development test sets. In interp-del-int, the A's are trained using the relaxed deleted interpolation technique described by Jelinek and Mercer, where one word is deleted at a time. In interp-del-int, we bucket an n-gram according to its count before deletion, as this turned out to significantly improve performance.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Jelinek-Mercer Smoothing",
"sec_num": "4.2.5"
},
{
"text": "(new-avg-count and new-one-count)",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Novel Smoothing Methods",
"sec_num": "4.2.6"
},
{
"text": "The implementation new-avg-count, corresponding to smoothing method average-count, is identical to interp-held-out except that we use the novel bucketing scheme described in section 3.1. In the implementation new-one-count, we have different parameters j3~ and 7~ in equation 4for each n.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Novel Smoothing Methods",
"sec_num": "4.2.6"
},
{
"text": "In Figure 2 , we display the performance of the interp-baseline method for bigram and trigram models on TIPSTER, Brown, and the WSJ subset of TIPSTER. In Figures 3-6 , we display the relative performance of various smoothing techniques with respect to the baseline method on these corpora, as measured by difference in entropy. In the graphs on the left of Figures 2-4 , each point represents an average over ten runs; the error bars represent the empirical standard deviation over these runs. Due to resource limitations, we only performed multiple runs for data sets of 50,000 sentences or less. Each point on the graphs on the right represents a single run, but we consider sizes up to the amount of data available. The graphs on the bottom of Figures 3-4 are close-ups of the graphs above, focusing on those algorithms that perform better than the baseline. To give an idea of how these cross-entropy differences translate to perplexity, each 0.014 bits correspond roughly to a 1% change in perplexity.",
"cite_spans": [],
"ref_spans": [
{
"start": 3,
"end": 11,
"text": "Figure 2",
"ref_id": "FIGREF4"
},
{
"start": 154,
"end": 165,
"text": "Figures 3-6",
"ref_id": null
},
{
"start": 357,
"end": 368,
"text": "Figures 2-4",
"ref_id": "FIGREF4"
},
{
"start": 747,
"end": 754,
"text": "Figures",
"ref_id": null
}
],
"eq_spans": [],
"section": "Results",
"sec_num": "5"
},
{
"text": "In each run except as noted below, optimal values for the parameters of the given technique were searched for using Powell's search algorithm as realized in Numerical Recipes in C (Press et al., 1988, pp. 309-317) . Parameters were chosen to optimize the cross-entropy of one of the development test sets associated with the given training set. To constrain the search, we searched only those parameters that were found to affect performance significantly, as verified through preliminary experiments over several data sizes. For katz and church-gale, we did not perform the parameter search for training sets over 50,000 sentences due to resource constraints, and instead manually extrapolated parameter val- ues from optimal values found on smaller data sizes. We ran interp-del-int only on sizes up to 50,000 sentences due to time constraints.",
"cite_spans": [
{
"start": 180,
"end": 213,
"text": "(Press et al., 1988, pp. 309-317)",
"ref_id": null
}
],
"ref_spans": [],
"eq_spans": [],
"section": "Results",
"sec_num": "5"
},
{
"text": "From these graphs, we see that additive smoothing performs poorly and that methods katz and interp-held-out consistently perform well. Our implementation church-gale performs poorly except on large bigram training sets, where it performs the best. The novel methods new-avg-count and new-one-count perform well uniformly across training data sizes, and are superior for trigram models. Notice that while performance is relatively consistent across corpora, it varies widely with respect to training set size and n-gram order.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Results",
"sec_num": "5"
},
{
"text": "The method interp-del-int performs significantly worse than interp-held-out, though they differ only in the data used to train the A's. However, we delete one word at a time in interp-del-int; we hypothesize that deleting larger chunks would lead to more similar performance.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Results",
"sec_num": "5"
},
{
"text": "In Figure 7 , we show how the values of the parameters 6 and Cmin affect the performance of methods katz and new-avg-count, respectively, over several training data sizes. Notice that poor parameter setting can lead to very significant losses in performance, and that optimal parameter settings depend on training set size.",
"cite_spans": [],
"ref_spans": [
{
"start": 3,
"end": 11,
"text": "Figure 7",
"ref_id": null
}
],
"eq_spans": [],
"section": "Results",
"sec_num": "5"
},
{
"text": "To give an informal estimate of the difficulty of implementation of each method, in Table 1 we display the number of lines of C++ code in each implementation excluding the core code common across techniques.",
"cite_spans": [],
"ref_spans": [
{
"start": 84,
"end": 91,
"text": "Table 1",
"ref_id": null
}
],
"eq_spans": [],
"section": "Results",
"sec_num": "5"
},
{
"text": "To our knowledge, this is the first empirical comparison of smoothing techniques in language modeling of such scope: no other study has used multiple training data sizes, corpora, or has performed parameter optimization. We show that in order to completely 3To implement the baseline method, we just used the interp-held-out code as it is a special case. Written anew, it probably would have been about 50 lines. .... ..~...y~ \"'\"'~\" -..+.,..,,., .,-\" .... o.-..-~'\"\"~.......~ '-. . . ",
"cite_spans": [
{
"start": 413,
"end": 480,
"text": ".... ..~...y~ \"'\"'~\" -..+.,..,,., .,-\" .... o.-..-~'\"\"~.......~ '-.",
"ref_id": null
}
],
"ref_spans": [],
"eq_spans": [],
"section": "Discussion",
"sec_num": "6"
}
],
"back_matter": [
{
"text": "The authors would like to thank Stuart Shieber and the anonymous reviewers for their comments on previous versions of this paper. We would also like to thank William Gale and Geoffrey Sampson for supplying us with code for \"Good-Turing frequency estimation without tears.\" This research was supported by the National Science Foundation under Grant No. IRI-93-50192 and Grant No. CDA-94-01024. The second author was also supported by a National Science Foundation Graduate Student Fellowship.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Acknowledgements",
"sec_num": null
},
{
"text": "In this study, we measure performance solely through the cross-entropy of test data; it would be interesting to see how these cross-entropy differences correlate with performance in end applications such as speech recognition. In addition, it would be interesting to see whether these results extend to fields other than language modeling where smoothing is used, such as prepositional phrase attachment (Collins and Brooks, 1995) , part-of-speech tagging (Church, 1988) , and stochastic parsing (Magerman, 1994) .",
"cite_spans": [
{
"start": 404,
"end": 430,
"text": "(Collins and Brooks, 1995)",
"ref_id": "BIBREF6"
},
{
"start": 456,
"end": 470,
"text": "(Church, 1988)",
"ref_id": "BIBREF4"
},
{
"start": 496,
"end": 512,
"text": "(Magerman, 1994)",
"ref_id": "BIBREF19"
}
],
"ref_spans": [],
"eq_spans": [],
"section": "annex",
"sec_num": null
}
],
"bib_entries": {
"BIBREF0": {
"ref_id": "b0",
"title": "A maximum likelihood approach to continuous speech recognition",
"authors": [
{
"first": "Lalit",
"middle": [
"R"
],
"last": "Bahl",
"suffix": ""
},
{
"first": "Frederick",
"middle": [],
"last": "Jelinek",
"suffix": ""
},
{
"first": "Robert",
"middle": [
"L"
],
"last": "Mercer",
"suffix": ""
}
],
"year": 1983,
"venue": "IEEE Transactions on Pattern Analysis and Machine Intelligence",
"volume": "5",
"issue": "2",
"pages": "179--190",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Bahl, Lalit R., Frederick Jelinek, and Robert L. Mercer. 1983. A maximum likelihood approach to continuous speech recognition. IEEE Trans- actions on Pattern Analysis and Machine Intelli- gence, PAMI-5(2):179-190, March.",
"links": null
},
"BIBREF1": {
"ref_id": "b1",
"title": "A statistical approach to machine translation",
"authors": [
{
"first": "Peter",
"middle": [
"F"
],
"last": "Brown",
"suffix": ""
},
{
"first": "John",
"middle": [],
"last": "Cocke",
"suffix": ""
},
{
"first": "Stephen",
"middle": [
"A"
],
"last": "Dellapietra",
"suffix": ""
},
{
"first": "Vincent",
"middle": [
"J"
],
"last": "Dellapietra",
"suffix": ""
},
{
"first": "Frederick",
"middle": [],
"last": "Jelinek",
"suffix": ""
},
{
"first": "John",
"middle": [
"D"
],
"last": "Lafferty",
"suffix": ""
},
{
"first": "Robert",
"middle": [
"L"
],
"last": "Mercer",
"suffix": ""
},
{
"first": "Paul",
"middle": [
"S"
],
"last": "Roossin",
"suffix": ""
}
],
"year": 1990,
"venue": "Computational Linguistics",
"volume": "16",
"issue": "2",
"pages": "79--85",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Brown, Peter F., John Cocke, Stephen A. DellaPi- etra, Vincent J. DellaPietra, Frederick Jelinek, John D. Lafferty, Robert L. Mercer, and Paul S. Roossin. 1990. A statistical approach to machine translation. Computational Linguistics, 16(2):79- 85, June.",
"links": null
},
"BIBREF2": {
"ref_id": "b2",
"title": "An estimate of an upper bound for the entropy of English",
"authors": [
{
"first": "Peter",
"middle": [
"F"
],
"last": "Brown",
"suffix": ""
},
{
"first": "A",
"middle": [],
"last": "Stephen",
"suffix": ""
},
{
"first": "Vincent",
"middle": [
"J"
],
"last": "Dellapietra",
"suffix": ""
},
{
"first": "Jennifer",
"middle": [
"C"
],
"last": "Dellapietra",
"suffix": ""
},
{
"first": "Robert",
"middle": [
"L"
],
"last": "Lai",
"suffix": ""
},
{
"first": "",
"middle": [],
"last": "Mercer",
"suffix": ""
}
],
"year": 1992,
"venue": "Computational Linguistics",
"volume": "18",
"issue": "1",
"pages": "31--40",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Brown, Peter F., Stephen A. DellaPietra, Vincent J. DellaPietra, Jennifer C. Lai, and Robert L. Mer- cer. 1992. An estimate of an upper bound for the entropy of English. Computational Linguis- tics, 18(1):31-40, March.",
"links": null
},
"BIBREF3": {
"ref_id": "b3",
"title": "Building Probabilistic Models for Natural Language",
"authors": [
{
"first": "Stanley",
"middle": [
"F"
],
"last": "Chen",
"suffix": ""
}
],
"year": 1996,
"venue": "",
"volume": "",
"issue": "",
"pages": "",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Chen, Stanley F. 1996. Building Probabilistic Mod- els for Natural Language. Ph.D. thesis, Harvard University. In preparation.",
"links": null
},
"BIBREF4": {
"ref_id": "b4",
"title": "A stochastic parts program and noun phrase parser for unrestricted text",
"authors": [
{
"first": "Kenneth",
"middle": [],
"last": "Church",
"suffix": ""
}
],
"year": 1988,
"venue": "Proceedings of the Second Conference on Applied Natural Language Processing",
"volume": "",
"issue": "",
"pages": "136--143",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Church, Kenneth. 1988. A stochastic parts program and noun phrase parser for unrestricted text. In Proceedings of the Second Conference on Applied Natural Language Processing, pages 136-143.",
"links": null
},
"BIBREF5": {
"ref_id": "b5",
"title": "A comparison of the enhanced Good-Turing and deleted estimation methods for estimating probabilities of English bigrams",
"authors": [
{
"first": "Kenneth",
"middle": [
"W"
],
"last": "Church",
"suffix": ""
},
{
"first": "William",
"middle": [
"A"
],
"last": "Gale",
"suffix": ""
}
],
"year": 1991,
"venue": "Computer Speech and Language",
"volume": "5",
"issue": "",
"pages": "19--54",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Church, Kenneth W. and William A. Gale. 1991. A comparison of the enhanced Good-Turing and deleted estimation methods for estimating proba- bilities of English bigrams. Computer Speech and Language, 5:19-54.",
"links": null
},
"BIBREF6": {
"ref_id": "b6",
"title": "Prepositional phrase attachment through a backed-off model",
"authors": [
{
"first": "Michael",
"middle": [],
"last": "Collins",
"suffix": ""
},
{
"first": "James",
"middle": [],
"last": "Brooks",
"suffix": ""
}
],
"year": 1995,
"venue": "Proceedings of the Third Workshop on Very Large Corpora",
"volume": "",
"issue": "",
"pages": "27--38",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Collins, Michael and James Brooks. 1995. Prepo- sitional phrase attachment through a backed-off model. In David Yarowsky and Kenneth Church, editors, Proceedings of the Third Workshop on Very Large Corpora, pages 27-38, Cambridge, MA, June.",
"links": null
},
"BIBREF7": {
"ref_id": "b7",
"title": "Estimation procedures for language context: poor estimates are worse than none",
"authors": [
{
"first": "William",
"middle": [
"A"
],
"last": "Gale",
"suffix": ""
},
{
"first": "Kenneth",
"middle": [
"W"
],
"last": "Church",
"suffix": ""
}
],
"year": 1990,
"venue": "COMP",
"volume": "",
"issue": "",
"pages": "",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Gale, William A. and Kenneth W. Church. 1990. Estimation procedures for language context: poor estimates are worse than none. In COMP-",
"links": null
},
"BIBREF8": {
"ref_id": "b8",
"title": "Proceedings in Computational Statistics, 9th Symposium",
"authors": [],
"year": null,
"venue": "",
"volume": "",
"issue": "",
"pages": "69--74",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "STAT, Proceedings in Computational Statistics, 9th Symposium, pages 69-74, Dubrovnik, Yu- goslavia, September.",
"links": null
},
"BIBREF9": {
"ref_id": "b9",
"title": "What's wrong with adding one",
"authors": [
{
"first": "William",
"middle": [
"A"
],
"last": "Gale",
"suffix": ""
},
{
"first": "Kenneth",
"middle": [
"W"
],
"last": "Church",
"suffix": ""
}
],
"year": 1994,
"venue": "Corpus-Based Research into Language",
"volume": "",
"issue": "",
"pages": "",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Gale, William A. and Kenneth W. Church. 1994. What's wrong with adding one? In N. Oostdijk and P. de Haan, editors, Corpus-Based Research into Language. Rodolpi, Amsterdam.",
"links": null
},
"BIBREF10": {
"ref_id": "b10",
"title": "Good-Turing frequency estimation without tears",
"authors": [
{
"first": "William",
"middle": [
"A"
],
"last": "Gale",
"suffix": ""
},
{
"first": "Geoffrey",
"middle": [],
"last": "Sampson",
"suffix": ""
}
],
"year": 1995,
"venue": "Journal of Quantitative Linguistics",
"volume": "2",
"issue": "3",
"pages": "",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Gale, William A. and Geoffrey Sampson. 1995. Good-Turing frequency estimation without tears. Journal of Quantitative Linguistics, 2(3). To ap- pear.",
"links": null
},
"BIBREF11": {
"ref_id": "b11",
"title": "The population frequencies of species and the estimation of population parameters",
"authors": [
{
"first": "I",
"middle": [
"J"
],
"last": "Good",
"suffix": ""
}
],
"year": 1953,
"venue": "Biometrika",
"volume": "40",
"issue": "",
"pages": "237--264",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Good, I.J. 1953. The population frequencies of species and the estimation of population parame- ters. Biometrika, 40(3 and 4):237-264.",
"links": null
},
"BIBREF12": {
"ref_id": "b12",
"title": "Theory of Probability",
"authors": [
{
"first": "H",
"middle": [],
"last": "Jeffreys",
"suffix": ""
}
],
"year": 1948,
"venue": "",
"volume": "",
"issue": "",
"pages": "",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Jeffreys, H. 1948. Theory of Probability. Clarendon Press, Oxford, second edition.",
"links": null
},
"BIBREF13": {
"ref_id": "b13",
"title": "Interpolated estimation of Markov source parameters from sparse data",
"authors": [
{
"first": "Frederick",
"middle": [],
"last": "Jelinek",
"suffix": ""
},
{
"first": "Robert",
"middle": [
"L"
],
"last": "Mercer",
"suffix": ""
}
],
"year": 1980,
"venue": "Proceedings of the Workshop on Pattern Recognition in Practice, Amsterdam",
"volume": "",
"issue": "",
"pages": "",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Jelinek, Frederick and Robert L. Mercer. 1980. In- terpolated estimation of Markov source parame- ters from sparse data. In Proceedings of the Work- shop on Pattern Recognition in Practice, Amster- dam, The Netherlands: North-Holland, May.",
"links": null
},
"BIBREF14": {
"ref_id": "b14",
"title": "Probability: deductive and inductive problems. Mind",
"authors": [
{
"first": "W",
"middle": [
"E"
],
"last": "Johnson",
"suffix": ""
}
],
"year": 1932,
"venue": "",
"volume": "41",
"issue": "",
"pages": "421--423",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Johnson, W.E. 1932. Probability: deductive and inductive problems. Mind, 41:421-423.",
"links": null
},
"BIBREF15": {
"ref_id": "b15",
"title": "Estimation of probabilities from sparse data for the language model component of a speech recognizer",
"authors": [
{
"first": "Slava",
"middle": [
"M"
],
"last": "Katz",
"suffix": ""
}
],
"year": 1987,
"venue": "IEEE Transactions on Acoustics, Speech and Signal Processing",
"volume": "35",
"issue": "3",
"pages": "400--401",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Katz, Slava M. 1987. Estimation of probabilities from sparse data for the language model com- ponent of a speech recognizer. IEEE Transac- tions on Acoustics, Speech and Signal Processing, ASSP-35(3):400-401, March.",
"links": null
},
"BIBREF16": {
"ref_id": "b16",
"title": "A spelling correction program based on a noisy channel model",
"authors": [
{
"first": "M",
"middle": [
"D"
],
"last": "Kernighan",
"suffix": ""
},
{
"first": "K",
"middle": [
"W"
],
"last": "Church",
"suffix": ""
},
{
"first": "W",
"middle": [
"A"
],
"last": "Gale",
"suffix": ""
}
],
"year": 1990,
"venue": "Proceedings of the Thirteenth International Conference on Computational Linguistics",
"volume": "",
"issue": "",
"pages": "205--210",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Kernighan, M.D., K.W. Church, and W.A. Gale. 1990. A spelling correction program based on a noisy channel model. In Proceedings of the Thirteenth International Conference on Compu- tational Linguistics, pages 205-210.",
"links": null
},
"BIBREF17": {
"ref_id": "b17",
"title": "Note on the general case of the Bayes-Laplace formula for inductive or a posteriori probabilities",
"authors": [
{
"first": "G",
"middle": [
"J"
],
"last": "Lidstone",
"suffix": ""
}
],
"year": 1920,
"venue": "Transactions of the Faculty of Actuaries",
"volume": "8",
"issue": "",
"pages": "182--192",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Lidstone, G.J. 1920. Note on the general case of the Bayes-Laplace formula for inductive or a posteri- ori probabilities. Transactions of the Faculty of Actuaries, 8:182-192.",
"links": null
},
"BIBREF18": {
"ref_id": "b18",
"title": "A hierarchical Dirichlet language model",
"authors": [
{
"first": "David",
"middle": [
"J C"
],
"last": "Mackay",
"suffix": ""
},
{
"first": "Linda",
"middle": [
"C"
],
"last": "Peto",
"suffix": ""
}
],
"year": 1995,
"venue": "Natural Language Engineering",
"volume": "1",
"issue": "3",
"pages": "1--19",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "MacKay, David J. C. and Linda C. Peto. 1995. A hi- erarchical Dirichlet language model. Natural Lan- guage Engineering, 1(3):1-19.",
"links": null
},
"BIBREF19": {
"ref_id": "b19",
"title": "Natural Language Parsing as Statistical Pattern Recognition",
"authors": [
{
"first": "David",
"middle": [
"M"
],
"last": "Magerman",
"suffix": ""
}
],
"year": 1994,
"venue": "",
"volume": "",
"issue": "",
"pages": "",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Magerman, David M. 1994. Natural Language Pars- ing as Statistical Pattern Recognition. Ph.D. the- sis, Stanford University, February.",
"links": null
},
"BIBREF20": {
"ref_id": "b20",
"title": "Estimation of probabilities in the language model of the IBM speech recognition system",
"authors": [
{
"first": "Arthur",
"middle": [],
"last": "Nadas",
"suffix": ""
}
],
"year": 1984,
"venue": "IEEE Transactions on Acoustics, Speech and Signal Processing",
"volume": "32",
"issue": "4",
"pages": "859--861",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Nadas, Arthur. 1984. Estimation of probabilities in the language model of the IBM speech recognition system. IEEE Transactions on Acoustics, Speech and Signal Processing, ASSP-32(4):859-861, Au- gust.",
"links": null
},
"BIBREF21": {
"ref_id": "b21",
"title": "Numerical Recipes in C",
"authors": [
{
"first": "W",
"middle": [
"H"
],
"last": "Press",
"suffix": ""
},
{
"first": "B",
"middle": [
"P"
],
"last": "Flannery",
"suffix": ""
},
{
"first": "S",
"middle": [
"A"
],
"last": "Teukolsky",
"suffix": ""
},
{
"first": "W",
"middle": [
"T"
],
"last": "Vetterling",
"suffix": ""
}
],
"year": 1988,
"venue": "",
"volume": "",
"issue": "",
"pages": "",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Press, W.H., B.P. Flannery, S.A. Teukolsky, and W.T. Vetterling. 1988. Numerical Recipes in C. Cambridge University Press, Cambridge.",
"links": null
}
},
"ref_entries": {
"FIGREF1": {
"text": ")~ values for old and new bucketing schemes for Jelinek-Mercer smoothing; each point represents a single bucket",
"type_str": "figure",
"num": null,
"uris": null
},
"FIGREF2": {
"text": "Wl--n-l-13.1 Method average-countThis scheme is an instance of Jelinek-Mercer smoothing. Referring to equation (3), recall that Bahl et al. suggest bucketing the A~!-I according i--1 to c(Wi_n+l). We have found that partitioning the ~!-~ according to the average number of counts *--~+1 per non-zero element ~(~--~\"+1) yields better Iwi:~(~:_.+~)>01 results. Intuitively, the less sparse the data for estimating i-1 PML(WilWi_n+l), the larger A~,~-~ should be. *--~-t-1",
"type_str": "figure",
"num": null,
"uris": null
},
"FIGREF4": {
"text": "Baseline cross-entropy on test data; graph on left displays averages over ten runs for training sets up to 50,000 sentences, graph on right displays single runs for training sets up to 10,000,000 sentences average over ten runs at each size, up to 50........ ~ ............ ~ ............ = ................ ....... +....--~ plus~ne .",
"type_str": "figure",
"num": null,
"uris": null
},
"FIGREF5": {
"text": ", interp-held-out, ~nterpdel-int, new-avg-count, new-one-count (see below) data (-25 words~sentence) average over ten runs at each size, up to 50,000 sentences single run at each size, Up to 10,000,000 sentences 0......................................................................................................................................................... .. ~t 1 ............ ~ ........... ~ ........... .............. ~ ................ :::::::::::::::::::::::::::::::::::::::::::::: data (-25 words/sentonce)",
"type_str": "figure",
"num": null,
"uris": null
},
"FIGREF6": {
"text": "Trigram model on TIPSTER data; relative performance of various methods with respect to baseline; graphs on left display averages over ten runs for training sets up to 50,000 sentences, graphs on right display single runs for training sets up to 10,000,000 sentences; top graphs show all algorithms, bottom graphs zoom in on those methods that perform better than the baseline method .~-................. \"~'\" plus-o'n~ ............. ~ .... ...~ ............ {. .............. -0.00 ..... ~ ............. -'~-'~\" :=L:taterp-he~-out., interp~, el~tat, ,~ew,~zvg~ou ' ...... ~ne.~ount ! ...... ,ow)Bigram model on TIPSTER data; relative performance of various methods with respect to baseline; graphs on left display averages over ten runs for training sets up to 50,000 sentences, graphs on right display single runs for training sets up to 10,000,000 sentences; top graphs show all algorithms, bottom graphs zoom in on those methods that perform better than the baseline method of training data (-21 words~sentence) tzigram model 0 ......................................................................................................................................................................... ...i r ~ t e t p..<1 el-~ip_t .......... .0.12 :::.-.. \"'~=.... ......... ~ ........... Q.. interp*held-out ............ 7~.=: =-P~::.... \" ......... e ............... o ........... e ........... ~ ............... data (-21 words/sentence) Bigram and trigram models on Brown corpus; relative performance of various methods with respect to baseline",
"type_str": "figure",
"num": null,
"uris": null
},
"TABREF1": {
"type_str": "table",
"num": null,
"text": "reserves a section of the training data for this purpose. Alternatively, aelinek and Mercer describe a technique called deleted interpolation where different parts of the training data rotate in training either PML or the A,o!-' ; the results are then averaged.",
"html": null,
"content": "<table/>"
}
}
}
} |