File size: 66,795 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 | {
"paper_id": "D13-1013",
"header": {
"generated_with": "S2ORC 1.0.0",
"date_generated": "2023-01-19T16:40:48.850798Z"
},
"title": "Joint Parsing and Disfluency Detection in Linear Time",
"authors": [
{
"first": "Mohammad",
"middle": [
"Sadegh"
],
"last": "Rasooli",
"suffix": "",
"affiliation": {},
"email": "rasooli@cs.columbia.edu"
},
{
"first": "Joel",
"middle": [],
"last": "Tetreault",
"suffix": "",
"affiliation": {},
"email": "joel.tetreault@nuance.com"
}
],
"year": "",
"venue": null,
"identifiers": {},
"abstract": "We introduce a novel method to jointly parse and detect disfluencies in spoken utterances. Our model can use arbitrary features for parsing sentences and adapt itself with out-ofdomain data. We show that our method, based on transition-based parsing, performs at a high level of accuracy for both the parsing and disfluency detection tasks. Additionally, our method is the fastest for the joint task, running in linear time.",
"pdf_parse": {
"paper_id": "D13-1013",
"_pdf_hash": "",
"abstract": [
{
"text": "We introduce a novel method to jointly parse and detect disfluencies in spoken utterances. Our model can use arbitrary features for parsing sentences and adapt itself with out-ofdomain data. We show that our method, based on transition-based parsing, performs at a high level of accuracy for both the parsing and disfluency detection tasks. Additionally, our method is the fastest for the joint task, running in linear time.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Abstract",
"sec_num": null
}
],
"body_text": [
{
"text": "Detecting disfluencies in spontaneous speech has been widely studied by researchers in different communities including natural language processing (e.g. Qian and Liu (2013) ), speech processing (e.g. Wang et al. (2013) ) and psycholinguistics (e.g. Finlayson and Corley (2012) ). While the percentage of spoken words which are disfluent is typically not more than ten percent (Bortfeld et al., 2001 ), this additional \"noise\" makes it much harder for spoken language systems to predict the correct structure of the sentence.",
"cite_spans": [
{
"start": 153,
"end": 172,
"text": "Qian and Liu (2013)",
"ref_id": "BIBREF16"
},
{
"start": 200,
"end": 218,
"text": "Wang et al. (2013)",
"ref_id": "BIBREF17"
},
{
"start": 249,
"end": 276,
"text": "Finlayson and Corley (2012)",
"ref_id": "BIBREF4"
},
{
"start": 376,
"end": 398,
"text": "(Bortfeld et al., 2001",
"ref_id": "BIBREF1"
}
],
"ref_spans": [],
"eq_spans": [],
"section": "Introduction",
"sec_num": "1"
},
{
"text": "Disfluencies can be filled pauses (e.g. \"uh\", \"um\", \"huh\"), discourse markers (e.g. \"you know\", \"I mean\") or edited words which are repeated or corrected by the speaker. For example, in the following sentence, an edited phrase or reparandum interval (\"to Boston\") occurs with its repair (\"to Denver\"), a filled pause (\"uh\") and discourse marker (\"I mean\" Filled pauses and discourse markers are to some extent a fixed and closed set. The main challenge in finding disfluencies is the case where the edited phrase is neither a rough copy of its repair or has any repair phrase (i.e. discarded edited phrase). Hence, in previous work, researchers report their method performance on detecting edited phrases (reparandum) (Johnson and Charniak, 2004) .",
"cite_spans": [
{
"start": 718,
"end": 746,
"text": "(Johnson and Charniak, 2004)",
"ref_id": "BIBREF7"
}
],
"ref_spans": [],
"eq_spans": [],
"section": "Introduction",
"sec_num": "1"
},
{
"text": "In contrast to most previous work which focuses solely on either detection or on parsing, we introduce a novel framework for jointly parsing sentences with disfluencies. To our knowledge, our work is the first model that is based on joint dependency and disfluency detection. We show that our model is robust enough to detect disfluencies with high accuracy, while still maintaining a high level of dependency parsing accuracy that approaches the upper bound. Additionally, our model outperforms prior work on joint parsing and disfluency detection on the disfluency detection task, and improves upon this prior work by running in linear time complexity.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Introduction",
"sec_num": "1"
},
{
"text": "The remainder of this paper is as follows. In \u00a72, we overview some the previous work on disfluency detection. \u00a73 describes our model. Experiments are described in \u00a74 and Conclusions are made in \u00a75.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Introduction",
"sec_num": "1"
},
{
"text": "Disfluency detection approaches can be divided into two different groups: text-first and speech first (Nakatani and Hirschberg, 1993) . In the first approach, all prosodic and acoustic cues are ignored while in the second approach both grammatical and acoustic features are considered. For this paper, we focus on developing a text-first approach but our model is easily flexible with speech-first features because there is no restriction on the number and types of features in our model.",
"cite_spans": [
{
"start": 102,
"end": 133,
"text": "(Nakatani and Hirschberg, 1993)",
"ref_id": "BIBREF12"
}
],
"ref_spans": [],
"eq_spans": [],
"section": "Related Work",
"sec_num": "2"
},
{
"text": "Among text-first approaches, the work is split between developing systems which focus specifically on disfluency detection and those which couple disfluency detection with parsing. For the former, Charniak and Johnson (2001) employ a linear classifier to predict the edited phrases in Switchboard corpus (Godfrey et al., 1992) . Johnson and Charniak (2004) use a TAG-based noisy channel model to detect disfluencies while parsing with getting nbest parses from each sentence and re-ranking with a language model. The original TAG parser is not used for parsing itself and it is used just to find rough copies in the sentence. Their method achieves promising results on detecting edited words but at the expense of speed (the parser has a complexity of O(N 5 ). Kahn et al. (2005) use the same TAG model and add semi-automatically extracted prosodic features. Zwarts and Johnson (2011) improve the performance of TAG model by adding external language modeling information from data sets such as Gigaword in addition to using minimal expected Floss in n-best re-ranking. Georgila (2009) uses integer linear programming combined with CRF for learning disfluencies. That work shows that ILP can learn local and global constraints to improve the performance significantly. Qian and Liu (2013) achieve the best performance on the Switchboard corpus (Godfrey et al., 1992) without any additional data. They use three steps for detecting disfluencies using weighted Max-Margin Markov (M 3 ) network: detecting fillers, detecting edited words, and refining errors in previous steps.",
"cite_spans": [
{
"start": 197,
"end": 224,
"text": "Charniak and Johnson (2001)",
"ref_id": "BIBREF2"
},
{
"start": 304,
"end": 326,
"text": "(Godfrey et al., 1992)",
"ref_id": "BIBREF6"
},
{
"start": 329,
"end": 356,
"text": "Johnson and Charniak (2004)",
"ref_id": "BIBREF7"
},
{
"start": 761,
"end": 779,
"text": "Kahn et al. (2005)",
"ref_id": "BIBREF8"
},
{
"start": 859,
"end": 884,
"text": "Zwarts and Johnson (2011)",
"ref_id": "BIBREF19"
},
{
"start": 1069,
"end": 1084,
"text": "Georgila (2009)",
"ref_id": "BIBREF5"
},
{
"start": 1268,
"end": 1287,
"text": "Qian and Liu (2013)",
"ref_id": "BIBREF16"
},
{
"start": 1343,
"end": 1365,
"text": "(Godfrey et al., 1992)",
"ref_id": "BIBREF6"
}
],
"ref_spans": [],
"eq_spans": [],
"section": "Related Work",
"sec_num": "2"
},
{
"text": "Some text-first approaches treat parsing and disfluency detection jointly, though the models differ in the type of parse formalism employed. Lease and Johnson (2006) use a PCFG-based parser to parse sentences along with finding edited phrases. Miller and Schuler (2008) use a right-corner transform of binary branching structures on bracketed sentences but their results are much worse than (Johnson and Charniak, 2004) . To date, none of the prior joint approaches have used a dependency formalism.",
"cite_spans": [
{
"start": 141,
"end": 165,
"text": "Lease and Johnson (2006)",
"ref_id": "BIBREF9"
},
{
"start": 244,
"end": 269,
"text": "Miller and Schuler (2008)",
"ref_id": "BIBREF11"
},
{
"start": 391,
"end": 419,
"text": "(Johnson and Charniak, 2004)",
"ref_id": "BIBREF7"
}
],
"ref_spans": [],
"eq_spans": [],
"section": "Related Work",
"sec_num": "2"
},
{
"text": "We model the problem using a deterministic transition-based parser (Nivre, 2008) . These parsers have the advantage of being very accurate while being able to parse a sentence in linear time. An additional advantage is that they can use as many nonlocal and local features as needed.",
"cite_spans": [
{
"start": 67,
"end": 80,
"text": "(Nivre, 2008)",
"ref_id": "BIBREF15"
}
],
"ref_spans": [],
"eq_spans": [],
"section": "Joint Parsing Model",
"sec_num": "3"
},
{
"text": "Arc-Eager Algorithm We use the arc-eager algorithm (Nivre, 2004) which is a bottom-up parsing strategy that is used in greedy and k-beam transitionbased parsers. One advantage of this strategy is that the words can get a head from their left side, before getting right dependents. This is particularly beneficial for our task, since we know that reparanda are similar to their repairs. Hence, a reparandum may get its head but whenever the parser faces a repair, it removes the reparandum from the sentence and continues its actions.",
"cite_spans": [
{
"start": 51,
"end": 64,
"text": "(Nivre, 2004)",
"ref_id": "BIBREF14"
}
],
"ref_spans": [],
"eq_spans": [],
"section": "Joint Parsing Model",
"sec_num": "3"
},
{
"text": "The actions in an arc-eager parsing algorithm are:",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Joint Parsing Model",
"sec_num": "3"
},
{
"text": "\u2022 Left-arc (LA): The first word in the buffer becomes the head of the top word in the stack. The top word is popped after this action. Figure 1 : A sample transition sequence for the sentence \"flight to Boston uh I mean to Denver\". In the third column, only the underlined parse actions are learned by the parser (second classifier). The first classifier uses all instances for training (learns fluent words with \"regular\" label).",
"cite_spans": [],
"ref_spans": [
{
"start": 135,
"end": 143,
"text": "Figure 1",
"ref_id": null
}
],
"eq_spans": [],
"section": "Joint Parsing Model",
"sec_num": "3"
},
{
"text": "\u2022 Interjection (Intj[i]): treat a phrase in the look-ahead buffer (first i words) as a filled pause and remove them from the sentence. 2",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Joint Parsing Model",
"sec_num": "3"
},
{
"text": "Our model has two classifiers. The first classifier decides between four possible actions and possible candidates in the current configuration of the sentence. These actions are the three new ones from above and a new action Regular (Reg): which means do one of the original arc-eager parser actions.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Joint Parsing Model",
"sec_num": "3"
},
{
"text": "At each configuration, there might be several candidates for being a prn, intj or reparandum, and one regular candidate. The candidates for being a reparandum are a set of words outside the lookahead buffer and the candidates for being an intj or prn are a set of words beginning from the head of the look-ahead buffer. If the parser decides regular as the correct action, the second classifier predicts the best parsing transition, based on arc-eager parsing (Nivre, 2004) .",
"cite_spans": [
{
"start": 460,
"end": 473,
"text": "(Nivre, 2004)",
"ref_id": "BIBREF14"
}
],
"ref_spans": [],
"eq_spans": [],
"section": "Joint Parsing Model",
"sec_num": "3"
},
{
"text": "For example, in the 4th state in Figure 1 Training A transition-based parser action (our second-level classifier) is sensitive to the words in the buffer and stack. The problem is that we do not have gold dependencies for edited words in our data. Therefore, we need a parser to remove reparandum words from the buffer and push them into the stack. Since our parser cannot be trained on disfluent sentences from scratch, the first step is to train it on clean treebank data.",
"cite_spans": [],
"ref_spans": [
{
"start": 33,
"end": 41,
"text": "Figure 1",
"ref_id": null
}
],
"eq_spans": [],
"section": "Joint Parsing Model",
"sec_num": "3"
},
{
"text": "In the second step, we adapt parser weights by training it on disfluent sentences. Our assumption is that we do not know the correct dependencies between disfluent words and other words in the sentence. At each configuration, the parser updates itself with new instances by traversing all configurations in the sentences. In this case, if at the head of the buffer there is an intj or prn tag, the parser allows them to be removed from the buffer. If a reparandum word is not completely outside the buffer (the first two states in Figure 1 ), the parser decides between the four regular arc-eager actions (i.e. leftarc, right-arc, shift, and reduce). If the last word pushed into the stack is a reparandum and the first word in the buffer is a regular word, the parser removes all reparanda at the same level (in the case of nested edited words), removes their dependencies to other words and push their dependents into the stack. Otherwise, the parser performs the oracle action and adds that action as its new instance. 3 With an adapted parser which is our second-level classifier, we can train our first-level classifier. The same procedure repeats, except that instances for disfluency detection are used for updating parameter weights for the first classifier for deciding the actions. In Figure 1 , only the oracle actions (underlined) are added to the instances for updating parser weights but all first-level actions are learned by the first level classifier.",
"cite_spans": [
{
"start": 1022,
"end": 1023,
"text": "3",
"ref_id": null
}
],
"ref_spans": [
{
"start": 531,
"end": 539,
"text": "Figure 1",
"ref_id": null
},
{
"start": 1295,
"end": 1303,
"text": "Figure 1",
"ref_id": null
}
],
"eq_spans": [],
"section": "Joint Parsing Model",
"sec_num": "3"
},
{
"text": "For our experiments, we use the Switchboard corpus (Godfrey et al., 1992) with the same train/dev/test split as Johnson and Charniak (2004) . As in that work, incomplete words and punctuations are removed from data (except that we do not remove incomplete words that are not disfluent 4 ) and all words are turned into lower-case. The main difference with previous work is that we use Switchboard mrg files for training and testing our model (since they contain parse trees) instead of the more commonly used Swithboard dps text files. Mrg files are a subset of dps files with about more than half of their size. Unfortunately, the disfluencies marked in the dps files are not exactly the same as those marked in the corresponding mrg files. Hence, our result is not completely comparable to previous work except for (Kahn et al., 2005; Lease and Johnson, 2006; Miller and Schuler, 2008) .",
"cite_spans": [
{
"start": 51,
"end": 73,
"text": "(Godfrey et al., 1992)",
"ref_id": "BIBREF6"
},
{
"start": 112,
"end": 139,
"text": "Johnson and Charniak (2004)",
"ref_id": "BIBREF7"
},
{
"start": 817,
"end": 836,
"text": "(Kahn et al., 2005;",
"ref_id": "BIBREF8"
},
{
"start": 837,
"end": 861,
"text": "Lease and Johnson, 2006;",
"ref_id": "BIBREF9"
},
{
"start": 862,
"end": 887,
"text": "Miller and Schuler, 2008)",
"ref_id": "BIBREF11"
}
],
"ref_spans": [],
"eq_spans": [],
"section": "Experiments and Evaluation",
"sec_num": "4"
},
{
"text": "We use Tsurgeon (Levy and Andrew, 2006) for extracting sentences from mrg files and use the Penn2Malt tool 5 to convert them to dependencies. Afterwards, we provide dependency trees with disfluent words being the dependent of nothing.",
"cite_spans": [
{
"start": 16,
"end": 39,
"text": "(Levy and Andrew, 2006)",
"ref_id": "BIBREF10"
}
],
"ref_spans": [],
"eq_spans": [],
"section": "Experiments and Evaluation",
"sec_num": "4"
},
{
"text": "Learning For the first classifier, we use averaged structured Perceptron (AP) (Collins, 2002) with a minor modification. Since the first classifier data is heavily biased towards the \"regular label\", we modify the weight updates in the original algorithm to 2 (original is 1) for the cases where a \"reparandum\" is wrongly recognized as another label. We call the modified version \"weighted averaged Perceptron (WAP)\". We see that this simple modification improves the model accuracy. 6 For the second classifier (parser), we use the original averaged structured Perceptron algorithm. We report results on both AP and WAP versions of the parser.",
"cite_spans": [
{
"start": 78,
"end": 93,
"text": "(Collins, 2002)",
"ref_id": "BIBREF3"
}
],
"ref_spans": [],
"eq_spans": [],
"section": "Experiments and Evaluation",
"sec_num": "4"
},
{
"text": "Features Since for every state in the parser configuration, there are many candidates for being disfluent; we use local features as well as global features for the first classifier. Global features are mostly useful for discriminating between the four actions and local features are mostly useful for choosing a phrase as a candidate for being a disfluent phrase. The features are described in Figure 2 . For the second classifier, we use the same features as (Zhang and Nivre, 2011 , Table 1 ) except that we train our research/Penn2Malt.html 6 This is similar to WM 3 N in (Qian and Liu, 2013) .",
"cite_spans": [
{
"start": 460,
"end": 482,
"text": "(Zhang and Nivre, 2011",
"ref_id": "BIBREF18"
},
{
"start": 575,
"end": 595,
"text": "(Qian and Liu, 2013)",
"ref_id": "BIBREF16"
}
],
"ref_spans": [
{
"start": 394,
"end": 402,
"text": "Figure 2",
"ref_id": null
},
{
"start": 483,
"end": 492,
"text": ", Table 1",
"ref_id": null
}
],
"eq_spans": [],
"section": "Experiments and Evaluation",
"sec_num": "4"
},
{
"text": "First n words inside/outside buffer (n=1:4) First n POS i/o buffer (n=1:6) Are n words i/o buffer equal? (n=1:4) Are n POS i/o buffer equal? (n=1:4) n last FG transitions (n=1:5) n last transitions (n=1:5) n last FG transitions + first POS in the buffer (n=1:5) n last transitions + first POS in the buffer (n=1:5) (n+m)-gram of m/n POS i/o buffer (n,m=1:4) Refined (n+m)-gram of m/n POS i/o buffer (n,m=1:4) Are n first words of i/o buffer equal? (n=1:4) Are n first POS of i/o buffer equal? (n=1:4) Number of common words i/o buffer words (n=1:6) Local Features First n words of the candidate phrase (n=1:4) First n POS of the candidate phrase (n=1:6) Distance between the candidate and first word in the buffer Figure 2 : Features used for learning the first classifier.",
"cite_spans": [],
"ref_spans": [
{
"start": 714,
"end": 722,
"text": "Figure 2",
"ref_id": null
}
],
"eq_spans": [],
"section": "Global Features",
"sec_num": null
},
{
"text": "Refined n-gram is the n-gram without considering words that are recognized as disfluent. Fine-grained (FG) transitions are enriched with parse actions (e.g. \"regular:leftarc\").",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Global Features",
"sec_num": null
},
{
"text": "parser in a similar manner as the MaltParser (Nivre et al., 2007) without k-beam training.",
"cite_spans": [
{
"start": 45,
"end": 65,
"text": "(Nivre et al., 2007)",
"ref_id": "BIBREF13"
}
],
"ref_spans": [],
"eq_spans": [],
"section": "Global Features",
"sec_num": null
},
{
"text": "Parser Evaluation We evaluate our parser with both unlabeled attachment accuracy of correct words and precision and recall of finding the dependencies of correct words. 7 The second classifier is trained with 3 iterations in the first step and 3 iterations in the second step. We use the attachment accuracy of the parse tree of the correct sentences (without disfluencies) as the upper-bound attachment score and parsed tree of the disfluent sentences (without disfluency detection) as our lower-bound attachment score. As we can see in Table 1 , WAP does a slightly better job parsing sentences. The upper-bound parsing accuracy shows that we do not lose too much information while jointly detecting disfluencies. Our parser is not comparable to (Johnson and Charniak, 2004) and (Miller and Schuler, 2008) , since we use dependency relations for evaluation instead of constituencies. Kahn et al., 2005) , LJ = (Lease and Johnson, 2006) , MS = (Miller and Schuler, 2008) and QL = (Qian and Liu, 2013) .",
"cite_spans": [
{
"start": 169,
"end": 170,
"text": "7",
"ref_id": null
},
{
"start": 748,
"end": 776,
"text": "(Johnson and Charniak, 2004)",
"ref_id": "BIBREF7"
},
{
"start": 781,
"end": 807,
"text": "(Miller and Schuler, 2008)",
"ref_id": "BIBREF11"
},
{
"start": 886,
"end": 904,
"text": "Kahn et al., 2005)",
"ref_id": "BIBREF8"
},
{
"start": 912,
"end": 937,
"text": "(Lease and Johnson, 2006)",
"ref_id": "BIBREF9"
},
{
"start": 945,
"end": 971,
"text": "(Miller and Schuler, 2008)",
"ref_id": "BIBREF11"
},
{
"start": 981,
"end": 1001,
"text": "(Qian and Liu, 2013)",
"ref_id": "BIBREF16"
}
],
"ref_spans": [
{
"start": 538,
"end": 545,
"text": "Table 1",
"ref_id": null
}
],
"eq_spans": [],
"section": "Global Features",
"sec_num": null
},
{
"text": "(words with \"EDITED\" tag in mrg files). As we see in Table 2 , WAP works better than the original method. As mentioned before, the numbers are not completely comparable to others except for (Kahn et al., 2005; Lease and Johnson, 2006; Miller and Schuler, 2008) which we outperform. For the sake of comparing to the state of the art, the best result for this task (Qian and Liu, 2013) is replicated from their available software 8 on the portion of dps files that have corresponding mrg files. For a fairer comparison, we also optimized the number of training iterations of (Qian and Liu, 2013) for the mrg set based on dev data (10 iterations instead of 30 iterations). As shown in the results, our model accuracy is slightly less than the state-of-the-art (which focuses solely on the disfluency detection task and does no parsing), but we believe that the performance can be improved through better features and by changing the model. Another characteristic of our model is that it operates at a very high precision, though at the expense of some recall.",
"cite_spans": [
{
"start": 190,
"end": 209,
"text": "(Kahn et al., 2005;",
"ref_id": "BIBREF8"
},
{
"start": 210,
"end": 234,
"text": "Lease and Johnson, 2006;",
"ref_id": "BIBREF9"
},
{
"start": 235,
"end": 260,
"text": "Miller and Schuler, 2008)",
"ref_id": "BIBREF11"
},
{
"start": 363,
"end": 383,
"text": "(Qian and Liu, 2013)",
"ref_id": "BIBREF16"
},
{
"start": 573,
"end": 593,
"text": "(Qian and Liu, 2013)",
"ref_id": "BIBREF16"
}
],
"ref_spans": [
{
"start": 53,
"end": 60,
"text": "Table 2",
"ref_id": "TABREF3"
}
],
"eq_spans": [],
"section": "Disfluency Detection Evaluation",
"sec_num": null
},
{
"text": "google.com/p/disfluency-detection/. Results from the first version are 81.4 and 82.1 for the default and optimized settings.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Disfluency Detection Evaluation",
"sec_num": null
},
{
"text": "In this paper, we have developed a fast, yet accurate, joint dependency parsing and disfluency detection model. Such a parser is useful for spoken dialogue systems which typically encounter disfluent speech and require accurate syntactic structures. The model is completely flexible with adding other features (either text or speech features).",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Conclusion",
"sec_num": "5"
},
{
"text": "There are still many ways of improving this framework such as using k-beam training and decoding, using prosodic and acoustic features, using out of domain data for improving the language and parsing models, and merging the two classifiers into one through better feature engineering. It is worth noting that we put the dummy root word in the first position of the sentence. Ballesteros and Nivre (2013) show that parser accuracy can improve by changing that position for English.",
"cite_spans": [
{
"start": 375,
"end": 403,
"text": "Ballesteros and Nivre (2013)",
"ref_id": "BIBREF0"
}
],
"ref_spans": [],
"eq_spans": [],
"section": "Conclusion",
"sec_num": "5"
},
{
"text": "One of the main challenges in this problem is that most of the training instances are not disfluent and thus the sample space is very sparse. As seen in the experiments, we can get further improvements by modifying the weight updates in the Perceptron learner. In future work, we will explore different learning algorithms which can help us address the sparsity problem and improve the model accuracy. Another challenge is related to the parser speed, since the number of candidates and features are much greater than the number used in classical dependency parsers.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Conclusion",
"sec_num": "5"
},
{
"text": "In the literature, edited words are also known as \"reparandum\", and the fillers are known as \"interregnum\". Filled pauses are also called \"Interjections\".",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "",
"sec_num": null
},
{
"text": "In the bracketed version of Switchboard corpus, reparandum is tagged with EDITED and discourse markers and paused fillers are tagged as PRN and INTJ respectively.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "",
"sec_num": null
},
{
"text": "The reason that we use a parser instead of expanding all possible transitions for an edited word is that, the number of regular actions will increase and the other actions become sparser than natural.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "",
"sec_num": null
},
{
"text": "E.g. I want t-go to school. 5 http://stp.lingfil.uu.se/\u02dcnivre/",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "",
"sec_num": null
},
{
"text": "The parser is actually trained to do labeled attachment and labeled accuracy is about 1-1.5% lower than UAS.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "",
"sec_num": null
},
{
"text": "We use the second version of the code: http://code.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "",
"sec_num": null
}
],
"back_matter": [
{
"text": "We would like to thank anonymous reviewers for their helpful comments on the paper. Additionally, we were aided by researchers by their prompt responses to our many questions: Mark Core, Luciana Ferrer, Kallirroi Georgila, Mark Johnson, Jeremy Kahn, Yang Liu, Xian Qian, Kenji Sagae, and Wen Wang. Finally, this work was conducted during the first author's summer internship at the Nuance Sunnyvale Research Lab. We would like to thank the researchers in the group for the helpful discussions and assistance on different aspects of the problem. In particular, we would like to thank Chris Brew, Ron Kaplan, Deepak Ramachandran and Adwait Ratnaparkhi.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Acknowledgements",
"sec_num": null
}
],
"bib_entries": {
"BIBREF0": {
"ref_id": "b0",
"title": "Going to the roots of dependency parsing",
"authors": [
{
"first": "Miguel",
"middle": [],
"last": "Ballesteros",
"suffix": ""
},
{
"first": "Joakim",
"middle": [],
"last": "Nivre",
"suffix": ""
}
],
"year": 2013,
"venue": "Computational Linguistics",
"volume": "39",
"issue": "1",
"pages": "5--13",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Miguel Ballesteros and Joakim Nivre. 2013. Going to the roots of dependency parsing. Computational Lin- guistics, 39(1):5-13.",
"links": null
},
"BIBREF1": {
"ref_id": "b1",
"title": "Disfluency rates in conversation: Effects of age, relationship, topic, role, and gender",
"authors": [
{
"first": "Heather",
"middle": [],
"last": "Bortfeld",
"suffix": ""
},
{
"first": "Silvia",
"middle": [
"D"
],
"last": "Leon",
"suffix": ""
},
{
"first": "Jonathan",
"middle": [
"E"
],
"last": "Bloom",
"suffix": ""
},
{
"first": "Michael",
"middle": [
"F"
],
"last": "Schober",
"suffix": ""
},
{
"first": "Susan",
"middle": [
"E"
],
"last": "Brennan",
"suffix": ""
}
],
"year": 2001,
"venue": "Language and Speech",
"volume": "44",
"issue": "2",
"pages": "123--147",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Heather Bortfeld, Silvia D. Leon, Jonathan E. Bloom, Michael F. Schober, and Susan E. Brennan. 2001. Disfluency rates in conversation: Effects of age, re- lationship, topic, role, and gender. Language and Speech, 44(2):123-147.",
"links": null
},
"BIBREF2": {
"ref_id": "b2",
"title": "Edit detection and parsing for transcribed speech",
"authors": [
{
"first": "Eugene",
"middle": [],
"last": "Charniak",
"suffix": ""
},
{
"first": "Mark",
"middle": [],
"last": "Johnson",
"suffix": ""
}
],
"year": 2001,
"venue": "NAACL-HLT",
"volume": "",
"issue": "",
"pages": "1--9",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Eugene Charniak and Mark Johnson. 2001. Edit detec- tion and parsing for transcribed speech. In NAACL- HLT, pages 1-9.",
"links": null
},
"BIBREF3": {
"ref_id": "b3",
"title": "Discriminative training methods for hidden markov models: Theory and experiments with perceptron algorithms",
"authors": [
{
"first": "Michael",
"middle": [],
"last": "Collins",
"suffix": ""
}
],
"year": 2002,
"venue": "ACL",
"volume": "",
"issue": "",
"pages": "1--8",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Michael Collins. 2002. Discriminative training methods for hidden markov models: Theory and experiments with perceptron algorithms. In ACL, pages 1-8.",
"links": null
},
"BIBREF4": {
"ref_id": "b4",
"title": "Disfluency in dialogue: an intentional signal from the speaker? Psychonomic bulletin & review",
"authors": [
{
"first": "Ian",
"middle": [
"R"
],
"last": "Finlayson",
"suffix": ""
},
{
"first": "Martin",
"middle": [],
"last": "Corley",
"suffix": ""
}
],
"year": 2012,
"venue": "",
"volume": "19",
"issue": "",
"pages": "921--928",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Ian R. Finlayson and Martin Corley. 2012. Disfluency in dialogue: an intentional signal from the speaker? Psychonomic bulletin & review, 19(5):921-928.",
"links": null
},
"BIBREF5": {
"ref_id": "b5",
"title": "Using integer linear programming for detecting speech disfluencies",
"authors": [
{
"first": "Kallirroi",
"middle": [],
"last": "Georgila",
"suffix": ""
}
],
"year": 2009,
"venue": "NAACL-HLT",
"volume": "",
"issue": "",
"pages": "109--112",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Kallirroi Georgila. 2009. Using integer linear program- ming for detecting speech disfluencies. In NAACL- HLT, pages 109-112.",
"links": null
},
"BIBREF6": {
"ref_id": "b6",
"title": "Switchboard: Telephone speech corpus for research and development",
"authors": [
{
"first": "John",
"middle": [
"J"
],
"last": "Godfrey",
"suffix": ""
},
{
"first": "Edward",
"middle": [
"C"
],
"last": "Holliman",
"suffix": ""
},
{
"first": "Jane",
"middle": [],
"last": "Mc-Daniel",
"suffix": ""
}
],
"year": 1992,
"venue": "ICASSP",
"volume": "1",
"issue": "",
"pages": "517--520",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "John J. Godfrey, Edward C. Holliman, and Jane Mc- Daniel. 1992. Switchboard: Telephone speech corpus for research and development. In ICASSP, volume 1, pages 517-520.",
"links": null
},
"BIBREF7": {
"ref_id": "b7",
"title": "A tag-based noisy channel model of speech repairs",
"authors": [
{
"first": "Mark",
"middle": [],
"last": "Johnson",
"suffix": ""
},
{
"first": "Eugene",
"middle": [],
"last": "Charniak",
"suffix": ""
}
],
"year": 2004,
"venue": "ACL",
"volume": "",
"issue": "",
"pages": "33--39",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Mark Johnson and Eugene Charniak. 2004. A tag-based noisy channel model of speech repairs. In ACL, pages 33-39.",
"links": null
},
"BIBREF8": {
"ref_id": "b8",
"title": "Effective use of prosody in parsing conversational speech",
"authors": [
{
"first": "Jeremy",
"middle": [
"G"
],
"last": "Kahn",
"suffix": ""
},
{
"first": "Matthew",
"middle": [],
"last": "Lease",
"suffix": ""
},
{
"first": "Eugene",
"middle": [],
"last": "Charniak",
"suffix": ""
},
{
"first": "Mark",
"middle": [],
"last": "Johnson",
"suffix": ""
},
{
"first": "Mari",
"middle": [],
"last": "Ostendorf",
"suffix": ""
}
],
"year": 2005,
"venue": "EMNLP",
"volume": "",
"issue": "",
"pages": "233--240",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Jeremy G. Kahn, Matthew Lease, Eugene Charniak, Mark Johnson, and Mari Ostendorf. 2005. Effective use of prosody in parsing conversational speech. In EMNLP, pages 233-240.",
"links": null
},
"BIBREF9": {
"ref_id": "b9",
"title": "Early deletion of fillers in processing conversational speech",
"authors": [
{
"first": "Matthew",
"middle": [],
"last": "Lease",
"suffix": ""
},
{
"first": "Mark",
"middle": [],
"last": "Johnson",
"suffix": ""
}
],
"year": 2006,
"venue": "NAACL-HLT",
"volume": "",
"issue": "",
"pages": "73--76",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Matthew Lease and Mark Johnson. 2006. Early dele- tion of fillers in processing conversational speech. In NAACL-HLT, pages 73-76.",
"links": null
},
"BIBREF10": {
"ref_id": "b10",
"title": "Tregex and tsurgeon: tools for querying and manipulating tree data structures",
"authors": [
{
"first": "Roger",
"middle": [],
"last": "Levy",
"suffix": ""
},
{
"first": "Galen",
"middle": [],
"last": "Andrew",
"suffix": ""
}
],
"year": 2006,
"venue": "LREC",
"volume": "",
"issue": "",
"pages": "2231--2234",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Roger Levy and Galen Andrew. 2006. Tregex and tsur- geon: tools for querying and manipulating tree data structures. In LREC, pages 2231-2234.",
"links": null
},
"BIBREF11": {
"ref_id": "b11",
"title": "A unified syntactic model for parsing fluent and disfluent speech",
"authors": [
{
"first": "Tim",
"middle": [],
"last": "Miller",
"suffix": ""
},
{
"first": "William",
"middle": [],
"last": "Schuler",
"suffix": ""
}
],
"year": 2008,
"venue": "ACL-HLT",
"volume": "",
"issue": "",
"pages": "105--108",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Tim Miller and William Schuler. 2008. A unified syn- tactic model for parsing fluent and disfluent speech. In ACL-HLT, pages 105-108.",
"links": null
},
"BIBREF12": {
"ref_id": "b12",
"title": "A speech-first model for repair detection and correction",
"authors": [
{
"first": "Christine",
"middle": [],
"last": "Nakatani",
"suffix": ""
},
{
"first": "Julia",
"middle": [],
"last": "Hirschberg",
"suffix": ""
}
],
"year": 1993,
"venue": "ACL",
"volume": "",
"issue": "",
"pages": "46--53",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Christine Nakatani and Julia Hirschberg. 1993. A speech-first model for repair detection and correction. In ACL, pages 46-53.",
"links": null
},
"BIBREF13": {
"ref_id": "b13",
"title": "Maltparser: A languageindependent system for data-driven dependency parsing",
"authors": [
{
"first": "Joakim",
"middle": [],
"last": "Nivre",
"suffix": ""
},
{
"first": "Johan",
"middle": [],
"last": "Hall",
"suffix": ""
},
{
"first": "Jens",
"middle": [],
"last": "Nilsson",
"suffix": ""
},
{
"first": "Atanas",
"middle": [],
"last": "Chanev",
"suffix": ""
},
{
"first": "G\u00fclsen",
"middle": [],
"last": "Eryigit",
"suffix": ""
},
{
"first": "Sandra",
"middle": [],
"last": "K\u00fcbler",
"suffix": ""
},
{
"first": "Svetoslav",
"middle": [],
"last": "Marinov",
"suffix": ""
},
{
"first": "Erwin",
"middle": [],
"last": "Marsi",
"suffix": ""
}
],
"year": 2007,
"venue": "Natural Language Engineering",
"volume": "13",
"issue": "2",
"pages": "95--135",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Joakim Nivre, Johan Hall, Jens Nilsson, Atanas Chanev, G\u00fclsen Eryigit, Sandra K\u00fcbler, Svetoslav Marinov, and Erwin Marsi. 2007. Maltparser: A language- independent system for data-driven dependency pars- ing. Natural Language Engineering, 13(2):95-135.",
"links": null
},
"BIBREF14": {
"ref_id": "b14",
"title": "Incrementality in deterministic dependency parsing",
"authors": [
{
"first": "Joakim",
"middle": [],
"last": "Nivre",
"suffix": ""
}
],
"year": 2004,
"venue": "the Workshop on Incremental Parsing: Bringing Engineering and Cognition Together",
"volume": "",
"issue": "",
"pages": "50--57",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Joakim Nivre. 2004. Incrementality in deterministic dependency parsing. In the Workshop on Incremen- tal Parsing: Bringing Engineering and Cognition To- gether, pages 50-57.",
"links": null
},
"BIBREF15": {
"ref_id": "b15",
"title": "Algorithms for deterministic incremental dependency parsing",
"authors": [
{
"first": "Joakim",
"middle": [],
"last": "Nivre",
"suffix": ""
}
],
"year": 2008,
"venue": "Computational Linguistics",
"volume": "34",
"issue": "4",
"pages": "513--553",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Joakim Nivre. 2008. Algorithms for deterministic incre- mental dependency parsing. Computational Linguis- tics, 34(4):513-553.",
"links": null
},
"BIBREF16": {
"ref_id": "b16",
"title": "Disfluency detection using multi-step stacked learning",
"authors": [
{
"first": "Xian",
"middle": [],
"last": "Qian",
"suffix": ""
},
{
"first": "Yang",
"middle": [],
"last": "Liu",
"suffix": ""
}
],
"year": 2013,
"venue": "NAACL-HLT",
"volume": "",
"issue": "",
"pages": "820--825",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Xian Qian and Yang Liu. 2013. Disfluency detection using multi-step stacked learning. In NAACL-HLT, pages 820-825.",
"links": null
},
"BIBREF17": {
"ref_id": "b17",
"title": "A cross-language study on automatic speech disfluency detection",
"authors": [
{
"first": "Wen",
"middle": [],
"last": "Wang",
"suffix": ""
},
{
"first": "Andreas",
"middle": [],
"last": "Stolcke",
"suffix": ""
},
{
"first": "Jiahong",
"middle": [],
"last": "Yuan",
"suffix": ""
},
{
"first": "Mark",
"middle": [],
"last": "Liberman",
"suffix": ""
}
],
"year": 2013,
"venue": "NAACL-HLT",
"volume": "",
"issue": "",
"pages": "703--708",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Wen Wang, Andreas Stolcke, Jiahong Yuan, and Mark Liberman. 2013. A cross-language study on auto- matic speech disfluency detection. In NAACL-HLT, pages 703-708.",
"links": null
},
"BIBREF18": {
"ref_id": "b18",
"title": "Transition-based dependency parsing with rich non-local features",
"authors": [
{
"first": "Yue",
"middle": [],
"last": "Zhang",
"suffix": ""
},
{
"first": "Joakim",
"middle": [],
"last": "Nivre",
"suffix": ""
}
],
"year": 2011,
"venue": "ACL (Short Papers)",
"volume": "",
"issue": "",
"pages": "188--193",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Yue Zhang and Joakim Nivre. 2011. Transition-based dependency parsing with rich non-local features. In ACL (Short Papers), pages 188-193.",
"links": null
},
"BIBREF19": {
"ref_id": "b19",
"title": "The impact of language models and loss functions on repair disfluency detection",
"authors": [
{
"first": "Simon",
"middle": [],
"last": "Zwarts",
"suffix": ""
},
{
"first": "Mark",
"middle": [],
"last": "Johnson",
"suffix": ""
}
],
"year": 2011,
"venue": "ACL",
"volume": "",
"issue": "",
"pages": "703--711",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Simon Zwarts and Mark Johnson. 2011. The impact of language models and loss functions on repair disflu- ency detection. In ACL, pages 703-711.",
"links": null
}
},
"ref_entries": {
"FIGREF0": {
"type_str": "figure",
"num": null,
"uris": null,
"text": "\u2022 Right-arc (RA): The top word in the stack becomes the head of the first word in the buffer.\u2022 Reduce (R): The top word in the stack is popped. \u2022 Shift (SH): The first word in the buffer goes to the top of the stack."
},
"FIGREF1": {
"type_str": "figure",
"num": null,
"uris": null,
"text": ", there are multiple candidates for the first classifier: regular, \"I\" as prn[1] or intj[1], \"I mean\" as prn[2] or intj[2], \"I mean to\" as prn[3] or intj[3], \"I mean to Denver\" as prn[4] or intj[4], \"Boston\" as rp[3:3], \"to Boston\" as rp[2:3], and \"flight to Boston\" as rp[1:3]."
},
"TABREF2": {
"type_str": "table",
"text": "Table 1: Parsing results. UB = upperbound (parsing clean sentences), LB = lowerbound (parsing disfluent sentences without disfluency correction). UAS is unlabeled attachment score (accuracy), Pr. is precision, Rec. is recall and F1 is f-score.",
"html": null,
"content": "<table><tr><td/><td>UAS LB</td><td>UB</td><td colspan=\"2\">Pr. Rec. F2</td></tr><tr><td>AP</td><td colspan=\"4\">88.6 70.7 90.2 86.8 88.0 87.4</td></tr><tr><td colspan=\"5\">WAP 88.1 70.7 90.2 87.2 88.0 87.6</td></tr><tr><td/><td/><td/><td colspan=\"2\">Pr. Rec. F1</td></tr><tr><td/><td>AP</td><td/><td colspan=\"2\">92.9 71.6 80.9</td></tr><tr><td/><td>WAP</td><td/><td colspan=\"2\">85.1 77.9 81.4</td></tr><tr><td/><td>KL (2005)</td><td/><td>-</td><td>-</td><td>78.2</td></tr><tr><td/><td>LJ (2006)</td><td/><td>-</td><td>-</td><td>62.4</td></tr><tr><td/><td>MS (2008)</td><td/><td>-</td><td>-</td><td>30.6</td></tr><tr><td/><td colspan=\"2\">QL (2013) -Default</td><td>-</td><td>-</td><td>81.7</td></tr><tr><td colspan=\"3\">QL (2013) -Optimized</td><td>-</td><td>-</td><td>82.1</td></tr><tr><td/><td/><td/><td/><td>We evaluate</td></tr><tr><td/><td/><td/><td/><td>our model on detecting edited words in the sentences</td></tr></table>",
"num": null
},
"TABREF3": {
"type_str": "table",
"text": "Disfluency results. Pr. is precision, Rec. is recall and F1 is f-score. KL = (",
"html": null,
"content": "<table/>",
"num": null
}
}
}
} |