File size: 76,972 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 | {
"paper_id": "P05-1003",
"header": {
"generated_with": "S2ORC 1.0.0",
"date_generated": "2023-01-19T09:37:36.337440Z"
},
"title": "Logarithmic Opinion Pools for Conditional Random Fields",
"authors": [
{
"first": "Andrew",
"middle": [],
"last": "Smith",
"suffix": "",
"affiliation": {
"laboratory": "",
"institution": "University of Edinburgh United Kingdom",
"location": {}
},
"email": ""
},
{
"first": "Trevor",
"middle": [],
"last": "Cohn",
"suffix": "",
"affiliation": {
"laboratory": "",
"institution": "University of Melbourne",
"location": {
"country": "Australia"
}
},
"email": "tacohn@csse.unimelb.edu.au"
},
{
"first": "Miles",
"middle": [],
"last": "Osborne",
"suffix": "",
"affiliation": {
"laboratory": "",
"institution": "University of Edinburgh",
"location": {
"country": "United Kingdom"
}
},
"email": ""
}
],
"year": "",
"venue": null,
"identifiers": {},
"abstract": "Recent work on Conditional Random Fields (CRFs) has demonstrated the need for regularisation to counter the tendency of these models to overfit. The standard approach to regularising CRFs involves a prior distribution over the model parameters, typically requiring search over a hyperparameter space. In this paper we address the overfitting problem from a different perspective, by factoring the CRF distribution into a weighted product of individual \"expert\" CRF distributions. We call this model a logarithmic opinion pool (LOP) of CRFs (LOP-CRFs). We apply the LOP-CRF to two sequencing tasks. Our results show that unregularised expert CRFs with an unregularised CRF under a LOP can outperform the unregularised CRF, and attain a performance level close to the regularised CRF. LOP-CRFs therefore provide a viable alternative to CRF regularisation without the need for hyperparameter search.",
"pdf_parse": {
"paper_id": "P05-1003",
"_pdf_hash": "",
"abstract": [
{
"text": "Recent work on Conditional Random Fields (CRFs) has demonstrated the need for regularisation to counter the tendency of these models to overfit. The standard approach to regularising CRFs involves a prior distribution over the model parameters, typically requiring search over a hyperparameter space. In this paper we address the overfitting problem from a different perspective, by factoring the CRF distribution into a weighted product of individual \"expert\" CRF distributions. We call this model a logarithmic opinion pool (LOP) of CRFs (LOP-CRFs). We apply the LOP-CRF to two sequencing tasks. Our results show that unregularised expert CRFs with an unregularised CRF under a LOP can outperform the unregularised CRF, and attain a performance level close to the regularised CRF. LOP-CRFs therefore provide a viable alternative to CRF regularisation without the need for hyperparameter search.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Abstract",
"sec_num": null
}
],
"body_text": [
{
"text": "In recent years, conditional random fields (CRFs) (Lafferty et al., 2001) have shown success on a number of natural language processing (NLP) tasks, including shallow parsing (Sha and Pereira, 2003) , named entity recognition (McCallum and Li, 2003) and information extraction from research papers (Peng and McCallum, 2004) . In general, this work has demonstrated the susceptibility of CRFs to overfit the training data during parameter estimation. As a consequence, it is now standard to use some form of overfitting reduction in CRF training.",
"cite_spans": [
{
"start": 50,
"end": 73,
"text": "(Lafferty et al., 2001)",
"ref_id": "BIBREF7"
},
{
"start": 175,
"end": 198,
"text": "(Sha and Pereira, 2003)",
"ref_id": "BIBREF15"
},
{
"start": 226,
"end": 249,
"text": "(McCallum and Li, 2003)",
"ref_id": "BIBREF9"
},
{
"start": 298,
"end": 323,
"text": "(Peng and McCallum, 2004)",
"ref_id": "BIBREF13"
}
],
"ref_spans": [],
"eq_spans": [],
"section": "Introduction",
"sec_num": "1"
},
{
"text": "Recently, there have been a number of sophisticated approaches to reducing overfitting in CRFs, including automatic feature induction (McCallum, 2003) and a full Bayesian approach to training and inference (Qi et al., 2005) . These advanced methods tend to be difficult to implement and are often computationally expensive. Consequently, due to its ease of implementation, the current standard approach to reducing overfitting in CRFs is the use of a prior distribution over the model parameters, typically a Gaussian. The disadvantage with this method, however, is that it requires adjusting the value of one or more of the distribution's hyperparameters. This usually involves manual or automatic tuning on a development set, and can be an expensive process as the CRF must be retrained many times for different hyperparameter values.",
"cite_spans": [
{
"start": 134,
"end": 150,
"text": "(McCallum, 2003)",
"ref_id": "BIBREF11"
},
{
"start": 206,
"end": 223,
"text": "(Qi et al., 2005)",
"ref_id": "BIBREF14"
}
],
"ref_spans": [],
"eq_spans": [],
"section": "Introduction",
"sec_num": "1"
},
{
"text": "In this paper we address the overfitting problem in CRFs from a different perspective. We factor the CRF distribution into a weighted product of individual expert CRF distributions, each focusing on a particular subset of the distribution. We call this model a logarithmic opinion pool (LOP) of CRFs (LOP-CRFs), and provide a procedure for learning the weight of each expert in the product. The LOP-CRF framework is \"parameter-free\" in the sense that it does not involve the requirement to adjust hyperparameter values.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Introduction",
"sec_num": "1"
},
{
"text": "LOP-CRFs are theoretically advantageous in that their Kullback-Leibler divergence with a given distribution can be explicitly represented as a function of the KL-divergence with each of their expert distributions. This provides a well-founded framework for designing new overfitting reduction schemes: look to factorise a CRF distribution as a set of diverse experts.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Introduction",
"sec_num": "1"
},
{
"text": "We apply LOP-CRFs to two sequencing tasks in NLP: named entity recognition and part-of-speech tagging. Our results show that combination of unregularised expert CRFs with an unregularised standard CRF under a LOP can outperform the unregularised standard CRF, and attain a performance level that rivals that of the regularised standard CRF. LOP-CRFs therefore provide a viable alternative to CRF regularisation without the need for hyperparameter search.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Introduction",
"sec_num": "1"
},
{
"text": "A linear chain CRF defines the conditional probability of a state or label sequence s given an observed sequence o via 1 :",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Conditional Random Fields",
"sec_num": "2"
},
{
"text": "p(s | o) = 1 Z(o) exp T +1 \u2211 t=1 \u2211 k \u03bb k f k (s t\u22121 , s t , o,t) (1)",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Conditional Random Fields",
"sec_num": "2"
},
{
"text": "where T is the length of both sequences, \u03bb k are parameters of the model and Z(o) is the partition function that ensures (1) represents a probability distribution. The functions f k are feature functions representing the occurrence of different events in the sequences s and o.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Conditional Random Fields",
"sec_num": "2"
},
{
"text": "The parameters \u03bb k can be estimated by maximising the conditional log-likelihood of a set of labelled training sequences. The log-likelihood is given by:",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Conditional Random Fields",
"sec_num": "2"
},
{
"text": "L (\u03bb ) = \u2211 o,sp (o, s) log p(s | o; \u03bb ) = \u2211 o,sp (o, s) T +1 \u2211 t=1 \u03bb \u2022 f(s, o,t) \u2212 \u2211 op (o) log Z(o; \u03bb )",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Conditional Random Fields",
"sec_num": "2"
},
{
"text": "wherep(o, s) andp(o) are empirical distributions defined by the training set. At the maximum likelihood solution the model satisfies a set of feature constraints, whereby the expected count of each feature under the model is equal to its empirical count on the training data:",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Conditional Random Fields",
"sec_num": "2"
},
{
"text": "Ep (o,s) [ f k ] \u2212 E p(s|o) [ f k ] = 0, \u2200k",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Conditional Random Fields",
"sec_num": "2"
},
{
"text": "In general this cannot be solved for the \u03bb k in closed form so numerical routines must be used. Malouf (2002) and Sha and Pereira (2003) show that gradient-based algorithms, particularly limited memory variable metric (LMVM), require much less time to reach convergence, for some NLP tasks, than the iterative scaling methods (Della Pietra et al., 1997) previously used for log-linear optimisation problems. In all our experiments we use the LMVM method to train the CRFs.",
"cite_spans": [
{
"start": 96,
"end": 109,
"text": "Malouf (2002)",
"ref_id": "BIBREF8"
},
{
"start": 114,
"end": 136,
"text": "Sha and Pereira (2003)",
"ref_id": "BIBREF15"
},
{
"start": 326,
"end": 353,
"text": "(Della Pietra et al., 1997)",
"ref_id": "BIBREF3"
}
],
"ref_spans": [],
"eq_spans": [],
"section": "Conditional Random Fields",
"sec_num": "2"
},
{
"text": "For CRFs with general graphical structure, calculation of E p(s|o) [ f k ] is intractable, but for the linear chain case Lafferty et al. (2001) describe an efficient dynamic programming procedure for inference, similar in nature to the forward-backward algorithm in hidden Markov models.",
"cite_spans": [
{
"start": 121,
"end": 143,
"text": "Lafferty et al. (2001)",
"ref_id": "BIBREF7"
}
],
"ref_spans": [],
"eq_spans": [],
"section": "Conditional Random Fields",
"sec_num": "2"
},
{
"text": "In this paper an expert model refers a probabilistic model that focuses on modelling a specific subset of some probability distribution. The concept of combining the distributions of a set of expert models via a weighted product has previously been used in a range of different application areas, including economics and management science (Bordley, 1982) , and NLP (Osborne and Baldridge, 2004) .",
"cite_spans": [
{
"start": 340,
"end": 355,
"text": "(Bordley, 1982)",
"ref_id": "BIBREF0"
},
{
"start": 366,
"end": 395,
"text": "(Osborne and Baldridge, 2004)",
"ref_id": "BIBREF12"
}
],
"ref_spans": [],
"eq_spans": [],
"section": "Logarithmic Opinion Pools",
"sec_num": "3"
},
{
"text": "In this paper we restrict ourselves to sequence models. Given a set of sequence model experts, indexed by \u03b1, with conditional distributions p \u03b1 (s | o) and a set of non-negative normalised weights w \u03b1 , a logarithmic opinion pool 2 is defined as the distribution:",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Logarithmic Opinion Pools",
"sec_num": "3"
},
{
"text": "EQUATION",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [
{
"start": 0,
"end": 8,
"text": "EQUATION",
"ref_id": "EQREF",
"raw_str": "p LOP (s | o) = 1 Z LOP (o) \u220f \u03b1 [p \u03b1 (s | o)] w \u03b1",
"eq_num": "(2)"
}
],
"section": "Logarithmic Opinion Pools",
"sec_num": "3"
},
{
"text": "with w \u03b1 \u2265 0 and \u2211 \u03b1 w \u03b1 = 1, and where Z LOP (o) is the normalisation constant:",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Logarithmic Opinion Pools",
"sec_num": "3"
},
{
"text": "EQUATION",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [
{
"start": 0,
"end": 8,
"text": "EQUATION",
"ref_id": "EQREF",
"raw_str": "Z LOP (o) = \u2211 s \u220f \u03b1 [p \u03b1 (s | o)] w \u03b1",
"eq_num": "(3)"
}
],
"section": "Logarithmic Opinion Pools",
"sec_num": "3"
},
{
"text": "The weight w \u03b1 encodes our confidence in the opinion of expert \u03b1.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Logarithmic Opinion Pools",
"sec_num": "3"
},
{
"text": "Suppose that there is a \"true\" conditional distribution q(s | o) which each p \u03b1 (s | o) is attempting to model. Heskes (1998) shows that the KL divergence between q(s | o) and the LOP, can be decomposed into two terms:",
"cite_spans": [
{
"start": 112,
"end": 125,
"text": "Heskes (1998)",
"ref_id": "BIBREF5"
}
],
"ref_spans": [],
"eq_spans": [],
"section": "Logarithmic Opinion Pools",
"sec_num": "3"
},
{
"text": "K(q, p LOP ) = E \u2212 A (4) = \u2211 \u03b1 w \u03b1 K (q, p \u03b1 ) \u2212 \u2211 \u03b1 w \u03b1 K (p LOP , p \u03b1 )",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Logarithmic Opinion Pools",
"sec_num": "3"
},
{
"text": "This tells us that the closeness of the LOP model to q(s | o) is governed by a trade-off between two terms: an E term, which represents the closeness of the individual experts to q(s | o), and an A term, which represents the closeness of the individual experts to the LOP, and therefore indirectly to each other. Hence for the LOP to model q well, we desire models p \u03b1 which are individually good models of q (having low E) and are also diverse (having large A).",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Logarithmic Opinion Pools",
"sec_num": "3"
},
{
"text": "Because CRFs are log-linear models, we can see from equation 2that CRF experts are particularly well suited to combination under a LOP. Indeed, the resulting LOP is itself a CRF, the LOP-CRF, with potential functions given by a log-linear combination of the potential functions of the experts, with weights w \u03b1 . As a consequence of this, the normalisation constant for the LOP-CRF can be calculated efficiently via the usual forward-backward algorithm for CRFs. Note that there is a distinction between normalisation constant for the LOP-CRF, Z LOP as given in equation 3, and the partition function of the LOP-CRF, Z. The two are related as follows:",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "LOPs for CRFs",
"sec_num": "3.1"
},
{
"text": "p LOP (s | o) = 1 Z LOP (o) \u220f \u03b1 [p \u03b1 (s | o)] w \u03b1 = 1 Z LOP (o) \u220f \u03b1 U \u03b1 (s | o) Z \u03b1 (o) w \u03b1 = \u220f \u03b1 [U \u03b1 (s | o)] w \u03b1 Z LOP (o) \u220f \u03b1 [Z \u03b1 (o)] w \u03b1 where U \u03b1 = exp \u2211 T +1 t=1 \u2211 k \u03bb \u03b1k f \u03b1k (s t\u22121 , s t , o,t) and so log Z(o) = log Z LOP (o) + \u2211 \u03b1 w \u03b1 log Z \u03b1 (o)",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "LOPs for CRFs",
"sec_num": "3.1"
},
{
"text": "This relationship will be useful below, when we describe how to train the weights w \u03b1 of a LOP-CRF. In this paper we will use the term LOP-CRF weights to refer to the weights w \u03b1 in the weighted product of the LOP-CRF distribution and the term parameters to refer to the parameters \u03bb \u03b1k of each expert CRF \u03b1.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "LOPs for CRFs",
"sec_num": "3.1"
},
{
"text": "In our LOP-CRF training procedure we first train the expert CRFs unregularised on the training data. Then, treating the experts as static pre-trained models, we train the LOP-CRF weights w \u03b1 to maximise the log-likelihood of the training data. This training process is \"parameter-free\" in that neither stage involves the use of a prior distribution over expert CRF parameters or LOP-CRF weights, and so avoids the requirement to adjust hyperparameter values.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Training LOP-CRFs",
"sec_num": "3.2"
},
{
"text": "The likelihood of a data set under a LOP-CRF, as a function of the LOP-CRF weights, is given by:",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Training LOP-CRFs",
"sec_num": "3.2"
},
{
"text": "L(w) = \u220f o,s p LOP (s | o; w)p (o,s) = \u220f o,s 1 Z LOP (o; w) \u220f \u03b1 p \u03b1 (s | o) w \u03b1 p(o,s)",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Training LOP-CRFs",
"sec_num": "3.2"
},
{
"text": "After taking logs and rearranging, the loglikelihood can be expressed as:",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Training LOP-CRFs",
"sec_num": "3.2"
},
{
"text": "L (w) = \u2211 o,sp (o, s) \u2211 \u03b1 w \u03b1 log p \u03b1 (s | o) \u2212 \u2211 op (o) log Z LOP (o; w) = \u2211 \u03b1 w \u03b1 \u2211 o,sp (o, s) log p \u03b1 (s | o) + \u2211 \u03b1 w \u03b1 \u2211 op (o) log Z \u03b1 (o) \u2212 \u2211 op (o) log Z(o; w)",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Training LOP-CRFs",
"sec_num": "3.2"
},
{
"text": "For the first two terms, the quantities that are multiplied by w \u03b1 inside the (outer) sums are independent of the weights, and can be evaluated once at the beginning of training. The third term involves the partition function for the LOP-CRF and so is a function of the weights. It can be evaluated efficiently as usual for a standard CRF.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Training LOP-CRFs",
"sec_num": "3.2"
},
{
"text": "Taking derivatives with respect to w \u03b2 and rearranging, we obtain:",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Training LOP-CRFs",
"sec_num": "3.2"
},
{
"text": "\u2202 L (w) \u2202 w \u03b2 = \u2211 o,sp (o, s) log p \u03b2 (s | o) + \u2211 op (o) log Z \u03b2 (o) \u2212 \u2211 op (o)E p LOP (s|o) \u2211 t logU \u03b2t (o, s)",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Training LOP-CRFs",
"sec_num": "3.2"
},
{
"text": "where U \u03b2t (o, s) is the value of the potential function for expert \u03b2 on clique t under the labelling s for observation o. In a way similar to the representation of the expected feature count in a standard CRF, the third term may be re-written as:",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Training LOP-CRFs",
"sec_num": "3.2"
},
{
"text": "\u2212 \u2211 o \u2211 t \u2211 s ,s p LOP (s t\u22121 = s , s t = s , o) logU \u03b2t (s , s , o)",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Training LOP-CRFs",
"sec_num": "3.2"
},
{
"text": "Hence the derivative is tractable because we can use dynamic programming to efficiently calculate the pairwise marginal distribution for the LOP-CRF. Using these expressions we can efficiently train the LOP-CRF weights to maximise the loglikelihood of the data set. 3 We make use of the LMVM method mentioned earlier to do this. We will refer to a LOP-CRF with weights trained using this procedure as an unregularised LOP-CRF.",
"cite_spans": [
{
"start": 266,
"end": 267,
"text": "3",
"ref_id": null
}
],
"ref_spans": [],
"eq_spans": [],
"section": "Training LOP-CRFs",
"sec_num": "3.2"
},
{
"text": "The \"parameter-free\" aspect of the training procedure we introduced in the previous section relies on the fact that we do not use regularisation when training the LOP-CRF weights w \u03b1 . However, there is a possibility that this may lead to overfitting of the training data. In order to investigate this, we develop a regularised version of the training procedure and compare the results obtained with each. We use a prior distribution over the LOP-CRF weights. As the weights are non-negative and normalised we use a Dirichlet distribution, whose density function is given by:",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Regularisation",
"sec_num": "3.2.1"
},
{
"text": "p(w) = \u0393(\u2211 \u03b1 \u03b8 \u03b1 ) \u220f \u03b1 \u0393(\u03b8 \u03b1 ) \u220f \u03b1 w \u03b8 \u03b1 \u22121 \u03b1",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Regularisation",
"sec_num": "3.2.1"
},
{
"text": "where the \u03b8 \u03b1 are hyperparameters.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Regularisation",
"sec_num": "3.2.1"
},
{
"text": "Under this distribution, ignoring terms that are independent of the weights, the regularised loglikelihood involves an additional term:",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Regularisation",
"sec_num": "3.2.1"
},
{
"text": "\u2211 \u03b1 (\u03b8 \u03b1 \u2212 1) log w \u03b1",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Regularisation",
"sec_num": "3.2.1"
},
{
"text": "We assume a single value \u03b8 across all weights. The derivative of the regularised log-likelihood with respect to weight w \u03b2 then involves an additional term 1 w \u03b2 (\u03b8 \u2212 1). In our experiments we use the development set to optimise the value of \u03b8 . We will refer to a LOP-CRF with weights trained using this procedure as a regularised LOP-CRF.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Regularisation",
"sec_num": "3.2.1"
},
{
"text": "In this paper we apply LOP-CRFs to two sequence labelling tasks in NLP: named entity recognition (NER) and part-of-speech tagging (POS tagging).",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "The Tasks",
"sec_num": "4"
},
{
"text": "NER involves the identification of the location and type of pre-defined entities within a sentence and is often used as a sub-process in information extraction systems. With NER the CRF is presented with a set of sentences and must label each word so as to indicate whether the word appears outside an entity (O), at the beginning of an entity of type X (B-X) or within the continuation of an entity of type X (I-X).",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Named Entity Recognition",
"sec_num": "4.1"
},
{
"text": "All our results for NER are reported on the CoNLL-2003 shared task dataset (Tjong Kim Sang and De Meulder, 2003) . For this dataset the entity types are: persons (PER), locations (LOC), organisations (ORG) and miscellaneous (MISC). The training set consists of 14, 987 sentences and 204, 567 tokens, the development set consists of 3, 466 sentences and 51, 578 tokens and the test set consists of 3, 684 sentences and 46, 666 tokens.",
"cite_spans": [
{
"start": 75,
"end": 112,
"text": "(Tjong Kim Sang and De Meulder, 2003)",
"ref_id": "BIBREF17"
}
],
"ref_spans": [],
"eq_spans": [],
"section": "Named Entity Recognition",
"sec_num": "4.1"
},
{
"text": "POS tagging involves labelling each word in a sentence with its part-of-speech, for example noun, verb, adjective, etc. For our experiments we use the CoNLL-2000 shared task dataset (Tjong Kim Sang and Buchholz, 2000) . This has 48 different POS tags. In order to make training time manageable 4 , we collapse the number of POS tags from 48 to 5 following the procedure used in . In summary:",
"cite_spans": [
{
"start": 182,
"end": 217,
"text": "(Tjong Kim Sang and Buchholz, 2000)",
"ref_id": "BIBREF16"
}
],
"ref_spans": [],
"eq_spans": [],
"section": "Part-of-Speech Tagging",
"sec_num": "4.2"
},
{
"text": "\u2022 All types of noun collapse to category N.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Part-of-Speech Tagging",
"sec_num": "4.2"
},
{
"text": "\u2022 All types of verb collapse to category V.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Part-of-Speech Tagging",
"sec_num": "4.2"
},
{
"text": "\u2022 All types of adjective collapse to category J.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Part-of-Speech Tagging",
"sec_num": "4.2"
},
{
"text": "\u2022 All types of adverb collapse to category R.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Part-of-Speech Tagging",
"sec_num": "4.2"
},
{
"text": "\u2022 All other POS tags collapse to category O.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Part-of-Speech Tagging",
"sec_num": "4.2"
},
{
"text": "The training set consists of 7, 300 sentences and 173, 542 tokens, the development set consists of 1, 636 sentences and 38, 185 tokens and the test set consists of 2, 012 sentences and 47, 377 tokens.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Part-of-Speech Tagging",
"sec_num": "4.2"
},
{
"text": "For each task we compare the performance of the LOP-CRF to that of the standard CRF by defining a single, complex CRF, which we call a monolithic CRF, and a range of expert sets.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Expert sets",
"sec_num": "4.3"
},
{
"text": "The monolithic CRF for NER comprises a number of word and POS tag features in a window of five words around the current word, along with a set of orthographic features defined on the current word. These are based on those found in (Curran and Clark, 2003) . Examples include whether the current word is capitalised, is an initial, contains a digit, contains punctuation, etc. The monolithic CRF for NER has 450, 345 features.",
"cite_spans": [
{
"start": 231,
"end": 255,
"text": "(Curran and Clark, 2003)",
"ref_id": "BIBREF2"
}
],
"ref_spans": [],
"eq_spans": [],
"section": "Expert sets",
"sec_num": "4.3"
},
{
"text": "The monolithic CRF for POS tagging comprises word and POS features similar to those in the NER monolithic model, but over a smaller number of orthographic features. The monolithic model for POS tagging has 188, 448 features.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Expert sets",
"sec_num": "4.3"
},
{
"text": "Each of our expert sets consists of a number of CRF experts. Usually these experts are designed to focus on modelling a particular aspect or subset of the distribution. As we saw earlier, the aim here is to define experts that model parts of the distribution well while retaining mutual diversity. The experts from a particular expert set are combined under a LOP-CRF and the weights are trained as described previously.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Expert sets",
"sec_num": "4.3"
},
{
"text": "We define our range of expert sets as follows:",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Expert sets",
"sec_num": "4.3"
},
{
"text": "\u2022 Simple consists of the monolithic CRF and a single expert comprising a reduced subset of the features in the monolithic CRF. This reduced CRF models the entire distribution rather than focusing on a particular aspect or subset, but is much less expressive than the monolithic model. The reduced model comprises 24, 818 features for NER and 47, 420 features for POS tagging.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Expert sets",
"sec_num": "4.3"
},
{
"text": "\u2022 Positional consists of the monolithic CRF and a partition of the features in the monolithic CRF into three experts, each consisting only of features that involve events either behind, at or ahead of the current sequence position.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Expert sets",
"sec_num": "4.3"
},
{
"text": "\u2022 Label consists of the monolithic CRF and a partition of the features in the monolithic CRF into five experts, one for each label. For NER an expert corresponding to label X consists only of features that involve labels B-X or I-X at the current or previous positions, while for POS tagging an expert corresponding to label X consists only of features that involve label X at the current or previous positions. These experts therefore focus on trying to model the distribution of a particular label.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Expert sets",
"sec_num": "4.3"
},
{
"text": "\u2022 Random consists of the monolithic CRF and a random partition of the features in the monolithic CRF into four experts. This acts as a baseline to ascertain the performance that can be expected from an expert set that is not defined via any linguistic intuition.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Expert sets",
"sec_num": "4.3"
},
{
"text": "To compare the performance of LOP-CRFs trained using the procedure we described previously to that of a standard CRF regularised with a Gaussian prior, we do the following for both NER and POS tagging:",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Experiments",
"sec_num": "5"
},
{
"text": "\u2022 Train a monolithic CRF with regularisation using a Gaussian prior. We use the development set to optimise the value of the variance hyperparameter.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Experiments",
"sec_num": "5"
},
{
"text": "\u2022 Train every expert CRF in each expert set without regularisation (each expert set includes the monolithic CRF, which clearly need only be trained once).",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Experiments",
"sec_num": "5"
},
{
"text": "\u2022 For each expert set, create a LOP-CRF from the expert CRFs and train the weights of the LOP-CRF without regularisation. We compare its performance to that of the unregularised and regularised monolithic CRFs.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Experiments",
"sec_num": "5"
},
{
"text": "\u2022 To investigate whether training the LOP-CRF weights contributes significantly to the LOP-CRF's performance, for each expert set we create a LOP-CRF with uniform weights and compare its performance to that of the LOP-CRF with trained weights.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Experiments",
"sec_num": "5"
},
{
"text": "\u2022 To investigate whether unregularised training of the LOP-CRF weights leads to overfitting, for each expert set we train the weights of the LOP-CRF with regularisation using a Dirichlet prior. We optimise the hyperparameter in the Dirichlet distribution on the development set. We then compare the performance of the LOP-CRF with regularised weights to that of the LOP-CRF with unregularised weights.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Experiments",
"sec_num": "5"
},
{
"text": "Before presenting results for the LOP-CRFs, we briefly give performance figures for the monolithic CRFs and expert CRFs in isolation. For illustration, we do this for NER models only. Table 1 shows F scores on the development set for the NER CRFs. We see that, as expected, the expert CRFs in isolation model the data relatively poorly compared to the monolithic CRFs. Some of the label experts, for example, attain relatively low F scores as they focus only on modelling one particular label. Similar behaviour was observed for the POS tagging models. ",
"cite_spans": [],
"ref_spans": [
{
"start": 184,
"end": 191,
"text": "Table 1",
"ref_id": "TABREF1"
}
],
"eq_spans": [],
"section": "Experts",
"sec_num": "6.1"
},
{
"text": "In this section we present results for LOP-CRFs with unregularised weights. Table 2 gives F scores for NER LOP-CRFs while Table 3 gives accuracies for the POS tagging LOP-CRFs. The monolithic CRF scores are included for comparison. Both tables illustrate the following points:",
"cite_spans": [],
"ref_spans": [
{
"start": 76,
"end": 83,
"text": "Table 2",
"ref_id": "TABREF3"
},
{
"start": 122,
"end": 129,
"text": "Table 3",
"ref_id": "TABREF4"
}
],
"eq_spans": [],
"section": "LOP-CRFs with unregularised weights",
"sec_num": "6.2"
},
{
"text": "\u2022 In every case the LOP-CRFs outperform the unregularised monolithic CRF",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "LOP-CRFs with unregularised weights",
"sec_num": "6.2"
},
{
"text": "\u2022 In most cases the performance of LOP-CRFs rivals that of the regularised monolithic CRF, and in some cases exceeds it.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "LOP-CRFs with unregularised weights",
"sec_num": "6.2"
},
{
"text": "We use McNemar's matched-pairs test (Gillick and Cox, 1989) on point-wise labelling errors to examine the statistical significance of these results. We test significance at the 5% level. At this threshold, all the LOP-CRFs significantly outperform the corresponding unregularised monolithic CRF. In addition, those marked with * show a significant improvement over the regularised monolithic CRF. Only the value marked with \u2020 in Table 3 significantly under performs the regularised monolithic. All other values a do not differ significantly from those of the regularised monolithic CRF at the 5% level.",
"cite_spans": [
{
"start": 36,
"end": 59,
"text": "(Gillick and Cox, 1989)",
"ref_id": "BIBREF4"
}
],
"ref_spans": [
{
"start": 429,
"end": 436,
"text": "Table 3",
"ref_id": "TABREF4"
}
],
"eq_spans": [],
"section": "LOP-CRFs with unregularised weights",
"sec_num": "6.2"
},
{
"text": "These results show that LOP-CRFs with unregularised weights can lead to performance improvements that equal or exceed those achieved from a conventional regularisation approach using a Gaussian prior. The important difference, however, is that the LOP-CRF approach is \"parameter-free\" in the As an illustration of a typical weight distribution resulting from the training procedure, the positional LOP-CRF for POS tagging attaches weight 0.45 to the monolithic model and roughly equal weights to the other three experts.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "LOP-CRFs with unregularised weights",
"sec_num": "6.2"
},
{
"text": "By training LOP-CRF weights using the procedure we introduce in this paper, we allow the weights to take on non-uniform values. This corresponds to letting the opinion of some experts take precedence over others in the LOP-CRF's decision making. An alternative, simpler, approach would be to combine the experts under a LOP with uniform weights, thereby avoiding the weight training stage. We would like to ascertain whether this approach will significantly reduce the LOP-CRF's performance. As an illustration, Table 4 gives accuracies for LOP-CRFs with uniform weights for POS tagging. A similar pattern is observed for NER. Comparing these values to those in Tables 2 and 3, we can see that in Table 4 : Accuracies for POS tagging uniform LOP-CRFs general LOP-CRFs with uniform weights, although still performing significantly better than the unregularised monolithic CRF, generally under perform LOP-CRFs with trained weights. This suggests that the choice of weights can be important, and justifies the weight training stage.",
"cite_spans": [],
"ref_spans": [
{
"start": 512,
"end": 519,
"text": "Table 4",
"ref_id": null
},
{
"start": 697,
"end": 704,
"text": "Table 4",
"ref_id": null
}
],
"eq_spans": [],
"section": "LOP-CRFs with uniform weights",
"sec_num": "6.3"
},
{
"text": "To investigate whether unregularised training of the LOP-CRF weights leads to overfitting, we train the LOP-CRF with regularisation using a Dirichlet prior. The results we obtain show that in most cases a LOP-CRF with regularised weights achieves an almost identical performance to that with unregularised weights, and suggests there is little to be gained by weight regularisation. This is probably due to the fact that in our LOP-CRFs the number of experts, and therefore weights, is generally small and so there is little capacity for overfitting. We conjecture that although other choices of expert set may comprise many more experts than in our examples, the numbers are likely to be relatively small in comparison to, for example, the number of parameters in the individual experts. We therefore suggest that any overfitting effect is likely to be limited.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "LOP-CRFs with regularised weights",
"sec_num": "6.4"
},
{
"text": "We can see from Tables 2 and 3 that the performance of a LOP-CRF varies with the choice of expert set. For example, in our tasks the simple and positional expert sets perform better than those for the label and random sets. For an explanation here, we refer back to our discussion of equation (5). We conjecture that the simple and positional expert sets achieve good performance in the LOP-CRF because they consist of experts that are diverse while simultaneously being reasonable models of the data. The label expert set exhibits greater diversity between the experts, because each expert focuses on modelling a particular label only, but each expert is a relatively poor model of the entire distribution and the corresponding LOP-CRF performs worse. Similarly, the random experts are in general better models of the entire distribution but tend to be less diverse because they do not focus on any one aspect or subset of it. Intuitively, then, we want to devise experts that provide diverse but accurate views on the data.",
"cite_spans": [],
"ref_spans": [
{
"start": 16,
"end": 30,
"text": "Tables 2 and 3",
"ref_id": "TABREF3"
}
],
"eq_spans": [],
"section": "Choice of Expert Sets",
"sec_num": "6.5"
},
{
"text": "The expert sets we present in this paper were motivated by linguistic intuition, but clearly many choices exist. It remains an important open question as to how to automatically construct expert sets for good performance on a given task, and we intend to pursue this avenue in future research.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Choice of Expert Sets",
"sec_num": "6.5"
},
{
"text": "In this paper we have introduced the logarithmic opinion pool of CRFs as a way to address overfitting in CRF models. Our results show that a LOP-CRF can provide a competitive alternative to conventional regularisation with a prior while avoiding the requirement to search a hyperparameter space.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Conclusion and future work",
"sec_num": "7"
},
{
"text": "We have seen that, for a variety of types of expert, combination of expert CRFs with an unregularised standard CRF under a LOP with optimised weights can outperform the unregularised standard CRF and rival the performance of a regularised standard CRF.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Conclusion and future work",
"sec_num": "7"
},
{
"text": "We have shown how these advantages a LOP-CRF provides have a firm theoretical foundation in terms of the decomposition of the KL-divergence between a LOP-CRF and a target distribution, and how this provides a framework for designing new overfitting reduction schemes in terms of constructing diverse experts.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Conclusion and future work",
"sec_num": "7"
},
{
"text": "In this work we have considered training the weights of a LOP-CRF using pre-trained, static experts. In future we intend to investigate cooperative training of LOP-CRF weights and the parameters of each expert in an expert set.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Conclusion and future work",
"sec_num": "7"
},
{
"text": "In this paper we assume there is a one-to-one mapping between states and labels, though this need not be the case.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "",
"sec_num": null
},
{
"text": "Hinton (1999) introduced a variant of the LOP idea called Product of Experts, in which expert distributions are multiplied under a uniform weight distribution.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "",
"sec_num": null
},
{
"text": "We must ensure that the weights are non-negative and normalised. We achieve this by parameterising the weights as functions of a set of unconstrained variables via a softmax transformation. The values of the log-likelihood and its derivatives with respect to the unconstrained variables can be derived from the corresponding values for the weights w \u03b1 .",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "",
"sec_num": null
},
{
"text": "See(Cohn et al., 2005) for a scaling method allowing the full POS tagging task with CRFs.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "",
"sec_num": null
}
],
"back_matter": [
{
"text": "We wish to thank Stephen Clark, our colleagues in Edinburgh and the anonymous reviewers for many useful comments.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Acknowledgements",
"sec_num": null
}
],
"bib_entries": {
"BIBREF0": {
"ref_id": "b0",
"title": "A multiplicative formula for aggregating probability assessments",
"authors": [
{
"first": "R",
"middle": [
"F"
],
"last": "Bordley",
"suffix": ""
}
],
"year": 1982,
"venue": "Management Science",
"volume": "",
"issue": "28",
"pages": "1137--1148",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "R. F. Bordley. 1982. A multiplicative formula for aggregating probability assessments. Management Science, (28):1137- 1148.",
"links": null
},
"BIBREF1": {
"ref_id": "b1",
"title": "Scaling conditional random fields using error-correcting codes",
"authors": [
{
"first": "T",
"middle": [],
"last": "Cohn",
"suffix": ""
},
{
"first": "A",
"middle": [],
"last": "Smith",
"suffix": ""
},
{
"first": "M",
"middle": [],
"last": "Osborne",
"suffix": ""
}
],
"year": 2005,
"venue": "Proc. ACL",
"volume": "",
"issue": "",
"pages": "",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "T. Cohn, A. Smith, and M. Osborne. 2005. Scaling conditional random fields using error-correcting codes. In Proc. ACL 2005.",
"links": null
},
"BIBREF2": {
"ref_id": "b2",
"title": "Language independent NER using a maximum entropy tagger",
"authors": [
{
"first": "J",
"middle": [],
"last": "Curran",
"suffix": ""
},
{
"first": "S",
"middle": [],
"last": "Clark",
"suffix": ""
}
],
"year": 2003,
"venue": "Proc. CoNLL-2003",
"volume": "",
"issue": "",
"pages": "",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "J. Curran and S. Clark. 2003. Language independent NER using a maximum entropy tagger. In Proc. CoNLL-2003.",
"links": null
},
"BIBREF3": {
"ref_id": "b3",
"title": "Inducing features of random fields",
"authors": [
{
"first": "S",
"middle": [],
"last": "",
"suffix": ""
},
{
"first": "Della",
"middle": [],
"last": "Pietra",
"suffix": ""
},
{
"first": "Della",
"middle": [],
"last": "Pietra",
"suffix": ""
},
{
"first": "V",
"middle": [],
"last": "",
"suffix": ""
},
{
"first": "J",
"middle": [],
"last": "Lafferty",
"suffix": ""
}
],
"year": 1997,
"venue": "IEEE PAMI",
"volume": "19",
"issue": "",
"pages": "380--393",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "S. Della Pietra, Della Pietra V., and J. Lafferty. 1997. Induc- ing features of random fields. In IEEE PAMI, volume 19(4), pages 380-393.",
"links": null
},
"BIBREF4": {
"ref_id": "b4",
"title": "Some statistical issues in the comparison of speech recognition algorithms",
"authors": [
{
"first": "L",
"middle": [],
"last": "Gillick",
"suffix": ""
},
{
"first": "S",
"middle": [],
"last": "Cox",
"suffix": ""
}
],
"year": 1989,
"venue": "International Conference on Acoustics, Speech and Signal Processing",
"volume": "1",
"issue": "",
"pages": "532--535",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "L. Gillick and S. Cox. 1989. Some statistical issues in the comparison of speech recognition algorithms. In Interna- tional Conference on Acoustics, Speech and Signal Process- ing, volume 1, pages 532-535.",
"links": null
},
"BIBREF5": {
"ref_id": "b5",
"title": "Selecting weighting factors in logarithmic opinion pools",
"authors": [
{
"first": "T",
"middle": [],
"last": "Heskes",
"suffix": ""
}
],
"year": 1998,
"venue": "Advances in Neural Information Processing Systems 10",
"volume": "",
"issue": "",
"pages": "",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "T. Heskes. 1998. Selecting weighting factors in logarithmic opinion pools. In Advances in Neural Information Process- ing Systems 10.",
"links": null
},
"BIBREF6": {
"ref_id": "b6",
"title": "Product of experts",
"authors": [
{
"first": "G",
"middle": [
"E"
],
"last": "Hinton",
"suffix": ""
}
],
"year": 1999,
"venue": "ICANN",
"volume": "1",
"issue": "",
"pages": "1--6",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "G. E. Hinton. 1999. Product of experts. In ICANN, volume 1, pages 1-6.",
"links": null
},
"BIBREF7": {
"ref_id": "b7",
"title": "Conditional random fields: Probabilistic models for segmenting and labeling sequence data",
"authors": [
{
"first": "J",
"middle": [],
"last": "Lafferty",
"suffix": ""
},
{
"first": "A",
"middle": [],
"last": "Mccallum",
"suffix": ""
},
{
"first": "F",
"middle": [],
"last": "Pereira",
"suffix": ""
}
],
"year": 2001,
"venue": "Proc. ICML",
"volume": "",
"issue": "",
"pages": "",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "J. Lafferty, A. McCallum, and F. Pereira. 2001. Conditional random fields: Probabilistic models for segmenting and la- beling sequence data. In Proc. ICML 2001.",
"links": null
},
"BIBREF8": {
"ref_id": "b8",
"title": "A comparison of algorithms for maximum entropy parameter estimation",
"authors": [
{
"first": "R",
"middle": [],
"last": "Malouf",
"suffix": ""
}
],
"year": 2002,
"venue": "Proc. CoNLL",
"volume": "",
"issue": "",
"pages": "",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "R. Malouf. 2002. A comparison of algorithms for maximum entropy parameter estimation. In Proc. CoNLL-2002.",
"links": null
},
"BIBREF9": {
"ref_id": "b9",
"title": "Early results for named entity recognition with conditional random fields, feature induction and web-enhanced lexicons",
"authors": [
{
"first": "A",
"middle": [],
"last": "Mccallum",
"suffix": ""
},
{
"first": "W",
"middle": [],
"last": "Li",
"suffix": ""
}
],
"year": 2003,
"venue": "Proc. CoNLL-2003",
"volume": "",
"issue": "",
"pages": "",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "A. McCallum and W. Li. 2003. Early results for named entity recognition with conditional random fields, feature induction and web-enhanced lexicons. In Proc. CoNLL-2003.",
"links": null
},
"BIBREF10": {
"ref_id": "b10",
"title": "Dynamic conditional random fields for jointly labeling multiple sequences",
"authors": [
{
"first": "A",
"middle": [],
"last": "Mccallum",
"suffix": ""
},
{
"first": "K",
"middle": [],
"last": "Rohanimanesh",
"suffix": ""
},
{
"first": "C",
"middle": [],
"last": "Sutton",
"suffix": ""
}
],
"year": 2003,
"venue": "NIPS-2003 Workshop on Syntax, Semantics and Statistics",
"volume": "",
"issue": "",
"pages": "",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "A. McCallum, K. Rohanimanesh, and C. Sutton. 2003. Dy- namic conditional random fields for jointly labeling multiple sequences. In NIPS-2003 Workshop on Syntax, Semantics and Statistics.",
"links": null
},
"BIBREF11": {
"ref_id": "b11",
"title": "Efficiently inducing features of conditional random fields",
"authors": [
{
"first": "A",
"middle": [],
"last": "Mccallum",
"suffix": ""
}
],
"year": 2003,
"venue": "Proc. UAI",
"volume": "",
"issue": "",
"pages": "",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "A. McCallum. 2003. Efficiently inducing features of condi- tional random fields. In Proc. UAI 2003.",
"links": null
},
"BIBREF12": {
"ref_id": "b12",
"title": "Ensemble-based active learning for parse selection",
"authors": [
{
"first": "M",
"middle": [],
"last": "Osborne",
"suffix": ""
},
{
"first": "J",
"middle": [],
"last": "Baldridge",
"suffix": ""
}
],
"year": 2004,
"venue": "Proc. NAACL",
"volume": "",
"issue": "",
"pages": "",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "M. Osborne and J. Baldridge. 2004. Ensemble-based active learning for parse selection. In Proc. NAACL 2004.",
"links": null
},
"BIBREF13": {
"ref_id": "b13",
"title": "Accurate information extraction from research papers using conditional random fields",
"authors": [
{
"first": "F",
"middle": [],
"last": "Peng",
"suffix": ""
},
{
"first": "A",
"middle": [],
"last": "Mccallum",
"suffix": ""
}
],
"year": 2004,
"venue": "Proc. HLT-NAACL",
"volume": "",
"issue": "",
"pages": "",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "F. Peng and A. McCallum. 2004. Accurate information extrac- tion from research papers using conditional random fields. In Proc. HLT-NAACL 2004.",
"links": null
},
"BIBREF14": {
"ref_id": "b14",
"title": "Bayesian conditional random fields",
"authors": [
{
"first": "Y",
"middle": [],
"last": "Qi",
"suffix": ""
},
{
"first": "M",
"middle": [],
"last": "Szummer",
"suffix": ""
},
{
"first": "T",
"middle": [
"P"
],
"last": "Minka",
"suffix": ""
}
],
"year": 2005,
"venue": "Proc. AISTATS",
"volume": "",
"issue": "",
"pages": "",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Y. Qi, M. Szummer, and T. P. Minka. 2005. Bayesian condi- tional random fields. In Proc. AISTATS 2005.",
"links": null
},
"BIBREF15": {
"ref_id": "b15",
"title": "Shallow parsing with conditional random fields",
"authors": [
{
"first": "F",
"middle": [],
"last": "Sha",
"suffix": ""
},
{
"first": "F",
"middle": [],
"last": "Pereira",
"suffix": ""
}
],
"year": 2003,
"venue": "Proc. HLT-NAACL",
"volume": "",
"issue": "",
"pages": "",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "F. Sha and F. Pereira. 2003. Shallow parsing with conditional random fields. In Proc. HLT-NAACL 2003.",
"links": null
},
"BIBREF16": {
"ref_id": "b16",
"title": "Introduction to the CoNLL-2000 shared task: Chunking",
"authors": [
{
"first": "E",
"middle": [
"F"
],
"last": "Tjong Kim Sang",
"suffix": ""
},
{
"first": "S",
"middle": [],
"last": "Buchholz",
"suffix": ""
}
],
"year": 2000,
"venue": "Proc. CoNLL",
"volume": "",
"issue": "",
"pages": "",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "E. F. Tjong Kim Sang and S. Buchholz. 2000. Introduction to the CoNLL-2000 shared task: Chunking. In Proc. CoNLL- 2000.",
"links": null
},
"BIBREF17": {
"ref_id": "b17",
"title": "Introduction to the CoNLL-2003 shared task: Language-independent named entity recognition",
"authors": [
{
"first": "E",
"middle": [
"F"
],
"last": "Tjong Kim Sang",
"suffix": ""
},
{
"first": "F. De",
"middle": [],
"last": "Meulder",
"suffix": ""
}
],
"year": 2003,
"venue": "Proc. CoNLL-2003",
"volume": "",
"issue": "",
"pages": "",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "E. F. Tjong Kim Sang and F. De Meulder. 2003. Introduction to the CoNLL-2003 shared task: Language-independent named entity recognition. In Proc. CoNLL-2003.",
"links": null
}
},
"ref_entries": {
"TABREF1": {
"text": "Development set F scores for NER experts",
"html": null,
"type_str": "table",
"num": null,
"content": "<table/>"
},
"TABREF3": {
"text": "F scores for NER unregularised LOP-CRFs",
"html": null,
"type_str": "table",
"num": null,
"content": "<table><tr><td>Expert set</td><td>Development set</td><td>Test set</td></tr><tr><td>Monolithic unreg.</td><td>97.92</td><td>97.65</td></tr><tr><td>Monolithic reg.</td><td>98.02</td><td>97.84</td></tr><tr><td>Simple</td><td>98.31 *</td><td>98.12 *</td></tr><tr><td>Positional</td><td>98.03</td><td>97.81</td></tr><tr><td>Label</td><td>97.99</td><td>97.77</td></tr><tr><td>Random</td><td>97.99</td><td>97.76 \u2020</td></tr></table>"
},
"TABREF4": {
"text": "",
"html": null,
"type_str": "table",
"num": null,
"content": "<table><tr><td>: Accuracies for POS tagging unregularised</td></tr><tr><td>LOP-CRFs</td></tr><tr><td>sense that each expert CRF in the LOP-CRF is un-</td></tr><tr><td>regularised and the LOP weight training is also un-</td></tr><tr><td>regularised. We are therefore not required to search</td></tr><tr><td>a hyperparameter space. As an illustration, to ob-</td></tr><tr><td>tain our best results for the POS tagging regularised</td></tr><tr><td>monolithic model, we re-trained using 15 different</td></tr><tr><td>values of the Gaussian prior variance. With the</td></tr><tr><td>LOP-CRF we trained each expert CRF and the LOP</td></tr><tr><td>weights only once.</td></tr></table>"
}
}
}
} |