File size: 94,389 Bytes
6fa4bc9 | 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360 361 362 363 364 365 366 367 368 369 370 371 372 373 374 375 376 377 378 379 380 381 382 383 384 385 386 387 388 389 390 391 392 393 394 395 396 397 398 399 400 401 402 403 404 405 406 407 408 409 410 411 412 413 414 415 416 417 418 419 420 421 422 423 424 425 426 427 428 429 430 431 432 433 434 435 436 437 438 439 440 441 442 443 444 445 446 447 448 449 450 451 452 453 454 455 456 457 458 459 460 461 462 463 464 465 466 467 468 469 470 471 472 473 474 475 476 477 478 479 480 481 482 483 484 485 486 487 488 489 490 491 492 493 494 495 496 497 498 499 500 501 502 503 504 505 506 507 508 509 510 511 512 513 514 515 516 517 518 519 520 521 522 523 524 525 526 527 528 529 530 531 532 533 534 535 536 537 538 539 540 541 542 543 544 545 546 547 548 549 550 551 552 553 554 555 556 557 558 559 560 561 562 563 564 565 566 567 568 569 570 571 572 573 574 575 576 577 578 579 580 581 582 583 584 585 586 587 588 589 590 591 592 593 594 595 596 597 598 599 600 601 602 603 604 605 606 607 608 609 610 611 612 613 614 615 616 617 618 619 620 621 622 623 624 625 626 627 628 629 630 631 632 633 634 635 636 637 638 639 640 641 642 643 644 645 646 647 648 649 650 651 652 653 654 655 656 657 658 659 660 661 662 663 664 665 666 667 668 669 670 671 672 673 674 675 676 677 678 679 680 681 682 683 684 685 686 687 688 689 690 691 692 693 694 695 696 697 698 699 700 701 702 703 704 705 706 707 708 709 710 711 712 713 714 715 716 717 718 719 720 721 722 723 724 725 726 727 728 729 730 731 732 733 734 735 736 737 738 739 740 741 742 743 744 745 746 747 748 749 750 751 752 753 754 755 756 757 758 759 760 761 762 763 764 765 766 767 768 769 770 771 772 773 774 775 776 777 778 779 780 781 782 783 784 785 786 787 788 789 790 791 792 793 794 795 796 797 798 799 800 801 802 803 804 805 806 807 808 809 810 811 812 813 814 815 816 817 818 819 820 821 822 823 824 825 826 827 828 829 830 831 832 833 834 835 836 837 838 839 840 841 842 843 844 845 846 847 848 849 850 851 852 853 854 855 856 857 858 859 860 861 862 863 864 865 866 867 868 869 870 871 872 873 874 875 876 877 878 879 880 881 882 883 884 885 886 887 888 889 890 891 892 893 894 895 896 897 898 899 900 901 902 903 904 905 906 907 908 909 910 911 912 913 914 915 916 917 918 919 920 921 922 923 924 925 926 927 928 929 930 931 932 933 934 935 936 937 938 939 940 941 942 943 944 945 946 947 948 949 950 951 952 953 954 955 956 957 958 959 960 961 962 963 964 965 966 967 968 969 970 971 972 973 974 975 976 977 978 979 980 981 982 983 984 985 986 987 988 989 990 991 992 993 994 995 996 997 998 999 1000 1001 1002 1003 1004 1005 1006 1007 1008 1009 1010 1011 1012 1013 1014 1015 1016 1017 1018 1019 1020 1021 1022 1023 1024 1025 1026 1027 1028 1029 1030 1031 1032 1033 1034 1035 1036 1037 1038 1039 1040 1041 1042 1043 1044 1045 1046 1047 1048 1049 1050 1051 1052 1053 1054 1055 1056 1057 1058 1059 1060 1061 1062 1063 1064 1065 1066 1067 1068 1069 1070 1071 1072 1073 1074 1075 1076 1077 1078 1079 1080 1081 1082 1083 1084 1085 1086 1087 1088 1089 1090 1091 1092 1093 1094 1095 1096 1097 1098 1099 1100 1101 1102 1103 1104 1105 1106 1107 1108 1109 1110 1111 1112 1113 1114 1115 1116 1117 1118 1119 1120 1121 1122 1123 1124 1125 1126 1127 1128 1129 1130 1131 1132 1133 1134 1135 1136 1137 1138 1139 1140 1141 1142 1143 1144 1145 1146 1147 1148 1149 1150 1151 1152 1153 1154 1155 1156 1157 1158 1159 1160 1161 1162 1163 1164 1165 1166 1167 1168 1169 1170 1171 1172 1173 1174 1175 1176 1177 1178 1179 1180 1181 1182 1183 1184 1185 1186 1187 1188 1189 1190 1191 1192 1193 1194 1195 1196 1197 1198 1199 1200 1201 1202 1203 1204 1205 1206 1207 1208 1209 1210 1211 1212 1213 1214 1215 1216 1217 1218 1219 1220 1221 1222 1223 1224 1225 1226 1227 1228 1229 1230 1231 1232 1233 1234 1235 1236 1237 1238 1239 1240 1241 1242 1243 1244 1245 1246 1247 1248 1249 1250 1251 1252 1253 1254 1255 1256 1257 1258 1259 1260 1261 1262 1263 1264 1265 1266 1267 1268 1269 1270 1271 1272 1273 1274 1275 1276 1277 1278 1279 1280 1281 1282 1283 1284 1285 1286 1287 1288 1289 1290 1291 1292 1293 1294 1295 1296 1297 1298 1299 1300 1301 1302 1303 1304 1305 1306 1307 1308 1309 1310 1311 1312 1313 1314 1315 1316 1317 1318 1319 1320 1321 1322 1323 1324 1325 1326 1327 1328 1329 1330 1331 1332 1333 1334 1335 1336 1337 1338 1339 1340 1341 1342 1343 1344 1345 1346 1347 1348 1349 1350 1351 1352 1353 1354 1355 1356 1357 1358 1359 1360 1361 1362 1363 1364 1365 1366 1367 1368 1369 1370 1371 1372 1373 1374 1375 1376 1377 1378 1379 1380 1381 1382 1383 1384 1385 1386 1387 1388 1389 1390 1391 1392 1393 1394 1395 1396 1397 1398 1399 1400 1401 1402 1403 1404 1405 1406 1407 1408 1409 1410 1411 1412 1413 1414 1415 1416 1417 1418 1419 1420 1421 1422 1423 1424 1425 1426 1427 1428 1429 1430 1431 1432 1433 1434 1435 1436 1437 1438 1439 1440 1441 1442 1443 1444 1445 1446 1447 1448 1449 1450 1451 1452 1453 1454 1455 1456 1457 1458 1459 1460 1461 1462 1463 1464 1465 1466 1467 1468 1469 1470 1471 1472 1473 1474 1475 1476 1477 1478 1479 1480 1481 1482 1483 1484 1485 1486 1487 1488 1489 1490 1491 1492 1493 1494 1495 1496 1497 1498 1499 1500 1501 1502 1503 1504 1505 1506 1507 1508 1509 1510 1511 1512 1513 1514 1515 1516 1517 1518 1519 1520 1521 1522 1523 1524 1525 1526 1527 1528 1529 1530 1531 1532 1533 1534 1535 1536 1537 1538 1539 1540 1541 1542 1543 1544 1545 1546 1547 1548 1549 1550 1551 1552 1553 1554 1555 1556 1557 1558 1559 1560 1561 1562 1563 1564 1565 1566 1567 1568 1569 1570 1571 1572 1573 1574 1575 1576 1577 1578 1579 1580 1581 1582 1583 1584 1585 1586 1587 1588 1589 1590 1591 1592 1593 1594 1595 1596 1597 1598 1599 1600 1601 1602 1603 1604 1605 1606 1607 1608 1609 1610 1611 1612 1613 1614 1615 1616 1617 1618 1619 1620 1621 1622 1623 1624 1625 1626 1627 1628 1629 1630 1631 1632 1633 1634 1635 1636 1637 1638 1639 1640 1641 1642 1643 1644 1645 1646 1647 1648 1649 1650 1651 1652 1653 1654 1655 1656 1657 1658 1659 1660 1661 1662 1663 1664 1665 1666 1667 1668 1669 1670 1671 1672 1673 1674 1675 1676 1677 1678 1679 1680 1681 1682 1683 1684 1685 1686 1687 1688 1689 1690 1691 1692 1693 1694 1695 1696 1697 1698 1699 1700 | {
"paper_id": "P16-1014",
"header": {
"generated_with": "S2ORC 1.0.0",
"date_generated": "2023-01-19T08:55:47.243286Z"
},
"title": "Pointing the Unknown Words",
"authors": [
{
"first": "Caglar",
"middle": [],
"last": "Gulcehre",
"suffix": "",
"affiliation": {},
"email": ""
},
{
"first": "Ramesh",
"middle": [],
"last": "Nallapati",
"suffix": "",
"affiliation": {},
"email": ""
},
{
"first": "Ibm",
"middle": [
"T J"
],
"last": "Watson Research",
"suffix": "",
"affiliation": {},
"email": ""
},
{
"first": "Bowen",
"middle": [],
"last": "Zhou",
"suffix": "",
"affiliation": {},
"email": ""
},
{
"first": "Yoshua",
"middle": [],
"last": "Bengio",
"suffix": "",
"affiliation": {},
"email": ""
}
],
"year": "",
"venue": null,
"identifiers": {},
"abstract": "The problem of rare and unknown words is an important issue that can potentially effect the performance of many NLP systems, including traditional count-based and deep learning models. We propose a novel way to deal with the rare and unseen words for the neural network models using attention. Our model uses two softmax layers in order to predict the next word in conditional language models: one predicts the location of a word in the source sentence, and the other predicts a word in the shortlist vocabulary. At each timestep, the decision of which softmax layer to use is adaptively made by an MLP which is conditioned on the context. We motivate this work from a psychological evidence that humans naturally have a tendency to point towards objects in the context or the environment when the name of an object is not known. Using our proposed model, we observe improvements on two tasks, neural machine translation on the Europarl English to French parallel corpora and text summarization on the Gigaword dataset.",
"pdf_parse": {
"paper_id": "P16-1014",
"_pdf_hash": "",
"abstract": [
{
"text": "The problem of rare and unknown words is an important issue that can potentially effect the performance of many NLP systems, including traditional count-based and deep learning models. We propose a novel way to deal with the rare and unseen words for the neural network models using attention. Our model uses two softmax layers in order to predict the next word in conditional language models: one predicts the location of a word in the source sentence, and the other predicts a word in the shortlist vocabulary. At each timestep, the decision of which softmax layer to use is adaptively made by an MLP which is conditioned on the context. We motivate this work from a psychological evidence that humans naturally have a tendency to point towards objects in the context or the environment when the name of an object is not known. Using our proposed model, we observe improvements on two tasks, neural machine translation on the Europarl English to French parallel corpora and text summarization on the Gigaword dataset.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Abstract",
"sec_num": null
}
],
"body_text": [
{
"text": "Words are the basic input/output units in most of the NLP systems, and thus the ability to cover a large number of words is a key to building a robust NLP system. However, considering that (i) the number of all words in a language including named entities is very large and that (ii) language itself is an evolving system (people create new words), this can be a challenging problem.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Introduction",
"sec_num": "1"
},
{
"text": "A common approach followed by the recent neural network based NLP systems is to use a softmax output layer where each of the output di-mension corresponds to a word in a predefined word-shortlist. Because computing high dimensional softmax is computationally expensive, in practice the shortlist is limited to have only top-K most frequent words in the training corpus. All other words are then replaced by a special word, called the unknown word (UNK).",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Introduction",
"sec_num": "1"
},
{
"text": "The shortlist approach has two fundamental problems. The first problem, which is known as the rare word problem, is that some of the words in the shortlist occur less frequently in the training set and thus are difficult to learn a good representation, resulting in poor performance. Second, it is obvious that we can lose some important information by mapping different words to a single dummy token UNK. Even if we have a very large shortlist including all unique words in the training set, it does not necessarily improve the test performance, because there still exists a chance to see an unknown word at test time. This is known as the unknown word problem. In addition, increasing the shortlist size mostly leads to increasing rare words due to Zipf's Law.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Introduction",
"sec_num": "1"
},
{
"text": "These two problems are particularly critical in language understanding tasks such as factoid question answering (Bordes et al., 2015) where the words that we are interested in are often named entities which are usually unknown or rare words.",
"cite_spans": [
{
"start": 112,
"end": 133,
"text": "(Bordes et al., 2015)",
"ref_id": "BIBREF2"
}
],
"ref_spans": [],
"eq_spans": [],
"section": "Introduction",
"sec_num": "1"
},
{
"text": "In a similar situation, where we have a limited information on how to call an object of interest, it seems that humans (and also some primates) have an efficient behavioral mechanism of drawing attention to the object: pointing (Matthews et al., 2012) . Pointing makes it possible to deliver information and to associate context to a particular object without knowing how to call it. In particular, human infants use pointing as a fundamental communication tool (Tomasello et al., 2007) .",
"cite_spans": [
{
"start": 228,
"end": 251,
"text": "(Matthews et al., 2012)",
"ref_id": null
},
{
"start": 462,
"end": 486,
"text": "(Tomasello et al., 2007)",
"ref_id": null
}
],
"ref_spans": [],
"eq_spans": [],
"section": "Introduction",
"sec_num": "1"
},
{
"text": "In this paper, inspired by the pointing behavior of humans and recent advances in the atten-tion mechanism (Bahdanau et al., 2014) and the pointer networks , we propose a novel method to deal with the rare or unknown word problem. The basic idea is that we can see many NLP problems as a task of predicting target text given context text, where some of the target words appear in the context as well. We observe that in this case we can make the model learn to point a word in the context and copy it to the target text, as well as when to point. For example, in machine translation, we can see the source sentence as the context, and the target sentence as what we need to predict. In Figure 1 , we show an example depiction of how words can be copied from source to target in machine translation. Although the source and target languages are different, many of the words such as named entities are usually represented by the same characters in both languages, making it possible to copy. Similarly, in text summarization, it is natural to use some words in the original text in the summarized text as well.",
"cite_spans": [
{
"start": 107,
"end": 130,
"text": "(Bahdanau et al., 2014)",
"ref_id": "BIBREF0"
}
],
"ref_spans": [
{
"start": 686,
"end": 694,
"text": "Figure 1",
"ref_id": null
}
],
"eq_spans": [],
"section": "Introduction",
"sec_num": "1"
},
{
"text": "Specifically, to predict a target word at each timestep, our model first determines the source of the word generation, that is, whether to take one from a predefined shortlist or to copy one from the context. For the former, we apply the typical softmax operation, and for the latter, we use the attention mechanism to obtain the pointing softmax probability over the context words and pick the one of high probability. The model learns this decision so as to use the pointing only when the context includes a word that can be copied to the target. This way, our model can predict even the words which are not in the shortlist, as long as it appears in the context. Although some of the words still need to be labeled as UNK, i.e., if it is neither in the shortlist nor in the context, in experiments we show that this learning when and where to point improves the performance in machine translation and text summarization. Guillaume and Cesar have a blue car in Lausanne.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Introduction",
"sec_num": "1"
},
{
"text": "Copy Copy",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Copy",
"sec_num": null
},
{
"text": "English:",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "French:",
"sec_num": null
},
{
"text": "Figure 1: An example of how copying can happen for machine translation. Common words that appear both in source and the target can directly be copied from input to source. The rest of the unknown in the target can be copied from the input after being translated with a dictionary.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "French:",
"sec_num": null
},
{
"text": "The rest of the paper is organized as follows. In the next section, we review the related works including pointer networks and previous approaches to the rare/unknown problem. In Section 3, we review the neural machine translation with attention mechanism which is the baseline in our experiments. Then, in Section 4, we propose our method dealing with the rare/unknown word problem, called the Pointer Softmax (PS). The experimental results are provided in the Section 5 and we conclude our work in Section 6.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "French:",
"sec_num": null
},
{
"text": "The attention-based pointing mechanism is introduced first in the pointer networks . In the pointer networks, the output space of the target sequence is constrained to be the observations in the input sequence (not the input space). Instead of having a fixed dimension softmax output layer, softmax outputs of varying dimension is dynamically computed for each input sequence in such a way to maximize the attention probability of the target input. However, its applicability is rather limited because, unlike our model, there is no option to choose whether to point or not; it always points. In this sense, we can see the pointer networks as a special case of our model where we always choose to point a context word.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Related Work",
"sec_num": "2"
},
{
"text": "Several approaches have been proposed towards solving the rare words/unknown words problem, which can be broadly divided into three categories. The first category of the approaches focuses on improving the computation speed of the softmax output so that it can maintain a very large vocabulary. Because this only increases the shortlist size, it helps to mitigate the unknown word problem, but still suffers from the rare word problem. The hierarchical softmax (Morin and Bengio, 2005) , importance sampling (Bengio and Sen\u00e9cal, 2008; Jean et al., 2014) , and the noise contrastive estimation (Gutmann and Hyv\u00e4rinen, 2012; Mnih and Kavukcuoglu, 2013) methods are in the class.",
"cite_spans": [
{
"start": 461,
"end": 485,
"text": "(Morin and Bengio, 2005)",
"ref_id": "BIBREF13"
},
{
"start": 508,
"end": 534,
"text": "(Bengio and Sen\u00e9cal, 2008;",
"ref_id": "BIBREF1"
},
{
"start": 535,
"end": 553,
"text": "Jean et al., 2014)",
"ref_id": "BIBREF10"
}
],
"ref_spans": [],
"eq_spans": [],
"section": "Related Work",
"sec_num": "2"
},
{
"text": "The second category, where our proposed method also belongs to, uses information from the context. Notable works are (Luong et al., 2015) and (Hermann et al., 2015) . In particular, applying to machine translation task, (Luong et al., 2015) learns to point some words in source sentence and copy it to the target sentence, similarly to our method. However, it does not use attention mechanism, and by having fixed sized soft-max output over the relative pointing range (e.g., -7, . . . , -1, 0, 1, . . . , 7), their model (the Positional All model) has a limitation in applying to more general problems such as summarization and question answering, where, unlike machine translation, the length of the context and the pointing locations in the context can vary dramatically. In question answering setting, (Hermann et al., 2015) have used placeholders on named entities in the context. However, the placeholder id is directly predicted in the softmax output rather than predicting its location in the context.",
"cite_spans": [
{
"start": 117,
"end": 137,
"text": "(Luong et al., 2015)",
"ref_id": "BIBREF11"
},
{
"start": 142,
"end": 164,
"text": "(Hermann et al., 2015)",
"ref_id": "BIBREF9"
},
{
"start": 220,
"end": 240,
"text": "(Luong et al., 2015)",
"ref_id": "BIBREF11"
},
{
"start": 806,
"end": 828,
"text": "(Hermann et al., 2015)",
"ref_id": "BIBREF9"
}
],
"ref_spans": [],
"eq_spans": [],
"section": "Related Work",
"sec_num": "2"
},
{
"text": "The third category of the approaches changes the unit of input/output itself from words to a smaller resolution such as characters (Graves, 2013) or bytecodes (Sennrich et al., 2015; Gillick et al., 2015) . Although this approach has the main advantage that it could suffer less from the rare/unknown word problem, the training usually becomes much harder because the length of sequences significantly increases.",
"cite_spans": [
{
"start": 131,
"end": 145,
"text": "(Graves, 2013)",
"ref_id": "BIBREF6"
},
{
"start": 159,
"end": 182,
"text": "(Sennrich et al., 2015;",
"ref_id": "BIBREF17"
},
{
"start": 183,
"end": 204,
"text": "Gillick et al., 2015)",
"ref_id": "BIBREF5"
}
],
"ref_spans": [],
"eq_spans": [],
"section": "Related Work",
"sec_num": "2"
},
{
"text": "Simultaneously to our work, (Gu et al., 2016) and (Cheng and Lapata, 2016) proposed models that learn to copy from source to target and both papers analyzed their models on summarization tasks.",
"cite_spans": [
{
"start": 28,
"end": 45,
"text": "(Gu et al., 2016)",
"ref_id": "BIBREF7"
},
{
"start": 50,
"end": 74,
"text": "(Cheng and Lapata, 2016)",
"ref_id": "BIBREF3"
}
],
"ref_spans": [],
"eq_spans": [],
"section": "Related Work",
"sec_num": "2"
},
{
"text": "As the baseline neural machine translation system, we use the model proposed by (Bahdanau et al., 2014) that learns to (soft-)align and translate jointly. We refer this model as NMT. The encoder of the NMT is a bidirectional RNN (Schuster and Paliwal, 1997). The forward RNN reads input sequence",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Neural Machine Translation Model with Attention",
"sec_num": "3"
},
{
"text": "x = (x 1 , . . . , x T ) in left-to-right direction, resulting in a sequence of hidden states ( \u2212 \u2192 h 1 , . . . , \u2212 \u2192 h T ).",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Neural Machine Translation Model with Attention",
"sec_num": "3"
},
{
"text": "The backward RNN reads x in the reversed direction and outputs",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Neural Machine Translation Model with Attention",
"sec_num": "3"
},
{
"text": "( \u2190 \u2212 h 1 , . . . , \u2190 \u2212 h T ).",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Neural Machine Translation Model with Attention",
"sec_num": "3"
},
{
"text": "We then concatenate the hidden states of forward and backward RNNs at each time step and obtain a sequence of annotation vectors",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Neural Machine Translation Model with Attention",
"sec_num": "3"
},
{
"text": "(h 1 , . . . , h T ) where h j = \u2212 \u2192 h j || \u2190 \u2212 h j .",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Neural Machine Translation Model with Attention",
"sec_num": "3"
},
{
"text": "Here, || denotes the concatenation operator. Thus, each annotation vector h j encodes information about the j-th word with respect to all the other surrounding words in both directions.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Neural Machine Translation Model with Attention",
"sec_num": "3"
},
{
"text": "In the decoder, we usually use gated recurrent unit (GRU) Chung et al., 2014) . Specifically, at each time-step t, the softalignment mechanism first computes the relevance weight e tj which determines the contribution of annotation vector h j to the t-th target word. We use a non-linear mapping f (e.g., MLP) which takes h j , the previous decoder's hidden state s t\u22121 and the previous output y t\u22121 as input:",
"cite_spans": [
{
"start": 58,
"end": 77,
"text": "Chung et al., 2014)",
"ref_id": "BIBREF4"
}
],
"ref_spans": [],
"eq_spans": [],
"section": "Neural Machine Translation Model with Attention",
"sec_num": "3"
},
{
"text": "e tj = f (s t\u22121 , h j , y t\u22121 ).",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Neural Machine Translation Model with Attention",
"sec_num": "3"
},
{
"text": "The outputs e tj are then normalized as follows:",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Neural Machine Translation Model with Attention",
"sec_num": "3"
},
{
"text": "EQUATION",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [
{
"start": 0,
"end": 8,
"text": "EQUATION",
"ref_id": "EQREF",
"raw_str": "l tj = exp(e tj ) T k=1 exp(e tk ) .",
"eq_num": "(1)"
}
],
"section": "Neural Machine Translation Model with Attention",
"sec_num": "3"
},
{
"text": "We call l tj as the relevance score, or the alignment weight, of the j-th annotation vector.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Neural Machine Translation Model with Attention",
"sec_num": "3"
},
{
"text": "The relevance scores are used to get the context vector c t of the t-th target word in the translation:",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Neural Machine Translation Model with Attention",
"sec_num": "3"
},
{
"text": "c t = T j=1 l tj h j ,",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Neural Machine Translation Model with Attention",
"sec_num": "3"
},
{
"text": "The hidden state of the decoder s t is computed based on the previous hidden state s t\u22121 , the context vector c t and the output word of the previous time-step y t\u22121 :",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Neural Machine Translation Model with Attention",
"sec_num": "3"
},
{
"text": "EQUATION",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [
{
"start": 0,
"end": 8,
"text": "EQUATION",
"ref_id": "EQREF",
"raw_str": "s t = f r (s t\u22121 , y t\u22121 , c t ),",
"eq_num": "(2)"
}
],
"section": "Neural Machine Translation Model with Attention",
"sec_num": "3"
},
{
"text": "where f r is GRU. We use a deep output layer (Pascanu et al., 2013) to compute the conditional distribution over words:",
"cite_spans": [
{
"start": 45,
"end": 67,
"text": "(Pascanu et al., 2013)",
"ref_id": "BIBREF15"
}
],
"ref_spans": [],
"eq_spans": [],
"section": "Neural Machine Translation Model with Attention",
"sec_num": "3"
},
{
"text": "EQUATION",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [
{
"start": 0,
"end": 8,
"text": "EQUATION",
"ref_id": "EQREF",
"raw_str": "p(y t = a|y <t , x) \u221d exp \u03c8 a (Wo,bo) f o (s t , y t\u22121 , c t ) ,",
"eq_num": "(3)"
}
],
"section": "Neural Machine Translation Model with Attention",
"sec_num": "3"
},
{
"text": "where W is a learned weight matrix and b is a bias of the output layer. f o is a single-layer feedforward neural network. \u03c8 (Wo,bo) (\u2022) is a function that performs an affine transformation on its input. And the superscript a in \u03c8 a indicates the a-th column vector of \u03c8. The whole model, including both the encoder and the decoder, is jointly trained to maximize the (conditional) log-likelihood of target sequences given input sequences, where the training corpus is a set of (x n , y n )'s. Figure 2 illustrates the architecture of the NMT. Figure 2 : A depiction of neural machine translation architecture with attention. At each timestep, the model generates the attention weights l t . We use l t the encoder's hidden state to obtain the context c t . The decoder uses c t to predict a vector of probabilities for the words w t by using softmax. applicable approach to many NLP tasks, because it resolves the limitations about unknown words for neural networks. It can be used in parallel with other existing techniques such as the large vocabulary trick (Jean et al., 2014) . Our model learns two key abilities jointly to make the pointing mechanism applicable in more general settings: (i) to predict whether it is required to use the pointing or not at each time step and (ii) to point any location of the context sequence whose length can vary widely over examples. Note that the pointer networks are in lack of the ability (i), and the ability (ii) is not achieved in the models by (Luong et al., 2015) .",
"cite_spans": [
{
"start": 1060,
"end": 1079,
"text": "(Jean et al., 2014)",
"ref_id": "BIBREF10"
},
{
"start": 1492,
"end": 1512,
"text": "(Luong et al., 2015)",
"ref_id": "BIBREF11"
}
],
"ref_spans": [
{
"start": 493,
"end": 501,
"text": "Figure 2",
"ref_id": null
},
{
"start": 543,
"end": 551,
"text": "Figure 2",
"ref_id": null
}
],
"eq_spans": [],
"section": "Neural Machine Translation Model with Attention",
"sec_num": "3"
},
{
"text": "To achieve this, our model uses two softmax output layers, the shortlist softmax and the location softmax. The shortlist softmax is the same as the typical softmax output layer where each dimension corresponds a word in the predefined word shortlist. The location softmax is a pointer network where each of the output dimension corresponds to the location of a word in the context sequence. Thus, the output dimension of the location softmax varies according to the length of the given context sequence.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "... st+1 Encoder",
"sec_num": null
},
{
"text": "At each time-step, if the model decides to use the shortlist softmax, we generate a word w t from the shortlist. Otherwise, if it is expected that the context sequence contains a word which needs to be generated at the time step, we obtain the location of the context word l t from the location softmax. The key to making this possible is deciding when to use the shortlist softmax or the lo-cation softmax at each time step. In order to accomplish this, we introduce a switching network to the model. The switching network, which is a multilayer perceptron in our experiments, takes the representation of the context sequence (similar to the input annotation in NMT) and the previous hidden state of the output RNN as its input. It outputs a binary variable z t which indicates whether to use the shortlist softmax (when z t = 1) or the location softmax (when z t = 0). Note that if the word that is expected to be generated at each timestep is neither in the shortlist nor in the context sequence, the switching network selects the shortlist softmax, and then the shortlist softmax predicts UNK. The details of the pointer softmax model can be seen in Figure 3 as Figure 3 : A simple depiction of the Pointer Softmax(PS) architecture. At each timestep as usuall t , c t and the w t for the words over the limited vocabulary(shortlist) is being generated. We have an additional switching variable z t that decides whether to use w t or copy the word from the input via l t . The final word prediction will be performed via pointer softmax f t which can either copy the word from the source or predict the word from the shortlist vocabulary.",
"cite_spans": [],
"ref_spans": [
{
"start": 1154,
"end": 1165,
"text": "Figure 3 as",
"ref_id": null
},
{
"start": 1166,
"end": 1174,
"text": "Figure 3",
"ref_id": null
}
],
"eq_spans": [],
"section": "... st+1 Encoder",
"sec_num": null
},
{
"text": "More specifically, our goal is to maximize the probability of observing the target word sequence y = (y 1 , y 2 , . . . , y Ty ) and the word generation source z = (z 1 , z 2 , . . . , z Ty ), given the context sequence",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "... st+1 Encoder",
"sec_num": null
},
{
"text": "EQUATION",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [
{
"start": 0,
"end": 8,
"text": "EQUATION",
"ref_id": "EQREF",
"raw_str": "x = (x 1 , x 2 , . . . , x Tx ): p \u03b8 (y, z|x) = Ty t=1 p \u03b8 (y t , z t |y <t , z <t , x).",
"eq_num": "(4)"
}
],
"section": "... st+1 Encoder",
"sec_num": null
},
{
"text": "Note that the word observation y t can be either a word w t from the shortlist softmax or a location l t from the location softmax, depending on the switching variable z t . Considering this, we can factorize the above equation further",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "... st+1 Encoder",
"sec_num": null
},
{
"text": "p(y, z|x) = t\u2208Tw p(w t , z t |(y, z) <t , x) \u00d7 t \u2208T l p(l t , z t |(y, z) <t , x). (5)",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "... st+1 Encoder",
"sec_num": null
},
{
"text": "Here, T w is a set of time steps where z t = 1, and T l is a set of time-steps where z t = 0. And, T w \u222aT l = {1, 2, . . . , T y } and T w \u2229 T l = \u2205. We denote all previous observations at step t by (y, z) <t . Note also that h t = f ((y, z) <t ).",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "... st+1 Encoder",
"sec_num": null
},
{
"text": "Then, the joint probabilities inside each product can be further factorized as follows:",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "... st+1 Encoder",
"sec_num": null
},
{
"text": "EQUATION",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [
{
"start": 0,
"end": 8,
"text": "EQUATION",
"ref_id": "EQREF",
"raw_str": "p(w t , z t |(y, z) <t ) = p(w t |z t = 1, (y, z) <t ) \u00d7 p(z t = 1|(y, z) <t ) (6) p(l t , z t |(y, z) <t ) = p(l t |z t = 0, (y, z) <t ) \u00d7 p(z t = 0|(y, z) <t )",
"eq_num": "(7)"
}
],
"section": "... st+1 Encoder",
"sec_num": null
},
{
"text": "here, we omitted x which is conditioned on all probabilities in the above.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "... st+1 Encoder",
"sec_num": null
},
{
"text": "The switch probability is modeled as a multilayer perceptron with binary output:",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "... st+1 Encoder",
"sec_num": null
},
{
"text": "p(z t = 1|(y, z) <t , x) = \u03c3(f (x, h t\u22121 ; \u03b8)) (8) p(z t = 0|(y, z) <t , x) = 1 \u2212 \u03c3(f (x, h t\u22121 ; \u03b8)). (9)",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "... st+1 Encoder",
"sec_num": null
},
{
"text": "And p(w t |z t = 1, (y, z) <t , x) is the shortlist softmax and p(l t |z t = 0, (y, z) <t , x) is the location softmax which can be a pointer network. \u03c3(\u2022) stands for the sigmoid function, \u03c3(x) = 1 exp(-x)+1 . Given N such context and target sequence pairs, our training objective is to maximize the following log likelihood w.r.t. the model parameter \u03b8",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "... st+1 Encoder",
"sec_num": null
},
{
"text": "EQUATION",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [
{
"start": 0,
"end": 8,
"text": "EQUATION",
"ref_id": "EQREF",
"raw_str": "arg max \u03b8 1 N N n=1 log p \u03b8 (y n , z n |x n ).",
"eq_num": "(10)"
}
],
"section": "... st+1 Encoder",
"sec_num": null
},
{
"text": "In this section, we discuss practical details of the three fundamental components of the pointer softmax. The interactions between these components and the model is depicted in Figure 3 .",
"cite_spans": [],
"ref_spans": [
{
"start": 177,
"end": 185,
"text": "Figure 3",
"ref_id": null
}
],
"eq_spans": [],
"section": "Basic Components of the Pointer Softmax",
"sec_num": "4.1"
},
{
"text": "Location Softmax l t : The location of the word to copy from source text to the target is predicted by the location softmax l t . The location softmax outputs the conditional probability distribution p(l t |z t = 0, (y, z) <t , x). For models using the attention mechanism such as NMT, we can reuse the probability distributions over the source words in order to predict the location of the word to point. Otherwise we can simply use a pointer network of the model to predict the location.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Basic Components of the Pointer Softmax",
"sec_num": "4.1"
},
{
"text": "Shortlist Softmax w t : The subset of the words in the vocabulary V is being predicted by the shortlist softmax w t .",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Basic Components of the Pointer Softmax",
"sec_num": "4.1"
},
{
"text": "Switching network d t : The switching network d t is an MLP with sigmoid output function that outputs a scalar probability of switching between l t and w t , and represents the conditional probability distribution p(z t |(y, z) <t , x). For NMT model, we condition the MLP that outputs the switching probability on the representation of the context of the source text c t and the hidden state of the decoder h t . Note that, during the training, d t is observed, and thus we do not have to sample. The output of the pointer softmax, f t will be the concatenation of the the two vectors, d t \u00d7 w t and",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Basic Components of the Pointer Softmax",
"sec_num": "4.1"
},
{
"text": "(1 \u2212 d t ) \u00d7 l t .",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Basic Components of the Pointer Softmax",
"sec_num": "4.1"
},
{
"text": "At test time, we compute Eqn. (6) and (7) for all shortlist word w t and all location l t , and pick the word or location of the highest probability.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Basic Components of the Pointer Softmax",
"sec_num": "4.1"
},
{
"text": "In this section, we provide our main experimental results with the pointer softmax on machine translation and summarization tasks. In our experiments, we have used the same baseline model and just replaced the softmax layer with pointer softmax layer at the language model. We use the Adadelta (Zeiler, 2012) learning rule for the training of NMT models. The code for pointer softmax model is available at https://github.com/ caglar/pointer_softmax.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Experiments",
"sec_num": "5"
},
{
"text": "We construct a synthetic task and run some preliminary experiments in order to compare the results with the pointer softmax and the regular softmax's performance for the rare-words. The vocabulary size of our synthetic task is |V |= 600 using sequences of length 7. The words in the sequences are sampled according to their unigram distribu-tion which has the form of a geometric distribution. The task is to predict the least frequent word in the sequence according to unigram distribution of the words. During the training, the sequences are generated randomly. Before the training, validation and test sets are constructed with a fixed seed.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "The Rarest Word Detection",
"sec_num": "5.1"
},
{
"text": "We use a GRU layer over the input sequence and take the last-hidden state, in order to get the summary c t of the input sequence. The w t , l t are only conditioned on c t , and the MLP predicting the d t is conditioned on the latent representations of w t and l t . We use minibatches of size 250 using adam adaptive learning rate algorithm (Kingma and Adam, 2015) using the learning rate of 8 \u00d7 10 \u22124 and hidden layers with 1000 units.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "The Rarest Word Detection",
"sec_num": "5.1"
},
{
"text": "We train a model with pointer softmax where we assign pointers for the rarest 60 words and the rest of the words are predicted from the shortlist softmax of size 540. We observe that increasing the inverse temperature of the sigmoid output of d t to 2, in other words making the decisions of d t to become sharper, works better, i.e. d t = \u03c3(2x).",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "The Rarest Word Detection",
"sec_num": "5.1"
},
{
"text": "At the end of training with pointer softmax we obtain the error rate of 17.4% and by using softmax over all 600 tokens, we obtain the error-rate of 48.2%.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "The Rarest Word Detection",
"sec_num": "5.1"
},
{
"text": "In these series of experiments, we use the annotated Gigaword corpus as described in (Rush et al., 2015) . Moreover, we use the scripts that are made available by the authors of (Rush et al., 2015) 1 to preprocess the data, which results to approximately 3.8M training examples. This script generates about 400K validation and an equal number of test examples, but we use a randomly sampled subset of 2000 examples each for validation and testing. We also have made small modifications to the script to extract not only the tokenized words, but also system-generated named-entity tags. We have created two different versions of training data for pointers, which we call UNK-pointers data and entity-pointers data respectively.",
"cite_spans": [
{
"start": 85,
"end": 104,
"text": "(Rush et al., 2015)",
"ref_id": "BIBREF16"
}
],
"ref_spans": [],
"eq_spans": [],
"section": "Summarization",
"sec_num": "5.2"
},
{
"text": "For the UNK-pointers data, we trim the vocabulary of the source and target data in the training set and replace a word by the UNK token whenever a word occurs less than 5 times in either source or target data separately. Then, we create pointers from each UNK token in the target data to the position in the corresponding source document where the same word occurs in the source, as seen in the data before UNKs were created. It is possible that the source can have an UNK in the matching position, but we still created a pointer in this scenario as well. The resulting data has 2.7 pointers per 100 examples in the training set and 9.1 pointers rate in the validation set.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Summarization",
"sec_num": "5.2"
},
{
"text": "In the entity-pointers data, we exploit the named-entity tags in the annotated corpus and first anonymize the entities by replacing them with an integer-id that always starts from 1 for each document and increments from left to right. Entities that occur more than once in a single document share the same id. We create the anonymization at token-level, so as to allow partial entity matches between the source and target for multi-token entities. Next, we create a pointer from the target to source on similar lines as before, but only for exact matches of the anonymized entities. The resulting data has 161 pointers per 100 examples in the training set and 139 pointers per 100 examples in the validation set.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Summarization",
"sec_num": "5.2"
},
{
"text": "If there are multiple matches in the source, either in the UNK-pointers data or the entitypointers data, we resolve the conflict in favor of the first occurrence of the matching word in the source document. In the UNK data, we model the UNK tokens on the source side using a single placeholder embedding that is shared across all documents, and in the entity-pointers data, we model each entity-id in the source by a distinct placeholder, each of which is shared across all documents.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Summarization",
"sec_num": "5.2"
},
{
"text": "In all our experiments, we use a bidirectional GRU- RNN (Chung et al., 2014) for the encoder and a uni-directional RNN for the decoder. To speed-up training, we use the large-vocabulary trick (Jean et al., 2014) where we limit the vocabulary of the softmax layer of the decoder to 2000 words dynamically chosen from the words in the source documents of each batch and the most common words in the target vocabulary. In both experiments, we fix the embedding size to 100 and the hidden state dimension to 200. We use pretrained word2vec vectors trained on the same corpus to initialize the embeddings, but we finetuned them by backpropagating through the embeddings during training. Our vocabulary sizes are fixed to 125K for source and 75K for target for both exper-iments.",
"cite_spans": [
{
"start": 52,
"end": 76,
"text": "RNN (Chung et al., 2014)",
"ref_id": null
},
{
"start": 192,
"end": 211,
"text": "(Jean et al., 2014)",
"ref_id": "BIBREF10"
}
],
"ref_spans": [],
"eq_spans": [],
"section": "Summarization",
"sec_num": "5.2"
},
{
"text": "We use the reference data for pointers for the model only at the training time. During the test time, the switch makes a decision at every timestep on which softmax layer to use.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Summarization",
"sec_num": "5.2"
},
{
"text": "For evaluation, we use full-length Rouge F1 using the official evaluation tool 2 . In their work, the authors of (Bahdanau et al., 2014) use full-length Rouge Recall on this corpus, since the maximum length of limited-length version of Rouge recall of 75 bytes (intended for DUC data) is already long for Gigaword summaries. However, since full-length Recall can unfairly reward longer summaries, we also use full-length F1 in our experiments for a fair comparison between our models, independent of the summary length.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Summarization",
"sec_num": "5.2"
},
{
"text": "The experimental results comparing the Pointer Softmax with NMT model are displayed in Table 1 for the UNK pointers data and in Table 2 for the entity pointers data. As the experiments show, pointer softmax improves over the baseline NMT on both UNK data and entities data. Our hope was that the improvement would be larger for the entities data since the incidence of pointers was much greater. However, it turns out this is not the case, and we suspect the main reason is anonymization of entities which removed datasparsity by converting all entities to integer-ids that are shared across all documents. We believe that on de-anonymized data, our model could help more, since the issue of data-sparsity is more acute in this case. In Table 3 , we provide the results for summarization on Gigaword corpus in terms of recall as also similar comparison done by (Rush et al., 2015) . We observe improvements on all the scores with the addition of pointer softmax. Let us note that, since the test set of (Rush et al., 2015) is not publicly available, we sample 2000 texts with their summaries without replacement from the validation set and used those examples as our test set.",
"cite_spans": [
{
"start": 861,
"end": 880,
"text": "(Rush et al., 2015)",
"ref_id": "BIBREF16"
},
{
"start": 1003,
"end": 1022,
"text": "(Rush et al., 2015)",
"ref_id": "BIBREF16"
}
],
"ref_spans": [
{
"start": 128,
"end": 135,
"text": "Table 2",
"ref_id": "TABREF2"
},
{
"start": 737,
"end": 744,
"text": "Table 3",
"ref_id": "TABREF3"
}
],
"eq_spans": [],
"section": "Summarization",
"sec_num": "5.2"
},
{
"text": "In Table 4 we present a few system generated summaries from the Pointer Softmax model trained on the UNK pointers data. From those examples, it is apparent that the model has learned to accurately point to the source positions whenever it needs to generate rare words in the summary.",
"cite_spans": [],
"ref_spans": [
{
"start": 3,
"end": 10,
"text": "Table 4",
"ref_id": null
}
],
"eq_spans": [],
"section": "Summarization",
"sec_num": "5.2"
},
{
"text": "In our neural machine translation (NMT) experiments, we train NMT models with attention over the Europarl corpus (Bahdanau et al., 2014) over the sequences of length up to 50 for English to French translation. 3 . All models are trained with early-stopping which is done based on the negative log-likelihood (NLL) on the development set. Our evaluations to report the performance of our models are done on newstest2011 by using BLUE score. 4 We use 30, 000 tokens for both the source and the target language shortlist vocabularies (1 of the token is still reserved for the unknown words). The whole corpus contains 134, 831 unique English words and 153, 083 unique French words. We have created a word-level dictionary from French to English which contains translation of 15,953 words that are neither in shortlist vocabulary nor dictionary of common words for both the source and the target. There are about 49, 490 words shared between English and French parallel corpora of Europarl. Table 4 : Generated summaries from NMT with PS. Boldface words are the words copied from the source.",
"cite_spans": [
{
"start": 440,
"end": 441,
"text": "4",
"ref_id": null
}
],
"ref_spans": [
{
"start": 987,
"end": 994,
"text": "Table 4",
"ref_id": null
}
],
"eq_spans": [],
"section": "Neural Machine Translation",
"sec_num": "5.3"
},
{
"text": "Source #1 china 's tang gonghong set a world record with a clean and jerk lift of ### kilograms to win the women 's over-## kilogram weightlifting title at the asian games on tuesday . Target #1 china 's tang <unk>,sets world weightlifting record NMT+PS #1 china 's tang gonghong wins women 's weightlifting weightlifting title at asian games Source #2 owing to criticism , nbc said on wednesday that it was ending a three-month-old experiment that would have brought the first liquor advertisements onto national broadcast network television . Target #2 advertising : nbc retreats from liquor commercials NMT+PS #2 nbc says it is ending a three-month-old experiment Source #3 a senior trade union official here wednesday called on ghana 's government to be \" mindful of the plight \" of the ordinary people in the country in its decisions on tax increases . Target #3 tuc official,on behalf of ordinary ghanaians NMT+PS #3 ghana 's government urged to be mindful of the plight During the training, in order to decide whether to pick a word from the source sentence using attention/pointers or to predict the word from the short-list vocabulary, we use a simple heuristic. If the word is not in the short-list vocabulary, we first check if the word y t itself appears in the source sentence. If it is not, we check if the word itself is in the source sentence by using the shared words lookup table for the source and the target language. If the word is in the source sentence, we then use the location of the word in the source as the target. Otherwise we check if one of the English senses from the cross-language dictionary of the French word is in the source. If it is in the source sentence, then we use the location of that word as our translation. Otherwise we just use the argmax of l t as the target.",
"cite_spans": [],
"ref_spans": [
{
"start": 185,
"end": 194,
"text": "Target #1",
"ref_id": "TABREF1"
},
{
"start": 858,
"end": 867,
"text": "Target #3",
"ref_id": "TABREF3"
}
],
"eq_spans": [],
"section": "Neural Machine Translation",
"sec_num": "5.3"
},
{
"text": "For switching network d t , we observed that using a two-layered MLP with noisy-tanh activation (Gulcehre et al., 2016) function with residual connection from the lower layer (He et al., 2015) activation function to the upper hidden layers improves the BLEU score about 1 points over the d t using ReLU activation function. We initialized the biases of the last sigmoid layer of d t to \u22121 such that if d t becomes more biased toward choosing the shortlist vocabulary at the beginning of the training. We renormalize the gradients if the norm of the gradients exceed 1 (Pascanu et al., 2012) .",
"cite_spans": [
{
"start": 175,
"end": 192,
"text": "(He et al., 2015)",
"ref_id": "BIBREF8"
},
{
"start": 568,
"end": 590,
"text": "(Pascanu et al., 2012)",
"ref_id": "BIBREF14"
}
],
"ref_spans": [],
"eq_spans": [],
"section": "Neural Machine Translation",
"sec_num": "5.3"
},
{
"text": "In Table 5 , we provided the result of NMT with pointer softmax and we observe about 3.6 BLEU score improvement over our baseline.",
"cite_spans": [],
"ref_spans": [
{
"start": 3,
"end": 10,
"text": "Table 5",
"ref_id": "TABREF4"
}
],
"eq_spans": [],
"section": "Neural Machine Translation",
"sec_num": "5.3"
},
{
"text": "In Figure 4 , we show the validation curves of the NMT model with attention and the NMT model with shortlist-softmax layer. Pointer softmax converges faster in terms of number of minibatch updates and achieves a lower validation negative-log-likelihood (NLL) (63.91) after 200k updates over the Europarl dataset than the NMT model with shortlist softmax trained for 400k minibatch updates (65.26). Pointer softmax converges faster than the model using the shortlist softmax, because the targets provided to the pointer softmax also acts like guiding hints to the attention.",
"cite_spans": [],
"ref_spans": [
{
"start": 3,
"end": 11,
"text": "Figure 4",
"ref_id": "FIGREF2"
}
],
"eq_spans": [],
"section": "Neural Machine Translation",
"sec_num": "5.3"
},
{
"text": "In this paper, we propose a simple extension to the traditional soft attention-based shortlist softmax by using pointers over the input sequence. We show that the whole model can be trained jointly with single objective function. We observe noticeable improvements over the baselines on machine translation and summarization tasks by using pointer softmax. By doing a very simple mod- ification over the NMT, our model is able to generalize to the unseen words and can deal with rarewords more efficiently. For the summarization task on Gigaword dataset, the pointer softmax was able to improve the results even when it is used together with the large-vocabulary trick. In the case of neural machine translation, we observed that the training with the pointer softmax is also improved the convergence speed of the model as well. For French to English machine translation on Europarl corpora, we observe that using the pointer softmax can also improve the training convergence of the model.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Conclusion",
"sec_num": "6"
},
{
"text": "The Pointer SoftmaxIn this section, we introduce our method, called as the pointer softmax (PS), to deal with the rare and unknown words. The pointer softmax can be",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "",
"sec_num": null
},
{
"text": "https://github.com/facebook/NAMAS",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "",
"sec_num": null
},
{
"text": "http://www.berouge.com/Pages/default. aspx",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "",
"sec_num": null
},
{
"text": "In our experiments, we use an existing code, provided in https://github.com/kyunghyuncho/ dl4mt-material, and on the original model we only changed the last softmax layer for our experiments4 We compute the BLEU score using the multi-blue.perl script from Moses on tokenized sentence pairs.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "",
"sec_num": null
}
],
"back_matter": [
{
"text": "We would also like to thank the developers of Theano 5 , for developing such a powerful tool 5 http://deeplearning.net/software/ theano/ for scientific computing (Theano Development Team, 2016) . We acknowledge the support of the following organizations for research funding and computing support: NSERC, Samsung, Calcul Qu\u00e9bec, Compute Canada, the Canada Research Chairs and CIFAR. C. G. thanks for IBM T.J. Watson Research for funding this research during his internship between October 2015 and January 2016.",
"cite_spans": [
{
"start": 162,
"end": 193,
"text": "(Theano Development Team, 2016)",
"ref_id": null
}
],
"ref_spans": [],
"eq_spans": [],
"section": "Acknowledgments",
"sec_num": "7"
}
],
"bib_entries": {
"BIBREF0": {
"ref_id": "b0",
"title": "Neural machine translation by jointly learning to align and translate",
"authors": [
{
"first": "[",
"middle": [],
"last": "References",
"suffix": ""
},
{
"first": "",
"middle": [],
"last": "Bahdanau",
"suffix": ""
}
],
"year": 2014,
"venue": "",
"volume": "",
"issue": "",
"pages": "",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "References [Bahdanau et al.2014] Dzmitry Bahdanau, Kyunghyun Cho, and Yoshua Bengio. 2014. Neural machine translation by jointly learning to align and translate. CoRR, abs/1409.0473.",
"links": null
},
"BIBREF1": {
"ref_id": "b1",
"title": "Adaptive importance sampling to accelerate training of a neural probabilistic language model",
"authors": [
{
"first": "Yoshua",
"middle": [],
"last": "Bengio",
"suffix": ""
},
{
"first": "Jean-S\u00e9bastien",
"middle": [],
"last": "Sen\u00e9cal",
"suffix": ""
}
],
"year": 2008,
"venue": "IEEE Transactions on",
"volume": "19",
"issue": "4",
"pages": "713--722",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "[Bengio and Sen\u00e9cal2008] Yoshua Bengio and Jean- S\u00e9bastien Sen\u00e9cal. 2008. Adaptive importance sampling to accelerate training of a neural proba- bilistic language model. Neural Networks, IEEE Transactions on, 19(4):713-722.",
"links": null
},
"BIBREF2": {
"ref_id": "b2",
"title": "Antoine Bordes, Nicolas Usunier, Sumit Chopra, and Jason Weston. 2015. Largescale simple question answering with memory networks",
"authors": [
{
"first": "",
"middle": [],
"last": "Bordes",
"suffix": ""
}
],
"year": 2015,
"venue": "",
"volume": "",
"issue": "",
"pages": "",
"other_ids": {
"arXiv": [
"arXiv:1506.02075"
]
},
"num": null,
"urls": [],
"raw_text": "[Bordes et al.2015] Antoine Bordes, Nicolas Usunier, Sumit Chopra, and Jason Weston. 2015. Large- scale simple question answering with memory net- works. arXiv preprint arXiv:1506.02075.",
"links": null
},
"BIBREF3": {
"ref_id": "b3",
"title": "Dzmitry Bahdanau, Fethi Bougares, Holger Schwenk, and Yoshua Bengio. 2014. Learning phrase representations using rnn encoder-decoder for statistical machine translation",
"authors": [
{
"first": "Jianpeng",
"middle": [],
"last": "Lapata2016",
"suffix": ""
},
{
"first": "Mirella",
"middle": [],
"last": "Cheng",
"suffix": ""
},
{
"first": "",
"middle": [],
"last": "Lapata ; Kyunghyun",
"suffix": ""
},
{
"first": "Bart",
"middle": [],
"last": "Cho",
"suffix": ""
},
{
"first": "Caglar",
"middle": [],
"last": "Van Merri\u00ebnboer",
"suffix": ""
},
{
"first": "",
"middle": [],
"last": "Gulcehre",
"suffix": ""
}
],
"year": 2016,
"venue": "",
"volume": "",
"issue": "",
"pages": "",
"other_ids": {
"arXiv": [
"arXiv:1603.07252"
]
},
"num": null,
"urls": [],
"raw_text": "[Cheng and Lapata2016] Jianpeng Cheng and Mirella Lapata. 2016. Neural summarization by ex- tracting sentences and words. arXiv preprint arXiv:1603.07252. [Cho et al.2014] Kyunghyun Cho, Bart Van Merri\u00ebnboer, Caglar Gulcehre, Dzmitry Bahdanau, Fethi Bougares, Holger Schwenk, and Yoshua Bengio. 2014. Learning phrase representations using rnn encoder-decoder for statistical machine translation. arXiv preprint arXiv:1406.1078.",
"links": null
},
"BIBREF4": {
"ref_id": "b4",
"title": "Empirical evaluation of gated recurrent neural networks on sequence modeling",
"authors": [
{
"first": "Chung",
"middle": [],
"last": "",
"suffix": ""
}
],
"year": 2014,
"venue": "",
"volume": "",
"issue": "",
"pages": "",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "[Chung et al.2014] Junyoung Chung, \u00c7 aglar G\u00fcl\u00e7ehre, KyungHyun Cho, and Yoshua Bengio. 2014. Em- pirical evaluation of gated recurrent neural networks on sequence modeling. CoRR, abs/1412.3555.",
"links": null
},
"BIBREF5": {
"ref_id": "b5",
"title": "Dan Gillick, Cliff Brunk, Oriol Vinyals, and Amarnag Subramanya",
"authors": [
{
"first": "[",
"middle": [],
"last": "Gillick",
"suffix": ""
}
],
"year": 2015,
"venue": "Multilingual language processing from bytes",
"volume": "",
"issue": "",
"pages": "",
"other_ids": {
"arXiv": [
"arXiv:1512.00103"
]
},
"num": null,
"urls": [],
"raw_text": "[Gillick et al.2015] Dan Gillick, Cliff Brunk, Oriol Vinyals, and Amarnag Subramanya. 2015. Mul- tilingual language processing from bytes. arXiv preprint arXiv:1512.00103.",
"links": null
},
"BIBREF6": {
"ref_id": "b6",
"title": "Generating sequences with recurrent neural networks",
"authors": [
{
"first": "Alex",
"middle": [],
"last": "Graves",
"suffix": ""
}
],
"year": 2013,
"venue": "",
"volume": "",
"issue": "",
"pages": "",
"other_ids": {
"arXiv": [
"arXiv:1308.0850"
]
},
"num": null,
"urls": [],
"raw_text": "Alex Graves. 2013. Generating se- quences with recurrent neural networks. arXiv preprint arXiv:1308.0850.",
"links": null
},
"BIBREF7": {
"ref_id": "b7",
"title": "Noise-contrastive estimation of unnormalized statistical models, with applications to natural image statistics",
"authors": [
{
"first": "[",
"middle": [],
"last": "Gu",
"suffix": ""
}
],
"year": 2012,
"venue": "Caglar Gulcehre, Marcin Moczulski, Misha Denil, and Yoshua Bengio. 2016. Noisy activation functions",
"volume": "13",
"issue": "",
"pages": "307--361",
"other_ids": {
"arXiv": [
"arXiv:1603.06393"
]
},
"num": null,
"urls": [],
"raw_text": "[Gu et al.2016] Jiatao Gu, Zhengdong Lu, Hang Li, and Victor OK Li. 2016. Incorporating copying mechanism in sequence-to-sequence learning. arXiv preprint arXiv:1603.06393. [Gulcehre et al.2016] Caglar Gulcehre, Marcin Moczulski, Misha Denil, and Yoshua Bengio. 2016. Noisy activation functions. arXiv preprint arXiv:1603.00391. [Gutmann and Hyv\u00e4rinen2012] Michael U Gutmann and Aapo Hyv\u00e4rinen. 2012. Noise-contrastive esti- mation of unnormalized statistical models, with ap- plications to natural image statistics. The Journal of Machine Learning Research, 13(1):307-361.",
"links": null
},
"BIBREF8": {
"ref_id": "b8",
"title": "Deep residual learning for mage recognition",
"authors": [
{
"first": "[",
"middle": [],
"last": "He",
"suffix": ""
}
],
"year": 2015,
"venue": "",
"volume": "",
"issue": "",
"pages": "",
"other_ids": {
"arXiv": [
"arXiv:1512.03385"
]
},
"num": null,
"urls": [],
"raw_text": "[He et al.2015] Kaiming He, Xiangyu Zhang, Shao- qing Ren, and Jian Sun. 2015. Deep resid- ual learning for mage recognition. arXiv preprint arXiv:1512.03385.",
"links": null
},
"BIBREF9": {
"ref_id": "b9",
"title": "Teaching machines to read and comprehend",
"authors": [
{
"first": "[",
"middle": [],
"last": "Hermann",
"suffix": ""
}
],
"year": 2015,
"venue": "Advances in Neural Information Processing Systems",
"volume": "",
"issue": "",
"pages": "1684--1692",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "[Hermann et al.2015] Karl Moritz Hermann, Tomas Kocisky, Edward Grefenstette, Lasse Espeholt, Will Kay, Mustafa Suleyman, and Phil Blunsom. 2015. Teaching machines to read and comprehend. In Ad- vances in Neural Information Processing Systems, pages 1684-1692.",
"links": null
},
"BIBREF10": {
"ref_id": "b10",
"title": "On using very large target vocabulary for neural machine translation",
"authors": [
{
"first": "[",
"middle": [],
"last": "Jean",
"suffix": ""
}
],
"year": 2014,
"venue": "International Conference on Learning Representation",
"volume": "",
"issue": "",
"pages": "",
"other_ids": {
"arXiv": [
"arXiv:1412.2007"
]
},
"num": null,
"urls": [],
"raw_text": "[Jean et al.2014] S\u00e9bastien Jean, Kyunghyun Cho, Roland Memisevic, and Yoshua Bengio. 2014. On using very large target vocabulary for neural ma- chine translation. arXiv preprint arXiv:1412.2007. [Kingma and Adam2015] Diederik P Kingma and Jimmy Ba Adam. 2015. A method for stochastic optimization. In International Conference on Learning Representation.",
"links": null
},
"BIBREF11": {
"ref_id": "b11",
"title": "Addressing the rare word problem in neural machine translation",
"authors": [
{
"first": "[",
"middle": [],
"last": "Luong",
"suffix": ""
}
],
"year": 2012,
"venue": "Proceedings of ACL",
"volume": "15",
"issue": "",
"pages": "817--829",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "[Luong et al.2015] Minh-Thang Luong, Ilya Sutskever, Quoc V Le, Oriol Vinyals, and Wojciech Zaremba. 2015. Addressing the rare word problem in neural machine translation. In Proceedings of ACL. [Matthews et al.2012] Danielle Matthews, Tanya Behne, Elena Lieven, and Michael Tomasello. 2012. Origins of the human pointing gesture: a training study. Developmental science, 15(6):817- 829.",
"links": null
},
"BIBREF12": {
"ref_id": "b12",
"title": "Learning word embeddings efficiently with noise-contrastive estimation",
"authors": [
{
"first": "Andriy",
"middle": [],
"last": "Mnih",
"suffix": ""
},
{
"first": "Koray",
"middle": [],
"last": "Kavukcuoglu",
"suffix": ""
}
],
"year": 2013,
"venue": "Advances in Neural Information Processing Systems",
"volume": "",
"issue": "",
"pages": "2265--2273",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "[Mnih and Kavukcuoglu2013] Andriy Mnih and Koray Kavukcuoglu. 2013. Learning word embeddings efficiently with noise-contrastive estimation. In Ad- vances in Neural Information Processing Systems, pages 2265-2273.",
"links": null
},
"BIBREF13": {
"ref_id": "b13",
"title": "Hierarchical probabilistic neural network language model",
"authors": [
{
"first": "Frederic",
"middle": [],
"last": "Morin",
"suffix": ""
},
{
"first": "Yoshua",
"middle": [],
"last": "Bengio",
"suffix": ""
}
],
"year": 2005,
"venue": "Aistats",
"volume": "5",
"issue": "",
"pages": "246--252",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "[Morin and Bengio2005] Frederic Morin and Yoshua Bengio. 2005. Hierarchical probabilistic neural net- work language model. In Aistats, volume 5, pages 246-252. Citeseer.",
"links": null
},
"BIBREF14": {
"ref_id": "b14",
"title": "Razvan Pascanu, Tomas Mikolov, and Yoshua Bengio. 2012. On the difficulty of training recurrent neural networks",
"authors": [
{
"first": "[",
"middle": [],
"last": "Pascanu",
"suffix": ""
}
],
"year": 2012,
"venue": "",
"volume": "",
"issue": "",
"pages": "",
"other_ids": {
"arXiv": [
"arXiv:1211.5063"
]
},
"num": null,
"urls": [],
"raw_text": "[Pascanu et al.2012] Razvan Pascanu, Tomas Mikolov, and Yoshua Bengio. 2012. On the difficulty of training recurrent neural networks. arXiv preprint arXiv:1211.5063.",
"links": null
},
"BIBREF15": {
"ref_id": "b15",
"title": "How to construct deep recurrent neural networks",
"authors": [
{
"first": "[",
"middle": [],
"last": "Pascanu",
"suffix": ""
}
],
"year": 2013,
"venue": "",
"volume": "",
"issue": "",
"pages": "",
"other_ids": {
"arXiv": [
"arXiv:1312.6026"
]
},
"num": null,
"urls": [],
"raw_text": "[Pascanu et al.2013] Razvan Pascanu, Caglar Gulcehre, Kyunghyun Cho, and Yoshua Bengio. 2013. How to construct deep recurrent neural networks. arXiv preprint arXiv:1312.6026.",
"links": null
},
"BIBREF16": {
"ref_id": "b16",
"title": "A neural attention model for abstractive sentence summarization. CoRR, abs/1509.00685",
"authors": [
{
"first": "Alexander",
"middle": [
"M"
],
"last": "Rush",
"suffix": ""
},
{
"first": "Sumit",
"middle": [],
"last": "Chopra",
"suffix": ""
},
{
"first": "Jason",
"middle": [],
"last": "Weston",
"suffix": ""
},
{
"first": ";",
"middle": [],
"last": "Kuldip",
"suffix": ""
},
{
"first": "K",
"middle": [],
"last": "Paliwal",
"suffix": ""
}
],
"year": 1997,
"venue": "IEEE Transactions on",
"volume": "45",
"issue": "11",
"pages": "2673--2681",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "[Rush et al.2015] Alexander M. Rush, Sumit Chopra, and Jason Weston. 2015. A neural attention model for abstractive sentence summarization. CoRR, abs/1509.00685. [Schuster and Paliwal1997] Mike Schuster and Kuldip K Paliwal. 1997. Bidirectional recur- rent neural networks. Signal Processing, IEEE Transactions on, 45(11):2673-2681.",
"links": null
},
"BIBREF17": {
"ref_id": "b17",
"title": "Neural machine translation of rare words with subword units",
"authors": [
{
"first": "Rico",
"middle": [],
"last": "Sennrich",
"suffix": ""
},
{
"first": "Barry",
"middle": [],
"last": "Haddow",
"suffix": ""
},
{
"first": "Alexandra",
"middle": [],
"last": "Birch",
"suffix": ""
}
],
"year": 2015,
"venue": "",
"volume": "",
"issue": "",
"pages": "",
"other_ids": {
"arXiv": [
"arXiv:1508.07909"
]
},
"num": null,
"urls": [],
"raw_text": "[Sennrich et al.2015] Rico Sennrich, Barry Haddow, and Alexandra Birch. 2015. Neural machine trans- lation of rare words with subword units. arXiv preprint arXiv:1508.07909.",
"links": null
},
"BIBREF18": {
"ref_id": "b18",
"title": "Theano: A Python framework for fast computation of mathematical expressions",
"authors": [],
"year": 2007,
"venue": "Child development",
"volume": "78",
"issue": "3",
"pages": "705--722",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "[Theano Development Team2016] Theano Develop- ment Team. 2016. Theano: A Python framework for fast computation of mathematical expressions. arXiv e-prints, abs/1605.02688, May. [Tomasello et al.2007] Michael Tomasello, Malinda Carpenter, and Ulf Liszkowski. 2007. A new look at infant pointing. Child development, 78(3):705-722.",
"links": null
},
"BIBREF19": {
"ref_id": "b19",
"title": "Oriol Vinyals, Meire Fortunato, and Navdeep Jaitly. 2015. Pointer networks. In Advances in Neural Information Processing Systems",
"authors": [
{
"first": "",
"middle": [],
"last": "Vinyals",
"suffix": ""
}
],
"year": 2015,
"venue": "",
"volume": "",
"issue": "",
"pages": "2674--2682",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "[Vinyals et al.2015] Oriol Vinyals, Meire Fortunato, and Navdeep Jaitly. 2015. Pointer networks. In Ad- vances in Neural Information Processing Systems, pages 2674-2682.",
"links": null
},
"BIBREF20": {
"ref_id": "b20",
"title": "Adadelta: an adaptive learning rate method",
"authors": [
{
"first": "D",
"middle": [],
"last": "Matthew",
"suffix": ""
},
{
"first": "",
"middle": [],
"last": "Zeiler",
"suffix": ""
}
],
"year": 2012,
"venue": "",
"volume": "",
"issue": "",
"pages": "",
"other_ids": {
"arXiv": [
"arXiv:1212.5701"
]
},
"num": null,
"urls": [],
"raw_text": "Matthew D Zeiler. 2012. Adadelta: an adaptive learning rate method. arXiv preprint arXiv:1212.5701.",
"links": null
}
},
"ref_entries": {
"FIGREF0": {
"uris": null,
"num": null,
"text": "Guillaume et Cesar ont une voiture bleue a Lausanne.",
"type_str": "figure"
},
"FIGREF2": {
"uris": null,
"num": null,
"text": "A comparison of the validation learningcurves of the same NMT model trained with pointer softmax and the regular softmax layer. As can be seen from the figures, the model trained with pointer softmax converges faster than the regular softmax layer. Switching network for pointer softmax in thisFigure usesReLU activation function.",
"type_str": "figure"
},
"TABREF0": {
"num": null,
"content": "<table><tr><td/><td/><td/><td>ft</td></tr><tr><td/><td/><td>p</td><td>1 -p</td></tr><tr><td/><td/><td>wt</td><td/></tr><tr><td>...</td><td>st-1</td><td>st</td><td>lt</td></tr><tr><td/><td/><td/><td>ct</td></tr><tr><td/><td>Encoder</td><td>h0</td><td>...</td><td>hk</td></tr></table>",
"type_str": "table",
"text": "well.",
"html": null
},
"TABREF1": {
"num": null,
"content": "<table><tr><td/><td colspan=\"3\">Rouge-1 Rouge-2 Rouge-L</td></tr><tr><td>NMT + lvt</td><td>34.87</td><td>16.54</td><td>32.27</td></tr><tr><td colspan=\"2\">NMT + lvt + PS 35.19</td><td>16.66</td><td>32.51</td></tr></table>",
"type_str": "table",
"text": "Results on Gigaword Corpus when pointers are used for UNKs in the training data, using Rouge-F1 as the evaluation metric.",
"html": null
},
"TABREF2": {
"num": null,
"content": "<table><tr><td colspan=\"4\">: Results on anonymized Gigaword Corpus</td></tr><tr><td colspan=\"4\">when pointers are used for entities, using Rouge-</td></tr><tr><td colspan=\"2\">F1 as the evaluation metric.</td><td/><td/></tr><tr><td/><td colspan=\"3\">Rouge-1 Rouge-2 Rouge-L</td></tr><tr><td>NMT + lvt</td><td>34.89</td><td>16.78</td><td>32.37</td></tr><tr><td colspan=\"2\">NMT + lvt + PS 35.11</td><td>16.76</td><td>32.55</td></tr></table>",
"type_str": "table",
"text": "",
"html": null
},
"TABREF3": {
"num": null,
"content": "<table><tr><td/><td colspan=\"3\">Rouge-1 Rouge-2 Rouge-L</td></tr><tr><td>NMT + lvt</td><td>36.45</td><td>17.41</td><td>33.90</td></tr><tr><td colspan=\"2\">NMT + lvt + PS 37.29</td><td>17.75</td><td>34.70</td></tr></table>",
"type_str": "table",
"text": "Results on Gigaword Corpus for modeling UNK's with pointers in terms of recall.",
"html": null
},
"TABREF4": {
"num": null,
"content": "<table><tr><td/><td>BLEU-4</td></tr><tr><td>NMT</td><td>20.19</td></tr><tr><td colspan=\"2\">NMT + PS 23.76</td></tr></table>",
"type_str": "table",
"text": "Europarl Dataset (EN-FR)",
"html": null
}
}
}
} |