File size: 48,474 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 |
{
"paper_id": "1994",
"header": {
"generated_with": "S2ORC 1.0.0",
"date_generated": "2023-01-19T02:08:06.076735Z"
},
"title": "Using Part-of-speech Information in Word Alignment",
"authors": [
{
"first": "Jyun-Sheng",
"middle": [],
"last": "Chang",
"suffix": "",
"affiliation": {
"laboratory": "",
"institution": "National Tsing Hua University Hsinchu",
"location": {
"postCode": "30043",
"country": "Taiwan"
}
},
"email": "jschang@cs.nthu.edu.tw"
},
{
"first": "Huey-Chyun",
"middle": [],
"last": "Chen",
"suffix": "",
"affiliation": {
"laboratory": "",
"institution": "National Tsing Hua University Hsinchu",
"location": {
"postCode": "30043",
"country": "Taiwan"
}
},
"email": ""
}
],
"year": "",
"venue": null,
"identifiers": {},
"abstract": "We have developed a new program called PosAlign for identifying word correspondence in parallel text by using a translation model based on part-of-speech. The program takes the result from the part-of-speech taggers, and applies the POS-level constraints to produce word alignments. The parallel text of 55 suits of representative English sentence patterns (454 English sentences and their Chinese translations), were used to test the feasibility of this approach. By aligning cognates first before applying Brown et al.'s Model 2, the program converges faster and more correctly. Advantages of using a POS-based translation model include: (1) only parallel text of modest size is required for training. (2) Smaller number of model parameters are used and less storage is required. (3) A high degree generality is obtained, and the model is likely to be more applicable to text of other domains.",
"pdf_parse": {
"paper_id": "1994",
"_pdf_hash": "",
"abstract": [
{
"text": "We have developed a new program called PosAlign for identifying word correspondence in parallel text by using a translation model based on part-of-speech. The program takes the result from the part-of-speech taggers, and applies the POS-level constraints to produce word alignments. The parallel text of 55 suits of representative English sentence patterns (454 English sentences and their Chinese translations), were used to test the feasibility of this approach. By aligning cognates first before applying Brown et al.'s Model 2, the program converges faster and more correctly. Advantages of using a POS-based translation model include: (1) only parallel text of modest size is required for training. (2) Smaller number of model parameters are used and less storage is required. (3) A high degree generality is obtained, and the model is likely to be more applicable to text of other domains.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Abstract",
"sec_num": null
}
],
"body_text": [
{
"text": "The statistical machine translation model proposed by Brown et al. (1990) provides researchers with a new way of attacking the MT problem and has established itself as a new paradigm in MT research. Basically, there are two components in the statistical MT model, a prior and post prior probabilities. The a prior probabilities measure the possibility of a translation without considering the source given. These probabilities are often called the language model and the n-gram model seems to be the model of choice. The post prior probabilities measure the possibility of a translation under the condition that the source is given. This view of post prior probability of machine translation was first proposed and called translation model in Brown (1990) . Later Brown et al. (1993) proposed 5 translation models with different levels of approximation.",
"cite_spans": [
{
"start": 54,
"end": 73,
"text": "Brown et al. (1990)",
"ref_id": "BIBREF1"
},
{
"start": 743,
"end": 755,
"text": "Brown (1990)",
"ref_id": "BIBREF1"
},
{
"start": 764,
"end": 783,
"text": "Brown et al. (1993)",
"ref_id": "BIBREF6"
}
],
"ref_spans": [],
"eq_spans": [],
"section": "Introduction",
"sec_num": "1."
},
{
"text": "In order to estimate the parameters in the translation model, research has been conducted to find alignment between sentences (Brown et al. 1992 , Gale and Church 1992, Chen 993) , syntactic structures (Mutsumoto 19930), noun phrases (Kupiec 1993) , collocations (Smadja 1992) , and words Church 1990, 1993) in parallel text. Sentences from the two languages can be aligned with error rate as low as 0.6% using simple dynamic programming technique. The reason for the low error rates is threefolded. The alignments are predominantly 1-to-l. The lengths of counterpart sentences correlate closely. And finally, in most cases there are no cross dependencies between pairs of aligned sentences.",
"cite_spans": [
{
"start": 126,
"end": 144,
"text": "(Brown et al. 1992",
"ref_id": "BIBREF4"
},
{
"start": 147,
"end": 178,
"text": "Gale and Church 1992, Chen 993)",
"ref_id": null
},
{
"start": 234,
"end": 247,
"text": "(Kupiec 1993)",
"ref_id": "BIBREF14"
},
{
"start": 263,
"end": 276,
"text": "(Smadja 1992)",
"ref_id": null
},
{
"start": 289,
"end": 307,
"text": "Church 1990, 1993)",
"ref_id": null
}
],
"ref_spans": [],
"eq_spans": [],
"section": "Introduction",
"sec_num": "1."
},
{
"text": "Words, on the other hand, are much more difficult to align reliably, because they are less predominantly 1-to-l and the non-crossing constrain does not hold. Brown et al. proposed an approximation implementation of the EM algorithm to estimate the word-to-word translation model and to find the most likely word alignments for the aligned sentence pairs (Brown et al. 1989) . Even with the limitation of small dictionaries (1,700 most frequently used French words with translation sentences covered completely by a 1,000-word English vocabulary) and availability of 117,000 sentences, their method does not seem to produce robust estimation of word-based translation model (Gale and Church 1990 , p. 154). Another EM-based algorithm Word_align (Ido, Church and Gale 1993) with character alignment as the starting point, was shown to align 60.5% percent of the words correctly, and in 84% of the cases the offset from the correct alignment is at most 3. Gale and Church (1990) proposed using an \u03c7 2 -like associate measure for the plausibility of one-to-one word alignment, instead of the conditional probability derived from EM algorithm. The association ratios are consequently used in a dynamic programming algorithm to align words partially (only words with association higher than a threshold are considered for alignment). A certain threshold allows the algorithm to align 60% of the words and 95% of the alignments are correct.",
"cite_spans": [
{
"start": 354,
"end": 373,
"text": "(Brown et al. 1989)",
"ref_id": null
},
{
"start": 673,
"end": 694,
"text": "(Gale and Church 1990",
"ref_id": null
},
{
"start": 953,
"end": 975,
"text": "Gale and Church (1990)",
"ref_id": null
}
],
"ref_spans": [],
"eq_spans": [],
"section": "Introduction",
"sec_num": "1."
},
{
"text": "These lower success rates for word alignment are due to the predicament of very large parameter space of word-based translation models. Apparently, difficulties in training word-based translation model have constituted one of the major obstacles of using the statistical machine translation. We are interested in statistical MT model that can be trained sufficiently with modest amount of data. In this paper, we propose to work with part-of-speech-based translation models.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Introduction",
"sec_num": "1."
},
{
"text": "In the following sections, we will describe an efficient algorithm for word alignment that use part-ofspeech information. The algorithm first lookups a small dictionary of cognates to obtain initial alignment for the sentences. The partially aligned sentences are used to in an EM algorithm to train a statistical machine translation model on the level of part-of-speech. A small parallel text is used to test the feasibility of the approach. The model converges rapidly for the test set with reasonably good performance.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Introduction",
"sec_num": "1."
},
{
"text": "We use an example to introduce our framework for alignment Consider the parallel text (C, E) in ",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "The Word Alignment Algorithm",
"sec_num": "2."
},
{
"text": "The correct word alignments for the this pair of sentences are shown in Fig. 2 . As with sentence alignments, word alignments are not necessarily 1-to-l. In general, in a so-called i-to-j alignment, i words in one sentence might correspond to j words in its translation sentence. For example, the correct alignments of the sentences in Fig. 1 , consists of the following word alignments: 1-to-l alignment: (I, wuo) 0-to-l alignment: (-, zhen) 2-to-l alignment: (feel like, xiang) 1-to-l alignment: (crying, ku)",
"cite_spans": [],
"ref_spans": [
{
"start": 72,
"end": 78,
"text": "Fig. 2",
"ref_id": null
},
{
"start": 336,
"end": 342,
"text": "Fig. 1",
"ref_id": "FIGREF0"
}
],
"eq_spans": [],
"section": "Figure 1. Parallel Text",
"sec_num": null
},
{
"text": "The POS information, position of words, and word alignments are shown in Fig. 2 . For more example sentence of the parallel text, see Fig. 3 If a word-based translation model is used, the optimal word alignments are obtained using the probability t ( c | e) of the Chinese word c being the translation of the English word e where c \u03b5 (-, wuo, zhen, xiang, ku } and e \u03b5 { -, I , feel, like, crying }. The problem is that there are so many of these t parameters and they are extremely difficult to estimate. So, we propose the alternative of using a POS-to-POS translation model. Under the POS-based translation model, we calculate optimal alignments for (E, C) using the probability t ( c | e ) of the Chinese word with POS c being the translation of the English word with POS e where c \u03b5 (-, PRON, ADV, VERB } and e \u03b5 (-, PRON, VERB, PREP }. Obviously, there are far less part-of-speeches than there are words.",
"cite_spans": [],
"ref_spans": [
{
"start": 73,
"end": 79,
"text": "Fig. 2",
"ref_id": null
},
{
"start": 134,
"end": 140,
"text": "Fig. 3",
"ref_id": "FIGREF2"
}
],
"eq_spans": [],
"section": "Figure 1. Parallel Text",
"sec_num": null
},
{
"text": "We can estimate t ( c | e ) for POS-based model much the same way as for word-based model However, there are two common problems associated with statistical estimation of parameters. The first problem has to do with fertility involving deletion. It is considerably more difficult to estimate parameters involving deletion in sentence alignment (Simard 1990 , Brown 1990 ). The second problem is related to errors introduced due to distribution regularity. Because of the very nature of reliance on distribution regularity, statistical methods such as EM algorithms and inside-outside reestimation algorithms tend to introduce errors due to distribution regularity. For example, under statistically learned stochastic context free grammar, the period tends to be grouped with the previous word as a constituent (Periera and Schabes 1992). Aligning words regardless of their being function or content words creates similar problems. Firstly, the translation probability of any POS with \"-\" (empty) often unrealistically high. Similarly, the translation probability of a content POS with a pronoun is also unrealistically high due to the abundance of pronouns in the parallel text",
"cite_spans": [
{
"start": 344,
"end": 356,
"text": "(Simard 1990",
"ref_id": null
},
{
"start": 357,
"end": 369,
"text": ", Brown 1990",
"ref_id": "BIBREF1"
}
],
"ref_spans": [],
"eq_spans": [],
"section": "Figure 1. Parallel Text",
"sec_num": null
},
{
"text": "We have found that through initial alignment of cognates, we can deal with the two problems at the same time very effectively. Since the bulk of the cognate dictionary is function words which tends to appear more often than normal (distribution regularity) and involve non 1-to-l translation (deletion on one language side or fertile l-to-2, 2-to-l and 2-to-2 alignments). By aligning these pairs first, we can improve the accuracy of parameter estimation for the rest of the data. Consequently, the content words will be less likely to align incorrectly with function words. The idea of initial alignment is similar to using anchors (Brown 1990 ) and cognates (Simard, Foster and Isabelle 1992) in sentence alignment to deal with the issue of recovering from deletion of sentences. It is also related to using a partially bracketed corpus in learning a stochastic context-free grammar (Periera and Schabes 1992).",
"cite_spans": [
{
"start": 634,
"end": 645,
"text": "(Brown 1990",
"ref_id": "BIBREF1"
},
{
"start": 661,
"end": 695,
"text": "(Simard, Foster and Isabelle 1992)",
"ref_id": null
}
],
"ref_spans": [],
"eq_spans": [],
"section": "Figure 1. Parallel Text",
"sec_num": null
},
{
"text": "The cognate dictionary contains pairs of words in the two languages which are very likely to be used as mutual translation. Cognates include such words as pronouns, proper nouns, numerical expressions, and punctuation marks. For the example in Fig. 1 , we obtain the following initial alignments shown, in Fig. 4 . ",
"cite_spans": [],
"ref_spans": [
{
"start": 244,
"end": 250,
"text": "Fig. 1",
"ref_id": "FIGREF0"
},
{
"start": 306,
"end": 313,
"text": "Fig. 4",
"ref_id": null
}
],
"eq_spans": [],
"section": "Figure 1. Parallel Text",
"sec_num": null
},
{
"text": "PosAlign was first evaluated on a representative set of 55 English sentence patterns with 454 sentences and their translation in Chinese. Two taggers for English and Chinese are used in the experiment.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Implementation and Evaluation",
"sec_num": "3."
},
{
"text": "The entries of cognate dictionary came from the function words of an electronic Chinese-English dictionary (BDC 1990) and a translator handbook on function words (Wong 1989) , totaling some 3000 entries with various kinds of fertility (1-to-l, l-to-0,0-to-l, 1-to-many, many-to-1, and many-to-many).",
"cite_spans": [
{
"start": 162,
"end": 173,
"text": "(Wong 1989)",
"ref_id": null
}
],
"ref_spans": [],
"eq_spans": [],
"section": "Implementation and Evaluation",
"sec_num": "3."
},
{
"text": "The sentences in both languages run through taggers for the two languages first. The tagged text is used unedited. See Fig. 3 for some examples of tagged sentences. To make the training of translation model more effective, we have combined tags. For example, various inflection forms of English verbs, VB, VBD, VBG, etc., are combined into a single tag VERB. Similar combination was also done for the Chinese tagset. Pairs of counterpart sentences were then partially aligned with the help of cognate lookup. Probabilities for both POS mapping and distortion in positions of mapped word are derived. . . C5 Table 3 . The alignment result",
"cite_spans": [],
"ref_spans": [
{
"start": 119,
"end": 125,
"text": "Fig. 3",
"ref_id": "FIGREF2"
},
{
"start": 607,
"end": 614,
"text": "Table 3",
"ref_id": null
}
],
"eq_spans": [],
"section": "Implementation and Evaluation",
"sec_num": "3."
},
{
"text": "Notice that under the assumption of Model 2, we are not able to produce a 2-to-l alignment (feel like, xiang) as shown in Fig. 2 .",
"cite_spans": [],
"ref_spans": [
{
"start": 122,
"end": 128,
"text": "Fig. 2",
"ref_id": null
}
],
"eq_spans": [],
"section": "Implementation and Evaluation",
"sec_num": "3."
},
{
"text": "Compared with other word alignment algorithms, PosAlign does not require large amount of data for training, and was shown to produce alignment with high precision in complete alignment (taking 0-1 and 1-0 mapping into consideration). The program is much more general in the sense that it has the potential to performs as well on unrestricted text in a different domain.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Concluding remarks",
"sec_num": "4."
},
{
"text": "PosAlign provide an example of the way how statistical machine translation can be made more general and modular by incorporating other language processing modules such as part-of-speech taggers. We are currently expanded our cognate dictionary for better results. A second evaluation of PosAlign is being carried out for some 70 short bilingual news reportage. Initial results from the experiments are quite encouraging.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Concluding remarks",
"sec_num": "4."
}
],
"back_matter": [],
"bib_entries": {
"BIBREF0": {
"ref_id": "b0",
"title": "BDC Chinese-English Electronic Dictionary",
"authors": [
{
"first": "",
"middle": [],
"last": "Behavior Design Co",
"suffix": ""
}
],
"year": 1990,
"venue": "",
"volume": "",
"issue": "",
"pages": "",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Behavior Design Co. (1990) \"BDC Chinese-English Electronic Dictionary,\" Hsinchu, Taiwan.",
"links": null
},
"BIBREF1": {
"ref_id": "b1",
"title": "A Statistical Approach to Machine Translation",
"authors": [
{
"first": "P",
"middle": [],
"last": "Brown",
"suffix": ""
}
],
"year": 1990,
"venue": "Computational Linguistics",
"volume": "16",
"issue": "",
"pages": "79--85",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Brown, P., et al., (1990) \"A Statistical Approach to Machine Translation,\" Computational Linguistics, vol. 16, pp. 79-85.",
"links": null
},
"BIBREF2": {
"ref_id": "b2",
"title": "Word Sense Disambiguation using Statistical Methods",
"authors": [
{
"first": "P",
"middle": [],
"last": "Brown",
"suffix": ""
}
],
"year": 1991,
"venue": "Proceedings of the 29th Annual ACL Meeting",
"volume": "",
"issue": "",
"pages": "264--270",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Brown, P., et. al. (1991), \"Word Sense Disambiguation using Statistical Methods,\" Proceedings of the 29th Annual ACL Meeting, pp. 264-270.",
"links": null
},
"BIBREF3": {
"ref_id": "b3",
"title": "Analysis, Statistical Transfer, and Synthesis in Machine Translation",
"authors": [
{
"first": "P",
"middle": [],
"last": "Brown",
"suffix": ""
}
],
"year": 1992,
"venue": "Proceedings of the 4th International Conference on Theorectical and Methodological Issues in Machine Translation",
"volume": "",
"issue": "",
"pages": "83--100",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Brown, P., et al. (1992), \"Analysis, Statistical Transfer, and Synthesis in Machine Translation\" Proceedings of the 4th International Conference on Theorectical and Methodological Issues in Machine Translation, pp. 83-100.",
"links": null
},
"BIBREF4": {
"ref_id": "b4",
"title": "Dividing and Conquering Long Sentences in a Translation System",
"authors": [
{
"first": "P",
"middle": [],
"last": "Brown",
"suffix": ""
},
{
"first": "S",
"middle": [
"A"
],
"last": "Pietra",
"suffix": ""
},
{
"first": "V",
"middle": [
"J D"
],
"last": "Pietra",
"suffix": ""
},
{
"first": "R",
"middle": [],
"last": "Mercer",
"suffix": ""
}
],
"year": 1992,
"venue": "Proceedings of the Speech and Natural Language Workshop",
"volume": "",
"issue": "",
"pages": "267--271",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Brown, P., S.A. Pietra, V.J.D. Pietra, and R. Mercer (1992), \"Dividing and Conquering Long Sentences in a Translation System,\" Proceedings of the Speech and Natural Language Workshop, pp. 267-271.",
"links": null
},
"BIBREF5": {
"ref_id": "b5",
"title": "Aligning Sentences in Parallel Text",
"authors": [
{
"first": "",
"middle": [],
"last": "Brown",
"suffix": ""
}
],
"year": 1991,
"venue": "Proceedings of the Annual ACL meeting pp",
"volume": "",
"issue": "",
"pages": "169--176",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Brown et al. (1991) \"Aligning Sentences in Parallel Text,\" Proceedings of the Annual ACL meeting pp. 169-176.",
"links": null
},
"BIBREF6": {
"ref_id": "b6",
"title": "The Mathematics of Machine Translation: Parameter Estimation",
"authors": [
{
"first": "P",
"middle": [],
"last": "Brown",
"suffix": ""
}
],
"year": 1993,
"venue": "Computational Linguistics",
"volume": "19",
"issue": "2",
"pages": "",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Brown P., et al. (1993) \"The Mathematics of Machine Translation: Parameter Estimation,\" Computational Linguistics, 19(2), pp. -.",
"links": null
},
"BIBREF7": {
"ref_id": "b7",
"title": "Aligning Sentences in Bilingual Corpora Using Lexical Information",
"authors": [
{
"first": "S",
"middle": [],
"last": "Chen",
"suffix": ""
}
],
"year": 1993,
"venue": "Proceedings of the Annual ACL meeting",
"volume": "",
"issue": "",
"pages": "9--17",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Chen, S. (1993) \"Aligning Sentences in Bilingual Corpora Using Lexical Information,\" Proceedings of the Annual ACL meeting pp. 9-17.",
"links": null
},
"BIBREF8": {
"ref_id": "b8",
"title": "A Stochastic Parts Program and Noun Phrase Parser for Unrestricted Text",
"authors": [
{
"first": "K",
"middle": [
"W"
],
"last": "Church",
"suffix": ""
}
],
"year": 1988,
"venue": "Proceedings of The Second Applied Natural Language Processing, ACL",
"volume": "",
"issue": "",
"pages": "136--143",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Church, K.W. (1988) \"A Stochastic Parts Program and Noun Phrase Parser for Unrestricted Text,\" Proceedings of The Second Applied Natural Language Processing, ACL, pp. 136-143.",
"links": null
},
"BIBREF9": {
"ref_id": "b9",
"title": "Robust Bilingual Word Alignment for Machine Aided Translation",
"authors": [
{
"first": "I",
"middle": [],
"last": "Dagan",
"suffix": ""
},
{
"first": "K",
"middle": [
"W"
],
"last": "Church",
"suffix": ""
},
{
"first": "W",
"middle": [
"A"
],
"last": "Gale",
"suffix": ""
}
],
"year": 1993,
"venue": "Proceedings of the Annual ACL meeting",
"volume": "",
"issue": "",
"pages": "1--8",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Dagan, I., K.W. Church, and W.A. Gale. (1993) \"Robust Bilingual Word Alignment for Machine Aided Translation,\" Proceedings of the Annual ACL meeting pp. 1-8.",
"links": null
},
"BIBREF10": {
"ref_id": "b10",
"title": "Identifying Word Correspondences in Parallel Texts",
"authors": [
{
"first": "W",
"middle": [],
"last": "Gale",
"suffix": ""
},
{
"first": "K",
"middle": [],
"last": "Church",
"suffix": ""
}
],
"year": 1991,
"venue": "Proceedings of Speech and Natural Language Workshop",
"volume": "",
"issue": "",
"pages": "152--158",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Gale, W. and K. Church. (1991) \"Identifying Word Correspondences in Parallel Texts,\" Proceedings of Speech and Natural Language Workshop, pp. 152-158.",
"links": null
},
"BIBREF11": {
"ref_id": "b11",
"title": "Using Bilingual Materials to Develop Word Sense Disambiguation Methods",
"authors": [
{
"first": "W",
"middle": [],
"last": "Gale",
"suffix": ""
},
{
"first": "K",
"middle": [],
"last": "Church",
"suffix": ""
},
{
"first": "D",
"middle": [],
"last": "Yarowsky",
"suffix": ""
}
],
"year": 1992,
"venue": "Proceedings of the 4th International Conference on Theoretic and Methodological Issues in Machine Translation",
"volume": "",
"issue": "",
"pages": "101--112",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Gale, W., K. Church, and D. Yarowsky (1992), \"Using Bilingual Materials to Develop Word Sense Disambiguation Methods, \"Proceedings of the 4th International Conference on Theoretic and Methodological Issues in Machine Translation, pp. 101-112.",
"links": null
},
"BIBREF12": {
"ref_id": "b12",
"title": "Word Correspondence and Sense Tagging of Parallel Text",
"authors": [
{
"first": "S",
"middle": [
"J"
],
"last": "Ker",
"suffix": ""
},
{
"first": "J",
"middle": [
"S"
],
"last": "Chang",
"suffix": ""
}
],
"year": 1994,
"venue": "",
"volume": "",
"issue": "",
"pages": "",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Ker, SJ. and J.S. Chang. (1994) \"Word Correspondence and Sense Tagging of Parallel Text,\" Technical Report NSC82-0408-E-007-195, National Science Council, Taiwan.",
"links": null
},
"BIBREF13": {
"ref_id": "b13",
"title": "Lexical Ambiguity and Information Retrieval",
"authors": [
{
"first": "R",
"middle": [],
"last": "Krovetz",
"suffix": ""
},
{
"first": "W",
"middle": [],
"last": "Croft",
"suffix": ""
}
],
"year": 1992,
"venue": "ACM Transaction on Information Systems",
"volume": "",
"issue": "",
"pages": "115--141",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Krovetz, R. and W. Croft. (1992) \"Lexical Ambiguity and Information Retrieval,\" ACM Transaction on Information Systems, pp. 115-141.",
"links": null
},
"BIBREF14": {
"ref_id": "b14",
"title": "An Algorithm for Finding Noun Phrase Correspondences in Bilingual Corpora",
"authors": [
{
"first": "",
"middle": [],
"last": "Kupiec",
"suffix": ""
}
],
"year": 1993,
"venue": "Proceedings of the Annual ACL meeting",
"volume": "",
"issue": "",
"pages": "17--22",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Kupiec. (1993) \"An Algorithm for Finding Noun Phrase Correspondences in Bilingual Corpora\" Proceedings of the Annual ACL meeting, pp. 17-22.",
"links": null
},
"BIBREF15": {
"ref_id": "b15",
"title": "Structural Matching of Parallel Texts",
"authors": [
{
"first": "Y",
"middle": [],
"last": "Matsumoto",
"suffix": ""
}
],
"year": 1993,
"venue": "Proceedings of the Annual ACL meeting",
"volume": "",
"issue": "",
"pages": "23--30",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Matsumoto, Y. et al. (1993) \"Structural Matching of Parallel Texts\" Proceedings of the Annual ACL meeting, pp. 23-30.",
"links": null
},
"BIBREF16": {
"ref_id": "b16",
"title": "Inside-Outside Reestimation from Partially Bracked Corpora",
"authors": [
{
"first": "F",
"middle": [],
"last": "Pereira",
"suffix": ""
},
{
"first": "Y",
"middle": [],
"last": "Schabes",
"suffix": ""
}
],
"year": 1992,
"venue": "Proceedings of the Annual ACL meeting",
"volume": "",
"issue": "",
"pages": "128--135",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Pereira, F. and Y.Schabes. (1992) \"Inside-Outside Reestimation from Partially Bracked Corpora,\" Proceedings of the Annual ACL meeting, pp. 128-135.",
"links": null
},
"BIBREF17": {
"ref_id": "b17",
"title": "How to Compile a Bilingual Collocational Lexicon Automatically",
"authors": [
{
"first": "",
"middle": [],
"last": "Smadja",
"suffix": ""
}
],
"year": 1993,
"venue": "Proceedings of the AAAI Workshop on Statistically-based Natural Language Programming Techniques",
"volume": "",
"issue": "",
"pages": "57--63",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Smadja. (1993) \"How to Compile a Bilingual Collocational Lexicon Automatically. Proceedings of the AAAI Workshop on Statistically-based Natural Language Programming Techniques, pp. 57-63.",
"links": null
},
"BIBREF18": {
"ref_id": "b18",
"title": "Using Cognates to Align Sentences in Parallel Corpora",
"authors": [
{
"first": "M",
"middle": [],
"last": "Simard",
"suffix": ""
},
{
"first": "G",
"middle": [],
"last": "Foster",
"suffix": ""
},
{
"first": "P",
"middle": [],
"last": "Isabelle",
"suffix": ""
}
],
"year": 1993,
"venue": "Proceedings of the 4th International Conference on Theoretical and Methodological Issues in Machine Translation",
"volume": "",
"issue": "",
"pages": "",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Simard, M., G. Foster and P. Isabelle. (1993) \"Using Cognates to Align Sentences in Parallel Corpora, Proceedings of the 4th International Conference on Theoretical and Methodological Issues in Machine Translation, Montreal, Canada",
"links": null
},
"BIBREF19": {
"ref_id": "b19",
"title": "A Translation Handbook of Chinese-English Function Words",
"authors": [
{
"first": "I",
"middle": [],
"last": "Wong",
"suffix": ""
}
],
"year": 1990,
"venue": "",
"volume": "",
"issue": "",
"pages": "",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Wong, I. (1990) \"A Translation Handbook of Chinese-English Function Words\", Joint Publishing Co., Hong Kong.",
"links": null
}
},
"ref_entries": {
"FIGREF0": {
"type_str": "figure",
"text": "Both the Chinese and English sentences are tagged with part-of-speech information.",
"num": null,
"uris": null
},
"FIGREF1": {
"type_str": "figure",
"text": "4.c Yuehan(John)/PRON shihu(seem)/ADV shi(be)/V ge(piece)/M congmi(clever)/ADJ de(CLITCS)/S xiaohai(kid)/N./. 5.e The/DT sun/NN rises/VBZ in/IN the/DT east/NN./. 5.c Taiyang(sun)/N you(from)/PREP dongfang(east)/N shen(rise)/V qi(up)/PART./.",
"num": null,
"uris": null
},
"FIGREF2": {
"type_str": "figure",
"text": "Tagged parallel text",
"num": null,
"uris": null
},
"FIGREF3": {
"type_str": "figure",
"text": "Cognates and partially aligned sentences The sequence of POS's {(-, VERB, PREP, VERB); (-, ADV, VERB, VERB)} is then used in training Brown et al.'s Model 2. The PosAlign algorithm is summarized as follows: Algorithm PosAlign Tag the parallel text with part-of-speeches 2. Initial alignment with the help of cognate lookup 3. Train the translation model iteratively using the unaligned part of the POS sequences",
"num": null,
"uris": null
},
"TABREF3": {
"html": null,
"num": null,
"type_str": "table",
"content": "<table><tr><td colspan=\"5\">_______ English POS ________________ Chinese PQS _______ Prob.</td></tr><tr><td>2</td><td>infinite to</td><td>VERB</td><td>verb</td><td>0.162</td></tr><tr><td>2</td><td>infinite to</td><td>-</td><td>empty</td><td>0.119</td></tr><tr><td>D</td><td>cardinal number</td><td>Q</td><td>numeral</td><td>0.371</td></tr><tr><td>D</td><td>cardinal number</td><td>M</td><td>measure noun</td><td>0.120</td></tr><tr><td>E</td><td>existential there</td><td>VERB</td><td>verb</td><td>0.162</td></tr><tr><td>E</td><td>existential there</td><td>-</td><td>empty</td><td>0.129</td></tr><tr><td colspan=\"2\">PREP preposition</td><td>PREP</td><td>preposition</td><td>0.358</td></tr><tr><td colspan=\"2\">PREP preposition</td><td>-</td><td>empty</td><td>0.135</td></tr><tr><td colspan=\"2\">PREP preposition</td><td>ADV</td><td>adverb</td><td>0.102</td></tr><tr><td colspan=\"2\">PREP preposition</td><td>VERB</td><td>verb</td><td>0.088</td></tr><tr><td colspan=\"2\">PREP preposition</td><td>N</td><td>noun</td><td>0.066</td></tr><tr><td>ADJ</td><td>adjective</td><td>ADJ</td><td>adjective</td><td>0.206</td></tr><tr><td>ADJ</td><td>adjective</td><td>VERB</td><td>verb</td><td>0.145</td></tr><tr><td>ADJ</td><td>adjective</td><td>-</td><td>empty</td><td>0.128</td></tr><tr><td>ADJ</td><td>adjective</td><td>ADV</td><td>adverb</td><td>0.100</td></tr><tr><td>M</td><td>modal auxiliary</td><td>X</td><td>modal auxiliary</td><td>0.177</td></tr><tr><td>M</td><td>modal auxiliary</td><td>VERB</td><td>verb</td><td>0.165</td></tr><tr><td>M</td><td>modal auxiliary</td><td>-</td><td>empty</td><td>0.109</td></tr><tr><td>N</td><td>noun</td><td>N</td><td>noun</td><td>0.218</td></tr><tr><td>N</td><td>noun</td><td>-</td><td>empty</td><td>0.128</td></tr><tr><td colspan=\"2\">ADV adverb</td><td>ADV</td><td>adverb</td><td>0.203</td></tr><tr><td colspan=\"2\">ADV adverb</td><td>-</td><td>empty</td><td>0.116</td></tr><tr><td>T</td><td>determiner/pronoun</td><td>-</td><td>empty</td><td>0.121</td></tr><tr><td>T</td><td>determiner/pronoun</td><td>D</td><td>determiner</td><td>0.057</td></tr><tr><td colspan=\"2\">VERB verb</td><td>VERB</td><td>verb</td><td>0.244</td></tr><tr><td colspan=\"2\">VERB verb</td><td>-</td><td>empty</td><td>0.134</td></tr><tr><td colspan=\"2\">VERB verb</td><td>ADV</td><td>adverb</td><td>0.097</td></tr></table>",
"text": "presents the trained translation model. Only POS-to-POS mapping with highest probabilities are shown. POS Translation Model Using the example inFig. 1, we show how alignment is done using the trained model. As is being done in the training phrase, the cognates are aligned first, resulting in the following alignment being identified:"
}
}
}
} |