File size: 67,217 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 | {
"paper_id": "P02-1043",
"header": {
"generated_with": "S2ORC 1.0.0",
"date_generated": "2023-01-19T09:30:26.435905Z"
},
"title": "Generative Models for Statistical Parsing with Combinatory Categorial Grammar",
"authors": [
{
"first": "Julia",
"middle": [],
"last": "Hockenmaier",
"suffix": "",
"affiliation": {
"laboratory": "",
"institution": "University of Edinburgh",
"location": {
"postCode": "EH8 9LW",
"settlement": "Edinburgh",
"country": "United Kingdom"
}
},
"email": ""
},
{
"first": "Mark",
"middle": [],
"last": "Steedman",
"suffix": "",
"affiliation": {
"laboratory": "",
"institution": "University of Edinburgh",
"location": {
"postCode": "EH8 9LW",
"settlement": "Edinburgh",
"country": "United Kingdom"
}
},
"email": "steedman@cogsci.ed.ac.uk"
}
],
"year": "",
"venue": null,
"identifiers": {},
"abstract": "This paper compares a number of generative probability models for a widecoverage Combinatory Categorial Grammar (CCG) parser. These models are trained and tested on a corpus obtained by translating the Penn Treebank trees into CCG normal-form derivations. According to an evaluation of unlabeled word-word dependencies, our best model achieves a performance of 89.9%, comparable to the figures given by Collins (1999) for a linguistically less expressive grammar. In contrast to Gildea (2001), we find a significant improvement from modeling wordword dependencies.",
"pdf_parse": {
"paper_id": "P02-1043",
"_pdf_hash": "",
"abstract": [
{
"text": "This paper compares a number of generative probability models for a widecoverage Combinatory Categorial Grammar (CCG) parser. These models are trained and tested on a corpus obtained by translating the Penn Treebank trees into CCG normal-form derivations. According to an evaluation of unlabeled word-word dependencies, our best model achieves a performance of 89.9%, comparable to the figures given by Collins (1999) for a linguistically less expressive grammar. In contrast to Gildea (2001), we find a significant improvement from modeling wordword dependencies.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Abstract",
"sec_num": null
}
],
"body_text": [
{
"text": "The currently best single-model statistical parser (Charniak, 1999) achieves Parseval scores of over 89% on the Penn Treebank. However, the grammar underlying the Penn Treebank is very permissive, and a parser can do well on the standard Parseval measures without committing itself on certain semantically significant decisions, such as predicting null elements arising from deletion or movement. The potential benefit of wide-coverage parsing with CCG lies in its more constrained grammar and its simple and semantically transparent capture of extraction and coordination.",
"cite_spans": [
{
"start": 51,
"end": 67,
"text": "(Charniak, 1999)",
"ref_id": "BIBREF2"
}
],
"ref_spans": [],
"eq_spans": [],
"section": "Introduction",
"sec_num": "1"
},
{
"text": "We present a number of models over syntactic derivations of Combinatory Categorial Grammar (CCG, see Steedman (2000) and , this conference, for introduction), estimated from and tested on a translation of the Penn Treebank to a corpus of CCG normal-form derivations. CCG grammars are characterized by much larger category sets than standard Penn Treebank grammars, distinguishing for example between many classes of verbs with different subcategorization frames. As a result, the categorial lexicon extracted for this purpose from the training corpus has 1207 categories, compared with the 48 POS-tags of the Penn Treebank. On the other hand, grammar rules in CCG are limited to a small number of simple unary and binary combinatory schemata such as function application and composition. This results in a smaller and less overgenerating grammar than standard PCFGs (ca. 3,000 rules when instantiated with the above categories in sections 02-21, instead of 12,400 in the original Treebank representation (Collins, 1999) ).",
"cite_spans": [
{
"start": 101,
"end": 116,
"text": "Steedman (2000)",
"ref_id": "BIBREF14"
},
{
"start": 1004,
"end": 1019,
"text": "(Collins, 1999)",
"ref_id": "BIBREF7"
}
],
"ref_spans": [],
"eq_spans": [],
"section": "Introduction",
"sec_num": "1"
},
{
"text": "Since CCG produces unary and binary branching trees with a very fine-grained category set, CCG Parseval scores cannot be compared with scores of standard Treebank parsers. Therefore, we also evaluate performance using a dependency evaluation reported by Collins (1999) , which counts wordword dependencies as determined by local trees and their labels. According to this metric, a local tree with parent node P, head daughter H and non-head daughter S (and position of S relative to P, ie. left or right, which is implicit in CCG categories) defines a P H S dependency between the head word of S, w S , and the head word of H, w H . This measure is neutral with respect to the branching factor. Furthermore, as noted by Hockenmaier (2001) , it does not penalize equivalent analyses of multiple modi- Pierre Vinken 61 years old will join the board as a nonexecutive director Nov 29 AE AE AE AE AE AE\u00b4\u00cb \u00d2AE\u00c8\u00b5\u00d2AE\u00c8 \u00b4\u00cb \u00d0 \u00d2AE\u00c8\u00b5 \u00b4\u00cb \u00d2AE\u00c8\u03bc \u00cb \u00d2AE\u00c8\u00b5 \u00c8\u00c8\u00b5 AE\u00c8 AE\u00c8 AE AE \u00c8\u00c8 AE\u00c8 AE\u00c8 AE AE AE AE\u00b4\u00b4\u00cb\u00d2AE\u00c8\u00b5\u00d2\u00b4\u00cb\u00d2AE\u00c8\u00b5\u00b5 AE AE AE AE AE AE\u00b4\u00cb\u00d2AE\u00c8\u00b5\u00d2\u00b4\u00cb\u00d2AE\u00c8\u00b5 AE\u00c8 AE\u00c8 AE\u00c8 AE\u00c8 AE\u00c8 \u00cb \u00d2AE\u00c8\u00b4\u00cb \u00d2AE\u00c8\u00b5 \u00c8\u00c8 \u00c8\u00c8 AE\u00c8\u00d2AE\u00c8 \u00cb \u00d2AE\u00c8 AE\u00c8 \u00cb \u00d2AE\u00c8 AE\u00c8 \u00cb \u00d0 \u00d2AE\u00c8 \u00cb \u00d0 Figure 1 : A CCG derivation in our corpus fiers. In the unlabeled case (where it only matters whether word a is a dependent of word b, not what the label of the local tree is which defines this dependency), scores can be compared across grammars with different sets of labels and different kinds of trees. In order to compare our performance with the parser of , we also evaluate our best model according to the dependency evaluation introduced for that parser. For further discussion we refer the reader to .",
"cite_spans": [
{
"start": 254,
"end": 268,
"text": "Collins (1999)",
"ref_id": "BIBREF7"
},
{
"start": 720,
"end": 738,
"text": "Hockenmaier (2001)",
"ref_id": "BIBREF9"
}
],
"ref_spans": [
{
"start": 800,
"end": 1128,
"text": "Pierre Vinken 61 years old will join the board as a nonexecutive director Nov 29 AE AE AE AE AE AE\u00b4\u00cb \u00d2AE\u00c8\u00b5\u00d2AE\u00c8 \u00b4\u00cb \u00d0 \u00d2AE\u00c8\u00b5 \u00b4\u00cb \u00d2AE\u00c8\u03bc \u00cb \u00d2AE\u00c8\u00b5 \u00c8\u00c8\u00b5 AE\u00c8 AE\u00c8 AE AE \u00c8\u00c8 AE\u00c8 AE\u00c8 AE AE AE AE\u00b4\u00b4\u00cb\u00d2AE\u00c8\u00b5\u00d2\u00b4\u00cb\u00d2AE\u00c8\u00b5\u00b5 AE AE AE AE AE AE\u00b4\u00cb\u00d2AE\u00c8\u00b5\u00d2\u00b4\u00cb\u00d2AE\u00c8\u00b5 AE\u00c8 AE\u00c8 AE\u00c8 AE\u00c8 AE\u00c8 \u00cb \u00d2AE\u00c8\u00b4\u00cb \u00d2AE\u00c8\u00b5 \u00c8\u00c8 \u00c8\u00c8 AE\u00c8\u00d2AE\u00c8 \u00cb \u00d2AE\u00c8 AE\u00c8 \u00cb \u00d2AE\u00c8 AE\u00c8",
"ref_id": "TABREF1"
},
{
"start": 1142,
"end": 1150,
"text": "Figure 1",
"ref_id": null
}
],
"eq_spans": [],
"section": "Evaluating a CCG parser",
"sec_num": "2"
},
{
"text": "CCGbank is a corpus of CCG normal-form derivations obtained by translating the Penn Treebank trees using an algorithm described by .",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "CCGbank-a CCG treebank",
"sec_num": "3"
},
{
"text": "Almost all types of construction-with the exception of gapping and UCP (\"Unlike Coordinate Phrases\") are covered by the translation procedure, which processes 98.3% of the sentences in the training corpus (WSJ sections 02-21) and 98.5% of the sentences in the test corpus (WSJ section 23). The grammar contains a set of type-changing rules similar to the lexical rules described in Carpenter (1992) . Figure 1 shows a derivation taken from CCGbank. Categories, such as\u00b4\u00b4\u00cb \u00d2AE\u00c8\u00b5 \u00c8\u00c8\u00b5 AE\u00c8, encode unsaturated subcat frames. The complement-adjunct distinction is made explicit; for instance as a nonexecutive director is marked up as PP-CLR in the Treebank, and hence treated as a PP-complement of join, whereas Nov. 29 is marked up as an NP-TMP and therefore analyzed as VP modifier. The -CLR tag is not in fact a very reliable indicator of whether a constituent should be treated as a complement, but the translation to CCG is automatic and must do the best it can with the information in the Treebank.",
"cite_spans": [
{
"start": 382,
"end": 398,
"text": "Carpenter (1992)",
"ref_id": "BIBREF1"
}
],
"ref_spans": [
{
"start": 401,
"end": 410,
"text": "Figure 1",
"ref_id": null
}
],
"eq_spans": [],
"section": "CCGbank-a CCG treebank",
"sec_num": "3"
},
{
"text": "The verbal categories in CCGbank carry features distinguishing declarative verbs (and auxiliaries) from past participles in past tense, past participles for passive, bare infinitives and ing-forms.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "CCGbank-a CCG treebank",
"sec_num": "3"
},
{
"text": "There is a separate level for nouns and noun phrases, but, like the nonterminal NP in the Penn Treebank, noun phrases do not carry any number agreement. The derivations in CCGbank are \"normal-form\" in the sense that analyses involving the combinatory rules of type-raising and composition are only used when syntactically necessary. The models described here are all extensions of a very simple model which models derivations by a top-down tree-generating process. This model was originally described in Hockenmaier (2001) , where it was applied to a preliminary version of CCGbank, and its definition is repeated here in the top row of Table 1 . Given a (parent) node with category P, choose the expansion exp of P, where exp can be leaf (for lexical categories), unary (for unary expansions such as type-raising), left (for binary trees where the head daughter is left) or right (binary trees, head right). If P is a leaf node, generate its head word w. Otherwise, generate the category of its head daughter H. If P is binary branching, generate the category of its non-head daughter S (a complement or modifier of H).",
"cite_spans": [
{
"start": 504,
"end": 522,
"text": "Hockenmaier (2001)",
"ref_id": "BIBREF9"
}
],
"ref_spans": [
{
"start": 637,
"end": 644,
"text": "Table 1",
"ref_id": "TABREF1"
}
],
"eq_spans": [],
"section": "CCGbank-a CCG treebank",
"sec_num": "3"
},
{
"text": "+ \u2206 P#\u2206 L R P P exp#\u2206 L R P P exp H#\u2206 L R P",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Generative models of CCG derivations",
"sec_num": "4"
},
{
"text": "The model itself includes no prior knowledge specific to CCG other than that it only allows unary and binary branching trees, and that the sets of nonterminals and terminals are not disjoint (hence the need to include leaf as a possible expansion, which acts as a stop probability).",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Generative models of CCG derivations",
"sec_num": "4"
},
{
"text": "All the experiments reported in this section were conducted using sections 02-21 of CCGbank as training corpus, and section 23 as test corpus. We replace all rare words in the training data with their POS-tag. For all experiments reported here and in section 5, the frequency threshold was set to 5. Like Collins (1999) , we assume that the test data is POStagged, and can therefore replace unknown words in the test data with their POS-tag, which is more appropriate for a formalism like CCG with a large set of lexical categories than one generic token for all unknown words.",
"cite_spans": [
{
"start": 305,
"end": 319,
"text": "Collins (1999)",
"ref_id": "BIBREF7"
}
],
"ref_spans": [],
"eq_spans": [],
"section": "Generative models of CCG derivations",
"sec_num": "4"
},
{
"text": "The performance of the baseline model is shown in the top row of table 3. For six out of the 2379 sentences in our test corpus we do not get a parse. 1 The reason is that a lexicon consisting of the wordcategory pairs observed in the training corpus does not contain all the entries required to parse the test corpus. We discuss a simple, but imperfect, solution to this problem in section 7.",
"cite_spans": [
{
"start": 150,
"end": 151,
"text": "1",
"ref_id": null
}
],
"ref_spans": [],
"eq_spans": [],
"section": "Generative models of CCG derivations",
"sec_num": "4"
},
{
"text": "State-of-the-art statistical parsers use many other features, or conditioning variables, such as head words, subcategorization frames, distance measures and grandparent nodes. We too can extend the baseline model described in the previous section by including more features. Like the models of Goodman (1997) , the additional features in our model are generated probabilistically, whereas in the parser of Collins (1997) distance measures are assumed to be a function of the already generated structure and are not generated explicitly.",
"cite_spans": [
{
"start": 294,
"end": 308,
"text": "Goodman (1997)",
"ref_id": "BIBREF11"
},
{
"start": 406,
"end": 420,
"text": "Collins (1997)",
"ref_id": "BIBREF6"
}
],
"ref_spans": [],
"eq_spans": [],
"section": "Extending the baseline model",
"sec_num": "5"
},
{
"text": "In order to estimate the conditional probabilities of our model, we recursively smooth empirical estimates\u00ea i of specific conditional distributions with (possible smoothed) estimates of less specific distributions\u1ebd i 1 , using linear interpolation:",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Extending the baseline model",
"sec_num": "5"
},
{
"text": "e i \u03bb\u00ea i \u2022 1 \u03bb\u00b5\u1ebd i 1",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Extending the baseline model",
"sec_num": "5"
},
{
"text": "\u03bb is a smoothing weight which depends on the particular distribution. 2 When defining models, we will indicate a backoff level with a # sign between conditioning variables, eg. A B # C # D means that we interpolat\u00ea P\u00b4 A B C D\u00b5 withP\u00b4 A B C\u00b5, which is an interpolation ofP\u00b4 A B C\u00b5 andP\u00b4 A B\u00b5.",
"cite_spans": [
{
"start": 70,
"end": 71,
"text": "2",
"ref_id": null
}
],
"ref_spans": [],
"eq_spans": [],
"section": "Extending the baseline model",
"sec_num": "5"
},
{
"text": "The coordination feature We define a boolean feature, conj, which is true for constituents which expand to coordinations on the head path.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Adding non-lexical information",
"sec_num": "5.1"
},
{
"text": "\u00cb, \u2022conj \u00cb AE\u00c8, \u2022conj \u00cb AE\u00c8, conj \u00cb \u00b4\u00cb\u00d2AE\u00c8\u00b5 AE\u00c8 IBM\u00b4\u00cb \u00d2AE\u00c8\u00b5 AE\u00c8 buys \u00cb AE\u00c8 , \u2022conj \u00d3\u00d2 but \u00cb AE\u00c8 , conj \u00cb \u00b4\u00cb\u00d2AE\u00c8\u00b5 AE\u00c8 Lotus\u00b4\u00cb \u00d2AE\u00c8\u00b5 AE\u00c8 sells AE\u00c8 shares",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Adding non-lexical information",
"sec_num": "5.1"
},
{
"text": "This feature is generated at the root of the sentence with P\u00b4conj TOP\u00b5. For binary expansions, conj H is generated with P\u00b4conj H H S con j P \u00b5 and conj S is generated with P\u00b4conj S S # P exp P H conj P \u00b5. Table 1 shows how conj is used as a conditioning variable. This is intended to allow the model to capture the fact that, for a sentence without extraction, a CCG derivation where the subject is type-raised and composed with the verb is much more likely in right node raising constructions like the above.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Adding non-lexical information",
"sec_num": "5.1"
},
{
"text": "The impact of the grandparent feature Johnson (1998) showed that a PCFG estimated from a version of the Penn Treebank in which the label of a node's parent is attached to the node's own label yields a substantial improvement (LP/LR: from 73.5%/69.7% to 80.0%/79.2%). The inclusion of an additional grandparent feature gives Charniak (1999) a slight improvement in the Maximum Entropy inspired model, but a slight decrease in performance for an MLE model. At the same time coverage is reduced (which might not be the case if this was an additional feature in the model rather than a change in the representation of the categories). Both of these results are to be expected-CCG categories encode more contextual information than Treebank labels, in particular about parents and grandparents; therefore the history feature might be expected to have less impact. Moreover, since our category set is much larger, appending the parent node will lead to an even more fine-grained partitioning of the data, which then results in sparse data problems.",
"cite_spans": [
{
"start": 38,
"end": 52,
"text": "Johnson (1998)",
"ref_id": "BIBREF12"
},
{
"start": 324,
"end": 339,
"text": "Charniak (1999)",
"ref_id": "BIBREF2"
}
],
"ref_spans": [],
"eq_spans": [],
"section": "Adding non-lexical information",
"sec_num": "5.1"
},
{
"text": "Distance measures for CCG Our distance measures are related to those proposed by Goodman (1997) , which are appropriate for binary trees (unlike those of Collins (1997) ). Every node has a left distance measure, \u2206 L , measuring the distance from the head word to the left frontier of the constituent. There is a similar right distance measure \u2206 R . We implemented three different ways of measuring distance: \u2206 Adjacency measures string adjacency (0, 1 or 2 and more intervening words); \u2206 Verb counts intervening verbs (0 or 1 and more); and \u2206 Pct counts intervening punctuation marks (0, 1, 2 or 3 and more). These \u2206s are generated by the model in the following manner: at the root of the sentence, generate \u2206 L with P\u00b4\u2206 L TOP\u00b5, and",
"cite_spans": [
{
"start": 81,
"end": 95,
"text": "Goodman (1997)",
"ref_id": "BIBREF11"
},
{
"start": 154,
"end": 168,
"text": "Collins (1997)",
"ref_id": "BIBREF6"
}
],
"ref_spans": [],
"eq_spans": [],
"section": "Adding non-lexical information",
"sec_num": "5.1"
},
{
"text": "\u2206 R with P\u00b4\u2206 R TOP \u2206 L \u00b5.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Adding non-lexical information",
"sec_num": "5.1"
},
{
"text": "Then, for each expansion, if it is a unary expansion, \u2206 L H \u2206 L P and \u2206 R H \u2206 R P with a probability of 1. If it is a binary expansion, only the \u2206 in the direction of the sister changes, with a probability of P\u00b4\u2206 L H \u2206 L P H#P S\u00b5 if exp \u00d6 \u00d8, and analogously for exp \u00d0 \u00d8. \u2206 L S and \u2206 R S are conditioned on S and the \u2206 of H and P in the direction of S:",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Adding non-lexical information",
"sec_num": "5.1"
},
{
"text": "P\u00b4\u2206 L S S#\u2206 R P \u2206 R H \u00b5 and P\u00b4\u2206 R S S \u2206 L S #\u2206 R P \u2206 R H \u00b5.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Adding non-lexical information",
"sec_num": "5.1"
},
{
"text": "They are then used as further conditioning variables for the other distributions as shown in table 1. Table 3 also gives the Parseval and dependency scores obtained with each of these measures. \u2206 Pct has the smallest effect. However, our model does not yet contain anything like the hard constraint on punctuation marks in Collins (1999) . Gildea (2001) shows that removing the lexical dependencies in Model 1 of Collins (1997) (that is, not conditioning on w h when generating w s ) decreases labeled precision and recall by only 0.5%. It can therefore be assumed that the main influence of lexical head features (words and preterminals) in Collins' Model 1 is on the structural probabilities.",
"cite_spans": [
{
"start": 323,
"end": 337,
"text": "Collins (1999)",
"ref_id": "BIBREF7"
},
{
"start": 340,
"end": 353,
"text": "Gildea (2001)",
"ref_id": "BIBREF8"
},
{
"start": 413,
"end": 427,
"text": "Collins (1997)",
"ref_id": "BIBREF6"
}
],
"ref_spans": [
{
"start": 102,
"end": 109,
"text": "Table 3",
"ref_id": "TABREF2"
}
],
"eq_spans": [],
"section": "Adding non-lexical information",
"sec_num": "5.1"
},
{
"text": "In CCG, by contrast, preterminals are lexical categories, encoding complete subcategorization information. They therefore encode more information about the expansion of a nonterminal than Treebank POS-tags and thus are more constraining.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Adding lexical information",
"sec_num": "5.2"
},
{
"text": "Generating a constituent's lexical category c at its maximal projection (ie. either at the root of the tree, TOP, or when generating a non-head daughter S), and using the lexical category as conditioning vari-able (LexCat) increases performance of the baseline model as measured by P H S by almost 3%. In this model, c S , the lexical category of S depends on the category S and on the local tree in which S is generated. However, slightly worse performance is obtained for LexCatDep, a model which is identical to the original LexCat model, except that c S is also conditioned on c H , the lexical category of the head node, which introduces a dependency between the lexical categories.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Adding lexical information",
"sec_num": "5.2"
},
{
"text": "Since there is so much information in the lexical categories, one might expect that this would reduce the effect of conditioning the expansion of a constituent on its head word w. However, we did find a substantial effect. Generating the head word at the maximal projection (HeadWord) increases performance by a further 2%. Finally, conditioning w S on w H , hence including word-word dependencies, (HWDep) increases performance even more, by another 3.5%, or 8.3% overall. This is in stark contrast to Gildea's findings for Collins' Model 1.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Adding lexical information",
"sec_num": "5.2"
},
{
"text": "We conjecture that the reason why CCG benefits more from word-word dependencies than Collins' Model 1 is that CCG allows a cleaner parametrization of these surface dependencies. In Collins' Model 1, w S is conditioned not only on the local tree P H S , c H and w H , but also on the distance \u2206 between the head and the modifier to be generated. However, Model 1 does not incorporate the notion of subcategorization frames. Instead, the distance measure was found to yield a good, if imperfect, approximation to subcategorization information.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Adding lexical information",
"sec_num": "5.2"
},
{
"text": "Using our notation, Collins' Model 1 generates w S with the following probability:",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Adding lexical information",
"sec_num": "5.2"
},
{
"text": "P Collins1\u00b4wS c S \u2206 P H S c H w H \u00b5 \u03bb 1P\u00b4wS c S \u2206 P H S c H w H \u00b5 \u2022\u00b41 \u03bb 1 \u00b5 \u00a2 \u03bb 2P\u00b4wS c S \u2206 P H S c H \u00b5 \u2022 1 \u03bb 2 \u00b5P\u00b4w S c S \u00b5 \u00a3",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Adding lexical information",
"sec_num": "5.2"
},
{
"text": "-whereas the CCG dependency model generates w S as follows: P\u00b4w TOP \u00b5 LexCat P c P P exp c P P exp H#c P S#H exp P P TOP --LexCatDep P c P P exp c P P exp H#c P S#H exp P#c P P TOP --HeadWord P c P #w P P exp c P #w P P exp H#c P #w P S#H exp P P TOP c S c P HWDep P c P #w P P exp c P #w P P exp H#c P #w P S#H exp P P TOP c S #P H S w P c P HWDep\u2206 P c P #\u2206 L R P #w P P exp c P #\u2206 L R P #w P P exp H#\u2206 L R P #c P #w P S#H exp P P TOP c S #P H S w P c P HWDepConj P c P conj P #w P P exp c P conj P #w P P exp H conj P #c P #w P S#H exp P P TOP c S #P H S w P c P The P H S labeled dependencies we report are not directly comparable with Collins (1999) , since CCG categories encode subcategorization frames. For instance, if the direct object of a verb has been recognized as such, but a PP has been mistaken as a complement (whereas the gold standard says it is an adjunct), the fully labeled dependency evaluation P H S will not award a point. Therefore, we also include in Table 3 a more comparable evaluation S which only takes the correctness of the non-head category into account. The reported figures are also deflated by retaining verb features like tensed/untensed. If this is done (by stripping off all verb features), an improvement of 0.6% on the P H S score for our best model is obtained.",
"cite_spans": [
{
"start": 639,
"end": 653,
"text": "Collins (1999)",
"ref_id": "BIBREF7"
}
],
"ref_spans": [
{
"start": 978,
"end": 985,
"text": "Table 3",
"ref_id": "TABREF2"
}
],
"eq_spans": [],
"section": "Adding lexical information",
"sec_num": "5.2"
},
{
"text": "P",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Adding lexical information",
"sec_num": "5.2"
},
{
"text": "When incorporating the adjacency distance measure or the coordination feature into the dependency model (HWDep\u2206 and HWDepConj), overall performance is lower than with the dependency model alone. We conjecture that this arises from data sparseness. It cannot be concluded from these results alone that the lexical dependencies make structural information redundant or superfluous. Instead, it is quite likely that we are facing an estimation problem similar to Charniak (1999) , who reports that the inclusion of the grandparent feature worsens performance of an MLE model, but improves performance if the individual distributions are modelled using Maximum Entropy. This intuition is strengthened by the fact that, on casual inspection of the scores for individual sentences, it is sometimes the case that the lexicalized models perform worse than the unlexicalized models.",
"cite_spans": [
{
"start": 460,
"end": 475,
"text": "Charniak (1999)",
"ref_id": "BIBREF2"
}
],
"ref_spans": [],
"eq_spans": [],
"section": "Combining lexical and non-lexical information",
"sec_num": "5.3"
},
{
"text": "All of the experiments described above use the POStags as given by CCGbank (which are the Treebank tags, with some corrections necessary to acquire correct features on categories). It is reasonable to assume that this input is of higher quality than can be produced by a POS-tagger. We therefore ran the dependency model on a test corpus tagged with the POS-tagger of Ratnaparkhi (1996) , which is trained on the original Penn Treebank (see HWDep (+ tagger) in Table 3 ). Performance degrades slightly, which is to be expected, since our approach makes so much use of the POS-tag information for unknown words. However, a POS-tagger trained on CCGbank might yield slightly better results.",
"cite_spans": [
{
"start": 368,
"end": 386,
"text": "Ratnaparkhi (1996)",
"ref_id": "BIBREF13"
}
],
"ref_spans": [
{
"start": 461,
"end": 468,
"text": "Table 3",
"ref_id": "TABREF2"
}
],
"eq_spans": [],
"section": "The impact of tagging errors",
"sec_num": "5.4"
},
{
"text": "Unlike , our parser does not always model the dependencies in the logical form. For example, in the interpretation of a coordinate structure like \"buy and sell shares\", shares will head an object of both buy and sell. Similarly, in examples like \"buy the company that wins\", the relative construction makes company depend upon both buy as object and wins as subject. As is well known (Abney, 1997), DAG-like dependencies cannot in general be modeled with a generative approach of the kind taken here 3 .",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Limitations of the current model",
"sec_num": "5.5"
},
{
"text": "Clark et al. 2002presents another statistical CCG parser, which is based on a conditional (rather than generative) model of the derived dependency structure, including non-surface dependencies. The following table compares the two parsers according to the evaluation of surface and deep dependencies given in . We use Clark et al.'s parser to generate these dependencies from the output of our parser (see ",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Comparison with Clark et al. (2002)",
"sec_num": "5.6"
},
{
"text": "One of the advantages of CCG is that it provides a simple, surface grammatical analysis of extraction and coordination. We investigate whether our best Coordination There are two instances of argument cluster coordination (constructions like cost $5,000 in July and $6,000 in August) in the development corpus. Of these, HWDep recovers none correctly. This is a shortcoming in the model, rather than in CCG: the relatively high probability both of the NP modifier analysis of PPs like in July and of NP coordination is enough to misdirect the parser.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Performance on specific constructions",
"sec_num": "6"
},
{
"text": "There are 203 instances of verb phrase coordination (\u00cb \u00d2AE\u00c8, with any verbal feature) in the development corpus. On these, we obtain a labeled recall and precision of 67.0%/67.3%. Interestingly, on the 24 instances of right node raising (coordination of\u00b4\u00cb \u00d2AE\u00c8\u00b5 AE\u00c8), our parser achieves higher performance, with labeled recall and precision of 79.2% and 73.1%. Figure 2 gives an example of the output of our parser on such a sentence.",
"cite_spans": [],
"ref_spans": [
{
"start": 362,
"end": 370,
"text": "Figure 2",
"ref_id": null
}
],
"eq_spans": [],
"section": "Performance on specific constructions",
"sec_num": "6"
},
{
"text": "Extraction Long-range dependencies are not captured by the evaluation used here. However, the accuracy for recovering lexical categories for words with \"extraction\" categories, such as relative pronouns, gives some indication of how well the model detects the presence of such dependencies. The most common category for subject relative Figure 2 : Right node raising output produced by our parser. Punishing and retaliating are unknown words.",
"cite_spans": [],
"ref_spans": [
{
"start": 337,
"end": 345,
"text": "Figure 2",
"ref_id": null
}
],
"eq_spans": [],
"section": "Performance on specific constructions",
"sec_num": "6"
},
{
"text": "times by the parser, out of which 48 times it corresponded to a rule in the gold standard (or 34 times, if the exact bracketing of the \u00cb \u00d0 AE\u00c8 is taken into account-this lower figure is due to attachment decisions made elsewhere in the tree).",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Performance on specific constructions",
"sec_num": "6"
},
{
"text": "These figures are difficult to compare with standard Treebank parsers. Despite the fact that the original Treebank does contain traces for movement, none of the existing parsers try to generate these traces (with the exception of Collins' Model 3, for which he only gives an overall score of 96.3%/98.8% P/R for subject extraction and 81.4%/59.4% P/R for other cases). The only \"long range\" dependency for which Collins gives numbers is subject extraction SBAR, WHNP, SG, R , which has labeled precision and recall of 90.56% and 90.56%, whereas the CCG model achieves a labeled precision and recall of 94.3% and 96.5% on the most frequency subject extraction dependency AE\u00c8\u00d2AE\u00c8, AE\u00c8\u00d2AE\u00c8\u00b5 \u00b4\u00cb \u00d0 \u00d2AE\u00c8\u00b5, \u00cb \u00d0 \u00d2AE\u00c8 , which occurs 262 times in the gold standard and was produced 256 times by our parser. However, out of the 15 cases of this relation in the gold standard that our parser did not return, 8 were in fact analyzed as subject extraction of bare infinitivals AE\u00c8\u00d2AE\u00c8, AE\u00c8\u00d2AE\u00c8\u00b5 \u00b4\u00cb \u00d2AE\u00c8\u00b5, \u00cb \u00d2AE\u00c8 , yielding a combined recall of 97.3%.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Performance on specific constructions",
"sec_num": "6"
},
{
"text": "The most serious problem facing parsers like the present one with large category sets is not so much the standard problem of unseen words, but rather the problem of words that have been seen, but not with the necessary category.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Lexical coverage",
"sec_num": "7"
},
{
"text": "For standard Treebank parsers, the latter problem does not have much impact, if any, since the Penn Treebank tagset is fairly small, and the grammar underlying the Treebank is very permissive. However, for CCG this is a serious problem: the first three rows in table 4 show a significant difference in performance for sentences with complete lexical coverage (\"No missing\") and sentences with missing lexical entries (\"Missing\").",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Lexical coverage",
"sec_num": "7"
},
{
"text": "Using the POS-tags in the corpus, we can estimate the lexical probabilities P\u00b4w c\u00b5 using a linear interpolation between the relative frequency estimate\u015d P\u00b4w c\u00b5 and the following approximation: 5",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Lexical coverage",
"sec_num": "7"
},
{
"text": "We smooth the lexical probabilities as follows: Table 4 shows the performance of the baseline model with a frequency cutoff of 5 and 10 for rare words and with a smoothed and non-smoothed lexicon. 6 This frequency cutoff plays an important role here -smoothing with a small cutoff yields worse performance than not smoothing, whereas smoothing with a cutoff of 10 does not have a significant impact on performance. Smoothing the lexicon in this way does make the parser more robust, resulting in complete coverage of the test set. However, it does not affect overall performance, nor does it alleviate the problem for sentences with missing lexical entries for seen words. Table 4 : The impact of lexical coverage, using a different cutoff for rare words and smoothing (section 23)",
"cite_spans": [
{
"start": 197,
"end": 198,
"text": "6",
"ref_id": null
}
],
"ref_spans": [
{
"start": 48,
"end": 55,
"text": "Table 4",
"ref_id": null
},
{
"start": 673,
"end": 680,
"text": "Table 4",
"ref_id": null
}
],
"eq_spans": [],
"section": "P tags\u00b4w c\u00b5 \u2211t\u00betagsP\u00b4w t\u00b5P\u00b4t c\u00b5",
"sec_num": null
},
{
"text": "P\u00b4w c\u00b5 \u03bbP\u00b4w c\u00b5 \u2022 1 \u03bb\u00b5P tags\u00b4w c\u00b5",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "P tags\u00b4w c\u00b5 \u2211t\u00betagsP\u00b4w t\u00b5P\u00b4t c\u00b5",
"sec_num": null
},
{
"text": "We have compared a number of generative probability models of CCG derivations, and shown that our best model recovers 89.9% of word-word dependencies on section 23 of CCGbank. On section 00, it recovers 89.7% of word-word dependencies. These figures are surprisingly close to the figure of 90.9% reported by Collins (1999) on section 00, given that, in order to allow a direct comparison, we have used the same interpolation technique and beam strategy as Collins (1999) , which are very unlikely to be as well-tuned to our kind of grammar. As is to be expected, a statistical model of a CCG extracted from the Treebank is less robust than a model with an overly permissive grammar such as Collins (1999) . This problem seems to stem mainly from the incomplete coverage of the lexicon. We have shown that smoothing can compensate for entirely unknown words. However, this approach does not help on sentences which require previously unseen entries for known words. We would expect a less naive approach such as applying morphological rules to the observed entries, together with better smoothing techniques, to yield better results.",
"cite_spans": [
{
"start": 308,
"end": 322,
"text": "Collins (1999)",
"ref_id": "BIBREF7"
},
{
"start": 456,
"end": 470,
"text": "Collins (1999)",
"ref_id": "BIBREF7"
},
{
"start": 690,
"end": 704,
"text": "Collins (1999)",
"ref_id": "BIBREF7"
}
],
"ref_spans": [],
"eq_spans": [],
"section": "Conclusion and future work",
"sec_num": "8"
},
{
"text": "We have also shown that a statistical model of CCG benefits from word-word dependencies to a much greater extent than a less linguistically motivated model such as Collins' Model 1. This indicates to us that, although the task faced by a CCG parser might seem harder prima facie, there are advantages to using a more linguistically adequate grammar.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Conclusion and future work",
"sec_num": "8"
},
{
"text": "We conjecture that the minor variations in coverage among the other models (except Grandparent) are artefacts of the beam.2 We compute \u03bb in the same way asCollins (1999), p. 185.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "",
"sec_num": null
},
{
"text": "It remains to be seen whether the more restricted reentrancies of CCG will ultimately support a generative model.4 Due to the smaller grammar and lexicon of Clark et al., our parser can only be evaluated on slightly over 94% of the sentences in section 23, whereas the figures for are on 97%. model, HWDep, predicts the correct analyses, using the development section 00.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "",
"sec_num": null
},
{
"text": "We compute \u03bb in the same way asCollins (1999), p. 185.6 Smoothing was only done for categories with a total frequency of 100 or more.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "",
"sec_num": null
}
],
"back_matter": [
{
"text": "Thanks to Stephen Clark, Miles Osborne and the ACL-02 referees for comments. Various parts of the research were funded by EPSRC grants GR/M96889 and GR/R02450 and an EPSRC studentship.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Acknowledgements",
"sec_num": null
}
],
"bib_entries": {
"BIBREF0": {
"ref_id": "b0",
"title": "Stochastic Attribute-Value Grammars",
"authors": [
{
"first": "Steven",
"middle": [],
"last": "Abney",
"suffix": ""
}
],
"year": 1997,
"venue": "Computational Linguistics",
"volume": "",
"issue": "4",
"pages": "",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Steven Abney. 1997. Stochastic Attribute-Value Grammars. Computational Linguistics, 23(4).",
"links": null
},
"BIBREF1": {
"ref_id": "b1",
"title": "Categorial Grammars, Lexical Rules, and the English Predicative",
"authors": [
{
"first": "Bob",
"middle": [],
"last": "Carpenter",
"suffix": ""
}
],
"year": 1992,
"venue": "Formal Grammar: Theory and Implementation",
"volume": "",
"issue": "",
"pages": "",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Bob Carpenter. 1992. Categorial Grammars, Lexical Rules, and the English Predicative. In R. Levine, ed., Formal Grammar: Theory and Implementation. OUP.",
"links": null
},
"BIBREF2": {
"ref_id": "b2",
"title": "A Maximum-Entropy-Inspired Parser",
"authors": [
{
"first": "Eugene",
"middle": [],
"last": "Charniak",
"suffix": ""
}
],
"year": 1999,
"venue": "",
"volume": "",
"issue": "",
"pages": "",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Eugene Charniak. 1999. A Maximum-Entropy-Inspired Parser. TR CS-99-12, Brown University.",
"links": null
},
"BIBREF3": {
"ref_id": "b3",
"title": "Statistical Parsing with an Automatically-Extracted Tree Adjoining Grammar 38th ACL",
"authors": [
{
"first": "David",
"middle": [],
"last": "Chiang",
"suffix": ""
}
],
"year": 2000,
"venue": "",
"volume": "",
"issue": "",
"pages": "456--463",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "David Chiang. 2000. Statistical Parsing with an Automatically- Extracted Tree Adjoining Grammar 38th ACL, Hong Kong, pp. 456-463.",
"links": null
},
"BIBREF4": {
"ref_id": "b4",
"title": "Evaluating a Wide-Coverage CCG Parser",
"authors": [
{
"first": "Stephen",
"middle": [],
"last": "Clark",
"suffix": ""
},
{
"first": "Julia",
"middle": [],
"last": "Hockenmaier",
"suffix": ""
}
],
"year": 2002,
"venue": "LREC Beyond PARSEVAL workshop",
"volume": "",
"issue": "",
"pages": "",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Stephen Clark and Julia Hockenmaier. 2002. Evaluating a Wide-Coverage CCG Parser. LREC Beyond PARSEVAL workshop, Las Palmas, Spain.",
"links": null
},
"BIBREF5": {
"ref_id": "b5",
"title": "Building Deep Dependency Structures Using a Wide-Coverage CCG Parser. 40th ACL",
"authors": [
{
"first": "Stephen",
"middle": [],
"last": "Clark",
"suffix": ""
},
{
"first": "Julia",
"middle": [],
"last": "Hockenmaier",
"suffix": ""
},
{
"first": "Mark",
"middle": [],
"last": "Steedman",
"suffix": ""
}
],
"year": 2002,
"venue": "",
"volume": "",
"issue": "",
"pages": "",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Stephen Clark, Julia Hockenmaier, and Mark Steedman. 2002. Building Deep Dependency Structures Using a Wide- Coverage CCG Parser. 40th ACL, Philadelphia.",
"links": null
},
"BIBREF6": {
"ref_id": "b6",
"title": "Three Generative Lexicalized Models for Statistical Parsing. 35th ACL",
"authors": [
{
"first": "Michael",
"middle": [],
"last": "Collins",
"suffix": ""
}
],
"year": 1997,
"venue": "",
"volume": "",
"issue": "",
"pages": "16--23",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Michael Collins. 1997. Three Generative Lexicalized Models for Statistical Parsing. 35th ACL, Madrid, pp. 16-23.",
"links": null
},
"BIBREF7": {
"ref_id": "b7",
"title": "Head-Driven Statistical Models for Natural Language Parsing",
"authors": [
{
"first": "Michael",
"middle": [],
"last": "Collins",
"suffix": ""
}
],
"year": 1999,
"venue": "",
"volume": "",
"issue": "",
"pages": "",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Michael Collins. 1999. Head-Driven Statistical Models for Natural Language Parsing. Ph.D. thesis, University of Pennsylvania.",
"links": null
},
"BIBREF8": {
"ref_id": "b8",
"title": "Corpus Variation and Parser Performance",
"authors": [
{
"first": "Daniel",
"middle": [],
"last": "Gildea",
"suffix": ""
}
],
"year": 2001,
"venue": "EMNLP",
"volume": "",
"issue": "",
"pages": "",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Daniel Gildea. 2001. Corpus Variation and Parser Perfor- mance. EMNLP, Pittsburgh, PA.",
"links": null
},
"BIBREF9": {
"ref_id": "b9",
"title": "Statistical Parsing for CCG with Simple Generative Models",
"authors": [
{
"first": "Julia",
"middle": [],
"last": "Hockenmaier",
"suffix": ""
}
],
"year": 2001,
"venue": "Student Workshop, 39th ACL/ 10th EACL",
"volume": "",
"issue": "",
"pages": "7--12",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Julia Hockenmaier. 2001. Statistical Parsing for CCG with Simple Generative Models. Student Workshop, 39th ACL/ 10th EACL, Toulouse, France, pp. 7-12.",
"links": null
},
"BIBREF10": {
"ref_id": "b10",
"title": "Acquiring Compact Lexicalized Grammars from a Cleaner Treebank",
"authors": [
{
"first": "Julia",
"middle": [],
"last": "Hockenmaier",
"suffix": ""
},
{
"first": "Mark",
"middle": [],
"last": "Steedman",
"suffix": ""
}
],
"year": 2002,
"venue": "Third LREC",
"volume": "",
"issue": "",
"pages": "",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Julia Hockenmaier and Mark Steedman 2002. Acquiring Com- pact Lexicalized Grammars from a Cleaner Treebank. Third LREC, Las Palmas, Spain.",
"links": null
},
"BIBREF11": {
"ref_id": "b11",
"title": "Probabilistic Feature Grammars. IWPT",
"authors": [
{
"first": "Joshua",
"middle": [],
"last": "Goodman",
"suffix": ""
}
],
"year": 1997,
"venue": "",
"volume": "",
"issue": "",
"pages": "",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Joshua Goodman. 1997. Probabilistic Feature Grammars. IWPT, Boston.",
"links": null
},
"BIBREF12": {
"ref_id": "b12",
"title": "PCFG Models of Linguistic Tree Representations",
"authors": [
{
"first": "Mark",
"middle": [],
"last": "Johnson",
"suffix": ""
}
],
"year": 1998,
"venue": "Computational Linguistics",
"volume": "",
"issue": "4",
"pages": "",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Mark Johnson. 1998. PCFG Models of Linguistic Tree Repre- sentations. Computational Linguistics, 24(4).",
"links": null
},
"BIBREF13": {
"ref_id": "b13",
"title": "A Maximum Entropy Part-Of-Speech Tagger",
"authors": [
{
"first": "Adwait",
"middle": [],
"last": "Ratnaparkhi",
"suffix": ""
}
],
"year": 1996,
"venue": "EMNLP",
"volume": "",
"issue": "",
"pages": "133--142",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Adwait Ratnaparkhi. 1996. A Maximum Entropy Part-Of- Speech Tagger. EMNLP, Philadelphia, pp. 133-142.",
"links": null
},
"BIBREF14": {
"ref_id": "b14",
"title": "The Syntactic Process",
"authors": [
{
"first": "Mark",
"middle": [],
"last": "Steedman",
"suffix": ""
}
],
"year": 2000,
"venue": "",
"volume": "",
"issue": "",
"pages": "",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Mark Steedman. 2000. The Syntactic Process. The MIT Press, Cambridge Mass.",
"links": null
}
},
"ref_entries": {
"TABREF1": {
"text": "The unlexicalized models",
"num": null,
"html": null,
"content": "<table/>",
"type_str": "table"
},
"TABREF2": {
"text": "Grandparent) shows that a grammar transformation like Johnson's does yield an improvement, but not as dramatic as in the Treebank-CFG case.",
"num": null,
"html": null,
"content": "<table/>",
"type_str": "table"
},
"TABREF4": {
"text": "The lexicalized models",
"num": null,
"html": null,
"content": "<table><tr><td>Model</td><td colspan=\"3\">NoParse LexCat LP</td><td>LR</td><td>BP</td><td>BR</td><td>P H S</td><td>S</td><td>CM on</td><td>2 CD</td></tr><tr><td>Baseline</td><td>6</td><td>87.7</td><td colspan=\"5\">72.8 72.4 78,3 77.9 75.7</td><td colspan=\"2\">81.1 84.3 23.0</td><td>51.1</td></tr><tr><td>Conj</td><td>9</td><td>87.8</td><td colspan=\"5\">73.8 73.9 79.3 79.3 76.7</td><td colspan=\"2\">82.0 85.1 24.3</td><td>53.2</td></tr><tr><td>Grandparent \u2206 Pct \u2206 Verb \u2206 Adjacency</td><td>91 6 6 6</td><td>88.8 88.1 88.0 88.6</td><td colspan=\"5\">77.1 77.6 82.4 82.9 79.9 73.7 73.1 79.2 78.6 76.5 75.9 75.5 81.6 81.1 76.9 77.5 77.3 82.9 82.8 78.9</td><td colspan=\"2\">84.7 87.9 30.9 81.8 84.9 23.1 82.3 85.3 25.2 83.8 86.9 24.8</td><td>63.8 53.2 55.1 59.6</td></tr><tr><td>LexCat</td><td>9</td><td>88.5</td><td colspan=\"5\">75.8 76.0 81.3 81.5 78.6</td><td colspan=\"2\">83.7 86.8 27.4</td><td>57.8</td></tr><tr><td>LexCatDep</td><td>9</td><td>88.5</td><td colspan=\"5\">75.7 75.9 81.2 81.4 78.4</td><td colspan=\"2\">83.5 86.6 26.3</td><td>57.9</td></tr><tr><td>HeadWord</td><td>8</td><td>89.6</td><td colspan=\"5\">77.9 78.0 83.0 83.1 80.5</td><td colspan=\"2\">85.2 88.3 30.4</td><td>63.0</td></tr><tr><td>HWDep</td><td>8</td><td>92.0</td><td colspan=\"5\">81.6 81.9 85.5 85.9 84.0</td><td colspan=\"2\">87.8 90.1 37.9</td><td>69.2</td></tr><tr><td>HWDep\u2206</td><td>8</td><td>90.9</td><td colspan=\"5\">81.4 81.6 86.1 86.3 83.0</td><td colspan=\"2\">87.0 89.8 35.7</td><td>68.7</td></tr><tr><td>HWDepConj</td><td>9</td><td>91.8</td><td colspan=\"5\">80.7 81.2 84.8 85.3 83.6</td><td colspan=\"2\">87.5 89.9 36.5</td><td>68.6</td></tr><tr><td>HWDep (+ tagger)</td><td>7</td><td>91.7</td><td colspan=\"5\">81.4 81.8 85.6 85.9 83.6</td><td colspan=\"2\">87.5 89.9 38.1</td><td>69.1</td></tr></table>",
"type_str": "table"
},
"TABREF5": {
"text": "Performance of the models: LexCat indicates accuracy of the lexical categories; LP, LR, BP and BR (the standard Parseval scores labeled/bracketed precision and recall) are not commensurate with other Treebank parsers. P H S , S , and are as defined in section 2. CM on is the percentage of sentences with complete match on , and 2 CD is the percentage of sentences with under 2 \"crossing dependencies\" as defined by .",
"num": null,
"html": null,
"content": "<table/>",
"type_str": "table"
}
}
}
} |