File size: 119,052 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 2217 2218 2219 2220 2221 2222 2223 2224 2225 2226 2227 2228 2229 2230 2231 2232 2233 2234 2235 2236 2237 2238 2239 2240 2241 2242 2243 2244 2245 2246 2247 2248 2249 2250 2251 2252 2253 2254 2255 2256 2257 2258 2259 2260 2261 2262 2263 2264 2265 2266 2267 2268 2269 2270 2271 2272 2273 2274 2275 2276 2277 2278 2279 2280 2281 2282 2283 2284 2285 2286 2287 2288 2289 2290 2291 |
{
"paper_id": "2021",
"header": {
"generated_with": "S2ORC 1.0.0",
"date_generated": "2023-01-19T02:09:54.512602Z"
},
"title": "Bayesian Argumentation-Scheme Networks: A Probabilistic Model of Argument Validity Facilitated by Argumentation Schemes",
"authors": [
{
"first": "Takahiro",
"middle": [],
"last": "Kondo",
"suffix": "",
"affiliation": {
"laboratory": "",
"institution": "The University of Tokyo",
"location": {}
},
"email": "takahiro_kondo@is.s.u-tokyo.ac.jp"
},
{
"first": "Koki",
"middle": [],
"last": "Washio",
"suffix": "",
"affiliation": {
"laboratory": "",
"institution": "The University of Tokyo",
"location": {}
},
"email": "kwashio@is.s.u-tokyo.ac.jp"
},
{
"first": "Katsuhiko",
"middle": [],
"last": "Hayashi",
"suffix": "",
"affiliation": {
"laboratory": "",
"institution": "Gunma University",
"location": {}
},
"email": "khayashi0201@gmail.com"
},
{
"first": "Yusuke",
"middle": [],
"last": "Miyao",
"suffix": "",
"affiliation": {
"laboratory": "",
"institution": "The University of Tokyo",
"location": {}
},
"email": "yusuke@is.s.u-tokyo.ac.jp"
}
],
"year": "",
"venue": null,
"identifiers": {},
"abstract": "We propose a methodology for representing the reasoning structure of arguments using Bayesian networks and predicate logic facilitated by argumentation schemes. We express the meaning of text segments using predicate logic and map the boolean values of predicate logic expressions to nodes in a Bayesian network. The reasoning structure among text segments is described with a directed acyclic graph. While our formalism is highly expressive and capable of describing the informal logic of human arguments, it is too open-ended to actually build a network for an argument. It is not at all obvious which segment of argumentative text should be considered as a node in a Bayesian network, and how to decide the dependencies among nodes. To alleviate the difficulty, we provide abstract network fragments, called idioms, which represent typical argument justification patterns derived from argumentation schemes. The network construction process is decomposed into idiom selection, idiom instantiation, and idiom combination. We define 17 idioms in total by referring to argumentation schemes as well as analyzing actual arguments and fitting idioms to them. We also create a dataset consisting of pairs of an argumentative text and a corresponding Bayesian network. Our dataset contains about 2,400 pairs, which is large in the research area of argumentation schemes.",
"pdf_parse": {
"paper_id": "2021",
"_pdf_hash": "",
"abstract": [
{
"text": "We propose a methodology for representing the reasoning structure of arguments using Bayesian networks and predicate logic facilitated by argumentation schemes. We express the meaning of text segments using predicate logic and map the boolean values of predicate logic expressions to nodes in a Bayesian network. The reasoning structure among text segments is described with a directed acyclic graph. While our formalism is highly expressive and capable of describing the informal logic of human arguments, it is too open-ended to actually build a network for an argument. It is not at all obvious which segment of argumentative text should be considered as a node in a Bayesian network, and how to decide the dependencies among nodes. To alleviate the difficulty, we provide abstract network fragments, called idioms, which represent typical argument justification patterns derived from argumentation schemes. The network construction process is decomposed into idiom selection, idiom instantiation, and idiom combination. We define 17 idioms in total by referring to argumentation schemes as well as analyzing actual arguments and fitting idioms to them. We also create a dataset consisting of pairs of an argumentative text and a corresponding Bayesian network. Our dataset contains about 2,400 pairs, which is large in the research area of argumentation schemes.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Abstract",
"sec_num": null
}
],
"body_text": [
{
"text": "We propose to formalize the reasoning structure of an argument towards quantitative validity assessment. To assess the validity, we need to understand what kind of reasoning is applied. The following example is a discussion on whether the death penalty deters crime.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Introduction",
"sec_num": "1"
},
{
"text": "We should introduce the death penalty to deter crime. Q: Criminologists do not believe that it deters crime. P justifies the introduction of the death penalty as a means to deter crime. Q's objection is justified based on expert opinion. By understanding how the argument is justified, we can identify more specific issues that need to be validated. For example, we can ask whether the goal of deterring crime is worth achieving, or whether the criminologist's area of expertise is really relevant to the assertion. We use Bayesian networks and predicate logic to represent the reasoning structure of an argument as shown in Fig. 1 . We express the meaning of text segments using predicate logic. For example, 'the death penalty is a means to deter crime' is expressed as means_to(the death penalty, dettering crime). Then, we represent the dependencies between text segments using a Bayesian network. For example, whether or not a crime can be deterred depends on whether or not the death penalty has a deterrent effect on crime and whether or not the death penalty is actually introduced. This formalism is expressive and flexible to represent a variety of human justifications as it just models conditional dependencies. However, due to its flexibility, it is too open-ended to build a network for an argument. Why do we use the predicate means_to for a node in the example network? How can we decide the dependencies among nodes?",
"cite_spans": [],
"ref_spans": [
{
"start": 625,
"end": 631,
"text": "Fig. 1",
"ref_id": "FIGREF0"
}
],
"eq_spans": [],
"section": "P:",
"sec_num": null
},
{
"text": "To facilitate a network construction, we prepare abstract network fragments, called idioms, which represent typical argument justification patterns derived from argumentation schemes (Walton et al., 2008) . For example, one of the argumentation schemes, practical reasoning, is defined as follows.",
"cite_spans": [
{
"start": 183,
"end": 204,
"text": "(Walton et al., 2008)",
"ref_id": null
}
],
"ref_spans": [],
"eq_spans": [],
"section": "P:",
"sec_num": null
},
{
"text": "Major Premise: I have a goal G. Minor Premise: Carrying out this action A is a means to realize G. Conclusion: Therefore, I ought (practically speaking) to carry out this action A.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "P:",
"sec_num": null
},
{
"text": "We define an idiom for the justification of means to an end referring to practical reasoning. The dotted line in Fig. 1 uses this idiom. By preparing idioms, the network construction process is decomposed into idiom selection, idiom instantiation, and idiom combination. We define 17 idioms in total by referring to argumentation schemes as well as analyzing actual arguments and fitting idioms to them.",
"cite_spans": [],
"ref_spans": [
{
"start": 113,
"end": 119,
"text": "Fig. 1",
"ref_id": "FIGREF0"
}
],
"eq_spans": [],
"section": "P:",
"sec_num": null
},
{
"text": "We also create a dataset consisting of pairs of an argumentative text and a corresponding Bayesian network. The problem is that annotating argumentation schemes in argumentative texts is too expensive. In addition to identifying justification patterns, we need to represent the reasoning structure with Bayesian networks, which is almost impossible. We solve the problem by adopting a procedure to generate a text from a built network inspired by OVERNIGHT, the data building method in semantic parsing (Wang et al., 2015) . As a result, we can make about 2,400 argument-network pairs, which is large in the research area of argumentation schemes. Our dataset is publicly available at https://github.com/mynlp/basn.",
"cite_spans": [
{
"start": 503,
"end": 522,
"text": "(Wang et al., 2015)",
"ref_id": "BIBREF27"
}
],
"ref_spans": [],
"eq_spans": [],
"section": "P:",
"sec_num": null
},
{
"text": "In summary, our contribution is twofold.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "P:",
"sec_num": null
},
{
"text": "\u2022 We propose a methodology for representing the reasoning structure of arguments using Bayesian networks and predicate logic facilitated by argumentation schemes. (Sec. 3)",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "P:",
"sec_num": null
},
{
"text": "\u2022 We create a dataset consisting of pairs of an argumentative text and a corresponding Bayesian network while reducing an annotation effort. (Sec. 4)",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "P:",
"sec_num": null
},
{
"text": "2 Related Work",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "P:",
"sec_num": null
},
{
"text": "Wachsmuth et al. (2017) divided argumentation quality into three main dimensions, Cogency, Effectiveness, and Reasonableness, through comprehensive survey. Our method mainly focuses on cogency which relates to the acceptability of premises and the relevance of a premise to a conclusion. Reasonableness is about acceptability in the dialectical context. Our method can serve in the dialectical context because we can model the validity of an argument and its counterargument together in the same Bayesian network. On the other hand, we do not focus on effectiveness which relates to an emotional appeal and the style of an argument. The previous models for quantitatively assessing the quality of an argument, no matter which dimension of quality is being addressed, predict a score with either feature-based machine learning (Persing et al., 2010; Ng, 2013, 2015; Ghosh et al., 2016; Wachsmuth et al., 2016; Ke et al., 2019; Wachsmuth and Werner, 2020) or neural networks (Ke et al., 2018; Lauscher et al., 2020) . In contrast, our method uses a probabilistic model to calculate the validity.",
"cite_spans": [
{
"start": 826,
"end": 848,
"text": "(Persing et al., 2010;",
"ref_id": "BIBREF11"
},
{
"start": 849,
"end": 864,
"text": "Ng, 2013, 2015;",
"ref_id": null
},
{
"start": 865,
"end": 884,
"text": "Ghosh et al., 2016;",
"ref_id": "BIBREF4"
},
{
"start": 885,
"end": 908,
"text": "Wachsmuth et al., 2016;",
"ref_id": "BIBREF22"
},
{
"start": 909,
"end": 925,
"text": "Ke et al., 2019;",
"ref_id": "BIBREF6"
},
{
"start": 926,
"end": 953,
"text": "Wachsmuth and Werner, 2020)",
"ref_id": "BIBREF25"
},
{
"start": 973,
"end": 990,
"text": "(Ke et al., 2018;",
"ref_id": "BIBREF5"
},
{
"start": 991,
"end": 1013,
"text": "Lauscher et al., 2020)",
"ref_id": "BIBREF7"
}
],
"ref_spans": [],
"eq_spans": [],
"section": "Argument Validity Assessment",
"sec_num": "2.1"
},
{
"text": "In argumentation theory, the probabilistic model for argument assessment is not mature. Baroni et al. (2018) provide a comprehensive survey of formal argumentation. In chapter 2 of the book, Prakken states 'systematic studies of the combination of argumentation with probability were sparse.' Our method is based on Bayesian networks. The use of Bayesian networks in argumentation has been studied through practice rather than theory, as in the case of the legal field discussed next.",
"cite_spans": [
{
"start": 88,
"end": 108,
"text": "Baroni et al. (2018)",
"ref_id": "BIBREF0"
}
],
"ref_spans": [],
"eq_spans": [],
"section": "Argument Validity Assessment",
"sec_num": "2.1"
},
{
"text": "Argumentation schemes describe typical justification patterns in an argument. Walton et al. (2008) lists 60 representative argumentation schemes from various literature, and many studies refer to their list. In the list, argumentation schemes are defined by premises and a conclusion as shown in Sec. 1 and express the claim that a conclusion is plausible if all premises are true. In the legal domain, argumentation schemes help to construct a Bayesian network though network construction is manual. In the analysis of court cases, it has been attempted to calculate the validity of hypotheses based on evidence or testimony using Bayesian networks (Fenton et al., 2013; Vlek et al., 2014) . Since the arbitrariness of the network construction procedure is an issue, Timmer (2017) transforms an argumentation scheme that can be used in court cases into a fragment of a Bayesian network which provides a typical network structure. While the research shows that Bayesian networks and argumentation schemes are compatible, argumentation schemes are used for assisting manual network construction by humans. It does not aim to automatically construct a network from the argumentative text. We combine Bayesian networks with predicate logic towards automatic network construction using natural language processing. In addition, we also prepare fragments for the argumentation schemes that are not covered in the legal domain, such as practical reasoning.",
"cite_spans": [
{
"start": 78,
"end": 98,
"text": "Walton et al. (2008)",
"ref_id": null
},
{
"start": 650,
"end": 671,
"text": "(Fenton et al., 2013;",
"ref_id": "BIBREF3"
},
{
"start": 672,
"end": 690,
"text": "Vlek et al., 2014)",
"ref_id": "BIBREF21"
}
],
"ref_spans": [],
"eq_spans": [],
"section": "Bayesian Networks with Argumentation Schemes",
"sec_num": "2.2"
},
{
"text": "A few datasets annotate argument structures in argument texts for the purpose of argument evaluation. Stab and Gurevych (2014) annotate argument components (major claim, claim, and premises) and relations (support and attack) in persuasive essays. Ghosh et al. (2016) apply the same annotation as Stab and Gurevych (2014) to TOEFL essays associated with holistic scores (high/medium/low). Our dataset consists of pairs of a short argumentative text and its corresponding Bayesian network for modeling the validity of arguments.",
"cite_spans": [
{
"start": 102,
"end": 126,
"text": "Stab and Gurevych (2014)",
"ref_id": "BIBREF16"
},
{
"start": 248,
"end": 267,
"text": "Ghosh et al. (2016)",
"ref_id": "BIBREF4"
},
{
"start": 297,
"end": 321,
"text": "Stab and Gurevych (2014)",
"ref_id": "BIBREF16"
}
],
"ref_spans": [],
"eq_spans": [],
"section": "Datasets",
"sec_num": "2.3"
},
{
"text": "The major challenge in dataset creation is the difficulty of annotating argumentation schemes. Among existing datasets, the AraucariaDB corpus is the largest as far as we know and includes approximately 660 manually annotated arguments (Feng and Hirst, 2011). Lawrence et al. (2019) state that annotated corpora of argumentation schemes are scarce, small, and unrepresentative. They provide an annotation tool to address the issue. We adopt a different approach inspired by Wang et al. (2015) , in which they create training data for a semantic parser by generating canonical utterances from logical forms and paraphrasing them using crowdsourcing to reduce annotation effort. We employ a procedure to generate an argument text from a pre-built Bayesian network.",
"cite_spans": [
{
"start": 260,
"end": 282,
"text": "Lawrence et al. (2019)",
"ref_id": "BIBREF8"
},
{
"start": 474,
"end": 492,
"text": "Wang et al. (2015)",
"ref_id": "BIBREF27"
}
],
"ref_spans": [],
"eq_spans": [],
"section": "Datasets",
"sec_num": "2.3"
},
{
"text": "Our goal is to represent the reasoning structure of arguments with Bayesian networks towards an argument validity assessment. Fig. 2 shows the overall process of network construction. First, we prepare abstract Bayesian network fragments which represent typical argument justification patterns derived from argumentation schemes. We call the abstract fragments network idioms. Then, through the analysis of an argument, appropriate idioms are selected and embodied by assigning logical variables. We call the embodied fragments network instances. Finally, by combining the network instances, we can construct a Bayesian network for the entire argument.",
"cite_spans": [],
"ref_spans": [
{
"start": 126,
"end": 132,
"text": "Fig. 2",
"ref_id": "FIGREF1"
}
],
"eq_spans": [],
"section": "Bayesian Argumentation-Scheme Networks",
"sec_num": "3"
},
{
"text": "In the following, we first describe a formalism of an argument network. Second, we explain how network idioms are defined and introduce three representative idioms. Third, we provide network combination rules to construct an entire argument network. Finally, we explain the advantages of our methodology.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Bayesian Argumentation-Scheme Networks",
"sec_num": "3"
},
{
"text": "We formalize the reasoning structure of arguments with Bayesian networks and typed predicate logic. We represent a network node using a boolean random variable of a predicate logical expression. Logical variables (i.e. predicate arguments) are typed. A type is formally equivalent to using a specific (pre-defined) predicate to impose constraints on a logical variable. Edges represent conditional dependencies among boolean random variables.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Argument Reasoning Structure with Bayesian Networks and Prediate Logic",
"sec_num": "3.1"
},
{
"text": "Idioms are abstract Bayesian network fragments that represent typical argument justification patterns. Argumentation schemes help us to identify justification patterns. We can determine the network structure of an idiom according to the characteristic of justification. For example, practical reasoning deals with the causal relationship between means and ends. We connect an edge from means to ends along the direction of causality. In addition, argumentation schemes can also be used as a reference to decide predicates of idioms. For example, one of the premises in practical reasoning, 'Carrying out this action A is a means to realize G', can be expressed as means_to(G, A) in predicate logic. Thus, we regard means_to as a candidate of predicates.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Network Idioms",
"sec_num": "3.2"
},
{
"text": "In defining idioms, we refer to Walton et al. (2008) Regarding the design of idioms, unlike annotation schemes of argument mining datasets Gurevych, 2017, 2014; Peldszus and Stede, 2015) , the direction of an edge is not necessarily from a premise to a conclusion. The reason is that an idiom is designed to express the reasoning structure behind the justification of an argument. Although the justifications described by argumentation schemes often involve logical leaps, we can sometimes find a deductive relationship among premises and a conclusion. For example, practical reasoning concludes 'I ought to carry out this action A' from the two premises, 'I have a goal G' and 'carrying out this action A is a means to realize G.' That justification is fallacious from a formal logic perspective. Instead, if the two propositions, 'I carry out this action A' and 'carrying out this action A is a means to realize G', are true, then 'G is realized' is deductively true. By expressing such a relationship among premises and a conclusion in argumentation schemes, we model what reasoning is behind them.",
"cite_spans": [
{
"start": 32,
"end": 52,
"text": "Walton et al. (2008)",
"ref_id": null
},
{
"start": 139,
"end": 160,
"text": "Gurevych, 2017, 2014;",
"ref_id": null
},
{
"start": 161,
"end": 186,
"text": "Peldszus and Stede, 2015)",
"ref_id": "BIBREF10"
}
],
"ref_spans": [],
"eq_spans": [],
"section": "Network Idioms",
"sec_num": "3.2"
},
{
"text": "In this section, we introduce three representative idioms. All idioms and the definition of variable types are shown in Appendix A, Appendix B respectively. In the following explanations, the types of logical variables are omitted. In the figures shown below, a node symbol p corresponds to a premise and a symbol c to a conclusion.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Network Idioms",
"sec_num": "3.2"
},
{
"text": "This idiom expresses the justification of means to an end. Fig. 3 shows the network structure. A conclusion node is c: bring_about(X). Two premise nodes are p 1 : brign_about(Y ) and p 2 : means_to(X, Y ). The edges are from c to p 1 and from p 2 to p 1 . It expresses the relationship that X causes Y . The description in the argumentation-schemelike format is as follows.",
"cite_spans": [],
"ref_spans": [
{
"start": 59,
"end": 65,
"text": "Fig. 3",
"ref_id": null
}
],
"eq_spans": [],
"section": "Idiom 1: Means for Goal",
"sec_num": null
},
{
"text": "Premise1: Bringing about Y is desirable. Premise2: Bringing about X is a means to realize Y . Conclusion: We should bring about X.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Idiom 1: Means for Goal",
"sec_num": null
},
{
"text": "In the list of Walton et al. (2008) , practical reasoning corresponds to this idiom. Moreover, argument from need for help and argument from distress also correspond by interpreting them as a justification of an action aimed at solving the problem.",
"cite_spans": [
{
"start": 15,
"end": 35,
"text": "Walton et al. (2008)",
"ref_id": null
}
],
"ref_spans": [],
"eq_spans": [],
"section": "Idiom 1: Means for Goal",
"sec_num": null
},
{
"text": "This idiom is about justification based on the knowledge of a source. Fig. 4 shows the network structure. This idiom has four essential nodes. A con- clusion node is c: acceptable(X). Three premise nodes are p 1 : familiar_with(S, X), p 2 : honest(S), and p 3 : assert(S, X). The edges are from c to p 3 , from p 1 to p 3 , and from p 2 to p 3 . It expresses the relationship between a hypothesis and evidence. A source's assertion is the evidence that a hypothesis is true. This idiom has optional premise nodes o 1 : position_to_know(S, X) and o 2 : expert_in(S, X). Both nodes back up the source's knowledge by linking them to p 1 . The description in the argumentation-schemelike format is as follows.",
"cite_spans": [],
"ref_spans": [
{
"start": 70,
"end": 76,
"text": "Fig. 4",
"ref_id": "FIGREF2"
}
],
"eq_spans": [],
"section": "Idiom 2: From Source Knowledge",
"sec_num": null
},
{
"text": "Premise1: S is familiar with X. Premise3: S is honest. Premise3: S asserts X. Conclusion: X is acceptable.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Idiom 2: From Source Knowledge",
"sec_num": null
},
{
"text": "In the list of Walton et al. (2008) , argument from position to know, argument from expert opinion, and argument from witness testimony correspond to this idiom.",
"cite_spans": [
{
"start": 15,
"end": 35,
"text": "Walton et al. (2008)",
"ref_id": null
}
],
"ref_spans": [],
"eq_spans": [],
"section": "Idiom 2: From Source Knowledge",
"sec_num": null
},
{
"text": "This idiom is about justification based on whether a claim is in accordance with rules or norms. Fig. 5 shows the network structure. A conclusion node is c: acceptable(X). Two premise nodes are p 1 : acceptable(R) and p 2 : follow(X, R). The edges are from p 1 to c and from p 2 to c. It expresses the relationship between a general rule and a specific case. If the case is subject to the rule, the acceptability of the case follows the acceptability of the rule.",
"cite_spans": [],
"ref_spans": [
{
"start": 97,
"end": 103,
"text": "Fig. 5",
"ref_id": "FIGREF3"
}
],
"eq_spans": [],
"section": "Idiom 3: Rule or Principle",
"sec_num": null
},
{
"text": "The description in the argumentation-schemelike format is as follows.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Idiom 3: Rule or Principle",
"sec_num": null
},
{
"text": "Premise1: R is acceptable. Premise2: X follows R. Conclusion: X is acceptable.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Idiom 3: Rule or Principle",
"sec_num": null
},
{
"text": "There is a negation variant for this idiom.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Idiom 3: Rule or Principle",
"sec_num": null
},
{
"text": "Premise1: R must be obeyed. Premise2: X does not follow R. Conclusion: X is not acceptable.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Idiom 3: Rule or Principle",
"sec_num": null
},
{
"text": "In the list of Walton et al. (2008) , no argumentation scheme corresponds to this idiom. One of the argumentation schemes, argument from rule assumes that a rule describes specific actions to follow. However, in the actual arguments we analyze, the main point at issue is whether individual cases fall under the rules like the Constitution.",
"cite_spans": [
{
"start": 15,
"end": 35,
"text": "Walton et al. (2008)",
"ref_id": null
}
],
"ref_spans": [],
"eq_spans": [],
"section": "Idiom 3: Rule or Principle",
"sec_num": null
},
{
"text": "A Bayesian network for an entire argument often contains multiple network instances because multiple justifications may be combined, or a counterargument may be listed together. To construct an entire network, we need to combine network instances. We define the following four rules and default inference to merge or link nodes.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Combination Rules",
"sec_num": "3.3"
},
{
"text": "Merge nodes if the logical expression of nodes is equivalent.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Rule 1: merge the nodes with the same logical expressions",
"sec_num": null
},
{
"text": "Rule 2: link a node to its negational form Link a node to the node with the negational form.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Rule 1: merge the nodes with the same logical expressions",
"sec_num": null
},
{
"text": "For example, link acceptable(X) to \u00acacceptable(X). The edge can be either direction. A boolean value is inverted through the edge.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Rule 1: merge the nodes with the same logical expressions",
"sec_num": null
},
{
"text": "The same content can be expressed in two ways, predicates or propositions. For example, means_to(X, Y ) has the same meaning as acceptable(P ) if P is a propotition that X is a means to Y . In such case, we link a propositional expression like acceptable(P ) to a predicate expression like means_to(X, Y ). A boolean value remains unchanged through the edge.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Rule 3: link a propositional expression to its predicate expression",
"sec_num": null
},
{
"text": "Link nodes if the predicates of two nodes are the same and logical variables are of synonymous meaning. For example, link bring_about(X) to bring_about(X ) if the meaning of X is synonymous to that of X . A boolean value remains unchanged through the edge.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Rule 4: link nodes whose variables are synonymous",
"sec_num": null
},
{
"text": "Besides the formal rules, we may need to link nodes according to the reasoning that does not fit into any of the justification patterns of prepared network idioms. When Visser et al. (2018) annotated argumentation schemes, they also define default inference as not fitting any of the 60 argumentation schemes.",
"cite_spans": [
{
"start": 169,
"end": 189,
"text": "Visser et al. (2018)",
"ref_id": "BIBREF20"
}
],
"ref_spans": [],
"eq_spans": [],
"section": "Default Inference",
"sec_num": null
},
{
"text": "The advantages of our method are threefold.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Advantages of Our Method",
"sec_num": "3.4"
},
{
"text": "Argumentation schemes are useful for validating arguments as they help to teach critical thinking skills in pedagogy (Reed and Walton, 2001 ). However, few NLP models apply argumentation schemes to argument validity assessment. One reason for this could be that the mathematical formalism of argumentation schemes had not been established. We hope that our proposal will encourage more research on incorporating argumentation schemes into NLP models.",
"cite_spans": [
{
"start": 117,
"end": 139,
"text": "(Reed and Walton, 2001",
"ref_id": "BIBREF14"
}
],
"ref_spans": [],
"eq_spans": [],
"section": "Incorporating Argumentation Schemes into NLP Models for Argument Evaluation",
"sec_num": null
},
{
"text": "NLP techniques for transforming the meaning of a text into a formal representation are mainly based on sentences as the unit of analysis. Empirical research for applying discourse-level formal representation to real-world examples has not progressed. Although our proposal is limited to the analysis of arguments, we extend existing techniques to discourse analysis.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Extending Existing Techniques of Formal Semantic Representation to Discourse Analysis",
"sec_num": null
},
{
"text": "Human reasoning to justify arguments differs from formal logic. In general, a conclusion is not a logical consequence of premises. Humans make a variety of justifications expanding on deductive, inductive, and hypothetical reasoning. It is difficult to explain them in a unified manner. Our formalism is flexible enough to describe human reasoning because it just expresses the relationship between premises and conclusions with conditional probabilities. Though due to its flexibility, there is a large degree of freedom to express argument structure, the network can be patterned according to justification patterns provided by argumentation schemes.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Flexibility to Describe Human's Informal Logic and Applicability to Actual Arguments",
"sec_num": null
},
{
"text": "We create a dataset consisting of pairs of relatively short English argument texts and their corresponding Bayesian networks. Our dataset not only helps us to train a machine learning model but also empirically shows that our approach can model the reasoning structure of actual arguments.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Dataset Creation",
"sec_num": "4"
},
{
"text": "As described in Sec. 2.3, annotating argumentation schemes is very costly. In addition, our proposed method requires annotating a network structure. Therefore, it is difficult to prepare a large amount of high-quality data by the procedure of giving an argument text and annotating a network. In this work, we adopt the procedure of generating an argument text from a pre-built network.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Dataset Creation",
"sec_num": "4"
},
{
"text": "Dataset creation is done in the following steps.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Dataset Creation Procedure",
"sec_num": "4.1"
},
{
"text": "Create a Bayesian network from a wide range of discussions on a particular topic.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Dataset Creation Procedure",
"sec_num": "4.1"
},
{
"text": "ProCon.org is a website that collects a wide range of opinions for and against controversial issues, mainly in the United States. We select six topics (the death penalty, gun control, minimum wage, legal abortion, school uniforms, and nuclear power) that are also selected in the existing argument mining dataset (Stab et al., 2018) . We divide each topic into several subtopics. For each subtopic, we summarize the main points of discussion on the site and create a Bayesian network by applying network idioms. For example, for the subtopic on the death penalty, 'Does the death penalty deter crime?', we create a network like Fig. 1 . We also create a text corresponding to the network we have created. This procedure is carried out by the author.",
"cite_spans": [
{
"start": 313,
"end": 332,
"text": "(Stab et al., 2018)",
"ref_id": "BIBREF18"
}
],
"ref_spans": [
{
"start": 628,
"end": 634,
"text": "Fig. 1",
"ref_id": "FIGREF0"
}
],
"eq_spans": [],
"section": "Dataset Creation Procedure",
"sec_num": "4.1"
},
{
"text": "Decompose the built network into network fragments.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Dataset Creation Procedure",
"sec_num": "4.1"
},
{
"text": "We decompose the built Bayesian network, which represents the entire argument of each subtopic, into network fragments so that one fragment contains only one network instance as much as possible. For example, from the network of Fig. 1 , only the part surrounded by the dotted line is cut out.",
"cite_spans": [],
"ref_spans": [
{
"start": 229,
"end": 235,
"text": "Fig. 1",
"ref_id": "FIGREF0"
}
],
"eq_spans": [],
"section": "Dataset Creation Procedure",
"sec_num": "4.1"
},
{
"text": "The predicate logical expression of each node in the network instance is converted into natural language and divided into premises and a conclusion as the description of argumentation schemes. For example, the predicate expression means_to(the death penalty, deterring crime) is converted to 'The death penalty is a means to deter crime' and classified as a premise. This pro- Means for Goal 4 7 2 5 0 2 20 Goal from Means 0 0 0 0 0 5 5 From Consequence 14 11 19 10 19 5 78 Source Knowledge 4 0 0 0 2 0 6 Source Authority 0 1 0 3 2 0 6 Rule or Principle 8 4 3 4 0 0 19 Others 5 3 0 0 1 4 13 Total 35 26 24 22 24 16 147 cedure is almost done by a simple templating generation like 'argument1 predicate argument2.' The author adjusts it so that it is grammatically natural. As an example, the following premises and conclusion are generated.",
"cite_spans": [],
"ref_spans": [
{
"start": 377,
"end": 682,
"text": "Means for Goal 4 7 2 5 0 2 20 Goal from Means 0 0 0 0 0 5 5 From Consequence 14 11 19 10 19 5 78 Source Knowledge 4 0 0 0 2 0 6 Source Authority 0 1 0 3 2 0 6 Rule or Principle 8 4 3 4 0 0 19 Others 5 3 0 0 1 4 13 Total 35 26 24 22 24 16 147",
"ref_id": "TABREF1"
}
],
"eq_spans": [],
"section": "Convert a network fragment into premises and a conclusion.",
"sec_num": null
},
{
"text": "The death penalty is a means to deter crime. Premise2: Deterring crime is desirable.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Premise1:",
"sec_num": null
},
{
"text": "The death penalty is necessary.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Conclusion:",
"sec_num": null
},
{
"text": "Generate an argument text from premises and a conclusion by crowdsourcing.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Conclusion:",
"sec_num": null
},
{
"text": "We present premises and a conclusion to crowdworkers and ask them to convert given contents into a natural language text while preserving the original meaning. In addition, the following two instructions are given to the crowdworkers.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Conclusion:",
"sec_num": null
},
{
"text": "You can omit phrases considered to be common sense like 'Saving lives is desirable.'",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Conclusion:",
"sec_num": null
},
{
"text": "Do not use the following phrases in your writing.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Conclusion:",
"sec_num": null
},
{
"text": "\u2022 follow, above, desirable, acceptable, necessary, true, questionable",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Conclusion:",
"sec_num": null
},
{
"text": "\u2022 be means to/for, lead to, some may argue, either A or B",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Conclusion:",
"sec_num": null
},
{
"text": "\u2022 X is evidence/example that Y, X is enough to achieve the goal",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Conclusion:",
"sec_num": null
},
{
"text": "\u2022 respect to, obey, assert, credible, familiar, popular",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Conclusion:",
"sec_num": null
},
{
"text": "The latter prohibits using idiom-derived keyphrases in the presented premises or conclusions. ",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Conclusion:",
"sec_num": null
},
{
"text": "We created networks for a total of 53 subtopics of 6 topics and extracted 147 network fragments from them. The subtopics of the death penalty are shown in Table 2 . We use Amazon Mechanical Turk and ask 20 crowd workers to generate an argument text per premises-conclusion pair made from the network fragments. Workers were paid 0.15$ per response. The number of subtopics per topic and the number of premises-conclusion patterns presented to the workers are shown in Table 1 . As a result of the crowdsourcing, 2,370 valid responses were obtained. 1,359 responses that did not follow the instructions were excluded as invalid responses. The vocabulary size and average sentence length (word counting) of the responses are shown in Table 3 . The following metrics are measured in order to verify the diversity and bias of the created data since the procedure of present- ing premises-conclusion pairs to the crowdworkers would result in a large bias. The results are shown in Table 4 .",
"cite_spans": [],
"ref_spans": [
{
"start": 155,
"end": 162,
"text": "Table 2",
"ref_id": "TABREF2"
},
{
"start": 468,
"end": 475,
"text": "Table 1",
"ref_id": "TABREF1"
},
{
"start": 732,
"end": 739,
"text": "Table 3",
"ref_id": "TABREF3"
},
{
"start": 976,
"end": 983,
"text": "Table 4",
"ref_id": "TABREF4"
}
],
"eq_spans": [],
"section": "Results",
"sec_num": "4.2"
},
{
"text": "We measure the cosine similarity of the Bag of Words (BoW) representation between a premisesconclusion pair presented to crowdworkers and their response.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "SIM-INOUT: Cosine Similarity between Inputs and Responses in Crowdsourcing",
"sec_num": null
},
{
"text": "We measure the BoW cosine similarity among the responses to the same premises-conclusion pair and average the values.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "SIM-RES: Cosine Similarity among Responses in Crowdsourcing",
"sec_num": null
},
{
"text": "A cloud worker may simply answer by combining premises with a conclusion using 'because.' We investigate the ratio of such responses. We divide responses into the first half and the second half separated by 'because.' Then, we calculate the BoW cosine similarity between the first half and a conclusion, and between the second half and premises. We compute the percentage of responses where both similarity values are more than 0.7.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "BECAUSE: Percentage of 'Conclusion because Premises' Response Pattern",
"sec_num": null
},
{
"text": "The data size is large in the research area of argumentation schemes. We create 2,370 pairs of argumentative text and a corresponding Bayesian network in total. The number of arguments is larger than the largest existing dataset annotated with argumentation schemes though the vocabulary size is smaller.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Characteristics of Our Dataset",
"sec_num": "4.3"
},
{
"text": "As for the diversity of the dataset, the bias is suppressed. Table 4 shows the values of metrics. The following example is the case where a SIM-INOUT (the BOW cosine similarity between a premisesconclusion pair and a response) is 0.67.",
"cite_spans": [],
"ref_spans": [
{
"start": 61,
"end": 68,
"text": "Table 4",
"ref_id": "TABREF4"
}
],
"eq_spans": [],
"section": "Characteristics of Our Dataset",
"sec_num": "4.3"
},
{
"text": "Premises: Gun ownership causes an accidental gun death. Accidental gun death is not desirable. Conclusion: Gun ownership is not necessary. Worker's Response: Lots of accidental gun death results from gun ownership which is completely avoidable.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Characteristics of Our Dataset",
"sec_num": "4.3"
},
{
"text": "Since this is an average example in terms of the metric, we can infer that the responses are not biased toward uniformity. We also come to the same conclusion based on our observations of all the responses. Furthermore, the BECAUSE metric is 6.3, which indicates few workers used the response pattern of 'Conclusion because Premises.'",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Characteristics of Our Dataset",
"sec_num": "4.3"
},
{
"text": "We set up a task to select idioms for argumentative texts to investigate how well the existing NLP model would work. We use idiom names listed in Table 1 as labels. Infrequent idioms are labeled as 'others.' We separate a dataset according to topics. The data for the death penalty is used as a test set and the rest as a training set.",
"cite_spans": [],
"ref_spans": [
{
"start": 146,
"end": 153,
"text": "Table 1",
"ref_id": "TABREF1"
}
],
"eq_spans": [],
"section": "A Baseline Model of Idiom Selection",
"sec_num": "4.4"
},
{
"text": "We apply BERT for sequence classification (Devlin et al., 2019; Wolf et al., 2020) as a baseline. The pre-trained model is BERT-Base Uncased and is not finetuned. Since learning without no sampling leads to always predicting the most frequent idiom in training data, we employ (Loshchilov and Hutter, 2019) . We perform 5fold cross-validation on the training set and determine hyperparameters. The learning rate is 10 \u22125 chosen from {10 \u22124 , 10 \u22125 , 10 \u22126 } and the weight decay is 10 \u22123 chosen from {10 \u22122 , 10 \u22123 , 10 \u22124 }. The max epoch is 400 and the batch size is 64. We repeat the experiment ten times. The averages of precision, recall, and f1 for each idiom are shown in Table 5 . The average of overall accuracies for the test set is 27.5%. It indicates just selecting a suitable idiom is not an easy task.",
"cite_spans": [
{
"start": 42,
"end": 63,
"text": "(Devlin et al., 2019;",
"ref_id": "BIBREF1"
},
{
"start": 64,
"end": 82,
"text": "Wolf et al., 2020)",
"ref_id": "BIBREF28"
},
{
"start": 277,
"end": 306,
"text": "(Loshchilov and Hutter, 2019)",
"ref_id": "BIBREF9"
}
],
"ref_spans": [
{
"start": 679,
"end": 686,
"text": "Table 5",
"ref_id": "TABREF5"
}
],
"eq_spans": [],
"section": "A Baseline Model of Idiom Selection",
"sec_num": "4.4"
},
{
"text": "We propose a methodology for representing the reasoning structure of arguments using Bayesian networks and predicate logic. We define 17 network idioms derived from argumentation schemes to facilitate the network construction by giving typical network patterns. We also create a large dataset consisting of 2,370 pairs of argument text and a corresponding Bayesian network. In future work, we train a model for automatic network construction. ",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Conclusion and Future Work",
"sec_num": "5"
},
{
"text": "The newtork structure of an idiom in this group is either Fig. 16 or Fig. 17 . The predicate CON-STRAINT is defined for each idiom. ",
"cite_spans": [],
"ref_spans": [
{
"start": 58,
"end": 65,
"text": "Fig. 16",
"ref_id": "FIGREF0"
},
{
"start": 69,
"end": 76,
"text": "Fig. 17",
"ref_id": "FIGREF0"
}
],
"eq_spans": [],
"section": "A.11 Constraint Idioms",
"sec_num": null
},
{
"text": "https://www.procon.org/",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "",
"sec_num": null
}
],
"back_matter": [
{
"text": "We thank Naoya Inoue for providing information about literature on argumentation schemes and web resources for argumentative texts.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Acknowledgements",
"sec_num": null
},
{
"text": "A situation or state of affairs. e.g. The death penalty has been introduced.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Case",
"sec_num": null
},
{
"text": "A rule, an assertion, a fact, etc. e.g. The death penalty deters crime.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Proposition",
"sec_num": null
},
{
"text": "A cencept or a term. e.g. Retribution.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Concept",
"sec_num": null
},
{
"text": "A characteristic of something. e.g. Punishment in proportion to serious crime.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Property",
"sec_num": null
}
],
"bib_entries": {
"BIBREF0": {
"ref_id": "b0",
"title": "Handbook of formal argumentation",
"authors": [
{
"first": "Pietro",
"middle": [],
"last": "Baroni",
"suffix": ""
},
{
"first": "M",
"middle": [],
"last": "Dov",
"suffix": ""
},
{
"first": "Massimiliano",
"middle": [],
"last": "Gabbay",
"suffix": ""
},
{
"first": "Leendert",
"middle": [],
"last": "Giacomin",
"suffix": ""
},
{
"first": "",
"middle": [],
"last": "Van Der Torre",
"suffix": ""
}
],
"year": 2018,
"venue": "",
"volume": "2",
"issue": "",
"pages": "",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Pietro Baroni, Dov M Gabbay, Massimiliano Gia- comin, and Leendert van der Torre. 2018. Handbook of formal argumentation. College Publications. 2 https://github.com/ufoym/ imbalanced-dataset-sampler",
"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 Technologies, Volume 1 (Long and Short Papers), pages 4171-4186, Minneapolis, Minnesota. Associ- ation for Computational Linguistics.",
"links": null
},
"BIBREF2": {
"ref_id": "b2",
"title": "Classifying arguments by scheme",
"authors": [
{
"first": "Vanessa",
"middle": [],
"last": "Wei Feng",
"suffix": ""
},
{
"first": "Graeme",
"middle": [],
"last": "Hirst",
"suffix": ""
}
],
"year": 2011,
"venue": "Proceedings of the 49th Annual Meeting of the Association for Computational Linguistics: Human Language Technologies",
"volume": "",
"issue": "",
"pages": "987--996",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Vanessa Wei Feng and Graeme Hirst. 2011. Classify- ing arguments by scheme. In Proceedings of the 49th Annual Meeting of the Association for Com- putational Linguistics: Human Language Technolo- gies, pages 987-996, Portland, Oregon, USA. Asso- ciation for Computational Linguistics.",
"links": null
},
"BIBREF3": {
"ref_id": "b3",
"title": "A general structure for legal arguments about evidence using bayesian networks",
"authors": [
{
"first": "Norman",
"middle": [],
"last": "Fenton",
"suffix": ""
},
{
"first": "Martin",
"middle": [],
"last": "Neil",
"suffix": ""
},
{
"first": "David",
"middle": [
"A"
],
"last": "Lagnado",
"suffix": ""
}
],
"year": 2013,
"venue": "Cognitive Science",
"volume": "37",
"issue": "1",
"pages": "61--102",
"other_ids": {
"DOI": [
"10.1111/cogs.12004"
]
},
"num": null,
"urls": [],
"raw_text": "Norman Fenton, Martin Neil, and David A. Lagnado. 2013. A general structure for legal arguments about evidence using bayesian networks. Cognitive Sci- ence, 37(1):61-102.",
"links": null
},
"BIBREF4": {
"ref_id": "b4",
"title": "Coarse-grained argumentation features for scoring persuasive essays",
"authors": [
{
"first": "Debanjan",
"middle": [],
"last": "Ghosh",
"suffix": ""
},
{
"first": "Aquila",
"middle": [],
"last": "Khanam",
"suffix": ""
},
{
"first": "Yubo",
"middle": [],
"last": "Han",
"suffix": ""
},
{
"first": "Smaranda",
"middle": [],
"last": "Muresan",
"suffix": ""
}
],
"year": 2016,
"venue": "Proceedings of the 54th Annual Meeting of the Association for Computational Linguistics",
"volume": "2",
"issue": "",
"pages": "549--554",
"other_ids": {
"DOI": [
"10.18653/v1/P16-2089"
]
},
"num": null,
"urls": [],
"raw_text": "Debanjan Ghosh, Aquila Khanam, Yubo Han, and Smaranda Muresan. 2016. Coarse-grained argumen- tation features for scoring persuasive essays. In Pro- ceedings of the 54th Annual Meeting of the Associa- tion for Computational Linguistics (Volume 2: Short Papers), pages 549-554, Berlin, Germany. Associa- tion for Computational Linguistics.",
"links": null
},
"BIBREF5": {
"ref_id": "b5",
"title": "Learning to give feedback: Modeling attributes affecting argument persuasiveness in student essays",
"authors": [
{
"first": "W",
"middle": [],
"last": "Zixuan Ke",
"suffix": ""
},
{
"first": "Nishant",
"middle": [],
"last": "Carlile",
"suffix": ""
},
{
"first": "Vincent",
"middle": [],
"last": "Gurrapadi",
"suffix": ""
},
{
"first": "",
"middle": [],
"last": "Ng",
"suffix": ""
}
],
"year": 2018,
"venue": "IJCAI",
"volume": "",
"issue": "",
"pages": "",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Zixuan Ke, W. Carlile, Nishant Gurrapadi, and Vincent Ng. 2018. Learning to give feedback: Modeling at- tributes affecting argument persuasiveness in student essays. In IJCAI.",
"links": null
},
"BIBREF6": {
"ref_id": "b6",
"title": "Give me more feedback II: Annotating thesis strength and related attributes in student essays",
"authors": [
{
"first": "Zixuan",
"middle": [],
"last": "Ke",
"suffix": ""
},
{
"first": "Hrishikesh",
"middle": [],
"last": "Inamdar",
"suffix": ""
},
{
"first": "Hui",
"middle": [],
"last": "Lin",
"suffix": ""
},
{
"first": "Vincent",
"middle": [],
"last": "Ng",
"suffix": ""
}
],
"year": 2019,
"venue": "Proceedings of the 57th Annual Meeting of the Association for Computational Linguistics",
"volume": "",
"issue": "",
"pages": "3994--4004",
"other_ids": {
"DOI": [
"10.18653/v1/P19-1390"
]
},
"num": null,
"urls": [],
"raw_text": "Zixuan Ke, Hrishikesh Inamdar, Hui Lin, and Vincent Ng. 2019. Give me more feedback II: Annotat- ing thesis strength and related attributes in student essays. In Proceedings of the 57th Annual Meet- ing of the Association for Computational Linguis- tics, pages 3994-4004, Florence, Italy. Association for Computational Linguistics.",
"links": null
},
"BIBREF7": {
"ref_id": "b7",
"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 Compu- tational Linguistics, pages 4563-4574, Barcelona, Spain (Online). International Committee on Compu- tational Linguistics.",
"links": null
},
"BIBREF8": {
"ref_id": "b8",
"title": "An online annotation assistant for argument schemes",
"authors": [
{
"first": "John",
"middle": [],
"last": "Lawrence",
"suffix": ""
},
{
"first": "Jacky",
"middle": [],
"last": "Visser",
"suffix": ""
},
{
"first": "Chris",
"middle": [],
"last": "Reed",
"suffix": ""
}
],
"year": 2019,
"venue": "Proceedings of the 13th Linguistic Annotation Workshop",
"volume": "",
"issue": "",
"pages": "100--107",
"other_ids": {
"DOI": [
"10.18653/v1/W19-4012"
]
},
"num": null,
"urls": [],
"raw_text": "John Lawrence, Jacky Visser, and Chris Reed. 2019. An online annotation assistant for argument schemes. In Proceedings of the 13th Linguistic An- notation Workshop, pages 100-107, Florence, Italy. Association for Computational Linguistics.",
"links": null
},
"BIBREF9": {
"ref_id": "b9",
"title": "Decoupled weight decay regularization",
"authors": [
{
"first": "I",
"middle": [],
"last": "Loshchilov",
"suffix": ""
},
{
"first": "F",
"middle": [],
"last": "Hutter",
"suffix": ""
}
],
"year": 2019,
"venue": "ICLR",
"volume": "",
"issue": "",
"pages": "",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "I. Loshchilov and F. Hutter. 2019. Decoupled weight decay regularization. In ICLR.",
"links": null
},
"BIBREF10": {
"ref_id": "b10",
"title": "An annotated corpus of argumentative microtexts",
"authors": [
{
"first": "Andreas",
"middle": [],
"last": "Peldszus",
"suffix": ""
},
{
"first": "Manfred",
"middle": [],
"last": "Stede",
"suffix": ""
}
],
"year": 2015,
"venue": "Proceedings of the First European Conference on Argumentation: Argumentation and Reasoned Action",
"volume": "",
"issue": "",
"pages": "",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Andreas Peldszus and Manfred Stede. 2015. An anno- tated corpus of argumentative microtexts. In Pro- ceedings of the First European Conference on Ar- gumentation: Argumentation and Reasoned Action, Lisbon, Portugal.",
"links": null
},
"BIBREF11": {
"ref_id": "b11",
"title": "Modeling organization in student essays",
"authors": [
{
"first": "Isaac",
"middle": [],
"last": "Persing",
"suffix": ""
},
{
"first": "Alan",
"middle": [],
"last": "Davis",
"suffix": ""
},
{
"first": "Vincent",
"middle": [],
"last": "Ng",
"suffix": ""
}
],
"year": 2010,
"venue": "Proceedings of the 2010 Conference on Empirical Methods in Natural Language Processing",
"volume": "",
"issue": "",
"pages": "229--239",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Isaac Persing, Alan Davis, and Vincent Ng. 2010. Mod- eling organization in student essays. In Proceedings of the 2010 Conference on Empirical Methods in Natural Language Processing, pages 229-239, Cam- bridge, MA. Association for Computational Linguis- tics.",
"links": null
},
"BIBREF12": {
"ref_id": "b12",
"title": "Modeling thesis clarity in student essays",
"authors": [
{
"first": "Isaac",
"middle": [],
"last": "Persing",
"suffix": ""
},
{
"first": "Vincent",
"middle": [],
"last": "Ng",
"suffix": ""
}
],
"year": 2013,
"venue": "Proceedings of the 51st Annual Meeting of the Association for Computational Linguistics",
"volume": "1",
"issue": "",
"pages": "260--269",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Isaac Persing and Vincent Ng. 2013. Modeling the- sis clarity in student essays. In Proceedings of the 51st Annual Meeting of the Association for Compu- tational Linguistics (Volume 1: Long Papers), pages 260-269, Sofia, Bulgaria. Association for Computa- tional Linguistics.",
"links": null
},
"BIBREF13": {
"ref_id": "b13",
"title": "Modeling argument strength in student essays",
"authors": [
{
"first": "Isaac",
"middle": [],
"last": "Persing",
"suffix": ""
},
{
"first": "Vincent",
"middle": [],
"last": "Ng",
"suffix": ""
}
],
"year": 2015,
"venue": "Proceedings of the 53rd Annual Meeting of the Association for Computational Linguistics and the 7th International Joint Conference on Natural Language Processing",
"volume": "1",
"issue": "",
"pages": "543--552",
"other_ids": {
"DOI": [
"10.3115/v1/P15-1053"
]
},
"num": null,
"urls": [],
"raw_text": "Isaac Persing and Vincent Ng. 2015. Modeling argu- ment strength in student essays. In Proceedings of the 53rd Annual Meeting of the Association for Computational Linguistics and the 7th International Joint Conference on Natural Language Processing (Volume 1: Long Papers), pages 543-552, Beijing, China. Association for Computational Linguistics.",
"links": null
},
"BIBREF14": {
"ref_id": "b14",
"title": "Applications of argumentation schemes",
"authors": [
{
"first": "C",
"middle": [],
"last": "Reed",
"suffix": ""
},
{
"first": "D",
"middle": [],
"last": "Walton",
"suffix": ""
}
],
"year": 2001,
"venue": "Proceedings of the 4th Conference of the Ontario Society for the Study of Argument",
"volume": "",
"issue": "",
"pages": "",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "C. Reed and D. Walton. 2001. Applications of argu- mentation schemes. In Proceedings of the 4th Con- ference of the Ontario Society for the Study of Argu- ment (OSSA2001).",
"links": null
},
"BIBREF15": {
"ref_id": "b15",
"title": "Language resources for studying argument",
"authors": [
{
"first": "Chris",
"middle": [],
"last": "Reed",
"suffix": ""
},
{
"first": "Raquel",
"middle": [
"Mochales"
],
"last": "Palau",
"suffix": ""
},
{
"first": "Glenn",
"middle": [],
"last": "Rowe",
"suffix": ""
},
{
"first": "Marie-Francine",
"middle": [],
"last": "Moens",
"suffix": ""
}
],
"year": 2008,
"venue": "Proceedings of the Sixth International Conference on Language Resources and Evaluation (LREC'08)",
"volume": "",
"issue": "",
"pages": "",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Chris Reed, Raquel Mochales Palau, Glenn Rowe, and Marie-Francine Moens. 2008. Language re- sources for studying argument. In Proceedings of the Sixth International Conference on Language Re- sources and Evaluation (LREC'08), Marrakech, Mo- rocco. European Language Resources Association (ELRA).",
"links": null
},
"BIBREF16": {
"ref_id": "b16",
"title": "Annotating argument components and relations in persuasive essays",
"authors": [
{
"first": "Christian",
"middle": [],
"last": "Stab",
"suffix": ""
},
{
"first": "Iryna",
"middle": [],
"last": "Gurevych",
"suffix": ""
}
],
"year": 2014,
"venue": "Proceedings of COLING 2014, the 25th International Conference on Computational Linguistics: Technical Papers",
"volume": "",
"issue": "",
"pages": "1501--1510",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Christian Stab and Iryna Gurevych. 2014. Annotating argument components and relations in persuasive es- says. In Proceedings of COLING 2014, the 25th In- ternational Conference on Computational Linguis- tics: Technical Papers, pages 1501-1510, Dublin, Ireland. Dublin City University and Association for Computational Linguistics.",
"links": null
},
"BIBREF17": {
"ref_id": "b17",
"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. 2017. Parsing ar- gumentation structures in persuasive essays. Com- putational Linguistics, 43(3):619-659.",
"links": null
},
"BIBREF18": {
"ref_id": "b18",
"title": "Crosstopic argument mining from heterogeneous sources",
"authors": [
{
"first": "Christian",
"middle": [],
"last": "Stab",
"suffix": ""
},
{
"first": "Tristan",
"middle": [],
"last": "Miller",
"suffix": ""
},
{
"first": "Benjamin",
"middle": [],
"last": "Schiller",
"suffix": ""
},
{
"first": "Pranav",
"middle": [],
"last": "Rai",
"suffix": ""
},
{
"first": "Iryna",
"middle": [],
"last": "Gurevych",
"suffix": ""
}
],
"year": 2018,
"venue": "Proceedings of the 2018 Conference on Empirical Methods in Natural Language Processing",
"volume": "",
"issue": "",
"pages": "3664--3674",
"other_ids": {
"DOI": [
"10.18653/v1/D18-1402"
]
},
"num": null,
"urls": [],
"raw_text": "Christian Stab, Tristan Miller, Benjamin Schiller, Pranav Rai, and Iryna Gurevych. 2018. Cross- topic argument mining from heterogeneous sources. In Proceedings of the 2018 Conference on Em- pirical Methods in Natural Language Processing, pages 3664-3674, Brussels, Belgium. Association for Computational Linguistics.",
"links": null
},
"BIBREF19": {
"ref_id": "b19",
"title": "Designing and Understanding Forensic Bayesian Networks using Argumentation",
"authors": [
{
"first": "T",
"middle": [],
"last": "Sjoerd",
"suffix": ""
},
{
"first": "",
"middle": [],
"last": "Timmer",
"suffix": ""
}
],
"year": 2017,
"venue": "",
"volume": "",
"issue": "",
"pages": "",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Sjoerd T. Timmer. 2017. Designing and Understand- ing Forensic Bayesian Networks using Argumenta- tion. Ph.D. thesis, Utrecht University.",
"links": null
},
"BIBREF20": {
"ref_id": "b20",
"title": "Revisiting computational models of argument schemes: Classification, annotation, comparison",
"authors": [
{
"first": "J",
"middle": [],
"last": "Visser",
"suffix": ""
},
{
"first": "John",
"middle": [],
"last": "Lawrence",
"suffix": ""
},
{
"first": "H",
"middle": [
"M"
],
"last": "Jean",
"suffix": ""
},
{
"first": "C",
"middle": [],
"last": "Wagemans",
"suffix": ""
},
{
"first": "",
"middle": [],
"last": "Reed",
"suffix": ""
}
],
"year": 2018,
"venue": "COMMA",
"volume": "",
"issue": "",
"pages": "",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "J. Visser, John Lawrence, Jean H. M. Wagemans, and C. Reed. 2018. Revisiting computational models of argument schemes: Classification, annotation, com- parison. In COMMA.",
"links": null
},
"BIBREF21": {
"ref_id": "b21",
"title": "Building bayesian networks for legal evidence with narratives: a case study evaluation",
"authors": [
{
"first": "Charlotte",
"middle": [],
"last": "Vlek",
"suffix": ""
},
{
"first": "Henry",
"middle": [],
"last": "Prakken",
"suffix": ""
},
{
"first": "Silja",
"middle": [],
"last": "Renooij",
"suffix": ""
},
{
"first": "Bart",
"middle": [],
"last": "Verheij",
"suffix": ""
}
],
"year": 2014,
"venue": "Artificial Intelligence and Law",
"volume": "",
"issue": "",
"pages": "",
"other_ids": {
"DOI": [
"10.1007/s10506-014-9161-7"
]
},
"num": null,
"urls": [],
"raw_text": "Charlotte Vlek, Henry Prakken, Silja Renooij, and Bart Verheij. 2014. Building bayesian networks for le- gal evidence with narratives: a case study evaluation. Artificial Intelligence and Law, 22.",
"links": null
},
"BIBREF22": {
"ref_id": "b22",
"title": "Using argument mining to assess the argumentation quality of essays",
"authors": [
{
"first": "Henning",
"middle": [],
"last": "Wachsmuth",
"suffix": ""
},
{
"first": "Khalid",
"middle": [],
"last": "Al-Khatib",
"suffix": ""
},
{
"first": "Benno",
"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 Confer- ence on Computational Linguistics: Technical Pa- pers, pages 1680-1691, Osaka, Japan. The COLING 2016 Organizing Committee.",
"links": null
},
"BIBREF23": {
"ref_id": "b23",
"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 of the Association for Computational Linguistics",
"volume": "1",
"issue": "",
"pages": "176--187",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Henning Wachsmuth, Nona Naderi, Yufang Hou, Yonatan Bilu, Vinodkumar Prabhakaran, Tim Al- berdingk Thijm, Graeme Hirst, and Benno Stein. 2017. Computational argumentation quality assess- ment in natural language. In Proceedings of the 15th Conference of the European Chapter of the Associa- tion for Computational Linguistics: Volume 1, Long Papers, pages 176-187, Valencia, Spain. Associa- tion for Computational Linguistics.",
"links": null
},
"BIBREF24": {
"ref_id": "b24",
"title": "A universal model for discourse-level argumentation analysis",
"authors": [
{
"first": "Henning",
"middle": [],
"last": "Wachsmuth",
"suffix": ""
},
{
"first": "Benno",
"middle": [],
"last": "Stein",
"suffix": ""
}
],
"year": 2017,
"venue": "ACM Transactions on Internet Technology (TOIT)",
"volume": "17",
"issue": "",
"pages": "1--24",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Henning Wachsmuth and Benno Stein. 2017. A univer- sal model for discourse-level argumentation analysis. ACM Transactions on Internet Technology (TOIT), 17:1 -24.",
"links": null
},
"BIBREF25": {
"ref_id": "b25",
"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. Intrin- sic quality assessment of arguments. In Proceed- ings of the 28th International Conference on Com- putational Linguistics, pages 6739-6745, Barcelona, Spain (Online). International Committee on Compu- tational Linguistics.",
"links": null
},
"BIBREF27": {
"ref_id": "b27",
"title": "Building a semantic parser overnight",
"authors": [
{
"first": "Yushi",
"middle": [],
"last": "Wang",
"suffix": ""
},
{
"first": "Jonathan",
"middle": [],
"last": "Berant",
"suffix": ""
},
{
"first": "Percy",
"middle": [],
"last": "Liang",
"suffix": ""
}
],
"year": 2015,
"venue": "Proceedings of the 53rd Annual Meeting of the Association for Computational Linguistics and the 7th International Joint Conference on Natural Language Processing",
"volume": "1",
"issue": "",
"pages": "1332--1342",
"other_ids": {
"DOI": [
"10.3115/v1/P15-1129"
]
},
"num": null,
"urls": [],
"raw_text": "Yushi Wang, Jonathan Berant, and Percy Liang. 2015. Building a semantic parser overnight. In Proceed- ings of the 53rd Annual Meeting of the Association for Computational Linguistics and the 7th Interna- tional Joint Conference on Natural Language Pro- cessing (Volume 1: Long Papers), pages 1332-1342, Beijing, China. Association for Computational Lin- guistics.",
"links": null
},
"BIBREF28": {
"ref_id": "b28",
"title": "Transformers: State-of-the-art natural language processing",
"authors": [
{
"first": "Thomas",
"middle": [],
"last": "Wolf",
"suffix": ""
},
{
"first": "Lysandre",
"middle": [],
"last": "Debut",
"suffix": ""
},
{
"first": "Victor",
"middle": [],
"last": "Sanh",
"suffix": ""
},
{
"first": "Julien",
"middle": [],
"last": "Chaumond",
"suffix": ""
},
{
"first": "Clement",
"middle": [],
"last": "Delangue",
"suffix": ""
},
{
"first": "Anthony",
"middle": [],
"last": "Moi",
"suffix": ""
},
{
"first": "Pierric",
"middle": [],
"last": "Cistac",
"suffix": ""
},
{
"first": "Tim",
"middle": [],
"last": "Rault",
"suffix": ""
},
{
"first": "Remi",
"middle": [],
"last": "Louf",
"suffix": ""
},
{
"first": "Morgan",
"middle": [],
"last": "Funtowicz",
"suffix": ""
},
{
"first": "Joe",
"middle": [],
"last": "Davison",
"suffix": ""
},
{
"first": "Sam",
"middle": [],
"last": "Shleifer",
"suffix": ""
},
{
"first": "Clara",
"middle": [],
"last": "Patrick Von Platen",
"suffix": ""
},
{
"first": "Yacine",
"middle": [],
"last": "Ma",
"suffix": ""
},
{
"first": "Julien",
"middle": [],
"last": "Jernite",
"suffix": ""
},
{
"first": "Canwen",
"middle": [],
"last": "Plu",
"suffix": ""
},
{
"first": "Teven",
"middle": [
"Le"
],
"last": "Xu",
"suffix": ""
},
{
"first": "Sylvain",
"middle": [],
"last": "Scao",
"suffix": ""
},
{
"first": "Mariama",
"middle": [],
"last": "Gugger",
"suffix": ""
},
{
"first": "",
"middle": [],
"last": "Drame",
"suffix": ""
}
],
"year": 2020,
"venue": "Proceedings of the 2020 Conference on Empirical Methods in Natural Language Processing: System Demonstrations",
"volume": "",
"issue": "",
"pages": "38--45",
"other_ids": {
"DOI": [
"10.18653/v1/2020.emnlp-demos.6"
]
},
"num": null,
"urls": [],
"raw_text": "Thomas Wolf, Lysandre Debut, Victor Sanh, Julien Chaumond, Clement Delangue, Anthony Moi, Pier- ric Cistac, Tim Rault, Remi Louf, Morgan Funtow- icz, Joe Davison, Sam Shleifer, Patrick von Platen, Clara Ma, Yacine Jernite, Julien Plu, Canwen Xu, Teven Le Scao, Sylvain Gugger, Mariama Drame, Quentin Lhoest, and Alexander Rush. 2020. Trans- formers: State-of-the-art natural language process- ing. In Proceedings of the 2020 Conference on Em- pirical Methods in Natural Language Processing: System Demonstrations, pages 38-45, Online. Asso- ciation for Computational Linguistics.",
"links": null
}
},
"ref_entries": {
"FIGREF0": {
"num": null,
"text": "A bayesian network on whether the death penalty deters crime. DP: introducing the death penalty, DC: deterring crime, S: criminologists.",
"uris": null,
"type_str": "figure"
},
"FIGREF1": {
"num": null,
"text": "Overall Process of Our Methodology patterns that appear in the actual argument we analyze. Our idiom covers 25 out of 60 argumentation schemes in theWalton et al. (2008)'s list.",
"uris": null,
"type_str": "figure"
},
"FIGREF2": {
"num": null,
"text": "From Source Knowledge Idiom",
"uris": null,
"type_str": "figure"
},
"FIGREF3": {
"num": null,
"text": "Rule or Principle Idiom",
"uris": null,
"type_str": "figure"
},
"FIGREF4": {
"num": null,
"text": "Cause to Effect Idiom Variable type: X is Action or Case. Y is Case. Referenced argumentation schemes: argument from cause to effect. From Consequence Idiom Variable type: X is Action or Case. Y is Case. Referenced argumentation schemes: argument from consequence, pragatic argument from alternatives, argument from threat, argument from fear appeal, argument from danger appeal. Goal from Means Idiom Variable type: X is Action or Case. Y is Case. Means for Goal Idiom Variable type: X is Action or Case. Y is Case. Referenced argumentation schemes: practical reasoning, argument fron need for help, argument from distress. From Source Knowledge Idiom Variable type: X is Proposition. S is Person(s) or Thing(s). Referenced argumentation schemes: argument from position to know, argument from expert opinion, argument from witness testimony, argument from popular practice. From Source Authority Idiom Variable type: X is Proposition. S is Person(s) or Thing(s). Referenced argumentation schemes: argument from popular opinion, argument from rules. Rule or Principle Idiom Variable type: R is Proposition. X is Action or Case. From Evidence Idiom Variable type: X is Proposition. E is Case. Referenced argumentation schemes: argument from sign, argument from evidence to a hypothesis. From Example Idiom Variable type: X is Proposition. E is Case. Referenced argumentation schemes: argument from example.",
"uris": null,
"type_str": "figure"
},
"FIGREF5": {
"num": null,
"text": ": Constraint Idiom1 Variable type: Both X and Y are Action or Case. Both X and Y are Action or Case. A.11.1 Either Constraint Use either as CONSTRAINT. Referenced argumentation schemes: argument from alternatives. A.11.2 Contradict Constraint Use contradict as CONSTRAINT. Referenced argumentation schemes: argument from oppositions. A.11.3 Consistent Constraint Use consistent as CONSTRAINT. Referenced argumentation schemes: pragmatic inconsistency, argument fron inconsistent commitment. Verbal Classification Idiom Variable type: X is Thins(s). V is Concept. F is Property. Referenced argumentation schemes: argument from verbal classification, argument from definition to verbal classification. Verbal Evaluation Idiom Variable type: X is Thins(s). V is Concept.",
"uris": null,
"type_str": "figure"
},
"TABREF0": {
"html": null,
"type_str": "table",
"text": "'s list of argumentation schemes as well as analyze actual arguments and fit idioms to them.In this work, we analyze arguments in ProCon.org 1 , which presents sourced pros and cons of debatable issues. As a result, we define a total of 17 idioms using a finite set of predicates and variable types. Currently, we define idioms for the justification",
"num": null,
"content": "<table><tr><td>Argumentation Schemes</td><td/><td colspan=\"3\">Network Idioms</td><td/><td>O</td><td>u r</td><td>P r o</td><td>p o</td><td>s a l</td></tr><tr><td>Practical Reasoning</td><td colspan=\"3\">bring_about(X)</td><td colspan=\"3\">means_to(X, Y)</td><td/></tr><tr><td>Premise: I have a ...</td><td/><td/><td/><td/><td/><td/><td/></tr><tr><td>Conclusion: Therefore</td><td>formalize</td><td colspan=\"3\">bring_about(Y)</td><td/><td/><td/></tr><tr><td/><td/><td/><td/><td/><td/><td/><td/><td>Idiom Combination</td></tr><tr><td>an argumentative text Should the Death Penalty</td><td colspan=\"3\">Idiom Selection</td><td/><td colspan=\"4\">bring_about(DP)</td><td>means_to(DP, DC)</td></tr><tr><td>Be Allowed?</td><td>analyze</td><td colspan=\"4\">Idiom Instantioation</td><td/><td/></tr><tr><td>We should introduce the death penalty. The death penalty deters crime. ...</td><td>X Y</td><td>\u2190 \u2190</td><td>D D C P : :</td><td>t d h e e t e d r e a t r i n g h</td><td>p e c r i m n a l t e</td><td>y</td><td colspan=\"2\">bring_about(DC)</td></tr></table>"
},
"TABREF1": {
"html": null,
"type_str": "table",
"text": "The number of premises-conclusion patterns. Patterns are counted by topic and by idiom.Idiom \\ Topic death-pena gun-ctrl min-wage abortion sc-uniform nuclear all",
"num": null,
"content": "<table/>"
},
"TABREF2": {
"html": null,
"type_str": "table",
"text": "Subtopics of the death penalty in our dataset.",
"num": null,
"content": "<table><tr><td>The death penalty</td></tr><tr><td>-Is the Death Penalty Immoral?</td></tr><tr><td>-Is the Death Penalty Unconstitutional?</td></tr><tr><td>-Has DNA Testing Led to Significant Im-</td></tr><tr><td>provements in the Criminal Justice System?</td></tr><tr><td>-Does the Death Penalty Deter Crime?</td></tr><tr><td>-Should the Death Penalty Be Used for Ret-</td></tr><tr><td>ribution?</td></tr><tr><td>-Does a Person's Race Affect the Likelihood</td></tr><tr><td>of Receiving the Death Penalty?</td></tr><tr><td>-Should the Death Penalty Be Abolished or</td></tr><tr><td>Paused Because of the Alleged --Possibility</td></tr><tr><td>of an Innocent Person Being Executed?</td></tr><tr><td>-Is Life in Prison without Parole a Better</td></tr><tr><td>Option Than the Death Penalty?</td></tr></table>"
},
"TABREF3": {
"html": null,
"type_str": "table",
"text": "Summary of crowdsourcing responses.",
"num": null,
"content": "<table><tr><td>Topic</td><td colspan=\"6\">death-pena gun-ctrl min-wage abortion sc-uniform nuclear</td><td>all</td></tr><tr><td>#responses</td><td>535</td><td>417</td><td>446</td><td>289</td><td>411</td><td>272</td><td>2,370</td></tr><tr><td>vocab size</td><td>1,052</td><td>776</td><td>549</td><td>692</td><td>735</td><td>609</td><td>2,493</td></tr><tr><td>sentence len</td><td>23.7</td><td>23.0</td><td>18.2</td><td>19.8</td><td>19.2</td><td>23.8</td><td>21.3</td></tr></table>"
},
"TABREF4": {
"html": null,
"type_str": "table",
"text": "Metrics of the diversity of our dataset.",
"num": null,
"content": "<table><tr><td>Metric</td><td colspan=\"7\">death-pena gun-ctrl min-wage abortion sc-uniform nuclear all</td></tr><tr><td>SIM-INOUT</td><td>0.69</td><td>0.66</td><td>0.69</td><td>0.62</td><td>0.65</td><td>0.68</td><td>0.67</td></tr><tr><td>SIM-RES</td><td>0.57</td><td>0.54</td><td>0.59</td><td>0.48</td><td>0.54</td><td>0.55</td><td>0.55</td></tr><tr><td>BECAUSE</td><td>6.4</td><td>5.8</td><td>10.3</td><td>1.7</td><td>5.1</td><td>7.4</td><td>6.3</td></tr></table>"
},
"TABREF5": {
"html": null,
"type_str": "table",
"text": "The results of a baseline idiom classifier for a test set. Some idioms do not apper in the test set.",
"num": null,
"content": "<table><tr><td>Idiom</td><td colspan=\"2\">Prec Recall</td><td>F1</td></tr><tr><td>Means for Goal</td><td colspan=\"3\">0.283 0.374 0.287</td></tr><tr><td>Goal from Means</td><td>-</td><td>-</td><td>-</td></tr><tr><td colspan=\"4\">From Consequence 0.636 0.169 0.230</td></tr><tr><td>Source Knowledge</td><td>0</td><td>0</td><td>NA</td></tr><tr><td>Source Authority</td><td>-</td><td>-</td><td>-</td></tr><tr><td>Rule or Principle</td><td colspan=\"3\">0.321 0.735 0.430</td></tr><tr><td>Others</td><td colspan=\"3\">0.137 0.061 0.077</td></tr><tr><td colspan=\"4\">both oversampling and undersampling 2 . Parame-</td></tr><tr><td colspan=\"2\">ters are optimized using AdamW</td><td/><td/></tr></table>"
}
}
}
} |