File size: 67,485 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 | {
"paper_id": "P07-1017",
"header": {
"generated_with": "S2ORC 1.0.0",
"date_generated": "2023-01-19T08:51:00.576142Z"
},
"title": "Generating Complex Morphology for Machine Translation",
"authors": [
{
"first": "Einat",
"middle": [],
"last": "Minkov",
"suffix": "",
"affiliation": {},
"email": "einatm@cs.cmu.edu"
},
{
"first": "Kristina",
"middle": [],
"last": "Toutanova",
"suffix": "",
"affiliation": {},
"email": ""
},
{
"first": "Hisami",
"middle": [],
"last": "Suzuki",
"suffix": "",
"affiliation": {},
"email": "hisamis@microsoft.com"
}
],
"year": "",
"venue": null,
"identifiers": {},
"abstract": "We present a novel method for predicting inflected word forms for generating morphologically rich languages in machine translation. We utilize a rich set of syntactic and morphological knowledge sources from both source and target sentences in a probabilistic model, and evaluate their contribution in generating Russian and Arabic sentences. Our results show that the proposed model substantially outperforms the commonly used baseline of a trigram target language model; in particular, the use of morphological and syntactic features leads to large gains in prediction accuracy. We also show that the proposed method is effective with a relatively small amount of data.",
"pdf_parse": {
"paper_id": "P07-1017",
"_pdf_hash": "",
"abstract": [
{
"text": "We present a novel method for predicting inflected word forms for generating morphologically rich languages in machine translation. We utilize a rich set of syntactic and morphological knowledge sources from both source and target sentences in a probabilistic model, and evaluate their contribution in generating Russian and Arabic sentences. Our results show that the proposed model substantially outperforms the commonly used baseline of a trigram target language model; in particular, the use of morphological and syntactic features leads to large gains in prediction accuracy. We also show that the proposed method is effective with a relatively small amount of data.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Abstract",
"sec_num": null
}
],
"body_text": [
{
"text": "Machine Translation (MT) quality has improved substantially in recent years due to applying data intensive statistical techniques. However, state-ofthe-art approaches are essentially lexical, considering every surface word or phrase in both the source sentence and the corresponding translation as an independent entity. A shortcoming of this word-based approach is that it is sensitive to data sparsity. This is an issue of importance as aligned corpora are an expensive resource, which is not abundantly available for many language pairs. This is particularly problematic for morphologically rich languages, where word stems are realized in many different surface forms, which exacerbates the sparsity problem.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Introduction",
"sec_num": "1"
},
{
"text": "In this paper, we explore an approach in which words are represented as a collection of morphological entities, and use this information to aid in MT for morphologically rich languages. Our goal is twofold: first, to allow generalization over morphology to alleviate the data sparsity problem in morphology generation. Second, to model syntactic coherence in the form of morphological agreement in the target language to improve the generation of morphologically rich languages. So far, this problem has been addressed in a very limited manner in MT, most typically by using a target language model.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Introduction",
"sec_num": "1"
},
{
"text": "In the framework suggested in this paper, we train a model that predicts the inflected forms of a sequence of word stems in a target sentence, given the corresponding source sentence. We use word and word alignment information, as well as lexical resources that provide morphological information about the words on both the source and target sides. Given a sentence pair, we also obtain syntactic analysis information for both the source and translated sentences. We generate the inflected forms of words in the target sentence using all of the available information, using a log-linear model that learns the relevant mapping functions.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Introduction",
"sec_num": "1"
},
{
"text": "As a case study, we focus on the English-Russian and English-Arabic language pairs. Unlike English, Russian and Arabic have very rich systems of morphology, each with distinct characteristics. Translating from a morphology-poor to a morphologyrich language is especially challenging since detailed morphological information needs to be decoded from a language that does not encode this information or does so only implicitly (Koehn, 2005) . We believe that these language pairs are represen-tative in this respect and therefore demonstrate the generality of our approach.",
"cite_spans": [
{
"start": 425,
"end": 438,
"text": "(Koehn, 2005)",
"ref_id": "BIBREF7"
}
],
"ref_spans": [],
"eq_spans": [],
"section": "Introduction",
"sec_num": "1"
},
{
"text": "There are several contributions of this work. First, we propose a general approach that shows promise in addressing the challenges of MT into morphologically rich languages. We show that the use of both syntactic and morphological information improves translation quality. We also show the utility of source language information in predicting the word forms of the target language. Finally, we achieve these results with limited morphological resources and training data, suggesting that the approach is generally useful for resource-scarce language pairs. Table 1 describes the morphological features relevant to Russian and Arabic, along with their possible values. The rightmost column in the table refers to the morphological features that are shared by Russian and Arabic, including person, number, gender and tense. While these features are fairly generic (they are also present in English), note that Russian includes an additional gender (neuter) and Arabic has a distinct number notion for two (dual). A central dimension of Russian morphology is case marking, realized as suffixation on nouns and nominal modifiers 1 . The Russian case feature includes six possible values, representing the notions of subject, direct object, location, etc. In Arabic, like other Semitic languages, word surface forms may include proclitics and enclitics (or prefixes and suffixes as we refer to them in this paper), concatenated to inflected stems. For nouns, prefixes include conjunctions (wa: \"and\", fa: \"and, so\"), prepositions (bi: \"by, with\", ka: \"like, such as\", li: \"for, to\") and a determiner, and suffixes include possessive pronouns. Verbal prefixes include conjunction and negation, and suffixes include object pronouns. Both object and possessive pronouns are captured by an indicator function for its presence or absence, as well as by the features that indicate their person, number and gender. As can be observed from the table, a large number of surface inflected forms can be generated by the combination of these features, making the morphological generation of these languages a non-trivial task.",
"cite_spans": [],
"ref_spans": [
{
"start": 557,
"end": 564,
"text": "Table 1",
"ref_id": "TABREF1"
}
],
"eq_spans": [],
"section": "Introduction",
"sec_num": "1"
},
{
"text": "Morphologically complex languages also tend to display a rich system of agreements. In Russian, for example, adjectives agree with head nouns in number, gender and case, and verbs agree with the subject noun in person and number (past tense verbs agree in gender and number). Arabic has a similarly rich system of agreement, with unique characteristics. For example, in addition to agreement involving person, number and gender, it also requires a determiner for each word in a definite noun phrase with adjectival modifiers; in a noun compound, a determiner is attached to the last noun in the chain. Also, non-human subject plural nouns require the verb to be inflected in a singular feminine form. Generating these morphologically complex languages is therefore more difficult than generating English in terms of capturing the agreement phenomena.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Russian and Arabic Morphology",
"sec_num": "2"
},
{
"text": "The use of morphological features in language modelling has been explored in the past for morphologyrich languages. For example, (Duh and Kirchhoff, 2004) showed that factored language models, which consider morphological features and use an optimized backoff policy, yield lower perplexity.",
"cite_spans": [
{
"start": 129,
"end": 154,
"text": "(Duh and Kirchhoff, 2004)",
"ref_id": "BIBREF2"
}
],
"ref_spans": [],
"eq_spans": [],
"section": "Related Work",
"sec_num": "3"
},
{
"text": "In the area of MT, there has been a large body of work attempting to modify the input to a translation system in order to improve the generated alignments for particular language pairs. For example, it has been shown (Lee, 2004) that determiner segmentation and deletion in Arabic sentences in an Arabic-to-English translation system improves sentence alignment, thus leading to improved overall translation quality. Another work (Koehn and Knight, 2003) showed improvements by splitting compounds in German. (Nie\u00dfen and Ney, 2004) demonstrated that a similar level of alignment quality can be achieved with smaller corpora applying morpho-syntactic source restructuring, using hierarchical lexicon models, in translating from German into English. (Popovi\u0107 and Ney, 2004) experimented successfully with translating from inflectional languages into English making use of POS tags, word stems and suffixes in the source language. More recently, (Goldwater and McClosky, 2005) In general, this line of work focused on translating from morphologically rich languages into English; there has been limited research in MT in the opposite direction. Koehn (2005) includes a survey of statistical MT systems in both directions for the Europarl corpus, and points out the challenges of this task. A recent work (El-Kahlout and Oflazer, 2006) experimented with English-to-Turkish translation with limited success, suggesting that inflection generation given morphological features may give positive results. In the current work, we suggest a probabilistic framework for morphology generation performed as post-processing. It can therefore be considered as complementary to the techniques described above. Our approach is general in that it is not specific to a particular language pair, and is novel in that it allows modelling of agreement on the target side. The framework suggested here is most closely related to (Suzuki and Toutanova, 2006) , which uses a probabilistic model to generate Japanese case markers for English-to-Japanese MT. This work can be viewed as a generalization of (Suzuki and Toutanova, 2006) in that our model generates inflected forms of words, and is not limited to generating a small, closed set of case markers. In addition, the morphology generation problem is more challenging in that it requires handling of complex agreement phenomena along multiple morphological dimensions.",
"cite_spans": [
{
"start": 217,
"end": 228,
"text": "(Lee, 2004)",
"ref_id": "BIBREF8"
},
{
"start": 430,
"end": 454,
"text": "(Koehn and Knight, 2003)",
"ref_id": "BIBREF6"
},
{
"start": 509,
"end": 531,
"text": "(Nie\u00dfen and Ney, 2004)",
"ref_id": "BIBREF11"
},
{
"start": 748,
"end": 771,
"text": "(Popovi\u0107 and Ney, 2004)",
"ref_id": "BIBREF13"
},
{
"start": 943,
"end": 973,
"text": "(Goldwater and McClosky, 2005)",
"ref_id": "BIBREF4"
},
{
"start": 1142,
"end": 1154,
"text": "Koehn (2005)",
"ref_id": "BIBREF7"
},
{
"start": 1906,
"end": 1934,
"text": "(Suzuki and Toutanova, 2006)",
"ref_id": "BIBREF15"
},
{
"start": 2079,
"end": 2107,
"text": "(Suzuki and Toutanova, 2006)",
"ref_id": "BIBREF15"
}
],
"ref_spans": [],
"eq_spans": [],
"section": "Related Work",
"sec_num": "3"
},
{
"text": "In this section, we define the task of of morphological generation as inflection prediction, as well as the lexical operations relevant for the task.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Inflection Prediction Framework",
"sec_num": "4"
},
{
"text": "Morphological analysis can be performed by applying language specific rules. These may include a full-scale morphological analysis with contextual disambiguation, or, when such resources are not available, simple heuristic rules, such as regarding the last few characters of a word as its morphogical suffix. In this work, we assume that lexicons L S and L T are available for the source and translation languages, respectively. Such lexicons can be created manually, or automatically from data. Given a lexicon L and a surface word w, we define the following operations:",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Morphology Analysis and Generation",
"sec_num": "4.1"
},
{
"text": "\u2022 Stemming -let S w = {s 1 , ..., s l } be the set of possible morphological stems (lemmas) of w according to L. 2",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Morphology Analysis and Generation",
"sec_num": "4.1"
},
{
"text": "\u2022 Inflection -let I w = {i 1 , ..., i m } be the set of surface form words that have the same stem as w. That is, i \u2208 I w iff S i S w = \u2205.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Morphology Analysis and Generation",
"sec_num": "4.1"
},
{
"text": "\u2022 Morphological analysis -let A w = {a 1 , ..., a v } be the set of possible morphological analyses for w. A morphological analysis a is a vector of categorical values, where the dimensions and possible values for each dimension in the vector representation space are defined by L.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Morphology Analysis and Generation",
"sec_num": "4.1"
},
{
"text": "We assume that we are given aligned sentence pairs, where a sentence pair includes a source and a tar- ..w n be the output of a MT system in the target language. This sentence can be converted into the corresponding stem set sequence S 1 , ...S t , ...S n , applying the stemming operation. Then the task is, for every stem set S t in the output sentence, to predict an inflection y t from its inflection set I t . The predicted inflections should both reflect the meaning conveyed by the source sentence, and comply with the agreement rules of the target language. 3 Figure 1 shows an example of an aligned English-Russian sentence pair: on the source (English) side, POS tags and word dependency structure are indicated by solid arcs. The alignments between English and Russian words are indicated by the dotted lines. The dependency structure on the Russian side, indicated by solid arcs, is given by a treelet MT system in our case (see Section 6.1), projected from the word dependency structure of English and word alignment information. Note that the Russian sentence displays agreement in number and gender between the subject noun (raspredelenie) and the predicate (zaversheno); note also that resursov is in genitive case, as it modifies the noun on its left.",
"cite_spans": [],
"ref_spans": [
{
"start": 568,
"end": 576,
"text": "Figure 1",
"ref_id": "FIGREF0"
}
],
"eq_spans": [],
"section": "The Task",
"sec_num": "4.2"
},
{
"text": "Our learning framework uses a Maximum Entropy Markov model (McCallum et al., 2000) . The model decomposes the overall probability of a predicted inflection sequence into a product of local probabilities for individual word predictions. The local probabilities are conditioned on the previous k predictions. The model implemented here is of second order: at any decision point t we condition the probability distribution over labels on the previous two predictions y t\u22121 and y t\u22122 in addition to the given (static) word context from both the source and target sentences. That is, the probability of a predicted inflection sequence is defined as follows:",
"cite_spans": [
{
"start": 59,
"end": 82,
"text": "(McCallum et al., 2000)",
"ref_id": "BIBREF10"
}
],
"ref_spans": [],
"eq_spans": [],
"section": "A Probabilistic Model",
"sec_num": "5.1"
},
{
"text": "p(y | x) = n t=1 p(y t | y t\u22121 , y t\u22122 , x t ), y t \u2208 I t",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "A Probabilistic Model",
"sec_num": "5.1"
},
{
"text": "where x t denotes the given context at position t and I t is the set of inflections corresponding to S t , from which the model should choose y t .",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "A Probabilistic Model",
"sec_num": "5.1"
},
{
"text": "The features we constructed pair up predicates on the context (x, y t\u22121 , y t\u22122 ) and the target label (y t ).",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "A Probabilistic Model",
"sec_num": "5.1"
},
{
"text": "In the suggested framework, it is straightforward to encode the morphological properties of a word, in addition to its surface inflected form. For example, for a particular inflected word form y t and its context, the derived paired features may include:",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "A Probabilistic Model",
"sec_num": "5.1"
},
{
"text": "\u03c6 k =",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "A Probabilistic Model",
"sec_num": "5.1"
},
{
"text": "1 if surface word yt is y \u2032 and s \u2032 \u2208 St+1 0 otherwise \u03c6 k+1 = 1 if Gender(yt) =\"Fem\" and Gender(yt\u22121) =\"Fem\" 0 otherwise",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "A Probabilistic Model",
"sec_num": "5.1"
},
{
"text": "In the first example, a given neighboring stem set S t+1 is used as a context feature for predicting the target word y t . The second feature captures the gender agreement with the previous word. This is possible because our model is of second order. Thus, we can derive context features describing the morphological properties of the two previous predictions. 4 Note that our model is not a simple multi-class classifier, because our features are shared across multiple target labels. For example, the gender feature above applies to many different inflected forms. Therefore, it is a structured prediction model, where the structure is defined by the morphological properties of the target predictions, in addition to the word sequence decomposition.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "A Probabilistic Model",
"sec_num": "5.1"
},
{
"text": "The information available for estimating the distribution over y t can be split into several categories, corresponding to feature source. The first major distinction is monolingual versus bilingual features: monolingual features refer only to the context (and predicted label) in the target language, while bilingual features have access to information in the source sentences, obtained by traversing the word alignment links from target words to a (set of) source words, as shown in Figure 1 . Both monolingual and bilingual features can be further split into three classes: lexical, morphological and syntactic. Lexical features refer to surface word forms, as well as their stems. Since our model is of second order, our monolingual lexical features include the features of a standard word trigram language model. Furthermore, since our model is discriminative (predicting word forms given their stems), the monolingual lexical model can use stems in addition to predicted words for the left and current position, as well as stems from the right context. Morphological features are those that refer to the features given in Table 1 . Morphological information is used in describing the target label as well as its context, and is intended to capture morphological generalizations. Finally, syntactic features can make use of syntactic analyses of the source and target sentences. Such analyses may be derived for the target language, using the pre-stemmed sentence. Without loss of generality, we will use here a dependency parsing paradigm. Given a syntactic analysis, one can construct syntactic features; for example, the stem of the parent word of y t . Syntactic features are expected to be useful in capturing agreement phenomena. Table 2 gives the full set of suggested features for Russian and Arabic, detailed by type. For monolingual lexical features, we consider the stems of the predicted word and its immediately adjacent words, in addition to traditional word bigram and trigram features. For monolingual morphological features, we consider the morphological attributes of the two previously predicted words and the current prediction; for monolingual syntactic features, we use the stem of the parent node.",
"cite_spans": [],
"ref_spans": [
{
"start": 484,
"end": 492,
"text": "Figure 1",
"ref_id": "FIGREF0"
},
{
"start": 1127,
"end": 1134,
"text": "Table 1",
"ref_id": "TABREF1"
},
{
"start": 1740,
"end": 1747,
"text": "Table 2",
"ref_id": "TABREF2"
}
],
"eq_spans": [],
"section": "Feature Categories",
"sec_num": "5.2"
},
{
"text": "The bilingual features include the set of words aligned to the focus word at position t, where they are treated as bag-of-words, i.e., each aligned word is assigned a separate feature. Bilingual lexical features can refer to words aligned to y t as all as words aligned to its immediate neighbors y t\u22121 and y t+1 . Bilingual morphological and syntactic features refer to the features of the source language, which are expected to be useful for predicting morphology in the target language. For example, the bilingual Det (determiner) feature is computed according to the source dependency tree: if a child of a word aligned to w t is a determiner, then the feature value is assigned its surface word form (such as a or the). The bilingual Prep feature is computed similarly, by checking the parent chain of the word aligned to w t for the existence of a preposition. This feature is hoped to be useful for predicting Arabic inflected forms with a prepositional prefix, as well as for predicting case marking in Russian. The bilingual ObjPron and PossPron features represent any object pronoun of the word aligned to w t and a preceding possessive pronoun, respectively. These features are expected to map to the object and possessive pronoun features in Arabic. Finally, the bilingual Compound feature checks whether a word appears as part of a noun compound in the English source. f this is the case, the feature is assigned the value of \"head\" or \"dependent\". This feature is relevant for predicting a genitive case in Russian and definiteness in Arabic.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Features",
"sec_num": "5.3"
},
{
"text": "In order to evaluate the effectiveness of the suggested approach, we performed reference experiments, that is, using the aligned sentence pairs of 132 reference translations rather than the output of an MT system as input. 5 This allows us to evaluate our method with a reduced noise level, as the words and word order are perfect in reference translations. These experiments thus constitute a preliminary step for tackling the real task of inflecting words in MT.",
"cite_spans": [
{
"start": 223,
"end": 224,
"text": "5",
"ref_id": null
}
],
"ref_spans": [],
"eq_spans": [],
"section": "Experimental Settings",
"sec_num": "6"
},
{
"text": "We used a corpus of approximately 1 million aligned sentence pairs for English-Russian, and 0.5 million pairs for English-Arabic. Both corpora are from a technical (software manual) domain, which we believe is somewhat restricted along some morphological dimensions, such as tense and person. We used 1,000 sentence pairs each for development and testing for both language pairs. The details of the datasets used are given in Table 3 . The sentence pairs were word-aligned using GIZA++ (Och and Ney, 2000) and submitted to a treelet-based MT system (Quirk et al., 2005) , which uses the word dependency structure of the source language and projects word dependency structure to the target language, creating the structure shown in Figure 1 above. Table 4 gives some relevant statistics of the lexicons we used. For Russian, a general-domain lexicon was available to us, consisting of about 80,000 lemmas (stems) and 9.4 inflected forms per stem. 6 Limiting the lexicon to word types that are seen in the training set reduces its size substantially to about 14,000 stems, and an average of 3.8 inflections per stem. We will use this latter \"domain-adapted\" lexicon in our experiments. 3.8 1.6 Ara. Lexicon \u2229 Train 12,670 7.0 1.7 Table 4 : Lexicon statistics For Arabic, as a full-size Arabic lexicon was not available to us, we used the Buckwalter morphological analyzer (Buckwalter, 2004) to derive a lexicon. To acquire the stemming and inflection operators, we submit all words in our training data to the Buckwalter analyzer. Note that Arabic displays a high level of ambiguity, each word corresponding to many possible segmentations and morphological analyses; we considered all of the different stems returned by the Buckwalter analyzer in creating a word's stem set. The lexicon created in this manner contains 12,670 distinct stems and 89,360 inflected forms.",
"cite_spans": [
{
"start": 486,
"end": 505,
"text": "(Och and Ney, 2000)",
"ref_id": "BIBREF12"
},
{
"start": 549,
"end": 569,
"text": "(Quirk et al., 2005)",
"ref_id": "BIBREF14"
},
{
"start": 946,
"end": 947,
"text": "6",
"ref_id": null
},
{
"start": 1370,
"end": 1388,
"text": "(Buckwalter, 2004)",
"ref_id": "BIBREF0"
}
],
"ref_spans": [
{
"start": 426,
"end": 433,
"text": "Table 3",
"ref_id": "TABREF4"
},
{
"start": 731,
"end": 739,
"text": "Figure 1",
"ref_id": "FIGREF0"
},
{
"start": 747,
"end": 754,
"text": "Table 4",
"ref_id": null
},
{
"start": 1228,
"end": 1235,
"text": "Table 4",
"ref_id": null
}
],
"eq_spans": [],
"section": "Data",
"sec_num": "6.1"
},
{
"text": "For the generation of word features, we only consider one dominant analysis for any surface word for simplicity. In case of ambiguity, we considered only the first (arbitrary) analysis for Russian. For Arabic, we apply the following heuristic: use the most frequent analysis estimated from the gold standard labels in the Arabic Treebank (Maamouri et al., 2005) ; if a word does not appear in the treebank, we choose the first analysis returned by the Buckwalter analyzer. Ideally, the best word analysis should be provided as a result of contextual disambiguation (e.g., (Habash and Rambow, 2005) ); we leave this for future work.",
"cite_spans": [
{
"start": 338,
"end": 361,
"text": "(Maamouri et al., 2005)",
"ref_id": "BIBREF9"
},
{
"start": 572,
"end": 597,
"text": "(Habash and Rambow, 2005)",
"ref_id": "BIBREF5"
}
],
"ref_spans": [],
"eq_spans": [],
"section": "Lexicon",
"sec_num": "6.2"
},
{
"text": "As a baseline, we pick a morphological inflection y t at random from I t . This random baseline serves as an indicator for the difficulty of the problem. Another more competitive baseline we implemented is a word trigram language model (LM). The LMs were trained using the CMU language modelling toolkit (Clarkson and Rosenfeld, 1997) with default settings on the training data described in Table 3 .",
"cite_spans": [
{
"start": 304,
"end": 334,
"text": "(Clarkson and Rosenfeld, 1997)",
"ref_id": "BIBREF1"
}
],
"ref_spans": [
{
"start": 391,
"end": 398,
"text": "Table 3",
"ref_id": "TABREF4"
}
],
"eq_spans": [],
"section": "Baseline",
"sec_num": "6.3"
},
{
"text": "In the experiments, our primary goal is to evaluate the effectiveness of the proposed model using all features available to us. Additionally, we are interested in knowing the contribution of each information source, namely of morpho-syntactic and bilingual features. Therefore, we study the performance of models including the full feature schemata as well as models that are restricted to feature subsets according to the feature types as described in Section 5.2. The models are as follows: Monolingual-Word, including LM-like and stem n-gram features only; Bilingual-Word, which also includes bilingual lexical features; 7 Monolingual-All, which has access to all the information available in the target language, including morphological and syntactic features; and finally, Bilingual-All, which includes all feature types from Table 2. For each model and language, we perform feature selection in the following manner. The features are represented as feature templates, such as \"POS=X\", which generate a set of binary features corresponding to different instantiations of the template, as in \"POS=NOUN\". In addition to individual features, conjunctions of up to three features are also considered for selection (e.g., \"POS=NOUN & Number=plural\"). Every conjunction of feature templates considered contains at least one predicate on the prediction y t , and up to two predicates on the context. The feature selection algorithm performs a greedy forward stepwise feature selection on the feature templates so as to maximize development set accuracy. The algorithm is similar to the one described in (Toutanova, 2006) . After this process, we performed some manual inspection of the selected templates, and finally obtained 11 and 36 templates for the Monolingual-All and Bilingual-All settings for Russian, respectively. These templates generated 7.9 million and 9.3 million binary feature instantiations in the final model, respectively. The corresponding numbers for Arabic were 27 feature templates (0.7 million binary instantiations) and 39 feature templates (2.3 million binary instantiations) for Monolingual-All and Bilingual-All, respectively. Table 5 shows the accuracy of predicting word forms for the baseline and proposed models. We report accuracy only on words that appear in our lexicons. Thus, punctuation, English words occurring in the target sentence, and words with unknown lemmas are excluded from the evaluation. The reported accuracy measure therefore abstracts away from the is-7 Overall, this feature set approximates the information that is available to a state-of-the-art statistical MT system.",
"cite_spans": [
{
"start": 1601,
"end": 1618,
"text": "(Toutanova, 2006)",
"ref_id": "BIBREF16"
}
],
"ref_spans": [
{
"start": 831,
"end": 839,
"text": "Table 2.",
"ref_id": "TABREF2"
},
{
"start": 2154,
"end": 2161,
"text": "Table 5",
"ref_id": null
}
],
"eq_spans": [],
"section": "Experiments",
"sec_num": "6.4"
},
{
"text": "Eng Table 5 : Accuracy (%) results by model sue of incomplete coverage of the lexicon. When we encounter these words in the true MT scenario, we will make no predictions about them, and simply leave them unmodified. In our current experiments, in Russian, 68.2% of all word tokens were in Cyrillic, of which 93.8% were included in our lexicon. In Arabic, 85.5% of all word tokens were in Arabic characters, of which 99.1% were in our lexicon. 8 The results in Table 5 show that the suggested models outperform the language model substantially for both languages. In particular, the contribution of both bilingual and non-lexical features is noteworthy: adding non-lexical features consistently leads to 1.5% to 2% absolute gain in both monolingual and bilingual settings in both language pairs. We obtain a particularly large gain in the Russian bilingual case, in which the absolute gain is more than 4%, translating to 34% error rate reduction. Adding bilingual features has a similar effect of gaining about 2% (and 4% for Russian non-lexical) in accuracy over monolingual models. The overall accuracy is lower in Arabic than in Russian, reflecting the inherent difficulty of the task, as indicated by the random baseline (31.7 in Russian vs. 16.3 in Arabic).",
"cite_spans": [
{
"start": 443,
"end": 444,
"text": "8",
"ref_id": null
}
],
"ref_spans": [
{
"start": 4,
"end": 11,
"text": "Table 5",
"ref_id": null
},
{
"start": 460,
"end": 467,
"text": "Table 5",
"ref_id": null
}
],
"eq_spans": [],
"section": "Model",
"sec_num": null
},
{
"text": "In order to evaluate the effectiveness of the model in alleviating the data sparsity problem in morphological generation, we trained inflection prediction models on various subsets of the training data described in Table 3 , and tested their accuracy. The results are given in Figure 2 . We can see that with as few as 5,000 training sentences pairs, the model obtains much better accuracy than the language model, which is trained on data that is larger by a few orders of magnitude. We also note that the learning curve Figure 2 : Accuracy, varying training data size becomes less steep as we use more training data, suggesting that the models are successfully learning generalizations.",
"cite_spans": [],
"ref_spans": [
{
"start": 215,
"end": 222,
"text": "Table 3",
"ref_id": "TABREF4"
},
{
"start": 277,
"end": 285,
"text": "Figure 2",
"ref_id": null
},
{
"start": 522,
"end": 530,
"text": "Figure 2",
"ref_id": null
}
],
"eq_spans": [],
"section": "Model",
"sec_num": null
},
{
"text": "We have also manually examined some representative cases where the proposed model failed to make a correct prediction. In both Russian and Arabic, a very common pattern was a mistake in predicting the gender (as well as number and person in Arabic) of pronouns. This may be attributed to the fact that the correct choice of the pronoun requires coreference resolution, which is not available in our model. A more thorough analysis of the results will be helpful to bring further improvements.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Model",
"sec_num": null
},
{
"text": "We presented a probabilistic framework for morphological generation given aligned sentence pairs, incorporating morpho-syntactic information from both the source and target sentences. The results, using reference translations, show that the proposed models achieve substantially better accuracy than language models, even with a relatively small amount of training data. Our models using morphosyntactic information also outperformed models using only lexical information by a wide margin. This result is very promising for achieving our ultimate goal of improving MT output by using a specialized model for target language morphological generation. Though this goal is clearly outside the scope of this paper, we conducted a preliminary experiment where an English-to-Russian MT system was trained on a stemmed version of the aligned data and used to generate stemmed word sequences, which were then inflected using the suggested framework. This simple integration of the proposed model with the MT system improved the BLEU score by 1.7. The most obvious next step of our research, therefore, is to further pursue the integration of the proposed model to the end-to-end MT scenario.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Conclusions and Future Work",
"sec_num": "8"
},
{
"text": "There are multiple paths for obtaining further improvements over the results presented here. These include refinement in feature design, word analysis disambiguation, morphological and syntactic analysis on the source English side (e.g., assigning semantic role tags), to name a few. Another area of investigation is capturing longer-distance agreement phenomena, which can be done by implementing a global statistical model, or by using features from dependency trees more effectively.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Conclusions and Future Work",
"sec_num": "8"
},
{
"text": "Case marking also exists in Arabic. However, in many instances, it is realized by diacritics which are ignored in standard orthography. In our experiments, we include case marking in Arabic only when it is reflected in the orthography.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "",
"sec_num": null
},
{
"text": "Multiple stems are possible due to ambiguity in morphological analysis.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "",
"sec_num": null
},
{
"text": "That is, assuming that the stem sequence that is output by the MT system is correct.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "",
"sec_num": null
},
{
"text": "Note that while we decompose the prediction task left-toright, an appealing alternative is to define a top-down decomposition, traversing the dependency tree of the sentence. However, this requires syntactic analysis of sufficient quality.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "",
"sec_num": null
},
{
"text": "In this case, yt should equal wt, according to the task definition.6 The averages reported inTable 4are by type and do not consider word frequencies in the data.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "",
"sec_num": null
},
{
"text": "For Arabic, the inflection ambiguity was extremely high: there were on average 39 inflected forms per stem set in our development corpus (per token), as opposed to 7 in Russian. We therefore limited the evaluation of Arabic to those stems that have up to 30 inflected forms, resulting in 17 inflected forms per stem set on average in the development data.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "",
"sec_num": null
}
],
"back_matter": [],
"bib_entries": {
"BIBREF0": {
"ref_id": "b0",
"title": "Buckwalter arabic morphological analyzer version 2.0",
"authors": [
{
"first": "Tim",
"middle": [],
"last": "Buckwalter",
"suffix": ""
}
],
"year": 2004,
"venue": "",
"volume": "",
"issue": "",
"pages": "",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Tim Buckwalter. 2004. Buckwalter arabic morphological ana- lyzer version 2.0.",
"links": null
},
"BIBREF1": {
"ref_id": "b1",
"title": "Statistical language modelling using the CMU cambridge toolkit",
"authors": [
{
"first": "Philip",
"middle": [],
"last": "Clarkson",
"suffix": ""
},
{
"first": "Roni",
"middle": [],
"last": "Rosenfeld",
"suffix": ""
}
],
"year": 1997,
"venue": "",
"volume": "",
"issue": "",
"pages": "",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Philip Clarkson and Roni Rosenfeld. 1997. Statistical language modelling using the CMU cambridge toolkit. In Eurospeech.",
"links": null
},
"BIBREF2": {
"ref_id": "b2",
"title": "Automatic learning of language model structure",
"authors": [
{
"first": "Kevin",
"middle": [],
"last": "Duh",
"suffix": ""
},
{
"first": "Kathrin",
"middle": [],
"last": "Kirchhoff",
"suffix": ""
}
],
"year": 2004,
"venue": "COLING",
"volume": "",
"issue": "",
"pages": "",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Kevin Duh and Kathrin Kirchhoff. 2004. Automatic learning of language model structure. In COLING.",
"links": null
},
"BIBREF3": {
"ref_id": "b3",
"title": "Initial explorations in English to Turkish statistical machine translation",
"authors": [],
"year": 2006,
"venue": "NAACL workshop on statistical machine translation",
"volume": "",
"issue": "",
"pages": "",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Ilknur Durgar El-Kahlout and Kemal Oflazer. 2006. Initial ex- plorations in English to Turkish statistical machine transla- tion. In NAACL workshop on statistical machine translation.",
"links": null
},
"BIBREF4": {
"ref_id": "b4",
"title": "Improving statistical MT through morphological analysis",
"authors": [
{
"first": "Sharon",
"middle": [],
"last": "Goldwater",
"suffix": ""
},
{
"first": "David",
"middle": [],
"last": "Mcclosky",
"suffix": ""
}
],
"year": 2005,
"venue": "EMNLP",
"volume": "",
"issue": "",
"pages": "",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Sharon Goldwater and David McClosky. 2005. Improving sta- tistical MT through morphological analysis. In EMNLP.",
"links": null
},
"BIBREF5": {
"ref_id": "b5",
"title": "Arabic tokenization, part-of-speech tagging and morphological disambiguation in one fell swoop",
"authors": [
{
"first": "Nizar",
"middle": [],
"last": "Habash",
"suffix": ""
},
{
"first": "Owen",
"middle": [],
"last": "Rambow",
"suffix": ""
}
],
"year": 2005,
"venue": "ACL",
"volume": "",
"issue": "",
"pages": "",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Nizar Habash and Owen Rambow. 2005. Arabic tokenization, part-of-speech tagging and morphological disambiguation in one fell swoop. In ACL.",
"links": null
},
"BIBREF6": {
"ref_id": "b6",
"title": "Empirical methods for compound splitting",
"authors": [
{
"first": "Philipp",
"middle": [],
"last": "Koehn",
"suffix": ""
},
{
"first": "Kevin",
"middle": [],
"last": "Knight",
"suffix": ""
}
],
"year": 2003,
"venue": "EACL",
"volume": "",
"issue": "",
"pages": "",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Philipp Koehn and Kevin Knight. 2003. Empirical methods for compound splitting. In EACL.",
"links": null
},
"BIBREF7": {
"ref_id": "b7",
"title": "Europarl: A parallel corpus for statistical machine translation",
"authors": [
{
"first": "Philipp",
"middle": [],
"last": "Koehn",
"suffix": ""
}
],
"year": 2005,
"venue": "MT Summit",
"volume": "",
"issue": "",
"pages": "",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Philipp Koehn. 2005. Europarl: A parallel corpus for statistical machine translation. In MT Summit.",
"links": null
},
"BIBREF8": {
"ref_id": "b8",
"title": "Morphological analysis for statistical machine translation",
"authors": [
{
"first": "Young-Suk",
"middle": [],
"last": "Lee",
"suffix": ""
}
],
"year": 2004,
"venue": "HLT-NAACL",
"volume": "",
"issue": "",
"pages": "",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Young-Suk Lee. 2004. Morphological analysis for statistical machine translation. In HLT-NAACL.",
"links": null
},
"BIBREF9": {
"ref_id": "b9",
"title": "Arabic Treebank: Part 1 v 3.0. Linguistic Data Consortium",
"authors": [
{
"first": "Mohamed",
"middle": [],
"last": "Maamouri",
"suffix": ""
},
{
"first": "Ann",
"middle": [],
"last": "Bies",
"suffix": ""
},
{
"first": "Tim",
"middle": [],
"last": "Buckwalter",
"suffix": ""
},
{
"first": "Hubert",
"middle": [],
"last": "Jin",
"suffix": ""
}
],
"year": 2005,
"venue": "",
"volume": "",
"issue": "",
"pages": "",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Mohamed Maamouri, Ann Bies, Tim Buckwalter, and Hubert Jin. 2005. Arabic Treebank: Part 1 v 3.0. Linguistic Data Consortium.",
"links": null
},
"BIBREF10": {
"ref_id": "b10",
"title": "Maximum entropy markov models for information extraction and segmentation",
"authors": [
{
"first": "Andrew",
"middle": [],
"last": "Mccallum",
"suffix": ""
},
{
"first": "Dayne",
"middle": [],
"last": "Freitag",
"suffix": ""
},
{
"first": "Fernando",
"middle": [
"C N"
],
"last": "Pereira",
"suffix": ""
}
],
"year": 2000,
"venue": "ICML",
"volume": "",
"issue": "",
"pages": "",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Andrew McCallum, Dayne Freitag, and Fernando C. N. Pereira. 2000. Maximum entropy markov models for information extraction and segmentation. In ICML.",
"links": null
},
"BIBREF11": {
"ref_id": "b11",
"title": "Statistical machine translation with scarce resources using morpho-syntactic information",
"authors": [
{
"first": "Sonja",
"middle": [],
"last": "Nie\u00dfen",
"suffix": ""
},
{
"first": "Hermann",
"middle": [],
"last": "Ney",
"suffix": ""
}
],
"year": 2004,
"venue": "Computational Linguistics",
"volume": "30",
"issue": "2",
"pages": "181--204",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Sonja Nie\u00dfen and Hermann Ney. 2004. Statistical machine translation with scarce resources using morpho-syntactic in- formation. Computational Linguistics, 30(2):181-204.",
"links": null
},
"BIBREF12": {
"ref_id": "b12",
"title": "Improved statistical alignment models",
"authors": [
{
"first": "Josef",
"middle": [],
"last": "Franz",
"suffix": ""
},
{
"first": "Hermann",
"middle": [],
"last": "Och",
"suffix": ""
},
{
"first": "",
"middle": [],
"last": "Ney",
"suffix": ""
}
],
"year": 2000,
"venue": "ACL",
"volume": "",
"issue": "",
"pages": "",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Franz Josef Och and Hermann Ney. 2000. Improved statistical alignment models. In ACL.",
"links": null
},
"BIBREF13": {
"ref_id": "b13",
"title": "Towards the use of word stems and suffixes for statistical machine translation",
"authors": [
{
"first": "Maja",
"middle": [],
"last": "Popovi\u0107",
"suffix": ""
},
{
"first": "Hermann",
"middle": [],
"last": "Ney",
"suffix": ""
}
],
"year": 2004,
"venue": "LREC",
"volume": "",
"issue": "",
"pages": "",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Maja Popovi\u0107 and Hermann Ney. 2004. Towards the use of word stems and suffixes for statistical machine translation. In LREC.",
"links": null
},
"BIBREF14": {
"ref_id": "b14",
"title": "Dependency tree translation: Syntactically informed phrasal SMT",
"authors": [
{
"first": "Chris",
"middle": [],
"last": "Quirk",
"suffix": ""
},
{
"first": "Arul",
"middle": [],
"last": "Menezes",
"suffix": ""
},
{
"first": "Colin",
"middle": [],
"last": "Cherry",
"suffix": ""
}
],
"year": 2005,
"venue": "ACL",
"volume": "",
"issue": "",
"pages": "",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Chris Quirk, Arul Menezes, and Colin Cherry. 2005. Depen- dency tree translation: Syntactically informed phrasal SMT. In ACL.",
"links": null
},
"BIBREF15": {
"ref_id": "b15",
"title": "Learning to predict case markers in Japanese",
"authors": [
{
"first": "Hisami",
"middle": [],
"last": "Suzuki",
"suffix": ""
},
{
"first": "Kristina",
"middle": [],
"last": "Toutanova",
"suffix": ""
}
],
"year": 2006,
"venue": "COLING-ACL",
"volume": "",
"issue": "",
"pages": "",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Hisami Suzuki and Kristina Toutanova. 2006. Learning to pre- dict case markers in Japanese. In COLING-ACL.",
"links": null
},
"BIBREF16": {
"ref_id": "b16",
"title": "Competitive generative models with structure learning for NLP classification tasks",
"authors": [
{
"first": "Kristina",
"middle": [],
"last": "Toutanova",
"suffix": ""
}
],
"year": 2006,
"venue": "EMNLP",
"volume": "",
"issue": "",
"pages": "",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Kristina Toutanova. 2006. Competitive generative models with structure learning for NLP classification tasks. In EMNLP.",
"links": null
}
},
"ref_entries": {
"FIGREF0": {
"uris": null,
"text": "Aligned English-Russian sentence pair with syntactic and morphological annotation get sentence, and lexicons L S and L T that support the operations described in the section above. Let a sentence w 1 , ...w t , .",
"num": null,
"type_str": "figure"
},
"FIGREF1": {
"uris": null,
"text": "f : POS, Person, Number, Gender, Tense f (Alt), f (Alt\u22121), Neg, Det, Prep, Conj, ObjPron, PossPron, f (Alt+1), f (Al HEAD(t) ) Comp",
"num": null,
"type_str": "figure"
},
"TABREF1": {
"text": "",
"content": "<table/>",
"num": null,
"type_str": "table",
"html": null
},
"TABREF2": {
"text": "The feature set suggested for English-Russian and English-Arabic pairs",
"content": "<table/>",
"num": null,
"type_str": "table",
"html": null
},
"TABREF4": {
"text": "Data set statistics: corpus size and average sentence length (in words)",
"content": "<table/>",
"num": null,
"type_str": "table",
"html": null
}
}
}
} |