File size: 135,084 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 2292 2293 2294 2295 2296 2297 2298 2299 2300 2301 2302 2303 2304 2305 2306 2307 2308 2309 2310 2311 2312 2313 2314 2315 2316 2317 2318 2319 2320 2321 2322 2323 2324 2325 2326 2327 2328 2329 2330 2331 2332 2333 2334 2335 2336 2337 2338 2339 2340 2341 2342 2343 2344 2345 2346 2347 2348 2349 2350 2351 2352 2353 2354 2355 2356 2357 2358 2359 2360 2361 2362 2363 2364 2365 2366 2367 2368 2369 2370 2371 2372 2373 2374 2375 2376 2377 2378 2379 2380 2381 2382 2383 2384 2385 2386 2387 2388 2389 2390 2391 2392 2393 2394 2395 2396 2397 2398 2399 2400 2401 2402 2403 2404 2405 2406 2407 2408 2409 2410 2411 2412 2413 2414 2415 2416 2417 2418 2419 2420 2421 2422 2423 2424 2425 2426 2427 2428 2429 2430 2431 2432 2433 2434 2435 2436 2437 2438 2439 2440 2441 2442 2443 2444 2445 2446 2447 2448 2449 2450 2451 2452 2453 2454 2455 2456 2457 2458 2459 2460 2461 2462 2463 2464 2465 2466 2467 2468 2469 2470 2471 2472 2473 2474 2475 2476 2477 2478 2479 2480 2481 2482 2483 2484 2485 2486 2487 2488 2489 2490 2491 2492 2493 2494 2495 2496 2497 2498 2499 2500 2501 2502 2503 |
{
"paper_id": "N15-1046",
"header": {
"generated_with": "S2ORC 1.0.0",
"date_generated": "2023-01-19T14:32:51.695780Z"
},
"title": "Using Summarization to Discover Argument Facets in Online Idealogical Dialog",
"authors": [
{
"first": "Amita",
"middle": [],
"last": "Misra",
"suffix": "",
"affiliation": {
"laboratory": "Dialogue Systems Lab",
"institution": "UC Santa Cruz Natural Language",
"location": {
"addrLine": "1156 N. High. SOE-3 Santa Cruz",
"postCode": "95064",
"region": "California",
"country": "USA"
}
},
"email": ""
},
{
"first": "Pranav",
"middle": [],
"last": "Anand",
"suffix": "",
"affiliation": {
"laboratory": "Dialogue Systems Lab",
"institution": "UC Santa Cruz Natural Language",
"location": {
"addrLine": "1156 N. High. SOE-3 Santa Cruz",
"postCode": "95064",
"region": "California",
"country": "USA"
}
},
"email": ""
},
{
"first": "Jean",
"middle": [
"Fox"
],
"last": "Tree",
"suffix": "",
"affiliation": {
"laboratory": "Dialogue Systems Lab",
"institution": "UC Santa Cruz Natural Language",
"location": {
"addrLine": "1156 N. High. SOE-3 Santa Cruz",
"postCode": "95064",
"region": "California",
"country": "USA"
}
},
"email": ""
},
{
"first": "Marilyn",
"middle": [],
"last": "Walker",
"suffix": "",
"affiliation": {
"laboratory": "Dialogue Systems Lab",
"institution": "UC Santa Cruz Natural Language",
"location": {
"addrLine": "1156 N. High. SOE-3 Santa Cruz",
"postCode": "95064",
"region": "California",
"country": "USA"
}
},
"email": ""
}
],
"year": "",
"venue": null,
"identifiers": {},
"abstract": "More and more of the information available on the web is dialogic, and a significant portion of it takes place in online forum conversations about current social and political topics. We aim to develop tools to summarize what these conversations are about. What are the CENTRAL PROPOSITIONS associated with different stances on an issue; what are the abstract objects under discussion that are central to a speaker's argument? How can we recognize that two CENTRAL PROPOSITIONS realize the same FACET of the argument? We hypothesize that the CENTRAL PROPOSITIONS are exactly those arguments that people find most salient, and use human summarization as a probe for discovering them. We describe our corpus of human summaries of opinionated dialogs, then show how we can identify similar repeated arguments, and group them into FACETS across many discussions of a topic. We define a new task, ARGUMENT FACET SIMILARITY (AFS), and show that we can predict AFS with a .54 correlation score, versus an ngram system baseline of .39 and a semantic textual similarity system baseline of .45.",
"pdf_parse": {
"paper_id": "N15-1046",
"_pdf_hash": "",
"abstract": [
{
"text": "More and more of the information available on the web is dialogic, and a significant portion of it takes place in online forum conversations about current social and political topics. We aim to develop tools to summarize what these conversations are about. What are the CENTRAL PROPOSITIONS associated with different stances on an issue; what are the abstract objects under discussion that are central to a speaker's argument? How can we recognize that two CENTRAL PROPOSITIONS realize the same FACET of the argument? We hypothesize that the CENTRAL PROPOSITIONS are exactly those arguments that people find most salient, and use human summarization as a probe for discovering them. We describe our corpus of human summaries of opinionated dialogs, then show how we can identify similar repeated arguments, and group them into FACETS across many discussions of a topic. We define a new task, ARGUMENT FACET SIMILARITY (AFS), and show that we can predict AFS with a .54 correlation score, versus an ngram system baseline of .39 and a semantic textual similarity system baseline of .45.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Abstract",
"sec_num": null
}
],
"body_text": [
{
"text": "In the wake of the Penn TreeBank, much progress has been achieved in processing the monologic, informational language characteristic of newswire text. But an increasing share of the text data on the web is unlike newswire in a variety of ways: it is dialogic, opinionated, argumentative. And while some of these dialogs may be a little more than flame wars, a significant portion involve contentful, rea-PostID:Turn S1:1 Agreed She is ignoring my religious freedom and trying to institute her religion into law. The law that will bar my family from legal protections. It won't protect her marriage but will bar me and my people from from being full citizens. She isn't protecting marriage but perserving her heterosexual privledge. S2:1 How on earth is she impeding on you religious freedom? She isn't trying to take away your right to any religious ceremony. With such a wide-open standard of what constitutes religious freedom that you seem to have, any legislation could be construed as imposing on religious freedom. S1:2 Because it is her religious belief that marriage is between a man and a woman. My religious belief is that marriage is between two people that love each other regardless of sex. She is tying to place her religious belif into law over mine. Who gets hurt here? If my religious belief is put into law she can still marry the person of her choice. If her religious belief gets put into law she can still marry the person of her choice but I do not get to. So I and my people are hurt by codifing her religious belief into law. She is trying to keep gay people out of marriage and thus preserve her heterosexual privledge. S2:2 But by that definition, either one could be viewed as impeding on religious freedom, including your view impeding on hers ! We don't define imposing on religious freedom on the basis of having different ideals. It doesn't effect your religion or religious freedom if you don't get benefits under gay marriages. You can argue in other ways, on other basis, but the idea that not giving gays marriage benefits is imposing on religious freedom is an empty \" argument \". soned disputes on important social and political topics, as exemplified by the forum snippets in Figs. 1 and 3. Studying data like this will undoubtedly help us to understand dialogic and informal argumentative language in general. And, indeed, previous work (Abbott et al., 2011; Somasundaran and Wiebe, 2010) has examined the structure of these discussions -e.g., the argumentative discourse relation a post bears to its parent (agreeing or disagreeing), or the stance that a person takes on an issue.",
"cite_spans": [
{
"start": 2376,
"end": 2397,
"text": "(Abbott et al., 2011;",
"ref_id": "BIBREF0"
},
{
"start": 2398,
"end": 2427,
"text": "Somasundaran and Wiebe, 2010)",
"ref_id": "BIBREF36"
}
],
"ref_spans": [],
"eq_spans": [],
"section": "Introduction",
"sec_num": "1"
},
{
"text": "Our goal here is to develop techniques to recognize the specific arguments and counterarguments people tend to advance, and group them across discussions into the FACETS on which that issue is ar- gued across the population at large. Recognizing the FACETS of an argument automatically entails at least two subtasks, as schematized in Fig. 2a .",
"cite_spans": [],
"ref_spans": [
{
"start": 335,
"end": 342,
"text": "Fig. 2a",
"ref_id": "FIGREF1"
}
],
"eq_spans": [],
"section": "Introduction",
"sec_num": "1"
},
{
"text": "PostID:Turn S1:1 Certainly not yours. You should know that I am for no marriage in government. It should be left to a religious institution where it will actually mean something. The states should then go back to doing something that actually makes sense and doesn't reward people like Britney Spears for being white trash. S2:1 That is all well and good, but it is not the religious ceremony and sanction that gays are looking for. They already have that; there are churches that perform same-sex marriages. It is the civil benefits that are at issue. Are you saying you would be in favor of foregoing ALL the legal rights and benefits you are afforded by marriage? For example: *Assumption of Spouse's Pension *Automatic Inheritance *Automatic Housing Lease Transfer *Bereavement Leave.... What do you say? S1:2 yeah I know. I'm saying that there should be a better system. For example, if you had a best friend who you are roommates with... both hetero for the sake of argument... and never wish to get married then could they get some of the benefits you described? First, there must be a system, the CENTRAL PROPOSITION detector, that can extract the most essential arguments in a particular conversation. Example CENTRAL PROPOSITIONS in Figs. 1 and 3 are provided in bold. Second, there must be another system, the ARGUMENT FACET inducer, that relates these conversation-specific arguments to each other in terms of FACETS, e.g. that identifies the two spe-cific central propositions in Figs. 1 and 3 about \"legal protections\" and \"civil benefits\" as the same (abstract) FACET, namely that same-sex marriage is about getting the civil rights benefits of marriage.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Introduction",
"sec_num": "1"
},
{
"text": "We first focus on the question of extracting reliable data for central propositions. See Fig. 2b . We propose that the CENTRAL PROPOSITIONS of a dialog are exactly those arguments that people find most salient, which is naturally reflected by their summarization behavior. We then apply the Pyramid method, by which the CENTRAL PROPO-SITIONS bubble up to the highest tiers of the pyramid, thereby allowing us to identify them. With the central propositions in hand, we proceed to build the argument facet inducer. We introduce a new task of ARGUMENT FACET SIMILARITY (AFS). We discuss how AFS is similar to, but different than SEMANTIC TEXTUAL SIMILARITY (STS) (Agirre et al., 2012; Jurgens et al., 2014; Agirre et al., 2013; Beltagy et al., 2014; Han et al., 2013) .",
"cite_spans": [
{
"start": 661,
"end": 682,
"text": "(Agirre et al., 2012;",
"ref_id": "BIBREF1"
},
{
"start": 683,
"end": 704,
"text": "Jurgens et al., 2014;",
"ref_id": "BIBREF20"
},
{
"start": 705,
"end": 725,
"text": "Agirre et al., 2013;",
"ref_id": "BIBREF2"
},
{
"start": 726,
"end": 747,
"text": "Beltagy et al., 2014;",
"ref_id": "BIBREF3"
},
{
"start": 748,
"end": 765,
"text": "Han et al., 2013)",
"ref_id": "BIBREF15"
}
],
"ref_spans": [
{
"start": 89,
"end": 96,
"text": "Fig. 2b",
"ref_id": "FIGREF1"
}
],
"eq_spans": [],
"section": "Introduction",
"sec_num": "1"
},
{
"text": "Sec. 2 provides a more detailed overview and description of our method, and the data that it produces. Sec. 3 describes our experimental setup for the AFS task and then presents our results. We describe a learning approach that achieves correlations of .54 on the AFS task, as compared to a baseline correlation of .45 using off-the-shelf modules that are competitive in STS tasks. We delay a detailed discussion of related work to Sec. 4 when we can compare it to our own approach. Sec. 5 summarizes the paper and discusses future work.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Introduction",
"sec_num": "1"
},
{
"text": "2 Experimental Method Fig. 2 summarizes our overall method for producing the summary corpus and then extracting arguments and clustering them into FACETS. Our method consists of the following steps: S1: Dialog Selection. S2: MT summarization of dialogs selected in S1. S3: Pyramid annotation of summaries produced by S2 and selection of top-tier pyramid labels as CENTRAL PROPOSITIONS for individual dialogs. S4: Clustering of CENTRAL PROPOSITIONS from S3. S5: MT ARGUMENT FACET SIMILARITY task, using clusters from S4. S6: Train and test a predictor for ARGUMENT FACET SIMILARITY (Sec. 3).",
"cite_spans": [],
"ref_spans": [
{
"start": 22,
"end": 28,
"text": "Fig. 2",
"ref_id": "FIGREF1"
}
],
"eq_spans": [],
"section": "Introduction",
"sec_num": "1"
},
{
"text": "We explain these steps in more detail below. S1: Dialog Selection. We use the publicly available Internet Argument Corpus (IAC) (Walker et al., 2012) . We use the links in the meta-data to extract a sequence of turns to build two-party dialog chains like those in Figs. 1 and 3. We extracted 85 dialogs for the topic gay marriage from an original corpus of 1292 discussion threads using these criteria:",
"cite_spans": [
{
"start": 128,
"end": 149,
"text": "(Walker et al., 2012)",
"ref_id": "BIBREF37"
}
],
"ref_spans": [],
"eq_spans": [],
"section": "Introduction",
"sec_num": "1"
},
{
"text": "\u2022 Number of turns per contributor: We want dialogs in which substantive issues were discussed, so we extract dialogs with at least 3 turns per conversant that present at least 2 different perspectives on an issue.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Introduction",
"sec_num": "1"
},
{
"text": "\u2022 Author: Some authors post frequently and would dominate the corpus if we use random selection. To get richer, more diverse dialogs expressing different perspectives, we only select a single dialog between any particular pair of authors from a discussion thread.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Introduction",
"sec_num": "1"
},
{
"text": "\u2022 Word Count in a post: Some posts are long. To make it practical to collect dialog summaries, we extract dialogs where the number of words per turn is less than 250.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Introduction",
"sec_num": "1"
},
{
"text": "S2: MT Summarization Task. The summarization task was run on Mechanical Turk. To get good S1 thinks that the government should stay out of marriage and that it should be left to religious institutions. He thinks there needs to be a better system and that single people are the ones that are harmed the most by marriage laws because they are unable to get any of the benefits that married people do even if they want them, or it is important to their situation. S2 says religious ceremonies aren't what gay people want because they already can have them via churches. They want the rights and to keep the government out would be to give up those rights. If single people want those rights they should get married, but he thinks you should be free to marry who you wish. The issue here is whether government or religion should decides the principles of marriage, and who is allowed to get married. Speaker one believes that leaving it up to religions groups does not satisfy what gays are looking for. They are searching for the civil benefits that come with a marriage and would like to be treated equally in that respect. The speaker believes gay should be able to marry a person of their choice and get equal rights.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Introduction",
"sec_num": "1"
},
{
"text": "Speaker two opinions that there should indeed be a better system for marriage benefits and that it is all \"single\" people that get screwed over by marriage's current stature. Speaker two believes that gay people should marry a woman if they want the same rights. quality summaries, workers completed a qualification test involving summarizing a sample dialog. Workers were instructed to summarize according to dialog length: dialogs under 750 words in 125 words, and those above 750 in 175 words. We use 45 dialogs in this study and save the other 40 for future work. We collect 5 summaries for each dialog resulting in a dataset of 225 summaries. S3: Pyramid Annotation. We trained three undergraduates to annotate summaries to produce pyramids. We hypothesize that we can use the Pyramid method to induce the FACETS of a topic across a set of dialogs (Nenkova and Passonneau, 2004) . The annotation of Pyramids seeks to uncover the common elements, or summary content units (SCUs), across several summaries (in our case, 5). Each SCU identifies a set of spans that are semantically equivalent. Each SCU also has a unique annotatorgenerated label that reflects the semantic meaning of the contributions. Because our aim here is to focus on argument propositional content, the annotators were instructed to keep only the main proposition in the SCU as the label, ignoring any attributions or other types of content. See Table 1 . Once annotation is complete, the SCUs are ranked based on their frequency across all of the summaries, as shown by the Tier in Fig. 5 , which includes data from the two summaries in Fig. 4 .",
"cite_spans": [
{
"start": 853,
"end": 883,
"text": "(Nenkova and Passonneau, 2004)",
"ref_id": "BIBREF30"
}
],
"ref_spans": [
{
"start": 1420,
"end": 1427,
"text": "Table 1",
"ref_id": "TABREF0"
},
{
"start": 1557,
"end": 1563,
"text": "Fig. 5",
"ref_id": null
},
{
"start": 1612,
"end": 1618,
"text": "Fig. 4",
"ref_id": "FIGREF3"
}
],
"eq_spans": [],
"section": "Introduction",
"sec_num": "1"
},
{
"text": "Contributor S1 points to the trend to legalize gay marriage in western countries such as Netherlands, Belgium, and most of Canada Contributor S1 refutes this assertion, citing a number of countries which recognize same-sex marriage. Contributor He states the US is more similar to Anglo nations and in many of those gay marriage is legal. Label A number of countries recognize samesex marriage. S4: SCUs to clusters. The pyramid structure directly reflects the content that the annotators deem most important in the original dialog. We are interested in the content that bubbles to the top across all the dialogs. We take the Tier 3 and above SCUs as our CENTRAL PROPOSITIONS, and extract the labels of those SCUs. This gives a total of 329 SCU labels. In what follows we treat a cluster of CENTRAL PROPOSITIONS as a FACET label, just as a synset concept in WordNet is labeled by its members. The purpose of AFS, then, is to provide a similarity metric on these SCU labels. As described below (and sketched in Fig. 2c ), we used Mechanical Turk to provide similarity scores between pairs of SCU central propositions. Although, in principle, we could have asked about all possible pairs of the 329 CENTRAL PROPOSITIONS, most pairs are likely to be unrelated, and so we used an initial clustering algorithm to help reduce the work and cost.",
"cite_spans": [],
"ref_spans": [
{
"start": 1010,
"end": 1017,
"text": "Fig. 2c",
"ref_id": "FIGREF1"
}
],
"eq_spans": [],
"section": "Introduction",
"sec_num": "1"
},
{
"text": "To group similar arguments, we performed clustering across our 329 labels. We performed Agglomerative Clustering using Scikit-learn (Agg Clustering in Fig. 2c ). (Pedregosa et al., 2011) . It recursively merges the pair of clusters that minimally increases a given linkage distance. We used cosine similarity as the distance measure with average linkage criteria. To focus on topic-specific cues, the clustering was performed using only nouns, verbs and adjectives. After generating all pairwise combinations within a cluster, this approach yielded 1131 argument pairs used in the Mechanical Turk AFS task. See Fig. 2c .",
"cite_spans": [
{
"start": 162,
"end": 186,
"text": "(Pedregosa et al., 2011)",
"ref_id": "BIBREF32"
}
],
"ref_spans": [
{
"start": 151,
"end": 158,
"text": "Fig. 2c",
"ref_id": "FIGREF1"
},
{
"start": 611,
"end": 618,
"text": "Fig. 2c",
"ref_id": "FIGREF1"
}
],
"eq_spans": [],
"section": "Introduction",
"sec_num": "1"
},
{
"text": "We would like you to classify each of the following sets of pairs based on your perception of how SIM-ILAR the arguments are, on the following scale, examples follow. (5) Completely equivalent, mean pretty much exactly the same thing, using different words. (4) Mostly equivalent, but some unimportant details differ. One argument may be more specific than another or include a relatively unimportant extra fact.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Instructions",
"sec_num": null
},
{
"text": "(3) Roughly equivalent, but some important information differs or is missing. This includes cases where the argument is about the same FACET but the authors have different stances on that facet.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Instructions",
"sec_num": null
},
{
"text": "(2) Not equivalent, but share some details. For example, talking about the same entities but making different arguments (different facets) (1) Not equivalent, but are on same topic (0) On a different topic Facet: A facet is a low level issue that often reoccurs in many arguments in support of the author's stance or in attacking the other author's position. There are many ways to argue for your stance on a topic. For example, in a discussion about the death penalty you may argue in favor of it by claiming that it deters crime. Alternatively, you may argue in favor of the death penalty because it gives victims of the crimes closure. On the other hand you may argue against the death penalty because some innocent people will be wrongfully executed or because it is a cruel and unusual punishment. Each of these specific points is a facet. For two utterances to be about the same facet, it is not necessary that the authors have the same belief toward the facet. For example, one author may believe that the death penalty is a cruel and unusual punishment while the other one attacks that position. However, in order to attack that position they must be discussing the same facet. shows the instructions defining AFS for the MT HIT. Inspired by the scale used for STS, we collected annotations on a 6 point scale. One crucial difference in our formulation was a desire to capture similarity in FACET and argument simultaneously. The use of the value 3 for 'same FACET, contradictory stance' was a well-thought decision in the definition of AFS.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Instructions",
"sec_num": null
},
{
"text": "Used by summarizer? Tier 1 2 3 4 5",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "SCU Label",
"sec_num": null
},
{
"text": "Gay couples are interested in the rights and benefits associated with marriage. 5 Gay people should be able to marry a person of their choice and get equal rights. 5 Government should not be involved in marriage and marriage should be left to religious institutions.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "SCU Label",
"sec_num": null
},
{
"text": "Discussion on the civil benefits of marriage and the rights of marriage. 4 Gay couples are unable to get any benefits that married people do. 4 There should be a better system for marriage benefits. 4 Religious ceremonies are not what gay people want.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "5",
"sec_num": null
},
{
"text": "3 Single people are the ones that are harmed the most by marriage laws.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "5",
"sec_num": null
},
{
"text": "3 Gay people should marry the opposite sex if they want the same rights.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "5",
"sec_num": null
},
{
"text": "2 Gays have religious ceremonies already can have them via churches 1 Relation to the issues by consideration of the case of a life-long bachelor uncle 1 Figure 5 : Pyramid for Dialog-2. SCU labels in Tiers 3-5 are assumed to be the CENTRAL PROPOSITIONS.",
"cite_spans": [],
"ref_spans": [
{
"start": 154,
"end": 162,
"text": "Figure 5",
"ref_id": null
}
],
"eq_spans": [],
"section": "5",
"sec_num": null
},
{
"text": "Just as two words can only be antonyms if they are in the same semantic field, two arguments can only be contradictory if they are about the same FACET. Thus, we instruct annotators to give a score of 3 to opposing arguments on the same FACET.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "5",
"sec_num": null
},
{
"text": "The task was put on Mechanical Turk using two separate batches. For the first batch we randomly selected 500 pairs from our pairs dataset of 1131 pairs. However, our subsequent impression was that the clustering had not filtered out enough of the unrelated pairs (score 0-1). For the second batch we selected the top 500 pairs according to the UMBC similarity score (Han et al., 2013) . This gave us a final pair dataset of 1000 pairs. Since AFS is a novel and subjective task, workers took a qualification test. Then each pair was annotated by 5 workers, and one of the authors provided gold standard labels. The HIT allowed 5 AFS judgements per hit, thus the number of pairs annotated by a worker varies from 5 to 1000.",
"cite_spans": [
{
"start": 366,
"end": 384,
"text": "(Han et al., 2013)",
"ref_id": "BIBREF15"
}
],
"ref_spans": [],
"eq_spans": [],
"section": "5",
"sec_num": null
},
{
"text": "To increase reliability, we removed the annotations from those workers who had attempted less than 4 hits (20 pairs) and had the lowest pairwise correlations with our gold standard annotation. Our final AFS score was the average score across all the annotators. The final AFS score correlated at .7 with our gold standard annotation, showing that the AFS similarity task is well-defined, and understandable by minimally trained annotators on MT. Table 4 provides typical examples of argument pairs and their MT AFS score, along with the predicted scores from some of our models. We discuss the AFS values and interesting cases in Sec. 3 below.",
"cite_spans": [],
"ref_spans": [
{
"start": 446,
"end": 453,
"text": "Table 4",
"ref_id": "TABREF6"
}
],
"eq_spans": [],
"section": "5",
"sec_num": null
},
{
"text": "Given the data collected above, we defined a supervised machine learning experiment with AFS as our dependent variable and different collections of features inspired from previous work as our independent variables.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Machine Learning Experiments and Results",
"sec_num": "3"
},
{
"text": "NGRAM overlap. This is our primary baseline. For each argument, we extracted all the unigrams, bigrams and trigrams, and then counted how many were in overlap across the two arguments. For unigrams we did not include stop words. Stemmed Ngrams were used to get better overlap. UMBC. This is our secondary baseline. This feature is the Semantic Textual Similarity obtained using UMBC Semantic Similarity tool (Han et al., 2013) DISCO Distributionally Similar Category. We used the distributional similarity tool DISCO with the pre-computed English Wikipedia word space (Kolb, 2008) . We extract the top 5 distributionally similar nouns, verbs, and adjectives for each argument. For each argument pair, three vector pairs (over nouns, verbs, and adjectives) are created with this extended vocabulary. Stemming was performed and cosine similarity between these vector pairs was calculated. LIWC Category. This feature set is based on the Linguistics Inquiry Word Count tool (Pennebaker et al., 2001 ). To tune these features, we first used a set of gay marriage posts from websites such as Creat-eDebate and ConvinceMe to extract relevant LIWC categories. We supplemented this data with gay marriage posts from 4forums, but excluded the discussion threads in our dialog corpus. From this data, we extracted the LIWC categories most frequent nouns, verbs and adjectives. For the verbs category, we excluded the verbs present in the NLTK stop word list. We retained only semantically rich categories such as Biological Processes, Causation, Cognitive Processes, Humans, Negative Emotion, Positive Emotion, Religion, Sexual, and Social Processes. The score for this set was the LIWC category overlap count across pairs for each category. ROUGE Scores. ROUGE is a family of metrics to determine the quality of a summary by comparing it to other ideal summaries (Lin, 2004) . It is based on a number of overlapping units such as n-gram, word sequences, and word pairs. This feature includes all of the rouge f-scores available via the package at https://pypi.python.org/pypi/pyrouge/0.1.0.",
"cite_spans": [
{
"start": 408,
"end": 426,
"text": "(Han et al., 2013)",
"ref_id": "BIBREF15"
},
{
"start": 568,
"end": 580,
"text": "(Kolb, 2008)",
"ref_id": "BIBREF21"
},
{
"start": 971,
"end": 995,
"text": "(Pennebaker et al., 2001",
"ref_id": "BIBREF33"
},
{
"start": 1854,
"end": 1865,
"text": "(Lin, 2004)",
"ref_id": "BIBREF22"
}
],
"ref_spans": [],
"eq_spans": [],
"section": "Features",
"sec_num": "3.1"
},
{
"text": "Our aim is to predict the similarity among repeated arguments across many discussions in online social and political debate forums, a task we have dubbed ARGUMENT FACET SIMILARITY (AFS). Given the CENTRAL PROPOSITIONS from the CP detector (see Fig. 2a ), we need to train an argument FACET inducer. We define AFS as a regression problem and evaluate support vector regression and linear regression for 10-fold cross validation using the Weka machine learning toolkit (Hall et al., 2005) . Table 2 shows that the results for support vector regression are worse than the linear regression model using our proposed features combined with UMBC, hence we focus hereon on linear regression. Table 3 provides the correlations, MAE, and RMS values for models produced using various sets of features. We considered two baselines, simple Ngram overlap and the off-the-shelf UMBC STS metric (Han et al., 2013) . In general, we found that Ngram overlap (Row 1) performed best alone of our features, but falls short of the UMBC base- We then tested combinations of of features to determine which feature sets are complementary. LIWC + NGRAM is significantly different than NGRAM alone ( p < 0.01), and ROUGE + NGRAM is significantly different than NGRAM alone ( p = 0.03), but DISCO does not add anything ( p = 0.2). This shows that LIWC and ROUGE features complement Ngram features. Other combinations of interest are NGRAM + LIWC (Row 7) which amazingly performs as well as UMBC while UMBC includes sentence alignment, a model of negation, and distributional measures (Han et al., 2013) . This suggests that AFS is clearly a different task that STS. Additionally we also combined our proposed set of features with UMBC. A comparison of Row 15 (our feature set) with Rows 16 and 17 of It is also interesting to examine the differences in model scores for particular argument pairs as shown in Table 4 . The best performing model for each row is in bold in Table 4 . As described in the HIT instructions in Fig. 6 , values of AFS near 0 (Row 1) indicate different topics and no similarity. Values near 1 indicate same topic but different arguments (Rows 2,3). Values of 3 and above indicate same FACET (Rows 7, 8) , and values near 5 are same facet and very similar argument (Rows 12 and 13). Both Arg1 and Arg2 in Row 10 makes the same argument but Arg1 includes additional argumentation. In Row 12, there is very low Ngram overlap, but strong AFS and NLRD performs better than the other models, and LIWC performs well by itself.",
"cite_spans": [
{
"start": 467,
"end": 486,
"text": "(Hall et al., 2005)",
"ref_id": "BIBREF14"
},
{
"start": 880,
"end": 898,
"text": "(Han et al., 2013)",
"ref_id": "BIBREF15"
},
{
"start": 1557,
"end": 1575,
"text": "(Han et al., 2013)",
"ref_id": "BIBREF15"
},
{
"start": 2189,
"end": 2197,
"text": "(Rows 7,",
"ref_id": null
},
{
"start": 2198,
"end": 2200,
"text": "8)",
"ref_id": null
}
],
"ref_spans": [
{
"start": 244,
"end": 251,
"text": "Fig. 2a",
"ref_id": "FIGREF1"
},
{
"start": 489,
"end": 496,
"text": "Table 2",
"ref_id": "TABREF2"
},
{
"start": 685,
"end": 692,
"text": "Table 3",
"ref_id": "TABREF4"
},
{
"start": 1881,
"end": 1888,
"text": "Table 4",
"ref_id": "TABREF6"
},
{
"start": 1944,
"end": 1951,
"text": "Table 4",
"ref_id": "TABREF6"
},
{
"start": 1994,
"end": 2000,
"text": "Fig. 6",
"ref_id": "FIGREF5"
}
],
"eq_spans": [],
"section": "Results",
"sec_num": "3.2"
},
{
"text": "In Row 1, UMBC performs the best with a predicted score of 0.37 as opposed to an AFS score of 0.00. Other rows where UMBC on its own provides the best performance are highlighted in the table with Arg1 and Arg2 in bold. The top performance of NLRD in Row 5 without UMBC perhaps arises from the semantic information that extermination and holocau are somehow related. NGRAM overlap does the best in Row 13 despite the fact that the phrase No one argues the point that does not participate in the NGRAM overlap.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Results",
"sec_num": "3.2"
},
{
"text": "Our approach draws on three different strands of related work: (1) argument mining; (2) semantic textual similarity; and (3) dialog summarization, which we discuss and compare with our work below. Argument Mining. The study of the structure of arguments has a long tradition in logic, rhetoric and psychology (Walton et al., 2008; Reed and Rowe, 2004; Walton, 2009; Gilbert, 1997; Jackson and Jacobs, 1980; Madnani et al., 2012) . Much of this work has been on formal (legal or political) argumentation, and the small computational literature that has applied the rhetorical categories of this research has likewise focused on formal, monologic text (Feng and Hirst, 2011; Palau and Moens, 2009; Goudas et al., 2014) . More recent work (Ghosh et al., 2014) has attempted to apply these theories to dialogic text in online forums. Ghosh et al. label spans in conversations with attacking moves (CALL-OUTS) and their corresponding argumentative TAR-GETS in another speaker's utterance, and they attempt to learn these callout-target pairs in a supervised framework. Other work attempts to identify general categories of speech-acts such as disagreements or justifications (Misra and Walker, 2015; Biran and Rambow, 2011) .",
"cite_spans": [
{
"start": 309,
"end": 330,
"text": "(Walton et al., 2008;",
"ref_id": "BIBREF38"
},
{
"start": 331,
"end": 351,
"text": "Reed and Rowe, 2004;",
"ref_id": "BIBREF34"
},
{
"start": 352,
"end": 365,
"text": "Walton, 2009;",
"ref_id": "BIBREF39"
},
{
"start": 366,
"end": 380,
"text": "Gilbert, 1997;",
"ref_id": "BIBREF11"
},
{
"start": 381,
"end": 406,
"text": "Jackson and Jacobs, 1980;",
"ref_id": "BIBREF18"
},
{
"start": 407,
"end": 428,
"text": "Madnani et al., 2012)",
"ref_id": "BIBREF25"
},
{
"start": 650,
"end": 672,
"text": "(Feng and Hirst, 2011;",
"ref_id": "BIBREF9"
},
{
"start": 673,
"end": 695,
"text": "Palau and Moens, 2009;",
"ref_id": "BIBREF31"
},
{
"start": 696,
"end": 716,
"text": "Goudas et al., 2014)",
"ref_id": "BIBREF12"
},
{
"start": 736,
"end": 756,
"text": "(Ghosh et al., 2014)",
"ref_id": "BIBREF10"
},
{
"start": 1170,
"end": 1194,
"text": "(Misra and Walker, 2015;",
"ref_id": "BIBREF28"
},
{
"start": 1195,
"end": 1218,
"text": "Biran and Rambow, 2011)",
"ref_id": "BIBREF4"
}
],
"ref_spans": [],
"eq_spans": [],
"section": "Related Work",
"sec_num": "4"
},
{
"text": "What unites all of the above approaches is an interest in understanding the detailed rheotrical structure of a particular linguistic interaction (monologic or dialogic). Our present work is focused instead on inducing the recurring FACETS in a particular topic domain via weakly supervised learning over several dialogic interactions. Several different threads of recent research on argument mining have strong parallels with this goal (Conrad et al., 2012; Boltuzic an\u010f Snajder, 2014; Hasan and Ng, 2014) .",
"cite_spans": [
{
"start": 436,
"end": 457,
"text": "(Conrad et al., 2012;",
"ref_id": "BIBREF8"
},
{
"start": 458,
"end": 485,
"text": "Boltuzic an\u010f Snajder, 2014;",
"ref_id": null
},
{
"start": 486,
"end": 505,
"text": "Hasan and Ng, 2014)",
"ref_id": "BIBREF16"
}
],
"ref_spans": [],
"eq_spans": [],
"section": "Related Work",
"sec_num": "4"
},
{
"text": "Conrad & Wiebe construct an argument mining system on monologic weblog and news data about universal healthcare. One component of their system identifies ARGUING SEGMENTS and the second component labels the segments with the relevant stance-specific ARGUMENT TAGS. They show that distributional similarity features help identify arguments that belong to the same tag set (notably, we did not find distributional similarity helpful for AFS.) Boltuzic & Snajder pursue argument mining on comment streams. Instead of hand-generating argument tags like Conrad & Wiebe, they select short sentential summaries of the key arguments for a given topic from a debate website, and then label comments on the same topic from a different website with the most closely matching summary. The same problem on debate posts is tackled as a \"reason classification\" problem (Hasan and Ng, 2014), with a probabilistic framework for argument recognition (reason classification) that operates jointly with the related task of stance classification.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Related Work",
"sec_num": "4"
},
{
"text": "All of these approaches differ from ours in three respects. First, they all assume a finite set of topicspecific labels that are determined in some form by the researchers themselves. In contrast, we seek to uncover popular facets via clustering the central propositions across the dialogs. After our own initial categorical efforts, we feel that the argument \"topics\" have such nuance that they resist clear labels or category membership. Instead, we feel that a scale such as AFS is a better fit, both for the diversity of the data itself and for the idea of inducing FACETS bottom up. Second, these approaches assume the labels are dependent on a particular stance towards an issue, whereas our facets are deliberately designed to unify across stance disagreement. Finally, all other approaches in argument mining work from the source text itself. We instead (to our knowledge, for the first time) work from human summaries of dialogs because it is an open question whether the CENTRAL PROPOSITIONS for a dialog are really identifiable as continuous spans of text in the dialog itself. (Indeed, our corpus will allow us to determine how true that assumption is.) Semantic Textual Similarity. There appears to be similarity between FACET induction and aspect learning in sentiment analysis (Brody and Elhadad, 2010) , but FACETS are propositional abstract objects, while aspects can usually be described as nouns or properties. Facet induction is more similar to work on STS (Mihalcea et al., 2006; Yeh et al., 2009; Agirre et al., 2012; Han et al., 2013; Jurgens et al., 2014) . Calculating similarity is a central aspect of AFS. Our scale and MT task for AFS was inspired by the STS task and definition. In addition, as a baseline we apply an off-the-shelf system that calculates STS (UMBC) and compare it with our own system (Han et al., 2013) . In order to avoid asking for judgements for many unrelated arguments (CEN-TRAL PROPOSITIONS), and to make the AFS task more doable for Turkers, we also use UMBC as a filter on pairs of CENTRAL PROPOSITIONS as part of making our HIT. This biases the distribution of the training set to having a much larger set of more similar pairs, which has been a problem for previous work (Boltuzic and\u0160najder, 2014) , where the vast majority of pairs that were labelled were unrelated. However the AFS task is clearly different than STS, partly because the data is dialogic and partly because it is argumentative. Our results show that we can improve on STS systems for the AFS task. Dialog Summarization. Much previous work on dialog summarization focused on extracting phenomena specific to meetings, such as action items or decisions (Murray et al., 2006; Hsueh and Moore, 2008; Whittaker et al., 2012; Janin et al., 2004; Carletta, 2007) . Other approaches, like our work, use semantic similarity metrics to identify the most central or important utterances of a spoken dialog (Gurevych and Strube, 2004) , but do not attempt to find the FACETS of a set of arguments across multiple dialogs. Another parallel may exist between work on nuclearity in RST and its use in summarization (Marcu, 1999) . However our notion of a CENTRAL PROPOSITION is different than nuclearity in RST, since FACETS are derived from CEN-TRAL PROPOSITIONS that rise to the top of the pyramid across summarizers, and then (via AFS) across many dialogs on a topic, while RST nuclearity is only defined for a span of text by a single speaker.",
"cite_spans": [
{
"start": 1292,
"end": 1317,
"text": "(Brody and Elhadad, 2010)",
"ref_id": "BIBREF6"
},
{
"start": 1477,
"end": 1500,
"text": "(Mihalcea et al., 2006;",
"ref_id": "BIBREF27"
},
{
"start": 1501,
"end": 1518,
"text": "Yeh et al., 2009;",
"ref_id": "BIBREF41"
},
{
"start": 1519,
"end": 1539,
"text": "Agirre et al., 2012;",
"ref_id": "BIBREF1"
},
{
"start": 1540,
"end": 1557,
"text": "Han et al., 2013;",
"ref_id": "BIBREF15"
},
{
"start": 1558,
"end": 1579,
"text": "Jurgens et al., 2014)",
"ref_id": "BIBREF20"
},
{
"start": 1830,
"end": 1848,
"text": "(Han et al., 2013)",
"ref_id": "BIBREF15"
},
{
"start": 2227,
"end": 2254,
"text": "(Boltuzic and\u0160najder, 2014)",
"ref_id": null
},
{
"start": 2676,
"end": 2697,
"text": "(Murray et al., 2006;",
"ref_id": "BIBREF29"
},
{
"start": 2698,
"end": 2720,
"text": "Hsueh and Moore, 2008;",
"ref_id": "BIBREF17"
},
{
"start": 2721,
"end": 2744,
"text": "Whittaker et al., 2012;",
"ref_id": "BIBREF40"
},
{
"start": 2745,
"end": 2764,
"text": "Janin et al., 2004;",
"ref_id": "BIBREF19"
},
{
"start": 2765,
"end": 2780,
"text": "Carletta, 2007)",
"ref_id": "BIBREF7"
},
{
"start": 2920,
"end": 2947,
"text": "(Gurevych and Strube, 2004)",
"ref_id": "BIBREF13"
},
{
"start": 3125,
"end": 3138,
"text": "(Marcu, 1999)",
"ref_id": "BIBREF26"
}
],
"ref_spans": [],
"eq_spans": [],
"section": "Related Work",
"sec_num": "4"
},
{
"text": "Other work examines how social phenomena affect summarization, such as a study of how the politeness level in computer-generated dialogs impacted summaries (Roman et al., 2006) . Emotion naturally occurs in the IAC, and summarizers' orientation to emotion is intriguing. Emotional information has been observed even in summaries of professional chats discussing technology (Zhou and Hovy, 2005) . However the instructions to our Pyramid annotators were to not include information of this type in the pyramids. We are currently collecting an additional summary corpus using a method that we expect to result in more evaluative and emotional assessments in summaries.",
"cite_spans": [
{
"start": 156,
"end": 176,
"text": "(Roman et al., 2006)",
"ref_id": "BIBREF35"
},
{
"start": 373,
"end": 394,
"text": "(Zhou and Hovy, 2005)",
"ref_id": "BIBREF42"
}
],
"ref_spans": [],
"eq_spans": [],
"section": "Related Work",
"sec_num": "4"
},
{
"text": "This paper presents a method and results for extracting FACETS of a topic, across multiple informal arguments on the same topic. We first use human summarization of dialogs as a probe to determine the CENTRAL PROPOSITIONS of each dialog. Then we use clustering in combination with measures of SEMANTIC SIMILARITY to group the CEN-TRAL PROPOSITIONS into the important FACETS of an argument across many different dialogs. Importantly, we do not attempt to enumerate the possible FACETS for an argument in advance, believing that bottom-up discovery of FACETS is a better fit to the problem.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Conclusion",
"sec_num": "5"
},
{
"text": "This paper contributes to the current state of knowledge in three ways: (1) we collected summaries of spontaneously-produced written dialog of high social and political importance (available from http://nlds.soe.ucsc.edu/summarycorpus).",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Conclusion",
"sec_num": "5"
},
{
"text": "(2) we proposed a novel application of the pyramid summarization scheme to the task of FACET induction; and (3) we introduce a new task of ARGUMENT FACET SIMILARITY (AFS) aimed at identifying FACETS across opinionated dialogs and show that we can identify AFS with a correlation of .54 as opposed to a baseline of .46 provided by a system designed for a similar task. We suspect that the summarize-and-collate approach used here could be promisingly applied to produce annotations on a range of subjective, holistic properties of dialog.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Conclusion",
"sec_num": "5"
},
{
"text": "In future work, we aim to expand on this work in several ways. First, we hope to expand summaries, similarity judgments, and systems to several topics beyond gay marriage. We believe, for example, that the features and the system we have trained for AFS will apply to other domains without retraining, since none of the features are topic specific, but we have not shown that. In addition, we aim to develop additional features and improve on the results reported here. For example, we believe that it is possible that other off-the-shelf systems, such as for example one for sentence specificity (Louis and Nenkova, 2011; Louis and Nenkova, 2012) , might possibly help with aspects of this task. In addition, in future, we aim to automatically identify CENTRAL PROPO-SITIONS without the mediation of human summarizers and evaluators. Given the summaries that we have collected for each dialog, we plan to examine the relationship between the contributors to the related pyramid and the original source text, to determine whether indeed there are surface features of the source that would allow us to treat CENTRAL PROPOSITION detection as an extractive task.",
"cite_spans": [
{
"start": 597,
"end": 622,
"text": "(Louis and Nenkova, 2011;",
"ref_id": "BIBREF23"
},
{
"start": 623,
"end": 647,
"text": "Louis and Nenkova, 2012)",
"ref_id": "BIBREF24"
}
],
"ref_spans": [],
"eq_spans": [],
"section": "Conclusion",
"sec_num": "5"
}
],
"back_matter": [
{
"text": "Acknowledgments This work was funded by NSF GRANT IIS-1302668, Grant NPS-BAA-03, and an IARPA Grant on Persuasion in Dialogue to UCSC by subcontract from the University of Maryland.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "acknowledgement",
"sec_num": null
}
],
"bib_entries": {
"BIBREF0": {
"ref_id": "b0",
"title": "How can you say such things?!?: Recognizing Disagreement in Informal Political Argument",
"authors": [
{
"first": "Rob",
"middle": [],
"last": "Abbott",
"suffix": ""
},
{
"first": "Marilyn",
"middle": [],
"last": "Walker",
"suffix": ""
},
{
"first": "Jean",
"middle": [
"E"
],
"last": "Fox Tree",
"suffix": ""
},
{
"first": "Pranav",
"middle": [],
"last": "Anand",
"suffix": ""
},
{
"first": "Robeson",
"middle": [],
"last": "Bowmani",
"suffix": ""
},
{
"first": "Joseph",
"middle": [],
"last": "King",
"suffix": ""
}
],
"year": 2011,
"venue": "Proceedings of the ACL Workshop on Language and Social Media",
"volume": "",
"issue": "",
"pages": "",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Rob Abbott, Marilyn Walker, Jean E. Fox Tree, Pranav Anand, Robeson Bowmani, and Joseph King. 2011. How can you say such things?!?: Recognizing Dis- agreement in Informal Political Argument. In Pro- ceedings of the ACL Workshop on Language and So- cial Media.",
"links": null
},
"BIBREF1": {
"ref_id": "b1",
"title": "Semeval-2012 task 6: A pilot on semantic textual similarity",
"authors": [
{
"first": "Eneko",
"middle": [],
"last": "Agirre",
"suffix": ""
},
{
"first": "Mona",
"middle": [],
"last": "Diab",
"suffix": ""
},
{
"first": "Daniel",
"middle": [],
"last": "Cer",
"suffix": ""
},
{
"first": "Aitor",
"middle": [],
"last": "Gonzalez-Agirre",
"suffix": ""
}
],
"year": 2012,
"venue": "Proceedings of the First Joint Conference on Lexical and Computational Semantics",
"volume": "1",
"issue": "",
"pages": "385--393",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Eneko Agirre, Mona Diab, Daniel Cer, and Aitor Gonzalez-Agirre. 2012. Semeval-2012 task 6: A pilot on semantic textual similarity. In Proceedings of the First Joint Conference on Lexical and Computational Semantics-Volume 1: Proceedings of the main confer- ence and the shared task, and Volume 2: Proceedings of the Sixth International Workshop on Semantic Eval- uation, pages 385-393. Association for Computational Linguistics.",
"links": null
},
"BIBREF2": {
"ref_id": "b2",
"title": "Sem 2013 shared task: Semantic textual similarity, including a pilot on typed-similarity",
"authors": [
{
"first": "Eneko",
"middle": [],
"last": "Agirre",
"suffix": ""
},
{
"first": "Daniel",
"middle": [],
"last": "Cer",
"suffix": ""
},
{
"first": "Mona",
"middle": [],
"last": "Diab",
"suffix": ""
},
{
"first": "Aitor",
"middle": [],
"last": "Gonzalez-Agirre",
"suffix": ""
},
{
"first": "Weiwei",
"middle": [],
"last": "Guo",
"suffix": ""
}
],
"year": 2013,
"venue": "* SEM 2013: The Second Joint Conference on Lexical and Computational Semantics",
"volume": "",
"issue": "",
"pages": "",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Eneko Agirre, Daniel Cer, Mona Diab, Aitor Gonzalez- Agirre, and Weiwei Guo. 2013. Sem 2013 shared task: Semantic textual similarity, including a pilot on typed-similarity. In In* SEM 2013: The Second Joint Conference on Lexical and Computational Semantics. Association for Computational Linguistics. Citeseer.",
"links": null
},
"BIBREF3": {
"ref_id": "b3",
"title": "Probabilistic soft logic for semantic textual similarity",
"authors": [
{
"first": "Islam",
"middle": [],
"last": "Beltagy",
"suffix": ""
},
{
"first": "Katrin",
"middle": [],
"last": "Erk",
"suffix": ""
},
{
"first": "Raymond",
"middle": [],
"last": "Mooney",
"suffix": ""
}
],
"year": 2014,
"venue": "Proceedings of Association for Computational Linguistics (ACL-14)",
"volume": "",
"issue": "",
"pages": "",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Islam Beltagy, Katrin Erk, and Raymond Mooney. 2014. Probabilistic soft logic for semantic textual similar- ity. Proceedings of Association for Computational Linguistics (ACL-14).",
"links": null
},
"BIBREF4": {
"ref_id": "b4",
"title": "Identifying justifications in written dialogs",
"authors": [
{
"first": "O",
"middle": [],
"last": "Biran",
"suffix": ""
},
{
"first": "O",
"middle": [],
"last": "Rambow",
"suffix": ""
}
],
"year": 2011,
"venue": "2011 Fifth IEEE International Conference on Semantic Computing (ICSC)",
"volume": "",
"issue": "",
"pages": "162--168",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "O. Biran and O. Rambow. 2011. Identifying justifica- tions in written dialogs. In 2011 Fifth IEEE Inter- national Conference on Semantic Computing (ICSC), pages 162-168. IEEE.",
"links": null
},
"BIBREF5": {
"ref_id": "b5",
"title": "Back up your stance: Recognizing arguments in online discussions",
"authors": [
{
"first": "Filip",
"middle": [],
"last": "Boltuzic",
"suffix": ""
},
{
"first": "Jan\u0161najder",
"middle": [],
"last": "",
"suffix": ""
}
],
"year": 2014,
"venue": "Proceedings of the First Workshop on Argumentation Mining",
"volume": "",
"issue": "",
"pages": "49--58",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Filip Boltuzic and Jan\u0160najder. 2014. Back up your stance: Recognizing arguments in online discussions. In Proceedings of the First Workshop on Argumenta- tion Mining, pages 49-58.",
"links": null
},
"BIBREF6": {
"ref_id": "b6",
"title": "An unsupervised aspect-sentiment model for online reviews",
"authors": [
{
"first": "Samuel",
"middle": [],
"last": "Brody",
"suffix": ""
},
{
"first": "Noemie",
"middle": [],
"last": "Elhadad",
"suffix": ""
}
],
"year": 2010,
"venue": "Human Language Technologies: The 2010 Annual Conference of the North American Chapter of the Association for Computational Linguistics",
"volume": "",
"issue": "",
"pages": "804--812",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Samuel Brody and Noemie Elhadad. 2010. An unsu- pervised aspect-sentiment model for online reviews. In Human Language Technologies: The 2010 Annual Conference of the North American Chapter of the As- sociation for Computational Linguistics, pages 804- 812. Association for Computational Linguistics.",
"links": null
},
"BIBREF7": {
"ref_id": "b7",
"title": "Unleashing the killer corpus: experiences in creating the multi-everything ami meeting corpus",
"authors": [
{
"first": "Jean",
"middle": [],
"last": "Carletta",
"suffix": ""
}
],
"year": 2007,
"venue": "Language Resources and Evaluation",
"volume": "41",
"issue": "2",
"pages": "181--190",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Jean Carletta. 2007. Unleashing the killer corpus: ex- periences in creating the multi-everything ami meet- ing corpus. Language Resources and Evaluation, 41(2):181-190.",
"links": null
},
"BIBREF8": {
"ref_id": "b8",
"title": "Recognizing arguing subjectivity and argument tags",
"authors": [
{
"first": "Alexander",
"middle": [],
"last": "Conrad",
"suffix": ""
},
{
"first": "Janyce",
"middle": [],
"last": "Wiebe",
"suffix": ""
}
],
"year": 2012,
"venue": "Proceedings of the Workshop on Extra-Propositional Aspects of Meaning in Computational Linguistics",
"volume": "",
"issue": "",
"pages": "80--88",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Alexander Conrad, Janyce Wiebe, et al. 2012. Recog- nizing arguing subjectivity and argument tags. In Pro- ceedings of the Workshop on Extra-Propositional As- pects of Meaning in Computational Linguistics, pages 80-88. Association for Computational Linguistics.",
"links": null
},
"BIBREF9": {
"ref_id": "b9",
"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": "1",
"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 Computational Linguistics: Human Language Technologies-Volume 1, pages 987-996. Association for Computational Lin- guistics.",
"links": null
},
"BIBREF10": {
"ref_id": "b10",
"title": "Analyzing argumentative discourse units in online interactions",
"authors": [
{
"first": "Debanjan",
"middle": [],
"last": "Ghosh",
"suffix": ""
},
{
"first": "Smaranda",
"middle": [],
"last": "Muresan",
"suffix": ""
},
{
"first": "Nina",
"middle": [],
"last": "Wacholder",
"suffix": ""
},
{
"first": "Mark",
"middle": [],
"last": "Aakhus",
"suffix": ""
},
{
"first": "Matthew",
"middle": [],
"last": "Mitsui",
"suffix": ""
}
],
"year": 2014,
"venue": "",
"volume": "",
"issue": "",
"pages": "",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Debanjan Ghosh, Smaranda Muresan, Nina Wacholder, Mark Aakhus, and Matthew Mitsui. 2014. Analyzing argumentative discourse units in online interactions. ACL 2014, page 39.",
"links": null
},
"BIBREF11": {
"ref_id": "b11",
"title": "Coalescent argumentation",
"authors": [
{
"first": "Michael",
"middle": [
"A"
],
"last": "Gilbert",
"suffix": ""
}
],
"year": 1997,
"venue": "",
"volume": "",
"issue": "",
"pages": "",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Michael A. Gilbert. 1997. Coalescent argumentation.",
"links": null
},
"BIBREF12": {
"ref_id": "b12",
"title": "Argument extraction from news, blogs, and social media",
"authors": [
{
"first": "Theodosis",
"middle": [],
"last": "Goudas",
"suffix": ""
},
{
"first": "Christos",
"middle": [],
"last": "Louizos",
"suffix": ""
}
],
"year": 2014,
"venue": "Artificial Intelligence: Methods and Applications",
"volume": "",
"issue": "",
"pages": "287--299",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Theodosis Goudas, Christos Louizos, Georgios Petasis, and Vangelis Karkaletsis. 2014. Argument extraction from news, blogs, and social media. In Artificial In- telligence: Methods and Applications, pages 287-299. Springer.",
"links": null
},
"BIBREF13": {
"ref_id": "b13",
"title": "Semantic similarity applied to spoken dialogue summarization",
"authors": [
{
"first": "I",
"middle": [],
"last": "Gurevych",
"suffix": ""
},
{
"first": "M",
"middle": [],
"last": "Strube",
"suffix": ""
}
],
"year": 2004,
"venue": "Proceedings of the 20th international conference on Computational Linguistics",
"volume": "",
"issue": "",
"pages": "764--771",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "I. Gurevych and M. Strube. 2004. Semantic similarity applied to spoken dialogue summarization. In Pro- ceedings of the 20th international conference on Com- putational Linguistics, pages 764-771. ACL.",
"links": null
},
"BIBREF14": {
"ref_id": "b14",
"title": "The weka data mining software: An update",
"authors": [
{
"first": "M",
"middle": [],
"last": "Hall",
"suffix": ""
},
{
"first": "F",
"middle": [],
"last": "Eibe",
"suffix": ""
},
{
"first": "G",
"middle": [],
"last": "Holms",
"suffix": ""
},
{
"first": "B",
"middle": [],
"last": "Pfahringer",
"suffix": ""
},
{
"first": "P",
"middle": [],
"last": "Reutemann",
"suffix": ""
},
{
"first": "I",
"middle": [],
"last": "Witten",
"suffix": ""
}
],
"year": 2005,
"venue": "SIGKDD Explorations",
"volume": "",
"issue": "1",
"pages": "",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "M. Hall, F. Eibe, G. Holms, B. Pfahringer, P. Reutemann, and I. Witten. 2005. The weka data mining software: An update. SIGKDD Explorations, 11(1).",
"links": null
},
"BIBREF15": {
"ref_id": "b15",
"title": "Umbc ebiquitycore: Semantic textual similarity systems",
"authors": [
{
"first": "Lushan",
"middle": [],
"last": "Han",
"suffix": ""
},
{
"first": "Abhay",
"middle": [],
"last": "Kashyap",
"suffix": ""
},
{
"first": "Tim",
"middle": [],
"last": "Finin",
"suffix": ""
},
{
"first": "James",
"middle": [],
"last": "Mayfield",
"suffix": ""
},
{
"first": "Jonathan",
"middle": [],
"last": "Weese",
"suffix": ""
}
],
"year": 2013,
"venue": "",
"volume": "",
"issue": "",
"pages": "",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Lushan Han, Abhay Kashyap, Tim Finin, James May- field, and Jonathan Weese. 2013. Umbc ebiquity- core: Semantic textual similarity systems. Atlanta, Georgia, USA, page 44.",
"links": null
},
"BIBREF16": {
"ref_id": "b16",
"title": "Why are you taking this stance? identifying and classifying reasons in ideological debates",
"authors": [
{
"first": "Saidul",
"middle": [],
"last": "Kazi",
"suffix": ""
},
{
"first": "Vincent",
"middle": [],
"last": "Hasan",
"suffix": ""
},
{
"first": "",
"middle": [],
"last": "Ng",
"suffix": ""
}
],
"year": 2014,
"venue": "Proceedings of the Conference on Empirical Methods in Natural Language Processing",
"volume": "",
"issue": "",
"pages": "",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Kazi Saidul Hasan and Vincent Ng. 2014. Why are you taking this stance? identifying and classifying reasons in ideological debates. In Proceedings of the Confer- ence on Empirical Methods in Natural Language Pro- cessing.",
"links": null
},
"BIBREF17": {
"ref_id": "b17",
"title": "Automatic decision detection in meeting speech",
"authors": [
{
"first": "P",
"middle": [
"Y"
],
"last": "Hsueh",
"suffix": ""
},
{
"first": "J",
"middle": [],
"last": "Moore",
"suffix": ""
}
],
"year": 2008,
"venue": "Machine Learning for Multimodal Interaction",
"volume": "",
"issue": "",
"pages": "168--179",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "P.Y. Hsueh and J. Moore. 2008. Automatic decision de- tection in meeting speech. Machine Learning for Mul- timodal Interaction, pages 168-179.",
"links": null
},
"BIBREF18": {
"ref_id": "b18",
"title": "Structure of conversational argument: Pragmatic bases for the enthymeme",
"authors": [
{
"first": "Sally",
"middle": [],
"last": "Jackson",
"suffix": ""
},
{
"first": "Scott",
"middle": [],
"last": "Jacobs",
"suffix": ""
}
],
"year": 1980,
"venue": "Quarterly Journal of Speech",
"volume": "66",
"issue": "3",
"pages": "251--265",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Sally Jackson and Scott Jacobs. 1980. Structure of conversational argument: Pragmatic bases for the en- thymeme. Quarterly Journal of Speech, 66(3):251- 265.",
"links": null
},
"BIBREF19": {
"ref_id": "b19",
"title": "The icsi meeting project: Resources and research",
"authors": [
{
"first": "A",
"middle": [],
"last": "Janin",
"suffix": ""
},
{
"first": "J",
"middle": [],
"last": "Ang",
"suffix": ""
},
{
"first": "S",
"middle": [],
"last": "Bhagat",
"suffix": ""
},
{
"first": "R",
"middle": [],
"last": "Dhillon",
"suffix": ""
},
{
"first": "J",
"middle": [],
"last": "Edwards",
"suffix": ""
},
{
"first": "J",
"middle": [],
"last": "Macias-Guarasa",
"suffix": ""
},
{
"first": "N",
"middle": [],
"last": "Morgan",
"suffix": ""
},
{
"first": "B",
"middle": [],
"last": "Peskin",
"suffix": ""
},
{
"first": "E",
"middle": [],
"last": "Shriberg",
"suffix": ""
},
{
"first": "A",
"middle": [],
"last": "Stolcke",
"suffix": ""
}
],
"year": 2004,
"venue": "Proceedings of the 2004 ICASSP NIST Meeting Recognition Workshop",
"volume": "",
"issue": "",
"pages": "",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "A. Janin, J. Ang, S. Bhagat, R. Dhillon, J. Edwards, J. Macias-Guarasa, N. Morgan, B. Peskin, E. Shriberg, A. Stolcke, et al. 2004. The icsi meeting project: Resources and research. In Proceedings of the 2004 ICASSP NIST Meeting Recognition Workshop.",
"links": null
},
"BIBREF20": {
"ref_id": "b20",
"title": "Semeval-2014 task 3: Cross-level semantic similarity",
"authors": [
{
"first": "David",
"middle": [],
"last": "Jurgens",
"suffix": ""
},
{
"first": "Mohammad",
"middle": [],
"last": "Taher Pilehvar",
"suffix": ""
},
{
"first": "Roberto",
"middle": [],
"last": "Navigli",
"suffix": ""
}
],
"year": 2014,
"venue": "SemEval",
"volume": "",
"issue": "",
"pages": "",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "David Jurgens, Mohammad Taher Pilehvar, and Roberto Navigli. 2014. Semeval-2014 task 3: Cross-level se- mantic similarity. SemEval 2014, page 17.",
"links": null
},
"BIBREF21": {
"ref_id": "b21",
"title": "Disco: A multilingual database of distributionally similar words",
"authors": [
{
"first": "P",
"middle": [],
"last": "Kolb",
"suffix": ""
}
],
"year": 2008,
"venue": "Proceedings of KONVENS",
"volume": "",
"issue": "",
"pages": "",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "P. Kolb. 2008. Disco: A multilingual database of distributionally similar words. In Proceedings of KONVENS-2008.",
"links": null
},
"BIBREF22": {
"ref_id": "b22",
"title": "Rouge: A package for automatic evaluation of summaries rouge: A package for automatic evaluation of summaries",
"authors": [
{
"first": "C.-Y",
"middle": [],
"last": "Lin",
"suffix": ""
}
],
"year": 2004,
"venue": "Proceedings of the Workshop on Text Summarization Branches Out",
"volume": "",
"issue": "",
"pages": "",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "C.-Y. Lin. 2004. Rouge: A package for automatic evaluation of summaries rouge: A package for auto- matic evaluation of summaries. In Proceedings of the Workshop on Text Summarization Branches Out (WAS 2004).",
"links": null
},
"BIBREF23": {
"ref_id": "b23",
"title": "Automatic identification of general and specific sentences by leveraging discourse annotations",
"authors": [
{
"first": "Annie",
"middle": [],
"last": "Louis",
"suffix": ""
},
{
"first": "Ani",
"middle": [],
"last": "Nenkova",
"suffix": ""
}
],
"year": 2011,
"venue": "IJCNLP",
"volume": "",
"issue": "",
"pages": "605--613",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Annie Louis and Ani Nenkova. 2011. Automatic identi- fication of general and specific sentences by leveraging discourse annotations. In IJCNLP, pages 605-613.",
"links": null
},
"BIBREF24": {
"ref_id": "b24",
"title": "A corpus of general and specific sentences from news",
"authors": [
{
"first": "Annie",
"middle": [],
"last": "Louis",
"suffix": ""
},
{
"first": "Ani",
"middle": [],
"last": "Nenkova",
"suffix": ""
}
],
"year": 2012,
"venue": "LREC",
"volume": "",
"issue": "",
"pages": "1818--1821",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Annie Louis and Ani Nenkova. 2012. A corpus of gen- eral and specific sentences from news. In LREC, pages 1818-1821.",
"links": null
},
"BIBREF25": {
"ref_id": "b25",
"title": "Identifying high-level organizational elements in argumentative discourse",
"authors": [
{
"first": "Nitin",
"middle": [],
"last": "Madnani",
"suffix": ""
},
{
"first": "Michael",
"middle": [],
"last": "Heilman",
"suffix": ""
},
{
"first": "Joel",
"middle": [],
"last": "Tetreault",
"suffix": ""
},
{
"first": "Martin",
"middle": [],
"last": "Chodorow",
"suffix": ""
}
],
"year": 2012,
"venue": "Proceedings of the 2012 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies, NAACL HLT '12",
"volume": "",
"issue": "",
"pages": "20--28",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Nitin Madnani, Michael Heilman, Joel Tetreault, and Martin Chodorow. 2012. Identifying high-level or- ganizational elements in argumentative discourse. In Proceedings of the 2012 Conference of the North American Chapter of the Association for Computa- tional Linguistics: Human Language Technologies, NAACL HLT '12, pages 20-28, Stroudsburg, PA, USA. Association for Computational Linguistics.",
"links": null
},
"BIBREF26": {
"ref_id": "b26",
"title": "Discourse trees are good indicators of importance in text",
"authors": [
{
"first": "Daniel",
"middle": [],
"last": "Marcu",
"suffix": ""
}
],
"year": 1999,
"venue": "Advances in automatic text summarization",
"volume": "",
"issue": "",
"pages": "123--136",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Daniel Marcu. 1999. Discourse trees are good indica- tors of importance in text. Advances in automatic text summarization, pages 123-136.",
"links": null
},
"BIBREF27": {
"ref_id": "b27",
"title": "Corpus-based and knowledge-based measures of text semantic similarity",
"authors": [
{
"first": "Rada",
"middle": [],
"last": "Mihalcea",
"suffix": ""
},
{
"first": "Courtney",
"middle": [],
"last": "Corley",
"suffix": ""
},
{
"first": "Carlo",
"middle": [],
"last": "Strapparava",
"suffix": ""
}
],
"year": 2006,
"venue": "AAAI",
"volume": "6",
"issue": "",
"pages": "775--780",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Rada Mihalcea, Courtney Corley, and Carlo Strapparava. 2006. Corpus-based and knowledge-based measures of text semantic similarity. In AAAI, volume 6, pages 775-780.",
"links": null
},
"BIBREF28": {
"ref_id": "b28",
"title": "Topic independent identification of agreement and disagreement in social media dialogue",
"authors": [
{
"first": "Amita",
"middle": [],
"last": "Misra",
"suffix": ""
},
{
"first": "Marilyn",
"middle": [
"A"
],
"last": "Walker",
"suffix": ""
}
],
"year": 2015,
"venue": "Proceedings of the SIG-DIAL 2013 Conference: The 15th Annual Meeting of the Special Interest Group on Discourse and Dialogue",
"volume": "",
"issue": "",
"pages": "",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Amita Misra and Marilyn A Walker. 2015. Topic inde- pendent identification of agreement and disagreement in social media dialogue. In Proceedings of the SIG- DIAL 2013 Conference: The 15th Annual Meeting of the Special Interest Group on Discourse and Dialogue.",
"links": null
},
"BIBREF29": {
"ref_id": "b29",
"title": "Incorporating speaker and discourse features into speech summarization",
"authors": [
{
"first": "G",
"middle": [],
"last": "Murray",
"suffix": ""
},
{
"first": "S",
"middle": [],
"last": "Renals",
"suffix": ""
},
{
"first": "J",
"middle": [],
"last": "Carletta",
"suffix": ""
},
{
"first": "J",
"middle": [],
"last": "Moore",
"suffix": ""
}
],
"year": 2006,
"venue": "Proceedings of the main conference on Human Language Technology Conference of the North American Chapter of the Association of Computational Linguistics",
"volume": "",
"issue": "",
"pages": "367--374",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "G. Murray, S. Renals, J. Carletta, and J. Moore. 2006. In- corporating speaker and discourse features into speech summarization. In Proceedings of the main conference on Human Language Technology Conference of the North American Chapter of the Association of Com- putational Linguistics, pages 367-374. Association for Computational Linguistics.",
"links": null
},
"BIBREF30": {
"ref_id": "b30",
"title": "Evaluating content selection in summarization: The pyramid method",
"authors": [
{
"first": "A",
"middle": [],
"last": "Nenkova",
"suffix": ""
},
{
"first": "R",
"middle": [],
"last": "Passonneau",
"suffix": ""
}
],
"year": 2004,
"venue": "Proceedings of HLT-NAACL",
"volume": "",
"issue": "",
"pages": "",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "A. Nenkova and R. Passonneau. 2004. Evaluating con- tent selection in summarization: The pyramid method. In Proceedings of HLT-NAACL, volume 2004.",
"links": null
},
"BIBREF31": {
"ref_id": "b31",
"title": "Argumentation mining: the detection, classification and structure of a rguments in text",
"authors": [
{
"first": "Raquel",
"middle": [],
"last": "Mochales Palau",
"suffix": ""
},
{
"first": "Marie-Francine",
"middle": [],
"last": "Moens",
"suffix": ""
}
],
"year": 2009,
"venue": "Proceedings of the 12th international conference on artificial int elligence and law",
"volume": "",
"issue": "",
"pages": "98--107",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Raquel Mochales Palau and Marie-Francine Moens. 2009. Argumentation mining: the detection, classifi- cation and structure of a rguments in text. In Proceed- ings of the 12th international conference on artificial int elligence and law, pages 98-107. ACM.",
"links": null
},
"BIBREF32": {
"ref_id": "b32",
"title": "Scikit-learn: Machine learning in Python",
"authors": [
{
"first": "F",
"middle": [],
"last": "Pedregosa",
"suffix": ""
},
{
"first": "G",
"middle": [],
"last": "Varoquaux",
"suffix": ""
},
{
"first": "A",
"middle": [],
"last": "Gramfort",
"suffix": ""
},
{
"first": "V",
"middle": [],
"last": "Michel",
"suffix": ""
},
{
"first": "B",
"middle": [],
"last": "Thirion",
"suffix": ""
},
{
"first": "O",
"middle": [],
"last": "Grisel",
"suffix": ""
},
{
"first": "M",
"middle": [],
"last": "Blondel",
"suffix": ""
},
{
"first": "P",
"middle": [],
"last": "Prettenhofer",
"suffix": ""
},
{
"first": "R",
"middle": [],
"last": "Weiss",
"suffix": ""
},
{
"first": "V",
"middle": [],
"last": "Dubourg",
"suffix": ""
},
{
"first": "J",
"middle": [],
"last": "Vanderplas",
"suffix": ""
},
{
"first": "A",
"middle": [],
"last": "Passos",
"suffix": ""
},
{
"first": "D",
"middle": [],
"last": "Cournapeau",
"suffix": ""
},
{
"first": "M",
"middle": [],
"last": "Brucher",
"suffix": ""
},
{
"first": "M",
"middle": [],
"last": "Perrot",
"suffix": ""
},
{
"first": "E",
"middle": [],
"last": "Duchesnay",
"suffix": ""
}
],
"year": 2011,
"venue": "Journal of Machine Learning Research",
"volume": "12",
"issue": "",
"pages": "2825--2830",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "F. Pedregosa, G. Varoquaux, A. Gramfort, V. Michel, B. Thirion, O. Grisel, M. Blondel, P. Prettenhofer, R. Weiss, V. Dubourg, J. Vanderplas, A. Passos, D. Cournapeau, M. Brucher, M. Perrot, and E. Duches- nay. 2011. Scikit-learn: Machine learning in Python. Journal of Machine Learning Research, 12:2825- 2830.",
"links": null
},
"BIBREF33": {
"ref_id": "b33",
"title": "LIWC: Linguistic Inquiry and Word Count",
"authors": [
{
"first": "J",
"middle": [
"W"
],
"last": "Pennebaker",
"suffix": ""
},
{
"first": "L",
"middle": [
"E"
],
"last": "Francis",
"suffix": ""
},
{
"first": "R",
"middle": [
"J"
],
"last": "Booth",
"suffix": ""
}
],
"year": 2001,
"venue": "",
"volume": "",
"issue": "",
"pages": "",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "J. W. Pennebaker, L. E. Francis, and R. J. Booth, 2001. LIWC: Linguistic Inquiry and Word Count.",
"links": null
},
"BIBREF34": {
"ref_id": "b34",
"title": "Araucaria: Software for argument analysis, diagramming and representation",
"authors": [
{
"first": "Chris",
"middle": [],
"last": "Reed",
"suffix": ""
},
{
"first": "Glenn",
"middle": [],
"last": "Rowe",
"suffix": ""
}
],
"year": 2004,
"venue": "International Journal on Artificial Intelligence Tools",
"volume": "13",
"issue": "04",
"pages": "961--979",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Chris Reed and Glenn Rowe. 2004. Araucaria: Software for argument analysis, diagramming and representa- tion. International Journal on Artificial Intelligence Tools, 13(04):961-979.",
"links": null
},
"BIBREF35": {
"ref_id": "b35",
"title": "Politeness and bias in dialogue summarization: two exploratory studies",
"authors": [
{
"first": "N",
"middle": [],
"last": "Roman",
"suffix": ""
},
{
"first": "P",
"middle": [],
"last": "Piwek",
"suffix": ""
},
{
"first": "P",
"middle": [],
"last": "Carvalho",
"suffix": ""
},
{
"first": "M",
"middle": [
"B R"
],
"last": "Ariadne",
"suffix": ""
}
],
"year": 2006,
"venue": "Computing attitude and affect in text: theory and applications",
"volume": "20",
"issue": "",
"pages": "",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "N. Roman, P. Piwek, P. Carvalho, and M. B. R. Ariadne. 2006. Politeness and bias in dialogue summarization: two exploratory studies. In J. Shanahan, Y. Qu, and J. Wiebe, editors, Computing attitude and affect in text: theory and applications, volume 20 of The In- formation Retrieval Series. Springer.",
"links": null
},
"BIBREF36": {
"ref_id": "b36",
"title": "Recognizing stances in ideological on-line debates",
"authors": [
{
"first": "S",
"middle": [],
"last": "Somasundaran",
"suffix": ""
},
{
"first": "J",
"middle": [],
"last": "Wiebe",
"suffix": ""
}
],
"year": 2010,
"venue": "Proceedings of the NAACL HLT 2010 Workshop on Computational Approaches to Analysis and Generation of Emotion in Text",
"volume": "",
"issue": "",
"pages": "116--124",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "S. Somasundaran and J. Wiebe. 2010. Recognizing stances in ideological on-line debates. In Proceedings of the NAACL HLT 2010 Workshop on Computational Approaches to Analysis and Generation of Emotion in Text, pages 116-124. Association for Computational Linguistics.",
"links": null
},
"BIBREF37": {
"ref_id": "b37",
"title": "That's your evidence?: Classifying stance in online political debate",
"authors": [
{
"first": "Marilyn",
"middle": [],
"last": "Walker",
"suffix": ""
},
{
"first": "Pranav",
"middle": [],
"last": "Anand",
"suffix": ""
},
{
"first": "Rob",
"middle": [],
"last": "Abbott",
"suffix": ""
},
{
"first": "Jean",
"middle": [
"E"
],
"last": "Fox Tree",
"suffix": ""
},
{
"first": "Craig",
"middle": [],
"last": "Martell",
"suffix": ""
},
{
"first": "Joseph",
"middle": [],
"last": "King",
"suffix": ""
}
],
"year": 2012,
"venue": "",
"volume": "",
"issue": "",
"pages": "",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Marilyn Walker, Pranav Anand, Rob Abbott, Jean E. Fox Tree, Craig Martell, and Joseph King. 2012. That's your evidence?: Classifying stance in online political debate. Decision Support Sciences.",
"links": null
},
"BIBREF38": {
"ref_id": "b38",
"title": "Argumentation Schemes",
"authors": [
{
"first": "Douglas",
"middle": [],
"last": "Walton",
"suffix": ""
},
{
"first": "Chris",
"middle": [],
"last": "Reed",
"suffix": ""
},
{
"first": "Fabrizio",
"middle": [],
"last": "Macagno",
"suffix": ""
}
],
"year": 2008,
"venue": "",
"volume": "",
"issue": "",
"pages": "",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Douglas Walton, Chris Reed, and Fabrizio Macagno. 2008. Argumentation Schemes. Cambridge University Press.",
"links": null
},
"BIBREF39": {
"ref_id": "b39",
"title": "Argumentation theory: A very short introduction",
"authors": [
{
"first": "Douglas",
"middle": [],
"last": "Walton",
"suffix": ""
}
],
"year": 2009,
"venue": "Argumentation in Artificial Intelligence",
"volume": "",
"issue": "",
"pages": "1--22",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Douglas Walton. 2009. Argumentation theory: A very short introduction. In Guillermo Simari and Iyad Rah- wan, editors, Argumentation in Artificial Intelligence, pages 1-22. Springer US.",
"links": null
},
"BIBREF40": {
"ref_id": "b40",
"title": "Markup as you talk: establishing effective memory cues while still contributing to a meeting",
"authors": [
{
"first": "S",
"middle": [],
"last": "Whittaker",
"suffix": ""
},
{
"first": "V",
"middle": [],
"last": "Kalnikait\u00e9",
"suffix": ""
},
{
"first": "P",
"middle": [],
"last": "Ehlen",
"suffix": ""
}
],
"year": 2012,
"venue": "Proceedings of the ACM 2012 conference on Computer Supported Cooperative Work",
"volume": "",
"issue": "",
"pages": "349--358",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "S. Whittaker, V. Kalnikait\u00e9, and P. Ehlen. 2012. Markup as you talk: establishing effective memory cues while still contributing to a meeting. In Proceedings of the ACM 2012 conference on Computer Supported Coop- erative Work, pages 349-358. ACM.",
"links": null
},
"BIBREF41": {
"ref_id": "b41",
"title": "Wikiwalk: random walks on wikipedia for semantic relatedness",
"authors": [
{
"first": "Eric",
"middle": [],
"last": "Yeh",
"suffix": ""
},
{
"first": "Daniel",
"middle": [],
"last": "Ramage",
"suffix": ""
},
{
"first": "D",
"middle": [],
"last": "Christopher",
"suffix": ""
},
{
"first": "Eneko",
"middle": [],
"last": "Manning",
"suffix": ""
},
{
"first": "Aitor",
"middle": [],
"last": "Agirre",
"suffix": ""
},
{
"first": "",
"middle": [],
"last": "Soroa",
"suffix": ""
}
],
"year": 2009,
"venue": "Proceedings of the 2009 Workshop on Graph-based Methods for Natural Language Processing",
"volume": "",
"issue": "",
"pages": "41--49",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Eric Yeh, Daniel Ramage, Christopher D Manning, Eneko Agirre, and Aitor Soroa. 2009. Wikiwalk: random walks on wikipedia for semantic relatedness. In Proceedings of the 2009 Workshop on Graph-based Methods for Natural Language Processing, pages 41- 49. Association for Computational Linguistics.",
"links": null
},
"BIBREF42": {
"ref_id": "b42",
"title": "Digesting virtual geek culture: The summarization of technical internet relay chats",
"authors": [
{
"first": "L",
"middle": [],
"last": "Zhou",
"suffix": ""
},
{
"first": "E",
"middle": [],
"last": "Hovy",
"suffix": ""
}
],
"year": 2005,
"venue": "Proceedings of the 43rd Annual Meeting on Association for Computational Linguistics",
"volume": "",
"issue": "",
"pages": "298--305",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "L. Zhou and E. Hovy. 2005. Digesting virtual geek culture: The summarization of technical internet re- lay chats. In Proceedings of the 43rd Annual Meeting on Association for Computational Linguistics, pages 298-305. ACL.",
"links": null
}
},
"ref_entries": {
"FIGREF0": {
"uris": null,
"num": null,
"text": "Gay Marriage Dialog-1.",
"type_str": "figure"
},
"FIGREF1": {
"uris": null,
"num": null,
"text": "The overall engineering architecture of our approach. (a) Basic engineering approach for extracting CENTRAL PROPOSITIONS and clustering them into argument FACETS across several dialogs; (b) Workflow for 'detecting' central propositions via pyramid evaluation of multiple summaries; (c) Workflow for obtaining gold-standard labels for AFS task.",
"type_str": "figure"
},
"FIGREF2": {
"uris": null,
"num": null,
"text": "Gay Marriage Dialog-2.",
"type_str": "figure"
},
"FIGREF3": {
"uris": null,
"num": null,
"text": "Two of the 5 Summaries for Dialog-2.",
"type_str": "figure"
},
"FIGREF4": {
"uris": null,
"num": null,
"text": "provides 2 of the 5 summaries collected for the dialog inFig. 3.",
"type_str": "figure"
},
"FIGREF5": {
"uris": null,
"num": null,
"text": "Instructions for AFS MT HIT. S5: MT Argument Facet Similarity HIT.Fig. 6",
"type_str": "figure"
},
"TABREF0": {
"content": "<table/>",
"type_str": "table",
"num": null,
"html": null,
"text": ""
},
"TABREF2": {
"content": "<table/>",
"type_str": "table",
"num": null,
"html": null,
"text": "Support Vector and Linear Regression."
},
"TABREF4": {
"content": "<table><tr><td>: Results for Different Individual Features</td></tr><tr><td>and Feature Combinations.</td></tr><tr><td>line (Row 2). It is interesting that Ngram alone out-</td></tr><tr><td>performs distributional measures (which Conrad &</td></tr><tr><td>Wiebe found most helpful) as well as Rouge (which</td></tr><tr><td>contains metrics insensitive to linear adjacency).</td></tr><tr><td>Table 3, Row 15, shows that the best correlation</td></tr><tr><td>that is achievable without UMBC is the combination</td></tr><tr><td>of Ngram, LIWC, ROUGE and DISCO (NLRD).</td></tr><tr><td>This combination significantly improves over the</td></tr><tr><td>UMBC baseline of 0.46 to 0.50 (paired t-test, p <</td></tr><tr><td>.05).</td></tr></table>",
"type_str": "table",
"num": null,
"html": null,
"text": ""
},
"TABREF5": {
"content": "<table><tr><td colspan=\"2\">Row N</td><td>L</td><td>U</td><td colspan=\"3\">NLRD NLRDU MT</td><td>Arg1</td><td/><td>Arg2</td><td/></tr><tr><td/><td/><td/><td/><td/><td/><td>AFS</td><td/><td/><td/><td/></tr><tr><td>1</td><td>1.38</td><td>1.50</td><td>0.37</td><td>1.31</td><td>0.40</td><td>0.00</td><td colspan=\"2\">everyone has the freedom of</td><td colspan=\"2\">service in the military</td></tr><tr><td/><td/><td/><td/><td/><td/><td/><td>speech</td><td/><td/><td/></tr><tr><td>2</td><td>2.00</td><td>2.02</td><td>1.55</td><td>2.33</td><td>1.86</td><td>1.14</td><td colspan=\"2\">gay people should be able to</td><td colspan=\"3\">referring to namecalling and vi-</td></tr><tr><td/><td/><td/><td/><td/><td/><td/><td colspan=\"2\">marry a person of their choice</td><td colspan=\"3\">olence from the original post</td></tr><tr><td/><td/><td/><td/><td/><td/><td/><td colspan=\"2\">and get equal rights</td><td colspan=\"3\">that was opposing gay rights</td></tr><tr><td>3</td><td>2.00</td><td>1.29</td><td>2.52</td><td>1.37</td><td>1.54</td><td>1.33</td><td colspan=\"2\">Constitutional right to be opposed</td><td colspan=\"3\">arguing about marriage benefits</td></tr><tr><td/><td/><td/><td/><td/><td/><td/><td colspan=\"2\">to gay marriage as well as gay</td><td colspan=\"3\">between single people and mar-</td></tr><tr><td/><td/><td/><td/><td/><td/><td/><td>people themselves</td><td/><td>ried</td><td/></tr><tr><td>4</td><td>2.00</td><td>1.70</td><td>2.74</td><td>1.77</td><td>1.98</td><td>1.80</td><td colspan=\"2\">people should not pick and choose</td><td colspan=\"3\">people did not want gay marriage</td></tr><tr><td/><td/><td/><td/><td/><td/><td/><td colspan=\"2\">what they want equal rights on.</td><td/><td/></tr><tr><td>5</td><td>1.38</td><td>1.92</td><td>0.88</td><td>1.94</td><td>1.64</td><td>2.50</td><td colspan=\"2\">the Republicans creating another</td><td colspan=\"3\">No republican in leadership</td></tr><tr><td/><td/><td/><td/><td/><td/><td/><td>Holocaust</td><td/><td colspan=\"3\">would call for the extermination</td></tr><tr><td/><td/><td/><td/><td/><td/><td/><td/><td/><td>of gays</td><td/></tr><tr><td>6</td><td>1.69</td><td>2.02</td><td>2.58</td><td>1.89</td><td>2.49</td><td>2.60</td><td colspan=\"2\">homosexuals have all the same</td><td colspan=\"3\">Opposition to equal rights for</td></tr><tr><td/><td/><td/><td/><td/><td/><td/><td colspan=\"2\">rights as heterosexuals</td><td>gay couples.</td><td/></tr><tr><td>7</td><td>1.83</td><td>2.40</td><td>1.46</td><td>2.81</td><td>2.51</td><td>3.00</td><td colspan=\"2\">There was prejudice against gays</td><td colspan=\"3\">it is prejudice as opposed to reli-</td></tr><tr><td/><td/><td/><td/><td/><td/><td/><td colspan=\"2\">in 1909 just as there is now</td><td colspan=\"3\">gious or moral beliefs which fuel</td></tr><tr><td/><td/><td/><td/><td/><td/><td/><td/><td/><td colspan=\"2\">the anti-gay agenda;</td></tr><tr><td>8</td><td>2.00</td><td>1.70</td><td>3.16</td><td>1.73</td><td>2.41</td><td>3.40</td><td>homosexual</td><td>relationships</td><td colspan=\"3\">marriage should be between a</td></tr><tr><td/><td/><td/><td/><td/><td/><td/><td colspan=\"2\">should not compare to het-</td><td colspan=\"2\">heterosexual couple</td></tr><tr><td/><td/><td/><td/><td/><td/><td/><td colspan=\"2\">erosexual marriages because</td><td/><td/></tr><tr><td/><td/><td/><td/><td/><td/><td/><td colspan=\"2\">only heterosexuals are legally</td><td/><td/></tr><tr><td/><td/><td/><td/><td/><td/><td/><td>allowed to marry</td><td/><td/><td/></tr><tr><td>9</td><td>2.00</td><td>2.70</td><td>2.09</td><td>2.83</td><td>3.03</td><td>3.50</td><td colspan=\"2\">it is prejudice as opposed to reli-</td><td colspan=\"3\">when people claim religion in do-</td></tr><tr><td/><td/><td/><td/><td/><td/><td/><td colspan=\"2\">gious or moral beliefs which fuel</td><td colspan=\"3\">ing prejudice they are actually</td></tr><tr><td/><td/><td/><td/><td/><td/><td/><td colspan=\"2\">the anti-gay agenda;</td><td colspan=\"2\">abandoning their morals</td></tr><tr><td>10</td><td>2.94</td><td>2.02</td><td>2.93</td><td>2.18</td><td>2.70</td><td>3.50</td><td colspan=\"2\">gay people should be able to</td><td colspan=\"3\">Gay couples are unable to get</td></tr><tr><td/><td/><td/><td/><td/><td/><td/><td colspan=\"2\">marry a person of their choice</td><td colspan=\"3\">any benefits that married peo-</td></tr><tr><td/><td/><td/><td/><td/><td/><td/><td colspan=\"2\">and get equal rights.</td><td>ple do.</td><td/></tr><tr><td>11</td><td>2.14</td><td>1.50</td><td>2.91</td><td>2.08</td><td>2.62</td><td>3.60</td><td colspan=\"2\">Paul Cameron is the voice of the</td><td>Conversation</td><td>about</td><td>Paul</td></tr><tr><td/><td/><td/><td/><td/><td/><td/><td>Republicans</td><td/><td>Cameron</td><td/></tr><tr><td>12</td><td>2.63</td><td>3.63</td><td>2.60</td><td>3.75</td><td>3.57</td><td>4.17</td><td colspan=\"2\">in opening this opportunity for</td><td colspan=\"3\">opponents of homosexual mar-</td></tr><tr><td/><td/><td/><td/><td/><td/><td/><td colspan=\"2\">gay marriage, the definition of</td><td colspan=\"3\">riage tend to argue that a change</td></tr><tr><td/><td/><td/><td/><td/><td/><td/><td colspan=\"2\">marriage will change</td><td colspan=\"3\">to marriage law would make it too</td></tr><tr><td/><td/><td/><td/><td/><td/><td/><td/><td/><td>open ended</td><td/></tr><tr><td>13</td><td>4.23</td><td>2.72</td><td>2.26</td><td>4.82</td><td>4.12</td><td>4.50</td><td colspan=\"2\">AIDs was initially spread in the</td><td colspan=\"3\">No one argues the point that AIDs</td></tr><tr><td/><td/><td/><td/><td/><td/><td/><td colspan=\"2\">United States primarily by homo-</td><td colspan=\"3\">was spread in the United States by</td></tr><tr><td/><td/><td/><td/><td/><td/><td/><td>sexuals.</td><td/><td>homosexuals.</td><td/></tr></table>",
"type_str": "table",
"num": null,
"html": null,
"text": "where we combine our features with UMBC shows that this improves the correlation further, from the UMBC baseline of 0.46 to 0.54 (p < 0.01.)"
},
"TABREF6": {
"content": "<table/>",
"type_str": "table",
"num": null,
"html": null,
"text": "Predicted Scores for each model and the Mechanical Turk AFS gold standard for selected argument pairs from the pairs dataset. Best performing model for each pair is shown in bold. The table is sorted by the AFS score (gold standard). The argument pairs shown in bold are cases where UMBC by itself beats our proposed model. KEY: Feature sets model. N = NGRAM, U = UMBC STS tool, L = Linguistic Inquiry and Word Count; R = Rouge, D = DISCO, AFS= Mean of Mechanical Turker AFS scores, our gold standard. For example, NLRD means a combination of NGRAM, LIWC, ROUGE and DISCO."
}
}
}
} |