File size: 117,715 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 1701 1702 1703 1704 1705 1706 1707 1708 1709 1710 1711 1712 1713 1714 1715 1716 1717 1718 1719 1720 1721 1722 1723 1724 1725 1726 1727 1728 1729 1730 1731 1732 1733 1734 1735 1736 1737 1738 1739 1740 1741 1742 1743 1744 1745 1746 1747 1748 1749 1750 1751 1752 1753 1754 1755 1756 1757 1758 1759 1760 1761 1762 1763 1764 1765 1766 1767 1768 1769 1770 1771 1772 1773 1774 1775 1776 1777 1778 1779 1780 1781 1782 1783 1784 1785 1786 1787 1788 1789 1790 1791 1792 1793 1794 1795 1796 1797 1798 1799 1800 1801 1802 1803 1804 1805 1806 1807 1808 1809 1810 1811 1812 1813 1814 1815 1816 1817 1818 1819 1820 1821 1822 1823 1824 1825 1826 1827 1828 1829 1830 1831 1832 1833 1834 1835 1836 1837 1838 1839 1840 1841 1842 1843 1844 1845 1846 1847 1848 1849 1850 1851 1852 1853 1854 1855 1856 1857 1858 1859 1860 1861 1862 1863 1864 1865 1866 1867 1868 1869 1870 1871 1872 1873 1874 1875 1876 1877 1878 1879 1880 1881 1882 1883 1884 1885 1886 1887 1888 1889 1890 1891 1892 1893 1894 1895 1896 1897 1898 1899 1900 1901 1902 1903 1904 1905 1906 1907 1908 1909 1910 1911 1912 1913 1914 1915 1916 1917 1918 1919 1920 1921 1922 1923 1924 1925 1926 1927 1928 1929 1930 1931 1932 1933 1934 1935 1936 1937 1938 1939 1940 1941 1942 1943 1944 1945 1946 1947 1948 1949 1950 1951 1952 1953 1954 1955 1956 1957 1958 1959 1960 1961 1962 1963 1964 1965 1966 1967 1968 1969 1970 1971 1972 1973 1974 1975 1976 1977 1978 1979 1980 1981 1982 1983 1984 1985 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 2018 2019 2020 2021 2022 2023 2024 2025 2026 2027 2028 2029 2030 2031 2032 2033 2034 2035 2036 2037 2038 2039 2040 2041 2042 2043 2044 2045 2046 2047 2048 2049 2050 2051 2052 2053 2054 2055 2056 2057 2058 2059 2060 2061 2062 2063 2064 2065 2066 2067 2068 2069 2070 2071 2072 2073 2074 2075 2076 2077 2078 2079 2080 2081 2082 2083 2084 2085 2086 2087 2088 2089 2090 2091 2092 2093 2094 2095 2096 2097 2098 2099 2100 2101 2102 2103 2104 2105 2106 2107 2108 2109 2110 2111 2112 2113 2114 2115 2116 2117 2118 2119 2120 2121 2122 2123 2124 2125 2126 2127 2128 2129 2130 2131 2132 2133 2134 2135 2136 2137 2138 2139 2140 2141 2142 2143 2144 2145 2146 2147 2148 2149 2150 2151 2152 2153 2154 2155 2156 2157 2158 2159 2160 2161 2162 2163 2164 2165 2166 2167 2168 2169 2170 2171 2172 2173 2174 2175 2176 2177 2178 2179 2180 2181 2182 2183 2184 2185 2186 2187 2188 2189 2190 2191 2192 2193 2194 2195 2196 2197 2198 2199 2200 2201 2202 2203 2204 2205 2206 2207 2208 2209 2210 2211 2212 2213 2214 2215 2216 |
{
"paper_id": "2021",
"header": {
"generated_with": "S2ORC 1.0.0",
"date_generated": "2023-01-19T02:10:02.097572Z"
},
"title": "Assessing the Sufficiency of Arguments through Conclusion Generation",
"authors": [
{
"first": "Timon",
"middle": [],
"last": "Gurcke",
"suffix": "",
"affiliation": {
"laboratory": "",
"institution": "Paderborn University",
"location": {
"settlement": "Paderborn",
"country": "Germany"
}
},
"email": ""
},
{
"first": "Milad",
"middle": [],
"last": "Alshomary",
"suffix": "",
"affiliation": {
"laboratory": "",
"institution": "Paderborn University",
"location": {
"settlement": "Paderborn",
"country": "Germany"
}
},
"email": ""
},
{
"first": "Henning",
"middle": [],
"last": "Wachsmuth",
"suffix": "",
"affiliation": {
"laboratory": "",
"institution": "Paderborn University",
"location": {
"settlement": "Paderborn",
"country": "Germany"
}
},
"email": ""
}
],
"year": "",
"venue": null,
"identifiers": {},
"abstract": "The premises of an argument give evidence or other reasons to support a conclusion. However, the amount of support required depends on the generality of a conclusion, the nature of the individual premises, and similar. An argument whose premises make its conclusion rationally worthy to be drawn is called sufficient in argument quality research. Previous work tackled sufficiency assessment as a standard text classification problem, not modeling the inherent relation of premises and conclusion. In this paper, we hypothesize that the conclusion of a sufficient argument can be generated from its premises. To study this hypothesis, we explore the potential of assessing sufficiency based on the output of large-scale pre-trained language models. Our best model variant achieves an F 1-score of .885, outperforming the previous state-of-the-art and being on par with human experts. While manual evaluation reveals the quality of the generated conclusions, their impact remains low ultimately.",
"pdf_parse": {
"paper_id": "2021",
"_pdf_hash": "",
"abstract": [
{
"text": "The premises of an argument give evidence or other reasons to support a conclusion. However, the amount of support required depends on the generality of a conclusion, the nature of the individual premises, and similar. An argument whose premises make its conclusion rationally worthy to be drawn is called sufficient in argument quality research. Previous work tackled sufficiency assessment as a standard text classification problem, not modeling the inherent relation of premises and conclusion. In this paper, we hypothesize that the conclusion of a sufficient argument can be generated from its premises. To study this hypothesis, we explore the potential of assessing sufficiency based on the output of large-scale pre-trained language models. Our best model variant achieves an F 1-score of .885, outperforming the previous state-of-the-art and being on par with human experts. While manual evaluation reveals the quality of the generated conclusions, their impact remains low ultimately.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Abstract",
"sec_num": null
}
],
"body_text": [
{
"text": "The quality assessment of natural language argumentation is nowadays studied extensively for various genres and text granularities, from entire news editorials (El Baff et al., 2020) to arguments in online forums (Lauscher et al., 2020) to single claims in social media discussions (Skitalinskaya et al., 2021) . The reason lies in its importance for driving downstream applications such as writing support (Stab, 2017) , argument search (Wachsmuth et al., 2017b) , and debating technologies (Slonim et al., 2021) . Wachsmuth et al. (2017a) organized quality dimensions of arguments into three complementary aspects: logic, rhetoric, and dialectic. Logical quality refers to the actual argument structure, that is, how strong an argument is in terms of the support of a claim (the argument's conclusion) by evidence and other reasons (the premises).",
"cite_spans": [
{
"start": 160,
"end": 182,
"text": "(El Baff et al., 2020)",
"ref_id": "BIBREF3"
},
{
"start": 213,
"end": 236,
"text": "(Lauscher et al., 2020)",
"ref_id": "BIBREF6"
},
{
"start": 282,
"end": 310,
"text": "(Skitalinskaya et al., 2021)",
"ref_id": "BIBREF10"
},
{
"start": 407,
"end": 419,
"text": "(Stab, 2017)",
"ref_id": "BIBREF13"
},
{
"start": 438,
"end": 463,
"text": "(Wachsmuth et al., 2017b)",
"ref_id": "BIBREF19"
},
{
"start": 492,
"end": 513,
"text": "(Slonim et al., 2021)",
"ref_id": "BIBREF11"
},
{
"start": 516,
"end": 540,
"text": "Wachsmuth et al. (2017a)",
"ref_id": "BIBREF18"
}
],
"ref_spans": [],
"eq_spans": [],
"section": "Introduction",
"sec_num": "1"
},
{
"text": "The first reason why education and preventative measures should receive a greater budget is the potential improvements in health system. I believe that decreasing the number of patients can lead hospitals and healthcare centers to be managed effectively which will result in better treatments for current patients. Therefore, society should be educated and became aware of health issues so that the potential precautions on the way of illnesses can be taken instead of trying to provide treatment for the increasing number of patients.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Introduction",
"sec_num": "1"
},
{
"text": "The second reason why governments should allocate more budget on prevention from illness and providing health education is the welfare of the society. In my opinion, there is nothing more important than health in a human's life and the happiness and welfare come with health. Therefore, a government's role should be providing means that lead its citizens to learn how to prevent from potential illness that can cause misery in people's lives. For example, the marketing campaign of Ministry of Health in Turkey which aimed smoking problem among the youth increased the well-being of those who quit smoking and adapted a better lifestyle after the campaign. Figure 1 : Two example arguments from a persuasive student essay, one classified as sufficient, the other as insufficient in the corpus of Stab and Gurevych (2017b) .",
"cite_spans": [
{
"start": 797,
"end": 822,
"text": "Stab and Gurevych (2017b)",
"ref_id": "BIBREF14"
}
],
"ref_spans": [
{
"start": 658,
"end": 666,
"text": "Figure 1",
"ref_id": null
}
],
"eq_spans": [],
"section": "Introduction",
"sec_num": "1"
},
{
"text": "A key dimension of logical quality is sufficiency, capturing whether an argument's premises together make it rationally worthy of drawing its conclusion (Johnson and Blair, 2006) . Consider, for example, the two arguments on health education in Figure 1 , taken from the argument-annotated essay corpus of Stab and Gurevych (2017a) . While the upper one was deemed sufficient by human experts (Stab and Gurevych, 2017b) , the lower one was not, likely because the second premise tries to reason from a single example. A reliable computational assessment of argument sufficiency would allow systems to determine those arguments that are well-reasoned.",
"cite_spans": [
{
"start": 153,
"end": 178,
"text": "(Johnson and Blair, 2006)",
"ref_id": "BIBREF5"
},
{
"start": 306,
"end": 331,
"text": "Stab and Gurevych (2017a)",
"ref_id": "BIBREF12"
},
{
"start": 393,
"end": 419,
"text": "(Stab and Gurevych, 2017b)",
"ref_id": "BIBREF14"
}
],
"ref_spans": [
{
"start": 245,
"end": 253,
"text": "Figure 1",
"ref_id": null
}
],
"eq_spans": [],
"section": "Introduction",
"sec_num": "1"
},
{
"text": "As detailed in Section 2, previous approaches to argument sufficiency assessment model the task as a standard text classification problem and tackle it with convolutional neural networks (Stab and Gurevych, 2017b) or traditional feature engineer-ing (Wachsmuth and Werner, 2020) . In the focused domain of persuasive student essays, Stab and Gurevych (2017b) obtained a macro F 1 -score of .827, not far away from human performance in their setting (.887). However, to further improve the state of the art, we expect the integration of knowledge beyond what is directly available in the text at hand is needed. In particular, we observe that existing work neither explicitly considers an argument's premises and conclusions, nor a property of their relationship. We hypothesize that only a sufficient argument makes it possible to infer the conclusion from the premises. Consequently, comparing the stated conclusion of an argument with one that is (automatically) generated from the premises could help the model to distinguish sufficient arguments from insufficient ones. This hypothesis raises the question of whether the knowledge encoded in large-scale pre-trained language models can be leveraged, a direction nearly unexplored so far in argument quality assessment.",
"cite_spans": [
{
"start": 187,
"end": 213,
"text": "(Stab and Gurevych, 2017b)",
"ref_id": "BIBREF14"
},
{
"start": 250,
"end": 278,
"text": "(Wachsmuth and Werner, 2020)",
"ref_id": "BIBREF20"
},
{
"start": 333,
"end": 358,
"text": "Stab and Gurevych (2017b)",
"ref_id": "BIBREF14"
}
],
"ref_spans": [],
"eq_spans": [],
"section": "Introduction",
"sec_num": "1"
},
{
"text": "In this paper, we study whether generating a conclusion from an argument's premises benefits the computational assessment of the argument's sufficiency. In particular, we first enrich the argument with structural annotations, highlighting which parts are the premises and which part is the conclusion. We propose in Section 4 to then mask the conclusion in order to learn to re-generate it using fine-tuned BART (Lewis et al., 2020) . Combining the generated conclusion with the original argument and its annotations, our approach learns to distinguish sufficient from insufficient arguments using a modified RoBERTa model (Liu et al., 2019) .",
"cite_spans": [
{
"start": 412,
"end": 432,
"text": "(Lewis et al., 2020)",
"ref_id": "BIBREF7"
},
{
"start": 623,
"end": 641,
"text": "(Liu et al., 2019)",
"ref_id": "BIBREF8"
}
],
"ref_spans": [],
"eq_spans": [],
"section": "Introduction",
"sec_num": "1"
},
{
"text": "Starting from ground-truth argument structure, we subsequently evaluate conclusion generation and sufficiency assessment on the merged annotations of the corpora of Stab and Gurevych (2017a) and Stab and Gurevych (2017b) , as described in Section 3. Our generation experiments indicate that fine-tuning BART leads to better conclusions, which are on par with human-written conclusions in terms of sufficiency, likeliness, and novelty (Section 5). To quantify the impact on sufficiency assessment, we explore various combinations of premises, original conclusion, and generated conclusion in systematic ablation tests, and we compare them to the state of the art and a human upper bound (Section 6). Our sufficiency experiments reveal that, even on the plain input text of an argument, RoBERTa already improves significantly over the state of the art. The addition of structural annotations and the generated conclusion lead to further improvements, although the benefit of generation ultimately remains limited, possibly due to the generally limited importance of knowing the conclusion on the given data. Finally, we discuss the results of our approaches in Section 7 in light of their implications for the field.",
"cite_spans": [
{
"start": 165,
"end": 190,
"text": "Stab and Gurevych (2017a)",
"ref_id": "BIBREF12"
},
{
"start": 195,
"end": 220,
"text": "Stab and Gurevych (2017b)",
"ref_id": "BIBREF14"
}
],
"ref_spans": [],
"eq_spans": [],
"section": "Introduction",
"sec_num": "1"
},
{
"text": "The main contributions of this paper are: 1",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Introduction",
"sec_num": "1"
},
{
"text": "\u2022 A language model that can generate humanlike argument conclusions.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Introduction",
"sec_num": "1"
},
{
"text": "\u2022 The new state-of-the-art approach to argument sufficiency assessment.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Introduction",
"sec_num": "1"
},
{
"text": "\u2022 Insights into the importance of mined and generated structure within argument assessment.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Introduction",
"sec_num": "1"
},
{
"text": "Computational argumentation research has assessed various dimensions of argument quality. Wachsmuth et al. (2017a) provide a theory-based taxonomy of 15 logical, rhetorical, and dialectical quality dimensions and of the work in natural language processing done in these directions. We focus on the (local) sufficiency dimension, which is key to logical cogency, representing that an argument's conclusion can rationally be drawn from its premises, given that these are acceptable and relevant (Johnson and Blair, 2006) . Few approaches tackled sufficiency computationally so far. Aside from Wachsmuth and Werner (2020) who assess it as one of the 15 dimensions above using traditional text-focused feature engineering, we are only aware of the work of Stab and Gurevych (2017b) who extend the argumentannotated essay corpus of Stab and Gurevych (2017a) with binary sufficiency annotations. On this basis, the authors compare a support vector machine using lexical, syntactic, and length features to a convolutional neural network (CNN) with word vectors, the latter achieving the best result with a macro F 1 -score of .827. In our experiments, we use their dataset and replicate their experiment settings, in order to compare to the CNN.",
"cite_spans": [
{
"start": 90,
"end": 114,
"text": "Wachsmuth et al. (2017a)",
"ref_id": "BIBREF18"
},
{
"start": 493,
"end": 518,
"text": "(Johnson and Blair, 2006)",
"ref_id": "BIBREF5"
},
{
"start": 591,
"end": 618,
"text": "Wachsmuth and Werner (2020)",
"ref_id": "BIBREF20"
},
{
"start": 752,
"end": 777,
"text": "Stab and Gurevych (2017b)",
"ref_id": "BIBREF14"
},
{
"start": 827,
"end": 852,
"text": "Stab and Gurevych (2017a)",
"ref_id": "BIBREF12"
}
],
"ref_spans": [],
"eq_spans": [],
"section": "Related Work",
"sec_num": "2"
},
{
"text": "Unlike Stab and Gurevych (2017a) , we rely on a transformer-based architecture, namely we adapt RoBERTa (Liu et al., 2019) to assess sufficiency. Approaches to argument quality assessment using such architectures are still limited, mostly focusing on a holistic view of quality (Gretz et al., 2020; Toledo et al., 2019) , although a few approaches used transformers for some of the dimensions of Wachsmuth et al. (2017a) , such as Lauscher et al. (2020) , or somewhat related dimensions in light of quality improvement (Skitalinskaya et al., 2021) .",
"cite_spans": [
{
"start": 7,
"end": 32,
"text": "Stab and Gurevych (2017a)",
"ref_id": "BIBREF12"
},
{
"start": 104,
"end": 122,
"text": "(Liu et al., 2019)",
"ref_id": "BIBREF8"
},
{
"start": 278,
"end": 298,
"text": "(Gretz et al., 2020;",
"ref_id": null
},
{
"start": 299,
"end": 319,
"text": "Toledo et al., 2019)",
"ref_id": "BIBREF16"
},
{
"start": 396,
"end": 420,
"text": "Wachsmuth et al. (2017a)",
"ref_id": "BIBREF18"
},
{
"start": 431,
"end": 453,
"text": "Lauscher et al. (2020)",
"ref_id": "BIBREF6"
},
{
"start": 519,
"end": 547,
"text": "(Skitalinskaya et al., 2021)",
"ref_id": "BIBREF10"
}
],
"ref_spans": [],
"eq_spans": [],
"section": "Related Work",
"sec_num": "2"
},
{
"text": "In contrast to the standard use of transformers for text classification, we leverage the structure of arguments for their assessment. Wachsmuth et al. (2016) provided evidence that mining the argumentative structure of persuasive essays helps to better assess four essay-level quality dimensions of argumentation. Similarly, we use annotations of the premises and conclusions of arguments for the sufficiency assessment, but we target the arguments. Moreover, we explore to benefit of conclusion generation for the assessment.",
"cite_spans": [
{
"start": 134,
"end": 157,
"text": "Wachsmuth et al. (2016)",
"ref_id": "BIBREF17"
}
],
"ref_spans": [],
"eq_spans": [],
"section": "Related Work",
"sec_num": "2"
},
{
"text": "The idea of reconstructing an argument's conclusion from its premises was introduced by Alshomary et al. 2020, but their approach focused on the inference of a conclusion's target. The actual generation of entire conclusions has so far only been studied by Syed et al. (2021) . The authors presented the first corpus for this task along with experiments where they adapted BART (Lewis et al., 2020 ) from summarization to conclusion generation. While they trained BART to directly generate a conclusion based on premises, we generate conclusions that fit the context of an entire argument. To this end, we leverage and finetune BART's inherent denoising capabilities obtained during pretraining to replace a mask token in an argument.",
"cite_spans": [
{
"start": 257,
"end": 275,
"text": "Syed et al. (2021)",
"ref_id": "BIBREF15"
},
{
"start": 378,
"end": 397,
"text": "(Lewis et al., 2020",
"ref_id": "BIBREF7"
}
],
"ref_spans": [],
"eq_spans": [],
"section": "Related Work",
"sec_num": "2"
},
{
"text": "To study our hypothesis that a sufficient argument's conclusion can be generated from its premises, we need data that is annotated for both argument structure and sufficiency. In this section, we describe how we employ existing corpora for this purpose.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Data",
"sec_num": "3"
},
{
"text": "The argument-annotated essay (AAE-v2) corpus (Stab and Gurevych, 2017a) contains structural annotations for 402 complete persuasive student essays. For conclusion generation (as well as for structure-based sufficiency assessment), we only need annotations of single arguments.",
"cite_spans": [
{
"start": 45,
"end": 71,
"text": "(Stab and Gurevych, 2017a)",
"ref_id": "BIBREF12"
}
],
"ref_spans": [],
"eq_spans": [],
"section": "Data for Conclusion Generation",
"sec_num": "3.1"
},
{
"text": "Our instance creation procedure resembles the one of Alshomary et al. 2020, but we work on argument level rather than conclusion level, since we approach conclusion generation as a language model denoising task (Lewis et al., 2020) . Concretely, instead of using premises-conclusion training pairs where the conclusion shall be generated given the premises, we rely on argument-argument pairs. The first argument here is a modified version of the second argument where the conclusion is masked. This way, we avoid conflicts with the argument-level sufficiency annotations of Stab and Gurevych (2017b) (see below). In total, we obtain 1506 argument-argument pairs relating to 1506 unique conclusions matched with 1029 unique arguments. On average, each argument has a length of 4.5 sentences and contains 94.6 tokens.",
"cite_spans": [
{
"start": 211,
"end": 231,
"text": "(Lewis et al., 2020)",
"ref_id": "BIBREF7"
}
],
"ref_spans": [],
"eq_spans": [],
"section": "Data for Conclusion Generation",
"sec_num": "3.1"
},
{
"text": "For training, we rely on 5-fold cross-validation, ensuring that the argument-argument pairs from one essay are never split between training, validation, and test data. This prevents possible data leakage that could artificially improve the final evaluation scores. For each folding, we use 70% training, 10% validation, and 20% test data.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Data for Conclusion Generation",
"sec_num": "3.1"
},
{
"text": "Stab and Gurevych (2017b) further classified each argument in the 402 essays of the AAE-v2 corpus as being sufficient or not. Following Johnson and Blair 2006, the authors defined that an \"argument complies with the sufficiency criterion if its premises provide enough evidence for accepting or rejecting the claim\" (we speak of \"conclusion\" here instead of \"claim\"). All 1029 arguments were labeled, of which 681 (66.2%) were considered sufficient and 348 (33.8%) insufficient.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Data for Sufficiency Assessment",
"sec_num": "3.2"
},
{
"text": "We use the provided corpus both in its original form and in a modified version where we replace the conclusion of an argument with two separator tokens, \"</s></s>\". This allows us to study a wide range of different approaches for sufficiency assessment by placing text in-between the two tokens as a replacement for the original conclusion.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Data for Sufficiency Assessment",
"sec_num": "3.2"
},
{
"text": "For training, we replicate the original 20-times 5-fold cross-validation setup of Stab and Gurevych (2017b) , with 70% training, 10% validation, and 20% test data, in order to ensure comparability.",
"cite_spans": [
{
"start": 82,
"end": 107,
"text": "Stab and Gurevych (2017b)",
"ref_id": "BIBREF14"
}
],
"ref_spans": [],
"eq_spans": [],
"section": "Data for Sufficiency Assessment",
"sec_num": "3.2"
},
{
"text": "This section describes our two-step approach to assess the sufficiency of a given argument through conclusion generation. First, we generate a conclusion from the argument's premises using a pretrained language model finetuned on the task of replacing the masked conclusion of an argument. Second, the generated conclusion is used to assess the argument's sufficiency by experimenting with eight modified versions of the original input argument (Section 6). An overview of the approach is shown in Figure 2 . In the following, we detail how we train the models for the two steps.",
"cite_spans": [],
"ref_spans": [
{
"start": 498,
"end": 506,
"text": "Figure 2",
"ref_id": "FIGREF1"
}
],
"eq_spans": [],
"section": "Approach",
"sec_num": "4"
},
{
"text": "Given an argument with a masked conclusion, the first task is to re-generate the conclusion. To tackle this task, we use BART-large (Lewis et al., 2020) and treat generation as a denoising task. We explore two model variants:",
"cite_spans": [
{
"start": 132,
"end": 152,
"text": "(Lewis et al., 2020)",
"ref_id": "BIBREF7"
}
],
"ref_spans": [],
"eq_spans": [],
"section": "Conclusion Generation using Denoising",
"sec_num": "4.1"
},
{
"text": "BART-unsupervised In this variant, we do not finetune BART on any data, but we use its vanilla denoising capabilities obtained in its pre-training procedure. Note that the masked conclusions usually do not represent entire sentences, thus leaving textual markers which trigger BART to generate a logical conclusion, for example, \"Thus, <mask>\" or \"This makes it clear that <mask>.\" We consider this model as a baseline.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Conclusion Generation using Denoising",
"sec_num": "4.1"
},
{
"text": "BART-supervised In this variant, we finetune BART on the data from Section 3, in order to tailor its denoising capabilities towards conclusion generation. In particular, we thereby adjust the language model towards the given domain and teach the model to replace the mask token with a conclusion (instead of just generating text that fits the context). We finetune BART using cross-entropy loss, as commonly done in text generation.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Conclusion Generation using Denoising",
"sec_num": "4.1"
},
{
"text": "The proper training settings of the two models are found via a hyperparameter search. For the evaluation described below, we ran 10 trials for each fold, testing batch sizes between 4 and 8 and learning rates between 5 \u2022 10 \u22126 and 5 \u2022 10 \u22125 . We fixed the number of epochs to 3 per fold, as we did not observe any improvements afterwards, and we used a cosine learning rate scheduler with 50 warmup steps to stabilize the training. At inference time, we employed a beam size of 5 to obtain the final generated conclusions. We considered the epoch out of three, which performs best on the validation data in terms of BERTScore (Zhang* et al., 2020) .",
"cite_spans": [
{
"start": 626,
"end": 647,
"text": "(Zhang* et al., 2020)",
"ref_id": "BIBREF21"
}
],
"ref_spans": [],
"eq_spans": [],
"section": "Conclusion Generation using Denoising",
"sec_num": "4.1"
},
{
"text": "Given a modified argument, the second task is to predict whether the premises in the argument are rationally worth drawing the conclusion. We use RoBERTa (Liu et al., 2019) for this task by adding a linear layer on top of the pooled output of the original model (Devlin et al., 2019) .",
"cite_spans": [
{
"start": 154,
"end": 172,
"text": "(Liu et al., 2019)",
"ref_id": "BIBREF8"
},
{
"start": 262,
"end": 283,
"text": "(Devlin et al., 2019)",
"ref_id": "BIBREF1"
}
],
"ref_spans": [],
"eq_spans": [],
"section": "Sufficiency Assessment using Structure",
"sec_num": "4.2"
},
{
"text": "Successfully optimizing RoBERTa using mean squared error (MSE) or cross-entropy loss functions would be difficult, as both of them do not align well with the target metric of sufficiency assessment (macro F 1 -score). We therefore follow ideas of Puthiya Parambath et al. 2014and Eban et al. (2017) who propose to optimize machine learning models on the F 1 -score directly. Accordingly, we allow the model to output probabilities instead of interpreting a single binary value. Analogous to Stab and Gurevych (2017b) , we allow our model to adjust hyperparameters between folds. In our experiments, we followed the same hyperparameter optimization procedure as before but for different parameters and ranges. In total, we ran 10 trials for each fold, and we adjusted the batch size to be between 16 and 32 and the learning rate between 10 \u22126 and 5 \u2022 10 \u22125 . We selected the epoch for each trial out of three, which performed best on the validation data in terms of macro F 1 -score. Table 1 : Automatic evaluation results of concluion generation: Rescaled F1-BERTScore and ROUGE-1/-2/-L scores of the two considered models on the full corpus.",
"cite_spans": [
{
"start": 280,
"end": 298,
"text": "Eban et al. (2017)",
"ref_id": "BIBREF2"
},
{
"start": 491,
"end": 516,
"text": "Stab and Gurevych (2017b)",
"ref_id": "BIBREF14"
}
],
"ref_spans": [
{
"start": 983,
"end": 990,
"text": "Table 1",
"ref_id": null
}
],
"eq_spans": [],
"section": "Sufficiency Assessment using Structure",
"sec_num": "4.2"
},
{
"text": "To study our hypothesis, we need to ensure that the generated conclusions are meaningful and fit in the context of a given argument, so they can be helpful in sufficiency assessment. In this section, we therefore evaluate the quality of the generated conclusions, both automatically and manually.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Evaluation of Conclusion Generation",
"sec_num": "5"
},
{
"text": "As indicated in Section 4, we compare two approaches: (1) BART-unsupervised, which replaces the mask token in an argument (denoising) with fitting text, as it is part of BART's training procedure (Lewis et al., 2020) ; and (2) BART-supervised, which finetunes BART on the argument-argument pairs from Section 3. For both approaches, we obtained the complete set of generated conclusions using the cross-validation setup described in Section 4. Matching these with the corresponding ground-truth conclusion, we then computed their quality in terms of BERTScore as well as ROUGE-1, ROUGE-2, and ROUGE-L.",
"cite_spans": [
{
"start": 196,
"end": 216,
"text": "(Lewis et al., 2020)",
"ref_id": "BIBREF7"
}
],
"ref_spans": [],
"eq_spans": [],
"section": "Automatic Evaluation",
"sec_num": "5.1"
},
{
"text": "Results Table 1 lists the results of the approaches. BART-unsupervised is a strong baseline in terms of lexical accuracy: Values such as 19.69 (ROUGE-1) and 16.40 (ROUGE-L) are comparable to those that Syed et al. (2021) achieved in similar domains with sophisticated approaches. However, finetuning on the argument-argument pairs does not only significantly increase the semantic similarity between generated and ground truth conclusions from 0.14 to 0.25 in terms of BERTScore, but it also leads to a slight increase in lexical accuracy.",
"cite_spans": [],
"ref_spans": [
{
"start": 8,
"end": 15,
"text": "Table 1",
"ref_id": null
}
],
"eq_spans": [],
"section": "Automatic Evaluation",
"sec_num": "5.1"
},
{
"text": "The metrics used to automatically evaluate conclusion generation are not ideal, since they expect a single correct result. The task of conclusion generation, in contrast, allows for multiple, possibly very different correct conclusions, for example, a different conclusion target may be derived from a single set of premises (Alshomary et al., 2020) . We thus conducted an additional manual annotation study to evaluate the quality of the conclusions generated by the two approaches in comparison to the human ground truth. We randomly chose 100 arguments from the given corpus, 50 labeled as sufficient and 50 labeled as insufficient. For each arguments, we additionally created two variants, replacing the original conclusion with the generated conclusion of either approach. In each case, we then presented the three arguments with their premises and conclusions highlighted to five annotators of different academic backgrounds (economics, computer science, health/medicine), none being an author of this paper. We asked each annotator three questions, Q1-Q3, on each argument, resulting 300 annotations for each model and 900 annotations in total. For consistency reasons, we used a 5-point Likert scale for each question:",
"cite_spans": [
{
"start": 325,
"end": 349,
"text": "(Alshomary et al., 2020)",
"ref_id": "BIBREF0"
}
],
"ref_spans": [],
"eq_spans": [],
"section": "Manual Evaluation",
"sec_num": "5.2"
},
{
"text": "\u2022 Q1: Are the premises sufficient to draw the conclusion? This question referred to the sufficiency of arguments, from \"not sufficient\" (score 1) to \"sufficient\" (score 5). We asked this question to see how the sufficiency of generated and human-written conclusions differs, thus directly evaluating our hypothesis.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Manual Evaluation",
"sec_num": "5.2"
},
{
"text": "\u2022 Q2: How likely is it that the conclusion will be inferred from the context? This question referred to the likelihood of a conclusion, from \"very unlikely\" (score 1) to \" very likely\" (score 5). We asked this question as an internal quality assurance, ruling out the possibility that our models generate conclusions unrelated to the given context of the argument.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Manual Evaluation",
"sec_num": "5.2"
},
{
"text": "\u2022 Q3: How can the conclusion be composed from the context? This question, finally, referred to the novelty of the generated conclusions in light of their context. The score range here is more complex; inspired by Syed et al. (2021) , who also ask annotators about the novelty of generated conclusions, it reflects the cognitive load required to infer the conclusions from the context of the argument: \"verbatim copying\" (1), \"synonymous copying\" (2), \"copying + fusion\" (3), \"inference\" (4), and \"can not be composed\" (5).",
"cite_spans": [
{
"start": 213,
"end": 231,
"text": "Syed et al. (2021)",
"ref_id": "BIBREF15"
}
],
"ref_spans": [],
"eq_spans": [],
"section": "Manual Evaluation",
"sec_num": "5.2"
},
{
"text": "Results For each question, Table 2 shows the inter-annotator agreement, the distribution of majority scores, and the resulting mean score of the three compared approaches (including the ground truth), and the mean rank. We obtained the rank by Table 2 : Manual evaluation results of conclusion generation on the 100 arguments of each of the three approaches: (a) Agreement of all five annotators in terms of Krippendorff's \u03b1 and majority. (b) Distribution of majority scores. For Q1/Q2, higher scores mean more sufficient/likely. For Q3, they mean less \"copying\" (see text for details). (c) Mean score of each approach and rank obtained by comparing the majority score for each argument in isolation.",
"cite_spans": [],
"ref_spans": [
{
"start": 27,
"end": 34,
"text": "Table 2",
"ref_id": null
},
{
"start": 244,
"end": 251,
"text": "Table 2",
"ref_id": null
}
],
"eq_spans": [],
"section": "Manual Evaluation",
"sec_num": "5.2"
},
{
"text": "treating each question as a ranking task where, for each argument, the approaches are ranked from 1 to 3 by decreasing highest majority score.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Manual Evaluation",
"sec_num": "5.2"
},
{
"text": "We find that the general agreement for the questions in terms of Krippendorff's \u03b1 is low, with values between .19 and .50, but comparable to other tasks in the realm of argumentation (Wachsmuth et al., 2017a) . On all three questions, the annotators agreed mostly for BART-unsupervised, followed by the BART-supervised, while having the least agreement for the ground truth. This may indicate a more apparent connection of the generated conclusions to the premises. In 57% to 64% of the cases, we observe majority agreement of the annotators for the first two questions, whereas this value goes up to 72%-78% for the last question.",
"cite_spans": [
{
"start": 183,
"end": 208,
"text": "(Wachsmuth et al., 2017a)",
"ref_id": "BIBREF18"
}
],
"ref_spans": [],
"eq_spans": [],
"section": "Manual Evaluation",
"sec_num": "5.2"
},
{
"text": "The ranking for Q1 shows that conclusions generated by our BART-supervised model overall ranked best in sufficiency (mean rank 1.37), even though the mean score is slightly better for the ground truth. While the differences are small, this behavior is expected as half of the provided arguments were initially labeled as insufficient.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Manual Evaluation",
"sec_num": "5.2"
},
{
"text": "Considering the likelihood of the premises (Q2), we see that conclusions of BART-supervised are on par with the ground truth conclusions (score rank 2.96 vs. 2.98, rank 1.39 vs. 1.42) and better than those generated by the baseline BARTunsupervised (1.54) . This suggests that the conclusions generated by our model both fit the context of the argument and are at least as likely to be drawn as the ones written by humans. This property is essential for our approach to sufficiency assessment to rule out the possibility of failure due to a low quality of the generated conclusions in general.",
"cite_spans": [
{
"start": 232,
"end": 255,
"text": "BARTunsupervised (1.54)",
"ref_id": null
}
],
"ref_spans": [],
"eq_spans": [],
"section": "Manual Evaluation",
"sec_num": "5.2"
},
{
"text": "Finally, we consider the cognitive load that is required to compose a conclusion given its context, as reflected by novelty (Q3). This information is vital to rule out that the generated conclusions are copied from the context of an argument instead of being inferred. We find the ground-truth conclusions to require the most cognitive load in this regard, having a clearly better mean rank (1.39) than the others (both 1.54). Thus, they potentially provide the most novelty to the context. The mean scores indicate an increase of novelty from BART-unsupervised (3.34) to BART-supervised (3.47) though.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Manual Evaluation",
"sec_num": "5.2"
},
{
"text": "The scores and ranks can be interpreted more easily when looking at the majority scores for the two BART models and the ground truth. As expected, we find that the amount of conclusions considered to be sufficient is approximately the same to those considered insufficient. The annotators did not agree on the highest sufficiency rank, which may be due to subjectivity in the perception of \"full\" sufficiency. We observe an analog behavior for score 5 for the likelihood of conclusions (Q2). Here, this may imply that, rarely, only a single conclusion would fit the premises of an argument. Regarding the novelty of the generated conclusions, Q3 reveals that the ground truth annotations are mostly inferences (score 4) and only rarely copied from the context (scores 1 and 2). While BART-unsupervised only somewhat follows this distribution, our BART-supervised model shows a similar behavior to the ground truth, though in less clear form. This is another indication that BART-supervised has learned to generate conclusions from premises.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Manual Evaluation",
"sec_num": "5.2"
},
{
"text": "Part Text (a) Sufficient Argument Second, <MASK>. Averagely, public transports use much less gasoline to carry people than private cars. It means that by using public transports, the less gas exhaust is pumped to the air and people will no longer have to bear the stuffy situation on the roads, which is always full of fumes.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "# Label",
"sec_num": null
},
{
"text": "Ground truth public transportation helps to solve the air pollution problems BART-unsupervised public transport is more efficient than private cars BART-supervised using public transports will help to reduce the amount of pollution in the air (b) Insufficient Argument Last, <MASK>. Playing musical instrument is a good way, I can play classical guitar. When I meet difficulties in studies, I will take my guitar and play the song Green Sleeves. It makes me feel better and gives me the confidence.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "# Label",
"sec_num": null
},
{
"text": "Ground truth we should develop at least one personal hobby, not to show off, but express our emotion when we feel depressed or pressured BART-unsupervised but not least, I love music BART-supervised playing musical instrument is very important to me (c) Insufficient Argument In addition to this, <MASK>. For instance, further enforcement banned smoking in capital in Sri Lanka has reduced this consumption related diseases and deaths, as per the ministry of health. As this shows, smoking restrictions has successfully daunted public from this bad puffing that put less strain on country's healthcare systems.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "# Label",
"sec_num": null
},
{
"text": "Ground truth introducing smoking ban in public places would greatly discourage people from engaging tobacco puffing BART-unsupervised Sri Lankan government has taken several measures to curb smoking BART-supervised smoking restrictions in Sri Lanka has brought a lot of benefits to the country Table 3 : Conclusions generated by the BART models for four arguments (with masked conclusion) compared to the ground truth conclusion: (a) BART-supervised almost reconstructs the ground truth. (b) Here, the two models increase sufficiency. (c) Sometimes, the generated conclusions remain rather vague and pick the wrong target.",
"cite_spans": [],
"ref_spans": [
{
"start": 294,
"end": 301,
"text": "Table 3",
"ref_id": null
}
],
"eq_spans": [],
"section": "# Label",
"sec_num": null
},
{
"text": "Error Analysis & Examples To better understand the differences between the two BART models and the ground truth, we analyzed the 100 examples from our annotation study manually. Table 3 compares the conclusions for three arguments.",
"cite_spans": [],
"ref_spans": [
{
"start": 178,
"end": 185,
"text": "Table 3",
"ref_id": null
}
],
"eq_spans": [],
"section": "# Label",
"sec_num": null
},
{
"text": "For the sufficient argument in Table 3 (a), BARTsupervised nearly perfectly reconstructs the ground truth, whereas BART-unsupervised generates a reasonable but less specific alternative. In Table 3 (b), the approaches appear to make the argument more sufficient, particularly BART-supervised. Both examples speak for the truth of our hypothesis that the conclusion of sufficient arguments can be generated from the premises. This tendency is further backed by our analysis in which we found that for BART-supervised 20% (10/50) of the sufficient arguments have perfect matching conclusions (Table 3(a)) 60% (30/50) either are less abstract or have a different conclusion target but are of equal quality, and 20% (10/50) have a conclusion that is of lower quality, compared to the ground truth. In contrast, only 10% (5/50) of the insufficient arguments are perfect matches, 60% (30/50) either are less abstract or have a different conclusion target of equal quality, and 30% (15/50) have lower quality generated conclusions (Table 3(b)) .",
"cite_spans": [],
"ref_spans": [
{
"start": 31,
"end": 38,
"text": "Table 3",
"ref_id": null
},
{
"start": 190,
"end": 197,
"text": "Table 3",
"ref_id": null
},
{
"start": 1024,
"end": 1036,
"text": "(Table 3(b))",
"ref_id": null
}
],
"eq_spans": [],
"section": "# Label",
"sec_num": null
},
{
"text": "Further analysis reveals that for 58% (29/50) of the insufficient arguments, BART-supervised generated a sufficient conclusion with a different target (16/29) or a different level of abstraction, either more specific (11/29) or more abstract (2/29) than the ground truth. In Table 3 (c), the BART models seem tricked by the anecdotal evidence given in the premises, mistakenly picking Sri Lanka as the conclusion's target.",
"cite_spans": [],
"ref_spans": [
{
"start": 275,
"end": 282,
"text": "Table 3",
"ref_id": null
}
],
"eq_spans": [],
"section": "# Label",
"sec_num": null
},
{
"text": "Regarding the difference of our models, we found that BART-supervised more often generated conclusions with targets that are equally likely to the target of the human ground truth (68/100 vs. 42/100).",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "# Label",
"sec_num": null
},
{
"text": "Using our conclusion generation model, BARTsupervised, we finally study our hypothesis on sufficiency assessment by experimenting with different input variations and testing on the corpus of Stab and Gurevych (2017b) . The first setting is by just using the RoBERTa model (direct sufficiency as- Table 4 : Results of argument sufficiency assessment: Accuracy as well as macro precision, recall, and F 1 -score of all evaluated approaches, averaged over twenty 5-fold cross-validations. Significant gains over Stab and Gurevych (2017a) and the RoBERTa approach without structural enrichment are marked with \u2020 and \u2021 , respectively (computed using Wilcoxon signed-rank test, p-value .05). The human upper bound is obtained on a subset of 432 arguments. sessment) and the second by introducing structural knowledge and our generated conclusions (indirect sufficiency assessment).",
"cite_spans": [
{
"start": 191,
"end": 216,
"text": "Stab and Gurevych (2017b)",
"ref_id": "BIBREF14"
},
{
"start": 509,
"end": 534,
"text": "Stab and Gurevych (2017a)",
"ref_id": "BIBREF12"
}
],
"ref_spans": [
{
"start": 296,
"end": 303,
"text": "Table 4",
"ref_id": null
}
],
"eq_spans": [],
"section": "Evaluation of Sufficiency Assessment",
"sec_num": "6"
},
{
"text": "First, we compare the \"base version\" of our approach, RoBERTa without additional structure annotations, to the human upper bound and the state of the art CNN of Stab and Gurevych (2017b). 2 Results The upper part of Table 4 shows the direct assessment results. Our RoBERTa model significantly outperforms the CNN both on accuracy (.889 vs. .846) and on macro F 1 score (.876 vs. .831), the latter being an improvement of whole 4.5 points. Our model also performs almost on par with the human upper bound, meaning it is approximately at the level of human performance. This underlines the potential of pre-trained transformer models in argument quality assessment.",
"cite_spans": [
{
"start": 161,
"end": 189,
"text": "Stab and Gurevych (2017b). 2",
"ref_id": null
}
],
"ref_spans": [
{
"start": 216,
"end": 223,
"text": "Table 4",
"ref_id": null
}
],
"eq_spans": [],
"section": "Direct Sufficiency Assessment",
"sec_num": "6.1"
},
{
"text": "As our conclusion generation model starts from structural annotations of a given argument, we systematically study the benefit of knowing the premises and the original conclusion, as well of having the generated conclusion. We consider the following eight input variants for the assessment:",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Indirect Sufficiency Assessment",
"sec_num": "6.2"
},
{
"text": "\u2022 RoBERTa-premises-only. Use the full argument as input, but replace the ground-truth conclusion with an <\\unk> token.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Indirect Sufficiency Assessment",
"sec_num": "6.2"
},
{
"text": "\u2022 RoBERTa-conclusion-only. Use only the ground-truth conclusion as input.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Indirect Sufficiency Assessment",
"sec_num": "6.2"
},
{
"text": "\u2022 RoBERTa-generated-only. Use only the generated conclusion as input.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Indirect Sufficiency Assessment",
"sec_num": "6.2"
},
{
"text": "\u2022 RoBERTa-premises+conclusion. Use the full argument as input and highlight the groundtruth conclusion using <\\s> tokens.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Indirect Sufficiency Assessment",
"sec_num": "6.2"
},
{
"text": "\u2022 RoBERTa-premises+generated. Use the full argument as input, but replace the groundtruth conclusion with its generated counterpart. Highlight the latter using <\\s> tokens.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Indirect Sufficiency Assessment",
"sec_num": "6.2"
},
{
"text": "\u2022 RoBERTa-conclusion+generated. Use only the ground-truth and the generated conclusion as input, separated with a <\\s> token",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Indirect Sufficiency Assessment",
"sec_num": "6.2"
},
{
"text": "\u2022 RoBERTa-all. Use the full argument as input, insert the highlight the generated conclusion after the ground-truth conclusion and highlight both together using <\\s> tokens.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Indirect Sufficiency Assessment",
"sec_num": "6.2"
},
{
"text": "Results The lower part of Table 4 shows that both RoBERTa-premises+conclusion and RoBERTa-all yield the best results overall, significantly outperforming our vanilla RoBERTa model by almost 1 point in terms of both accuracy (.896 vs. .889) and macro F 1 -score (.885 vs. .876) . The results suggest that using a generated conclusion for the argument does not really help, but also not hurts the model performance. This is also supported by RoBERTa-premises-only, which also matches the performance of RoBERTa-premises+generated. In general, however, bringing structural knowledge to the model gives a slight but significant improvement in sufficiency assessment.",
"cite_spans": [
{
"start": 261,
"end": 276,
"text": "(.885 vs. .876)",
"ref_id": null
}
],
"ref_spans": [
{
"start": 26,
"end": 33,
"text": "Table 4",
"ref_id": null
}
],
"eq_spans": [],
"section": "Indirect Sufficiency Assessment",
"sec_num": "6.2"
},
{
"text": "Looking at the weak performance of RoBERTa-conclusion+generated (macro F 1 -score .571), we see that an opposition of the two conclusions alone is not enough four sufficiency assessment. Even though adding the generated one improves over having the ground-truth conclusion only, all variants that include the premises perform much better.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Indirect Sufficiency Assessment",
"sec_num": "6.2"
},
{
"text": "To better understand the role of premises and conclusions in sufficiency assessment, we trained the three RoBERTa-<xy>-only variants. The high performance of RoBERTa-premises-only (macro F 1 -score .875) clearly reveals that the conclusion is of almost no importance on the data of Stab and Gurevych (2017b), being not significantly worse than vanilla RoBERTa. The low scores of RoBERTa-conclusion-only further support this hypothesis, suggesting that the knowledge obtained from the conclusion can be inferred from the argument without its conclusion alone. This result is very insightful in that it displays that the currently available data barely enables a study of sufficiency assessment in terms of its actual definition. Instead, we suppose that models mainly learn a correlation between the nature and the quality of a given set of premises and the possible sufficiency evolving from this. In particular, students who provide \"good\" premises for an argument in their essays can also deliver an inferrable conclusion.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Indirect Sufficiency Assessment",
"sec_num": "6.2"
},
{
"text": "Our results suggest that large-scale pre-trained transformer models can help assess the quality of arguments, here their sufficiency. They even nearly matched human performance. However, an accurate understanding of the argument sufficiency task in terms of the actual definition of sufficiency seems barely possible on the available data.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Discussion",
"sec_num": "7"
},
{
"text": "Employing knowledge about argumentative structure can benefit sufficiency assessment, in line with findings on predicting essay-level argument quality (Wachsmuth et al., 2016) . Our results suggest that there is at least some additional knowledge in an argument's conclusion that our model could not learn itself. However, we did not actually mine argumentative structure here, but we resorted to the human-annotated ground truth, which is usually not available in a real-world setting. Thus, the improvements obtained by the structure could vanish as soon as we resort to computational methods. We note, though, that we obtained state-of-the-art results also using RoBERTa on the plain text only.",
"cite_spans": [
{
"start": 151,
"end": 175,
"text": "(Wachsmuth et al., 2016)",
"ref_id": "BIBREF17"
}
],
"ref_spans": [],
"eq_spans": [],
"section": "Discussion",
"sec_num": "7"
},
{
"text": "Regarding the central hypothesis of this work, we study an example of a more task-aligned approach. However, our results show that answering the question of conclusion inferability by generating conclusions for an argument is difficult, as generated conclusions may be perceived as sufficient, likely, and novel, but may still not be unique. That is, for many premises, it may be possible to generated multiple sufficient conclusions, which naturally limits the impact of quality assessment through generation.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Discussion",
"sec_num": "7"
},
{
"text": "Consequently, although the task of sufficiency assessment appears to be solved on the given data, we argue for the need for more sophisticated corpora that better reflect the actual definition of the task, to ultimately allow studying whether approaches such as ours are needed in real-world applications.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Discussion",
"sec_num": "7"
},
{
"text": "In this work we have studied the task of argument sufficiency assessment based on auto-generated argument conclusions. According to our findings, traditional approaches can be improved by using largescale pre-trained transformer models and by incorporating knowledge about argumentative structure. The effect of our proposed idea to leverage generation for the assessment turned out low though. However, this may likely be caused by the available data, wehere sufficiency seems to barely depend on the arguments' conclusions, thus preventing our and previous approaches from actually tackling the task as intended by its definition.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Conclusion",
"sec_num": "8"
},
{
"text": "In general, the insights of this paper lay the foundation for more task-oriented approaches towards the assessment of argument quality dimensions, that are tailored towards the properties in scope (here the relation between premises and conclusion). To adequately evaluate such approaches, also refined corpora may be needed.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Conclusion",
"sec_num": "8"
},
{
"text": "The experiment code can be found under: https://gi thub.com/webis-de/ArgMining-21",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "",
"sec_num": null
},
{
"text": "Note that the human upper bound ofStab and Gurevych (2017b) was computed on a subset of 433 arguments annotated by three annotators only. Thus, it is only an approximation of the actual human performance. The human scores are based on pairwise comparisons of the three annotators.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "",
"sec_num": null
}
],
"back_matter": [
{
"text": "We thank Katharina Brennig, Simon Seidl, Abdullah Burak, Frederike Gurcke and Dr. Maurice Gurcke for their feedback. We gratefully acknowledge the computing time provided the described experiments by the Paderborn Center for Parallel Computing (PC 2 ). This project has been partially funded by the German Research Foundation (DFG) within the project OASiS, project number 455913891, as part of the Priority Program \"Robust Argumentation Machines (RATIO)\" (SPP-1999).",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Acknowledgments",
"sec_num": null
}
],
"bib_entries": {
"BIBREF0": {
"ref_id": "b0",
"title": "Target inference in argument conclusion generation",
"authors": [
{
"first": "Milad",
"middle": [],
"last": "Alshomary",
"suffix": ""
},
{
"first": "Shahbaz",
"middle": [],
"last": "Syed",
"suffix": ""
},
{
"first": "Martin",
"middle": [],
"last": "Potthast",
"suffix": ""
},
{
"first": "Henning",
"middle": [],
"last": "Wachsmuth",
"suffix": ""
}
],
"year": 2020,
"venue": "Proceedings of the 58th Annual Meeting of the Association for Computational Linguistics",
"volume": "",
"issue": "",
"pages": "4334--4345",
"other_ids": {
"DOI": [
"10.18653/v1/2020.acl-main.399"
]
},
"num": null,
"urls": [],
"raw_text": "Milad Alshomary, Shahbaz Syed, Martin Potthast, and Henning Wachsmuth. 2020. Target inference in argu- ment conclusion generation. In Proceedings of the 58th Annual Meeting of the Association for Compu- tational Linguistics, pages 4334-4345, Online. Asso- ciation for Computational Linguistics.",
"links": null
},
"BIBREF1": {
"ref_id": "b1",
"title": "BERT: Pre-training of deep bidirectional transformers for language understanding",
"authors": [
{
"first": "Jacob",
"middle": [],
"last": "Devlin",
"suffix": ""
},
{
"first": "Ming-Wei",
"middle": [],
"last": "Chang",
"suffix": ""
},
{
"first": "Kenton",
"middle": [],
"last": "Lee",
"suffix": ""
},
{
"first": "Kristina",
"middle": [],
"last": "Toutanova",
"suffix": ""
}
],
"year": 2019,
"venue": "Proceedings of the 2019 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies",
"volume": "1",
"issue": "",
"pages": "4171--4186",
"other_ids": {
"DOI": [
"10.18653/v1/N19-1423"
]
},
"num": null,
"urls": [],
"raw_text": "Jacob Devlin, Ming-Wei Chang, Kenton Lee, and Kristina Toutanova. 2019. BERT: Pre-training of deep bidirectional transformers for language under- standing. In Proceedings of the 2019 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Tech- nologies, Volume 1 (Long and Short Papers), pages 4171-4186, Minneapolis, Minnesota. Association for Computational Linguistics.",
"links": null
},
"BIBREF2": {
"ref_id": "b2",
"title": "Scalable learning of non-decomposable objectives",
"authors": [
{
"first": "Elad",
"middle": [],
"last": "Eban",
"suffix": ""
},
{
"first": "Mariano",
"middle": [],
"last": "Schain",
"suffix": ""
},
{
"first": "Alan",
"middle": [],
"last": "Mackey",
"suffix": ""
},
{
"first": "Ariel",
"middle": [],
"last": "Gordon",
"suffix": ""
},
{
"first": "Ryan",
"middle": [],
"last": "Rifkin",
"suffix": ""
},
{
"first": "Gal",
"middle": [],
"last": "Elidan",
"suffix": ""
}
],
"year": 2017,
"venue": "Artificial intelligence and statistics",
"volume": "",
"issue": "",
"pages": "832--840",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Elad Eban, Mariano Schain, Alan Mackey, Ariel Gor- don, Ryan Rifkin, and Gal Elidan. 2017. Scalable learning of non-decomposable objectives. In Ar- tificial intelligence and statistics, pages 832-840. PMLR.",
"links": null
},
"BIBREF3": {
"ref_id": "b3",
"title": "Analyzing the persuasive effect of style in news editorial argumentation",
"authors": [
{
"first": "Roxanne",
"middle": [
"El"
],
"last": "Baff",
"suffix": ""
},
{
"first": "Henning",
"middle": [],
"last": "Wachsmuth",
"suffix": ""
},
{
"first": "Khalid",
"middle": [
"Al"
],
"last": "Khatib",
"suffix": ""
},
{
"first": "Benno",
"middle": [],
"last": "Stein",
"suffix": ""
}
],
"year": 2020,
"venue": "Proceedings of the 58th Annual Meeting of the Association for Computational Linguistics",
"volume": "",
"issue": "",
"pages": "3154--3160",
"other_ids": {
"DOI": [
"10.18653/v1/2020.acl-main.287"
]
},
"num": null,
"urls": [],
"raw_text": "Roxanne El Baff, Henning Wachsmuth, Khalid Al Khatib, and Benno Stein. 2020. Analyzing the persuasive effect of style in news editorial argumen- tation. In Proceedings of the 58th Annual Meeting of the Association for Computational Linguistics, pages 3154-3160, Online. Association for Computational Linguistics.",
"links": null
},
"BIBREF4": {
"ref_id": "b4",
"title": "Ranit Aharonov, and Noam Slonim. 2020. A large-scale dataset for argument quality ranking: Construction and analysis",
"authors": [
{
"first": "Shai",
"middle": [],
"last": "Gretz",
"suffix": ""
},
{
"first": "Roni",
"middle": [],
"last": "Friedman",
"suffix": ""
},
{
"first": "Edo",
"middle": [],
"last": "Cohen-Karlik",
"suffix": ""
},
{
"first": "Assaf",
"middle": [],
"last": "Toledo",
"suffix": ""
},
{
"first": "Dan",
"middle": [],
"last": "Lahav",
"suffix": ""
}
],
"year": null,
"venue": "Proceedings of the AAAI Conference on Artificial Intelligence",
"volume": "34",
"issue": "",
"pages": "7805--7813",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Shai Gretz, Roni Friedman, Edo Cohen-Karlik, As- saf Toledo, Dan Lahav, Ranit Aharonov, and Noam Slonim. 2020. A large-scale dataset for argument quality ranking: Construction and analysis. In Pro- ceedings of the AAAI Conference on Artificial Intelli- gence, volume 34, pages 7805-7813.",
"links": null
},
"BIBREF5": {
"ref_id": "b5",
"title": "Logical self-defense",
"authors": [
{
"first": "Ralph",
"middle": [],
"last": "Henry",
"suffix": ""
},
{
"first": "Johnson",
"middle": [],
"last": "",
"suffix": ""
},
{
"first": "J Anthony",
"middle": [],
"last": "Blair",
"suffix": ""
}
],
"year": 2006,
"venue": "",
"volume": "",
"issue": "",
"pages": "",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Ralph Henry Johnson and J Anthony Blair. 2006. Logi- cal self-defense. Idea.",
"links": null
},
"BIBREF6": {
"ref_id": "b6",
"title": "Rhetoric, logic, and dialectic: Advancing theory-based argument quality assessment in natural language processing",
"authors": [
{
"first": "Anne",
"middle": [],
"last": "Lauscher",
"suffix": ""
},
{
"first": "Lily",
"middle": [],
"last": "Ng",
"suffix": ""
},
{
"first": "Courtney",
"middle": [],
"last": "Napoles",
"suffix": ""
},
{
"first": "Joel",
"middle": [],
"last": "Tetreault",
"suffix": ""
}
],
"year": 2020,
"venue": "Proceedings of the 28th International Conference on Computational Linguistics",
"volume": "",
"issue": "",
"pages": "4563--4574",
"other_ids": {
"DOI": [
"10.18653/v1/2020.coling-main.402"
]
},
"num": null,
"urls": [],
"raw_text": "Anne Lauscher, Lily Ng, Courtney Napoles, and Joel Tetreault. 2020. Rhetoric, logic, and dialectic: Ad- vancing theory-based argument quality assessment in natural language processing. In Proceedings of the 28th International Conference on Computational Linguistics, pages 4563-4574, Barcelona, Spain (On- line). International Committee on Computational Lin- guistics.",
"links": null
},
"BIBREF7": {
"ref_id": "b7",
"title": "BART: Denoising sequence-to-sequence pre-training for natural language generation, translation, and comprehension",
"authors": [
{
"first": "Mike",
"middle": [],
"last": "Lewis",
"suffix": ""
},
{
"first": "Yinhan",
"middle": [],
"last": "Liu",
"suffix": ""
},
{
"first": "Naman",
"middle": [],
"last": "Goyal",
"suffix": ""
},
{
"first": "Marjan",
"middle": [],
"last": "Ghazvininejad",
"suffix": ""
},
{
"first": "Abdelrahman",
"middle": [],
"last": "Mohamed",
"suffix": ""
},
{
"first": "Omer",
"middle": [],
"last": "Levy",
"suffix": ""
},
{
"first": "Veselin",
"middle": [],
"last": "Stoyanov",
"suffix": ""
},
{
"first": "Luke",
"middle": [],
"last": "Zettlemoyer",
"suffix": ""
}
],
"year": 2020,
"venue": "Proceedings of the 58th Annual Meeting of the Association for Computational Linguistics",
"volume": "",
"issue": "",
"pages": "7871--7880",
"other_ids": {
"DOI": [
"10.18653/v1/2020.acl-main.703"
]
},
"num": null,
"urls": [],
"raw_text": "Mike Lewis, Yinhan Liu, Naman Goyal, Marjan Ghazvininejad, Abdelrahman Mohamed, Omer Levy, Veselin Stoyanov, and Luke Zettlemoyer. 2020. BART: Denoising sequence-to-sequence pre-training for natural language generation, translation, and com- prehension. In Proceedings of the 58th Annual Meet- ing of the Association for Computational Linguistics, pages 7871-7880, Online. Association for Computa- tional Linguistics.",
"links": null
},
"BIBREF8": {
"ref_id": "b8",
"title": "Roberta: A robustly optimized bert pretraining approach",
"authors": [
{
"first": "Yinhan",
"middle": [],
"last": "Liu",
"suffix": ""
},
{
"first": "Myle",
"middle": [],
"last": "Ott",
"suffix": ""
},
{
"first": "Naman",
"middle": [],
"last": "Goyal",
"suffix": ""
},
{
"first": "Jingfei",
"middle": [],
"last": "Du",
"suffix": ""
},
{
"first": "Mandar",
"middle": [],
"last": "Joshi",
"suffix": ""
},
{
"first": "Danqi",
"middle": [],
"last": "Chen",
"suffix": ""
},
{
"first": "Omer",
"middle": [],
"last": "Levy",
"suffix": ""
},
{
"first": "M",
"middle": [],
"last": "Lewis",
"suffix": ""
},
{
"first": "Luke",
"middle": [],
"last": "Zettlemoyer",
"suffix": ""
},
{
"first": "Veselin",
"middle": [],
"last": "Stoyanov",
"suffix": ""
}
],
"year": 2019,
"venue": "",
"volume": "",
"issue": "",
"pages": "",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Yinhan Liu, Myle Ott, Naman Goyal, Jingfei Du, Man- dar Joshi, Danqi Chen, Omer Levy, M. Lewis, Luke Zettlemoyer, and Veselin Stoyanov. 2019. Roberta: A robustly optimized bert pretraining approach. ArXiv, abs/1907.11692.",
"links": null
},
"BIBREF9": {
"ref_id": "b9",
"title": "Optimizing f-measures by cost-sensitive classification",
"authors": [
{
"first": "Nicolas",
"middle": [],
"last": "Shameem Puthiya Parambath",
"suffix": ""
},
{
"first": "Yves",
"middle": [],
"last": "Usunier",
"suffix": ""
},
{
"first": "",
"middle": [],
"last": "Grandvalet",
"suffix": ""
}
],
"year": 2014,
"venue": "Advances in Neural Information Processing Systems",
"volume": "27",
"issue": "",
"pages": "2123--2131",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Shameem Puthiya Parambath, Nicolas Usunier, and Yves Grandvalet. 2014. Optimizing f-measures by cost-sensitive classification. Advances in Neural In- formation Processing Systems, 27:2123-2131.",
"links": null
},
"BIBREF10": {
"ref_id": "b10",
"title": "Learning from revisions: Quality assessment of claims in argumentation at scale",
"authors": [
{
"first": "Gabriella",
"middle": [],
"last": "Skitalinskaya",
"suffix": ""
},
{
"first": "Jonas",
"middle": [],
"last": "Klaff",
"suffix": ""
},
{
"first": "Henning",
"middle": [],
"last": "Wachsmuth",
"suffix": ""
}
],
"year": 2021,
"venue": "Proceedings of the 16th Conference of the European Chapter of the Association for Computational Linguistics: Main Volume",
"volume": "",
"issue": "",
"pages": "1718--1729",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Gabriella Skitalinskaya, Jonas Klaff, and Henning Wachsmuth. 2021. Learning from revisions: Quality assessment of claims in argumentation at scale. In Proceedings of the 16th Conference of the European Chapter of the Association for Computational Lin- guistics: Main Volume, pages 1718-1729, Online. Association for Computational Linguistics.",
"links": null
},
"BIBREF11": {
"ref_id": "b11",
"title": "An autonomous debating system",
"authors": [
{
"first": "Noam",
"middle": [],
"last": "Slonim",
"suffix": ""
},
{
"first": "Yonatan",
"middle": [],
"last": "Bilu",
"suffix": ""
},
{
"first": "Carlos",
"middle": [],
"last": "Alzate",
"suffix": ""
},
{
"first": "Roy",
"middle": [],
"last": "Bar-Haim",
"suffix": ""
},
{
"first": "Ben",
"middle": [],
"last": "Bogin",
"suffix": ""
},
{
"first": "Francesca",
"middle": [],
"last": "Bonin",
"suffix": ""
},
{
"first": "Leshem",
"middle": [],
"last": "Choshen",
"suffix": ""
},
{
"first": "Edo",
"middle": [],
"last": "Cohen-Karlik",
"suffix": ""
},
{
"first": "Lena",
"middle": [],
"last": "Dankin",
"suffix": ""
},
{
"first": "Lilach",
"middle": [],
"last": "Edelstein",
"suffix": ""
}
],
"year": 2021,
"venue": "Nature",
"volume": "591",
"issue": "7850",
"pages": "379--384",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Noam Slonim, Yonatan Bilu, Carlos Alzate, Roy Bar-Haim, Ben Bogin, Francesca Bonin, Leshem Choshen, Edo Cohen-Karlik, Lena Dankin, Lilach Edelstein, et al. 2021. An autonomous debating sys- tem. Nature, 591(7850):379-384.",
"links": null
},
"BIBREF12": {
"ref_id": "b12",
"title": "Parsing argumentation structures in persuasive essays",
"authors": [
{
"first": "Christian",
"middle": [],
"last": "Stab",
"suffix": ""
},
{
"first": "Iryna",
"middle": [],
"last": "Gurevych",
"suffix": ""
}
],
"year": 2017,
"venue": "Computational Linguistics",
"volume": "43",
"issue": "3",
"pages": "619--659",
"other_ids": {
"DOI": [
"10.1162/COLI_a_00295"
]
},
"num": null,
"urls": [],
"raw_text": "Christian Stab and Iryna Gurevych. 2017a. Parsing argumentation structures in persuasive essays. Com- putational Linguistics, 43(3):619-659.",
"links": null
},
"BIBREF13": {
"ref_id": "b13",
"title": "Argumentative writing support by means of natural language processing",
"authors": [
{
"first": "Christian Matthias Edwin",
"middle": [],
"last": "Stab",
"suffix": ""
}
],
"year": 2017,
"venue": "",
"volume": "",
"issue": "",
"pages": "",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Christian Matthias Edwin Stab. 2017. Argumentative writing support by means of natural language pro- cessing. Ph.D. thesis, Technische Universit\u00e4t Darm- stadt.",
"links": null
},
"BIBREF14": {
"ref_id": "b14",
"title": "Recognizing insufficiently supported arguments in argumentative essays",
"authors": [
{
"first": "Christian",
"middle": [
"Matthias"
],
"last": "",
"suffix": ""
},
{
"first": "Edwin",
"middle": [],
"last": "Stab",
"suffix": ""
},
{
"first": "Iryna",
"middle": [],
"last": "Gurevych",
"suffix": ""
}
],
"year": 2017,
"venue": "Proceedings of the 15th Conference of the European Chapter",
"volume": "1",
"issue": "",
"pages": "980--990",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Christian Matthias Edwin Stab and Iryna Gurevych. 2017b. Recognizing insufficiently supported argu- ments in argumentative essays. In Proceedings of the 15th Conference of the European Chapter of the Association for Computational Linguistics: Volume 1, Long Papers, pages 980-990.",
"links": null
},
"BIBREF15": {
"ref_id": "b15",
"title": "Generating informative conclusions for argumentative texts",
"authors": [
{
"first": "Shahbaz",
"middle": [],
"last": "Syed",
"suffix": ""
},
{
"first": "Khalid",
"middle": [
"Al"
],
"last": "Khatib",
"suffix": ""
},
{
"first": "Milad",
"middle": [],
"last": "Alshomary",
"suffix": ""
},
{
"first": "Henning",
"middle": [],
"last": "Wachsmuth",
"suffix": ""
},
{
"first": "Martin",
"middle": [],
"last": "Potthast",
"suffix": ""
}
],
"year": 2021,
"venue": "Findings of the Association for Computational Linguistics: ACL-IJCNLP 2021",
"volume": "",
"issue": "",
"pages": "3482--3493",
"other_ids": {
"DOI": [
"10.18653/v1/2021.findings-acl.306"
]
},
"num": null,
"urls": [],
"raw_text": "Shahbaz Syed, Khalid Al Khatib, Milad Alshomary, Henning Wachsmuth, and Martin Potthast. 2021. Generating informative conclusions for argumenta- tive texts. In Findings of the Association for Com- putational Linguistics: ACL-IJCNLP 2021, pages 3482-3493, Online. Association for Computational Linguistics.",
"links": null
},
"BIBREF16": {
"ref_id": "b16",
"title": "Automatic argument quality assessment -new datasets and methods",
"authors": [
{
"first": "Assaf",
"middle": [],
"last": "Toledo",
"suffix": ""
},
{
"first": "Shai",
"middle": [],
"last": "Gretz",
"suffix": ""
},
{
"first": "Edo",
"middle": [],
"last": "Cohen-Karlik",
"suffix": ""
},
{
"first": "Roni",
"middle": [],
"last": "Friedman",
"suffix": ""
},
{
"first": "Elad",
"middle": [],
"last": "Venezian",
"suffix": ""
},
{
"first": "Dan",
"middle": [],
"last": "Lahav",
"suffix": ""
},
{
"first": "Michal",
"middle": [],
"last": "Jacovi",
"suffix": ""
},
{
"first": "Ranit",
"middle": [],
"last": "Aharonov",
"suffix": ""
},
{
"first": "Noam",
"middle": [],
"last": "Slonim",
"suffix": ""
}
],
"year": 2019,
"venue": "Proceedings of the 2019 Conference on Empirical Methods in Natural Language Processing and the 9th International Joint Conference on Natural Language Processing (EMNLP-IJCNLP)",
"volume": "",
"issue": "",
"pages": "5625--5635",
"other_ids": {
"DOI": [
"10.18653/v1/D19-1564"
]
},
"num": null,
"urls": [],
"raw_text": "Assaf Toledo, Shai Gretz, Edo Cohen-Karlik, Roni Friedman, Elad Venezian, Dan Lahav, Michal Jacovi, Ranit Aharonov, and Noam Slonim. 2019. Auto- matic argument quality assessment -new datasets and methods. In Proceedings of the 2019 Confer- ence on Empirical Methods in Natural Language Pro- cessing and the 9th International Joint Conference on Natural Language Processing (EMNLP-IJCNLP), pages 5625-5635, Hong Kong, China. Association for Computational Linguistics.",
"links": null
},
"BIBREF17": {
"ref_id": "b17",
"title": "Using argument mining to assess the argumentation quality of essays",
"authors": [
{
"first": "Henning",
"middle": [],
"last": "Wachsmuth",
"suffix": ""
},
{
"first": "Al",
"middle": [],
"last": "Khalid",
"suffix": ""
},
{
"first": "Benno",
"middle": [],
"last": "Khatib",
"suffix": ""
},
{
"first": "",
"middle": [],
"last": "Stein",
"suffix": ""
}
],
"year": 2016,
"venue": "Proceedings of COLING 2016, the 26th International Conference on Computational Linguistics: Technical Papers",
"volume": "",
"issue": "",
"pages": "1680--1691",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Henning Wachsmuth, Khalid Al Khatib, and Benno Stein. 2016. Using argument mining to assess the argumentation quality of essays. In Proceedings of COLING 2016, the 26th International Conference on Computational Linguistics: Technical Papers, pages 1680-1691.",
"links": null
},
"BIBREF18": {
"ref_id": "b18",
"title": "Computational argumentation quality assessment in natural language",
"authors": [
{
"first": "Henning",
"middle": [],
"last": "Wachsmuth",
"suffix": ""
},
{
"first": "Nona",
"middle": [],
"last": "Naderi",
"suffix": ""
},
{
"first": "Yufang",
"middle": [],
"last": "Hou",
"suffix": ""
},
{
"first": "Yonatan",
"middle": [],
"last": "Bilu",
"suffix": ""
},
{
"first": "Vinodkumar",
"middle": [],
"last": "Prabhakaran",
"suffix": ""
},
{
"first": "Tim",
"middle": [
"Alberdingk"
],
"last": "Thijm",
"suffix": ""
},
{
"first": "Graeme",
"middle": [],
"last": "Hirst",
"suffix": ""
},
{
"first": "Benno",
"middle": [],
"last": "Stein",
"suffix": ""
}
],
"year": 2017,
"venue": "Proceedings of the 15th Conference of the European Chapter",
"volume": "1",
"issue": "",
"pages": "176--187",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Henning Wachsmuth, Nona Naderi, Yufang Hou, Yonatan Bilu, Vinodkumar Prabhakaran, Tim Alberd- ingk Thijm, Graeme Hirst, and Benno Stein. 2017a. Computational argumentation quality assessment in natural language. In Proceedings of the 15th Confer- ence of the European Chapter of the Association for Computational Linguistics: Volume 1, Long Papers, pages 176-187.",
"links": null
},
"BIBREF19": {
"ref_id": "b19",
"title": "Building an argument search engine for the web",
"authors": [
{
"first": "Henning",
"middle": [],
"last": "Wachsmuth",
"suffix": ""
},
{
"first": "Martin",
"middle": [],
"last": "Potthast",
"suffix": ""
},
{
"first": "Khalid",
"middle": [
"Al"
],
"last": "Khatib",
"suffix": ""
},
{
"first": "Yamen",
"middle": [],
"last": "Ajjour",
"suffix": ""
},
{
"first": "Jana",
"middle": [],
"last": "Puschmann",
"suffix": ""
},
{
"first": "Jiani",
"middle": [],
"last": "Qu",
"suffix": ""
},
{
"first": "Jonas",
"middle": [],
"last": "Dorsch",
"suffix": ""
},
{
"first": "Viorel",
"middle": [],
"last": "Morari",
"suffix": ""
},
{
"first": "Janek",
"middle": [],
"last": "Bevendorff",
"suffix": ""
},
{
"first": "Benno",
"middle": [],
"last": "Stein",
"suffix": ""
}
],
"year": 2017,
"venue": "Proceedings of the 4th Workshop on Argument Mining",
"volume": "",
"issue": "",
"pages": "49--59",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Henning Wachsmuth, Martin Potthast, Khalid Al Khatib, Yamen Ajjour, Jana Puschmann, Jiani Qu, Jonas Dorsch, Viorel Morari, Janek Bevendorff, and Benno Stein. 2017b. Building an argument search engine for the web. In Proceedings of the 4th Workshop on Argument Mining, pages 49-59.",
"links": null
},
"BIBREF20": {
"ref_id": "b20",
"title": "Intrinsic quality assessment of arguments",
"authors": [
{
"first": "Henning",
"middle": [],
"last": "Wachsmuth",
"suffix": ""
},
{
"first": "Till",
"middle": [],
"last": "Werner",
"suffix": ""
}
],
"year": 2020,
"venue": "Proceedings of the 28th International Conference on Computational Linguistics",
"volume": "",
"issue": "",
"pages": "6739--6745",
"other_ids": {
"DOI": [
"10.18653/v1/2020.coling-main.592"
]
},
"num": null,
"urls": [],
"raw_text": "Henning Wachsmuth and Till Werner. 2020. Intrinsic quality assessment of arguments. In Proceedings of the 28th International Conference on Computational Linguistics, pages 6739-6745, Barcelona, Spain (On- line). International Committee on Computational Lin- guistics.",
"links": null
},
"BIBREF21": {
"ref_id": "b21",
"title": "Bertscore: Evaluating text generation with bert",
"authors": [
{
"first": "Tianyi",
"middle": [],
"last": "Zhang",
"suffix": ""
},
{
"first": "*",
"middle": [],
"last": "",
"suffix": ""
},
{
"first": "Varsha",
"middle": [],
"last": "Kishore",
"suffix": ""
},
{
"first": "*",
"middle": [],
"last": "",
"suffix": ""
},
{
"first": "Felix",
"middle": [],
"last": "Wu",
"suffix": ""
},
{
"first": "*",
"middle": [],
"last": "",
"suffix": ""
},
{
"first": "Kilian",
"middle": [
"Q"
],
"last": "Weinberger",
"suffix": ""
},
{
"first": "Yoav",
"middle": [],
"last": "Artzi",
"suffix": ""
}
],
"year": 2020,
"venue": "International Conference on Learning Representations",
"volume": "",
"issue": "",
"pages": "",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Tianyi Zhang*, Varsha Kishore*, Felix Wu*, Kilian Q. Weinberger, and Yoav Artzi. 2020. Bertscore: Eval- uating text generation with bert. In International Conference on Learning Representations.",
"links": null
}
},
"ref_entries": {
"FIGREF1": {
"type_str": "figure",
"uris": null,
"text": "Illustration of our sufficiency assessment approach through generation: (1) BART is used to generate the masked conclusion in an argument.(2)The generated conclusion is combined with the ground truth annotation of the argument. (3) RoBERTa classifies the enriched argument as sufficient/insufficient. Several ablations of the annotations are tested in our experiments.",
"num": null
}
}
}
} |