File size: 102,138 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 | {
"paper_id": "P12-1015",
"header": {
"generated_with": "S2ORC 1.0.0",
"date_generated": "2023-01-19T09:28:03.030617Z"
},
"title": "Distributional Semantics in Technicolor",
"authors": [
{
"first": "Elia",
"middle": [],
"last": "Bruni",
"suffix": "",
"affiliation": {
"laboratory": "",
"institution": "University of Trento",
"location": {}
},
"email": "elia.bruni@unitn.it"
},
{
"first": "Gemma",
"middle": [],
"last": "Boleda",
"suffix": "",
"affiliation": {
"laboratory": "",
"institution": "University of Texas at Austin",
"location": {}
},
"email": "gemma.boleda@utcompling.com"
},
{
"first": "Marco",
"middle": [],
"last": "Baroni",
"suffix": "",
"affiliation": {
"laboratory": "",
"institution": "University of Trento",
"location": {}
},
"email": ""
},
{
"first": "Nam-Khanh",
"middle": [],
"last": "Tran",
"suffix": "",
"affiliation": {
"laboratory": "",
"institution": "University of Trento",
"location": {}
},
"email": ""
}
],
"year": "",
"venue": null,
"identifiers": {},
"abstract": "Our research aims at building computational models of word meaning that are perceptually grounded. Using computer vision techniques, we build visual and multimodal distributional models and compare them to standard textual models. Our results show that, while visual models with state-of-the-art computer vision techniques perform worse than textual models in general tasks (accounting for semantic relatedness), they are as good or better models of the meaning of words with visual correlates such as color terms, even in a nontrivial task that involves nonliteral uses of such words. Moreover, we show that visual and textual information are tapping on different aspects of meaning, and indeed combining them in multimodal models often improves performance.",
"pdf_parse": {
"paper_id": "P12-1015",
"_pdf_hash": "",
"abstract": [
{
"text": "Our research aims at building computational models of word meaning that are perceptually grounded. Using computer vision techniques, we build visual and multimodal distributional models and compare them to standard textual models. Our results show that, while visual models with state-of-the-art computer vision techniques perform worse than textual models in general tasks (accounting for semantic relatedness), they are as good or better models of the meaning of words with visual correlates such as color terms, even in a nontrivial task that involves nonliteral uses of such words. Moreover, we show that visual and textual information are tapping on different aspects of meaning, and indeed combining them in multimodal models often improves performance.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Abstract",
"sec_num": null
}
],
"body_text": [
{
"text": "Traditional semantic space models represent meaning on the basis of word co-occurrence statistics in large text corpora (Turney and Pantel, 2010) . These models (as well as virtually all work in computational lexical semantics) rely on verbal information only, while human semantic knowledge also relies on non-verbal experience and representation (Louwerse, 2011) , crucially on the information gathered through perception. Recent developments in computer vision make it possible to computationally model one vital human perceptual channel: vision (Mooney, 2008) . A few studies have begun to use visual information extracted from images as part of distributional semantic models (Bergsma and Van Durme, 2011; Bergsma and Goebel, 2011; Bruni et al., 2011; Feng and Lapata, 2010; Leong and Mihalcea, 2011) . These preliminary studies all focus on how vision may help text-based models in general terms, by evaluating performance on, for instance, word similarity datasets such as WordSim353.",
"cite_spans": [
{
"start": 120,
"end": 145,
"text": "(Turney and Pantel, 2010)",
"ref_id": "BIBREF31"
},
{
"start": 348,
"end": 364,
"text": "(Louwerse, 2011)",
"ref_id": "BIBREF19"
},
{
"start": 549,
"end": 563,
"text": "(Mooney, 2008)",
"ref_id": "BIBREF24"
},
{
"start": 681,
"end": 710,
"text": "(Bergsma and Van Durme, 2011;",
"ref_id": "BIBREF3"
},
{
"start": 711,
"end": 736,
"text": "Bergsma and Goebel, 2011;",
"ref_id": "BIBREF2"
},
{
"start": 737,
"end": 756,
"text": "Bruni et al., 2011;",
"ref_id": "BIBREF6"
},
{
"start": 757,
"end": 779,
"text": "Feng and Lapata, 2010;",
"ref_id": "BIBREF12"
},
{
"start": 780,
"end": 805,
"text": "Leong and Mihalcea, 2011)",
"ref_id": "BIBREF18"
}
],
"ref_spans": [],
"eq_spans": [],
"section": "Introduction",
"sec_num": "1"
},
{
"text": "This paper contributes to connecting language and perception, focusing on how to exploit visual information to build better models of word meaning, in three ways: (1) We carry out a systematic comparison of models using textual, visual, and both types of information. (2) We evaluate the models on general semantic relatedness tasks and on two specific tasks where visual information is highly relevant, as they focus on color terms. (3) Unlike previous work, we study the impact of using different kinds of visual information for these semantic tasks.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Introduction",
"sec_num": "1"
},
{
"text": "Our results show that, while visual models with state-of-the-art computer vision techniques perform worse than textual models in general semantic tasks, they are as good or better models of the meaning of words with visual correlates such as color terms, even in a nontrivial task that involves nonliteral uses of such words. Moreover, we show that visual and textual information are tapping on different aspects of meaning, such that they are complementary sources of information, and indeed combining them in multimodal models often improves performance. We also show that \"hybrid\" models exploiting the patterns of co-occurrence of words as tags of the same images can be a powerful surrogate of visual information under certain circumstances.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Introduction",
"sec_num": "1"
},
{
"text": "The rest of the paper is structured as follows. Section 2 introduces the textual, visual, multimodal, and hybrid models we use for our experiments. We present our experiments in sections 3 to 5. Section 6 reviews related work, and section 7 finishes with conclusions and future work.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Introduction",
"sec_num": "1"
},
{
"text": "For the current project, we constructed a set of textual distributional models that implement various standard ways to extract them from a corpus, chosen to be representative of the state of the art. In all cases, occurrence and co-occurrence statistics are extracted from the freely available ukWaC and Wackypedia corpora combined (size: 1.9B and 820M tokens, respectively). 1 Moreover, in all models the raw co-occurrence counts are transformed into nonnegative Local Mutual Information (LMI) scores. 2 Finally, in all models we harvest vector representations for the same words (lemmas), namely the top 20K most frequent nouns, 5K most frequent adjectives and 5K most frequent verbs in the combined corpora (for coherence with the vision-based models, that cannot exploit contextual information to distinguish nouns and adjectives, we merge nominal and adjectival usages of the color adjectives in the text-based models as well). The same 30K target nouns, verbs and adjectives are also employed as contextual elements.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Distributional semantic models 2.1 Textual models",
"sec_num": "2"
},
{
"text": "The Window2 and Window20 models are based on counting co-occurrences with collocates within a window of fixed width, in the tradition of HAL (Lund and Burgess, 1996) . Window2 records sentence-internal co-occurrence with the nearest 2 content words to the left and right of each target concept, a narrow context definition expected to capture taxonomic relations. Window20 considers a larger window of 20 words to the left and right of the target, and should capture broader topical relations. The Document model corresponds to a \"topic-based\" approach in which words are represented as distributions over documents. It is based on a word-bydocument matrix, recording the distribution of the 30K target words across the 30K documents in the concatenated corpus that have the largest cumulative LMI mass. This model is thus akin to traditional Latent Semantic Analysis (Landauer and Dumais, 1997) , without dimensionality reduction.",
"cite_spans": [
{
"start": 141,
"end": 165,
"text": "(Lund and Burgess, 1996)",
"ref_id": "BIBREF22"
},
{
"start": 868,
"end": 895,
"text": "(Landauer and Dumais, 1997)",
"ref_id": "BIBREF16"
}
],
"ref_spans": [],
"eq_spans": [],
"section": "Distributional semantic models 2.1 Textual models",
"sec_num": "2"
},
{
"text": "We add to the models we constructed the freely available Distributional Memory (DM) model, 3 that has been shown to reach state-of-the-art performance in many semantic tasks (Baroni and Lenci, 2010) . DM is an example of a more complex textbased model that exploits lexico-syntactic and dependency relations between words (see Baroni and Lenci's article for details), and we use it as an instance of a grammar-based model. DM is based on the same corpora we used plus the 100M-word British National Corpus, 4 and it also uses LMI scores.",
"cite_spans": [
{
"start": 174,
"end": 198,
"text": "(Baroni and Lenci, 2010)",
"ref_id": "BIBREF1"
},
{
"start": 327,
"end": 337,
"text": "Baroni and",
"ref_id": null
}
],
"ref_spans": [],
"eq_spans": [],
"section": "Distributional semantic models 2.1 Textual models",
"sec_num": "2"
},
{
"text": "The visual models use information extracted from images instead of textual corpora. We use image data where each image is associated with one or more words or tags (we use \"tag\" for each word associated to the image, and \"label\" for the set of tags of an image). We use the ESP-Game dataset, 5 containing 100K images labeled through a game with a purpose in which two people partnered online must independently and rapidly agree on an appropriate word to label randomly selected images. Once a word is entered by both partners in a certain number of game matches, that word is added to the label for that image, and it becomes a taboo word for the following rounds of the game (von Ahn and Dabbish, 2004 ). There are 20,515 distinct tags in the dataset, with an average of 4 tags per image. We build one vector with visual features for each tag in the dataset.",
"cite_spans": [
{
"start": 690,
"end": 703,
"text": "Dabbish, 2004",
"ref_id": "BIBREF34"
}
],
"ref_spans": [],
"eq_spans": [],
"section": "Visual models",
"sec_num": "2.2"
},
{
"text": "The visual features are extracted with the use of a standard bag-of-visual-words (BoVW) representation of images, inspired by NLP (Sivic and Zisserman, 2003; Csurka et al., 2004; Nister and Stewenius, 2006; Bosch et al., 2007; Yang et al., 2007) . This approach relies on the notion of a common vocabulary of \"visual words\" that can serve as discrete representations for all images. Contrary to what hap-pens in NLP, where words are (mostly) discrete and easy to identify, in vision the visual words need to be first defined. The process is completely inductive. In a nutshell, BoVW works as follows. From every image in a dataset, relevant areas are identified and a low-level feature vector (called a \"descriptor\") is built to represent each area. These vectors, living in what is sometimes called a descriptor space, are then grouped into a number of clusters. Each cluster is treated as a discrete visual word, and the clusters will be the vocabulary of visual words used to represent all the images in the collection. Now, given a new image, the nearest visual word is identified for each descriptor extracted from it, such that the image can be represented as a BoVW feature vector, by counting the instances of each visual word in the image (note that an occurrence of a low-level descriptor vector in an image, after mapping to the nearest cluster, will increment the count of a single dimension of the higher-level BoVW vector). In our work, the representation of each word (tag) is a also a BoVW vector. The values of each dimension are obtained by summing the occurrences of the relevant visual word in all the images tagged with the word. Again, raw counts are transformed into Local Mutual Information scores. The process to extract visual words and use them to create image-based vectors to represent (real) words is illustrated in Figure 1 , for a hypothetical example in which there is only one image in the collection labeled with the word horse. We extract descriptor features of two types. 6 First, the standard Scale-Invariant Feature Transform (SIFT) feature vectors (Lowe, 1999; Lowe, 2004) , good at characterizing parts of objects. Second, LAB features (Fairchild, 2005) , which encode only color information. We also experimented with other visual features, such as those focusing on edges (Canny, 1986 ), texture (Zhu et al., 2002 , and shapes (Oliva and Torralba, 2001 ), but they were not useful for the color tasks. Moreover, we experimented also with different color scales, such as LUV, HSV and RGB, obtaining significantly worse performance compared to LAB. Further details on feature extraction follow.",
"cite_spans": [
{
"start": 130,
"end": 157,
"text": "(Sivic and Zisserman, 2003;",
"ref_id": "BIBREF29"
},
{
"start": 158,
"end": 178,
"text": "Csurka et al., 2004;",
"ref_id": "BIBREF8"
},
{
"start": 179,
"end": 206,
"text": "Nister and Stewenius, 2006;",
"ref_id": "BIBREF25"
},
{
"start": 207,
"end": 226,
"text": "Bosch et al., 2007;",
"ref_id": "BIBREF5"
},
{
"start": 227,
"end": 245,
"text": "Yang et al., 2007)",
"ref_id": "BIBREF35"
},
{
"start": 2089,
"end": 2101,
"text": "(Lowe, 1999;",
"ref_id": "BIBREF20"
},
{
"start": 2102,
"end": 2113,
"text": "Lowe, 2004)",
"ref_id": "BIBREF21"
},
{
"start": 2178,
"end": 2195,
"text": "(Fairchild, 2005)",
"ref_id": "BIBREF10"
},
{
"start": 2316,
"end": 2328,
"text": "(Canny, 1986",
"ref_id": "BIBREF7"
},
{
"start": 2329,
"end": 2357,
"text": "), texture (Zhu et al., 2002",
"ref_id": null
},
{
"start": 2371,
"end": 2396,
"text": "(Oliva and Torralba, 2001",
"ref_id": "BIBREF26"
}
],
"ref_spans": [
{
"start": 1846,
"end": 1855,
"text": "Figure 1",
"ref_id": "FIGREF0"
}
],
"eq_spans": [],
"section": "Visual models",
"sec_num": "2.2"
},
{
"text": "! ! !\"#$%&'()*&!$(+%# !! !!!\"#$% ,*&$# - . / . !!!!0#%)*&!&#(&#$#1)+)'*1 !!!!!!!!2345!.6.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Visual models",
"sec_num": "2.2"
},
{
"text": "SIFT features are designed to be invariant to image scale and rotation, and have been shown to provide a robust matching across affine distortion, noise and change in illumination. The version of SIFT features that we use is sensitive to color (RGB scale; LUV, LAB and OPPONENT gave worse results). We automatically identified keypoints for each image and extracted SIFT features on a regular grid defined around the keypoint with five pixels spacing, at four multiple scales (10, 15, 20, 25 pixel radii), zeroing the low contrast ones. To obtain the visual word vocabulary, we cluster the SIFT feature vectors with the standardly used k-means clustering algorithm. We varied the number k of visual words between 500 and 2,500 in steps of 500.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Visual models",
"sec_num": "2.2"
},
{
"text": "For the SIFT-based representation of images, we used spatial histograms to introduce weak geometry (Grauman and Darrell, 2005; Lazebnik et al., 2006) , dividing the image into several (spatial) regions, representing each region in terms of BoVW, and then concatenating the vectors. In our experiments, the spatial regions were obtained by dividing the image in 4 \u00d7 4, for a total of 16 regions (other values and a global representation did not perform as well). Note that, following standard practice, descriptor clustering was performed ignoring the region partition, but the resulting visual words correspond to different dimensions in the concatenated BoVW vectors, depending on the region in which they occur. Consequently, a vocabulary of k visual words results in BoVW vectors with k \u00d7 16 dimensions.",
"cite_spans": [
{
"start": 99,
"end": 126,
"text": "(Grauman and Darrell, 2005;",
"ref_id": "BIBREF14"
},
{
"start": 127,
"end": 149,
"text": "Lazebnik et al., 2006)",
"ref_id": "BIBREF17"
}
],
"ref_spans": [],
"eq_spans": [],
"section": "Visual models",
"sec_num": "2.2"
},
{
"text": "The LAB color space plots image data in 3 dimensions along 3 independent (orthogonal) axes, one for brightness (luminance) and two for color (chrominance). Luminance corresponds closely to brightness as recorded by the brain-eye system; the chrominance (red-green and yellow-blue) axes mimic the oppositional color sensations the retina reports to the brain (Szeliski, 2010) . LAB features are densely sampled for each pixel. Also here we use the k-means algorithm to build the descriptor space. We varied the number of k visual words between 128 and 1,024 in steps of 128.",
"cite_spans": [
{
"start": 358,
"end": 374,
"text": "(Szeliski, 2010)",
"ref_id": "BIBREF30"
}
],
"ref_spans": [],
"eq_spans": [],
"section": "Visual models",
"sec_num": "2.2"
},
{
"text": "To assemble the textual and visual representations in multimodal semantic spaces, we concatenate the two vectors after normalizing them. We use the linear weighted combination function proposed by Bruni et al. (2011) : Given a word that is present both in the textual model and in the visual model, we separately normalize the two vectors F t and F v and we combine them as follows:",
"cite_spans": [
{
"start": 197,
"end": 216,
"text": "Bruni et al. (2011)",
"ref_id": "BIBREF6"
}
],
"ref_spans": [],
"eq_spans": [],
"section": "Multimodal models",
"sec_num": "2.3"
},
{
"text": "F = \u03b1 \u00d7 F t \u2295 (1 \u2212 \u03b1) \u00d7 F v",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Multimodal models",
"sec_num": "2.3"
},
{
"text": "where \u2295 is the vector concatenate operator. The weighting parameter \u03b1 (0 \u2264 \u03b1 \u2264 1) is tuned on the MEN development data (2,000 word pairs; details on the MEN dataset in the next section). We find the optimal value to be close to \u03b1 = 0.5 for most model combinations, suggesting that textual and visual information should have similar weight. Our implementation of the proposed method is open source and publicly available. 7",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Multimodal models",
"sec_num": "2.3"
},
{
"text": "We further introduce hybrid models that exploit the patterns of co-occurrence of words as tags of the same images. Like textual models, these models are based on word co-occurrence; like visual models, they consider co-occurrence in images (image labels). In one model (ESP-Win, analogous to window-based models), words tagging an image were represented in terms of co-occurrence with the other tags in the image label (Baroni and Lenci (2008) are a precedent for the use of ESP-Win). The other (ESP-Doc, analogous to document-based models) represented words in terms of their cooccurrence with images, using each image as a different dimension. This information is very easy to extract, as it does not require the sophisticated techniques used in computer vision. We expected these models to perform very bad; however, as we will show, they perform relatively well in all but one of the tasks tested.",
"cite_spans": [
{
"start": 419,
"end": 443,
"text": "(Baroni and Lenci (2008)",
"ref_id": "BIBREF0"
}
],
"ref_spans": [],
"eq_spans": [],
"section": "Hybrid models",
"sec_num": "2.4"
},
{
"text": "We test the models just presented in two different ways: First, as general models of word meaning, testing their correlation to human judgements on word similarity and relatedness (this section). Second, as models of the meaning of color terms (sections 4 and 5).",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Textual and visual models as general semantic models",
"sec_num": "3"
},
{
"text": "We use one standard dataset (WordSim353) and one new dataset (MEN). WordSim353 (Finkelstein et al., 2002) is a widely used benchmark constructed by asking 16 subjects to rate a set of 353 word pairs on a 10-point similarity scale and averaging the ratings (dollar/buck receives a high 9.22 average rating, professor/cucumber a low 0.31). MEN is a new evaluation benchmark with a better coverage of our multimodal semantic models. 8 It contains 3,000 pairs of randomly selected words that occur as ESP tags (pairs sampled to ensure a balanced range of relatedness levels according to a text-based semantic score). Each pair is scored on a [0, 1]-normalized semantic relatedness scale via ratings obtained by crowdsourcing on the Amazon Mechanical Turk (refer to the online MEN documentation for more details). For example, cold/frost has a high 0.9 MEN score, eat/hair a low 0.1. We evaluate the models in terms of their Spearman correlation to the human ratings. Our models have a perfect MEN coverage and a coverage of 252 WordSim pairs.",
"cite_spans": [
{
"start": 79,
"end": 105,
"text": "(Finkelstein et al., 2002)",
"ref_id": "BIBREF13"
}
],
"ref_spans": [],
"eq_spans": [],
"section": "Textual and visual models as general semantic models",
"sec_num": "3"
},
{
"text": "We used the development set of MEN to test the effect of varying the number k of visual words in SIFT and LAB. We restrict the discussion to SIFT with the optimal k (2.5K words) and to LAB with the optimal (256), lowest (128), and highest k (1024). We report the results of the multimodal models built with these visual models and the best textual models (Window2 and Window20).",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Textual and visual models as general semantic models",
"sec_num": "3"
},
{
"text": "Columns WS and MEN in Table 1 report correlations with the WordSim and MEN ratings, respectively. As expected, because they are more mature and capture a broader range of semantic information, textual models perform much better than purely visual models. Also as expected, SIFT features outperform the simpler LAB features for this task.",
"cite_spans": [],
"ref_spans": [
{
"start": 22,
"end": 29,
"text": "Table 1",
"ref_id": null
}
],
"eq_spans": [],
"section": "Textual and visual models as general semantic models",
"sec_num": "3"
},
{
"text": "A first indication that visual information helps is the fact that, for MEN, multimodal models perform best. Note that all models that are sensitive to visual information perform better for MEN than for WordSim, and the reverse is true for textual models. Because of its design, word pairs in MEN can be expected to be more imageable than those in Word-Sim, so the visual information is more relevant for this dataset. Also recall that we did some parameter tuning on held-out MEN data.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Textual and visual models as general semantic models",
"sec_num": "3"
},
{
"text": "Surprisingly, hybrid models perform quite well: They are around 10 points worse than textual and multimodal models for WordSim, and only slightly worse than multimodal models for MEN.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Textual and visual models as general semantic models",
"sec_num": "3"
},
{
"text": "In Experiment 1, we test the hypothesis that the relation between words denoting concrete things and words denoting their typical color is reflected by the distance of the corresponding vectors better when the models are sensitive to visual information.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Experiment 1: Discovering the color of concrete objects",
"sec_num": "4"
},
{
"text": "Two authors labeled by consensus a list of concrete nouns (extracted from the BLESS dataset 9 and the nouns in the BNC occurring with color terms more than 100 times) with one of the 11 colors from the basic set proposed by Berlin and Kay (1969) : black, blue, brown, green, grey, orange, pink, purple, red, white, yellow. Objects that do not have an obvious characteristic color (computer) and those with more than one characteristic color (zebra, bear) were eliminated. Moreover, only nouns covered by all the models were preserved. The final list con- Table 1 : Results of the textual, visual, multimodal, and hybrid models on the general semantic tasks (first two columns, section 3; Pearson \u03c1) and Experiments 1 (E1, section 4) and 2 (E2, section 5). E1 reports the median rank of the correct color and the number of top matches (in parentheses), and E2 the average difference in normalized cosines between literal and nonliteral adjectivenoun phrases, with the significance of a t-test (*** for p< 0.001, ** < 0.01, * < 0.05).",
"cite_spans": [
{
"start": 224,
"end": 245,
"text": "Berlin and Kay (1969)",
"ref_id": null
}
],
"ref_spans": [
{
"start": 555,
"end": 562,
"text": "Table 1",
"ref_id": null
}
],
"eq_spans": [],
"section": "Method",
"sec_num": "4.1"
},
{
"text": "tains 52 nouns. 10 Some random examples are foggrey, crow-black, wood-brown, parsley-green, and grass-green. For evaluation, we measured the cosine of each noun with the 11 basic color words in the space produced by each model, and recorded the rank of the correct color in the resulting ordered list.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Method",
"sec_num": "4.1"
},
{
"text": "Column E1 in Table 1 reports the median rank for each model (the smaller the rank, the better the model), as well as the number of exact matches (that is, number of nouns for which the model ranks the correct color first).",
"cite_spans": [],
"ref_spans": [
{
"start": 13,
"end": 20,
"text": "Table 1",
"ref_id": null
}
],
"eq_spans": [],
"section": "Results",
"sec_num": "4.2"
},
{
"text": "Discovering knowledge such that grass is green is arguably a simple task but Experiment 1 shows that textual models fail this simple task, with median ranks around 3. 11 This is consistent with the findings in Baroni and Lenci (2008) that standard distributional models do not capture the association between concrete concepts and their typical attributes. Visual models, as expected, are better at capturing the association between concepts and visual attributes. In fact, all models that are sensitive to visual information achieve median rank 1.",
"cite_spans": [
{
"start": 210,
"end": 233,
"text": "Baroni and Lenci (2008)",
"ref_id": "BIBREF0"
}
],
"ref_spans": [],
"eq_spans": [],
"section": "Results",
"sec_num": "4.2"
},
{
"text": "Multimodal models do not increase performance with respect to visual models: For instance, both W2-LAB 128 and W20-LAB 128 have the same median rank and number of exact matches as LAB 128 alone. Textual information in this case is not complementary to visual information, but simply poorer. Also note that LAB features do better than SIFT features. This is probably due to the fact that Experiment 1 is basically about identifying a large patch of color. The SIFT features we are using are also sensitive to color, but they seem to be misguided by the other cues that they extract from images. For example, pigs are pink in LAB space but brown in SIFT space, perhaps because SIFT focused on the color of the typical environment of a pig. We can thus confirm that, by limiting multimodal spaces to SIFT features, as has been done until now in the literature, we are missing important semantic information, such as the color information that we can mine with LAB.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Results",
"sec_num": "4.2"
},
{
"text": "Again we find that hybrid models do very well, in fact in this case they have the top performance, as they perform better than LAB 128 (the difference, which can be noticed in the number of exact matches, is highly significant according to a paired Mann-Whitney test, with p<0.001).",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Results",
"sec_num": "4.2"
},
{
"text": "Experiment 2 requires more sophisticated information than Experiment 1, as it involves distinguishing between literal and nonliteral uses of color terms.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Experiment 2",
"sec_num": "5"
},
{
"text": "We test the performance of the different models with a dataset consisting of color adjective-noun phrases, randomly drawn from the most frequent 8K nouns and 4K adjectives in the concatenated ukWaC, Wackypedia, and BNC corpora (four color terms are not among these, so the dataset includes phrases for black, blue, brown, green, red, white, and yellow only). These were tagged by consensus by two human judges as literal (white towel, black feather) or nonliteral (white wine, white musician, green future). Some phrases had both literal and nonliteral uses, such as blue book in \"book that is blue\" vs. \"automobile price guide\". In these cases, only the most common sense (according to the judges) was taken into account for the present experiment. The dataset consists of 370 phrases, of which our models cover 342, 227 literal and 115 nonliteral. 12 The prediction is that, in good semantic models, literal uses will in general result in a higher similarity between the noun and color term vectors: A white towel is white, while wine or musicians are not white in the same manner. We test this prediction by comparing the average cosine between the color term and the nouns across the literal and nonliteral pairs (similar results were obtained in an evaluation in terms of prediction accuracy of a simple classifier).",
"cite_spans": [
{
"start": 850,
"end": 852,
"text": "12",
"ref_id": null
}
],
"ref_spans": [],
"eq_spans": [],
"section": "Method",
"sec_num": "5.1"
},
{
"text": "Column E2 in Table 1 summarizes the results of the experiment, reporting the mean difference between the normalized cosines (that is, how large the difference is between the literal and nonliteral uses of color terms), as well as the significance of the differences according to a t-test. Window-based models perform best among textual models, particularly Window20, while the rest can't discriminate between the two uses. This is particularly striking for the Document model, which performs quite well in general semantic tasks but bad in visual tasks.",
"cite_spans": [],
"ref_spans": [
{
"start": 13,
"end": 20,
"text": "Table 1",
"ref_id": null
}
],
"eq_spans": [],
"section": "Results",
"sec_num": "5.2"
},
{
"text": "Visual models are all able to discriminate between the two uses, suggesting that indeed visual information can capture nonliteral aspects of meaning. However, in this case SIFT features perform much better than LAB features, as Experiment 2 involves tackling much more sophisticated information than Experiment 1. This is consistent with the fact that, for LAB, a lower k (lower granularity of the information) performs better for Experiment 1 and a higher k (higher granularity) for Experiment 2.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Results",
"sec_num": "5.2"
},
{
"text": "One crucial question to ask, given the goals of our research, is whether textual and visual models are doing essentially the same job, only using different types of information. Note that, in this case, multimodal models increase performance over the individual modalities, and are the best models for this task. This suggests that the information used in the individual models is complementary, and indeed there is no correlation between the cosines obtained with the best textual and visual models (Pearson's \u03c1 = .09, p = .11). Figure 2 depicts the results broken down by color. 13 Both modalities can capture the differences for black and green, probably because nonliteral uses of these color terms have also clear textual correlates (more concretely, topical correlates, as they are related to race and ecology, respectively). 14 Significantly, however, vision can capture nonliteral uses of blue and red, while text can't. Note that these uses (blue note, shark, shield, red meat, district, face) do not have a clear topical correlate, and thus it makes sense that vision does a better job.",
"cite_spans": [
{
"start": 581,
"end": 583,
"text": "13",
"ref_id": null
}
],
"ref_spans": [
{
"start": 530,
"end": 538,
"text": "Figure 2",
"ref_id": "FIGREF1"
}
],
"eq_spans": [],
"section": "Results",
"sec_num": "5.2"
},
{
"text": "Finally, note that for this more sophisticated task, hybrid models perform quite bad, which shows their limitations as models of word meaning. 15 Overall, 13 Yellow and brown are excluded because the dataset contains only one and two instances of nonliteral cases for these terms, respectively. The significance of the differences as explained in the text has been tested via t-tests.",
"cite_spans": [
{
"start": 155,
"end": 157,
"text": "13",
"ref_id": null
}
],
"ref_spans": [],
"eq_spans": [],
"section": "Results",
"sec_num": "5.2"
},
{
"text": "14 It's not entirely clear why neither modality can capture the differences for white; for text, it may be because the nonliteral cases are not so tied to race as is the cases for black, but they also contain many other types of nonliteral uses, such as type-referring (white wine/rice/cell) or metonymical ones (white smile).",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Results",
"sec_num": "5.2"
},
{
"text": "15 The hybrid model that performs best in the color tasks is ESP-Doc. This model can only detect a relation between an adjective and a noun if they directly co-occur in the label of at least one image (a \"document\" in this setting). The more direct cooccurrences there are, the more related the words will be for the model. This works for Exp. 1: Since the ESP labels are lists of what subjects saw in a picture, and the adjectives of Exp. 1 are typical colors of objects, there is a high co-occurrence, as all but one adjective-noun pairs co-occur in at least one ESP label. For the model to perform well in Exp. 2 too, literal phrases should occur in the same labels and non-literal pairs should not. We our results suggest that co-occurrence in an image label can be used as a surrogate of true visual information to some extent, but the behavior of hybrid models depends on ad-hoc aspects of the labeled dataset, and, from an empirical perspective, they are more limited than truly multimodal models, because they require large amounts of rich verbal picture descriptions to reach good coverage.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Results",
"sec_num": "5.2"
},
{
"text": "There is an increasing amount of work in computer vision that exploits text-derived information for image retrieval and annotation tasks (Farhadi et al., 2010; Kulkarni et al., 2011) . One particular techinque inspired by NLP that has acted as a very effective proxy from CV to NLP is precisely the BoVW. Recently, NLPers have begun exploiting BoVW to enrich distributional models that represent word meaning with visual features automatically extracted from images (Feng and Lapata, 2010; Bruni et al., 2011; Leong and Mihalcea, 2011) . Previous work in this area relied on SIFT features only, whereas we have enriched the visual representation of words with other kinds of features from computer vision, namely, color-related features (LAB). Moreover, earlier evaluation of multimodal models has focused only on standard word similarity tasks (using mainly WordSim353), whereas we have tested them on both general semantic tasks and specific tasks that tap directly into aspects of semantics (such as color) where we expect visual information to be crucial.",
"cite_spans": [
{
"start": 137,
"end": 159,
"text": "(Farhadi et al., 2010;",
"ref_id": "BIBREF11"
},
{
"start": 160,
"end": 182,
"text": "Kulkarni et al., 2011)",
"ref_id": "BIBREF15"
},
{
"start": 466,
"end": 489,
"text": "(Feng and Lapata, 2010;",
"ref_id": "BIBREF12"
},
{
"start": 490,
"end": 509,
"text": "Bruni et al., 2011;",
"ref_id": "BIBREF6"
},
{
"start": 510,
"end": 535,
"text": "Leong and Mihalcea, 2011)",
"ref_id": "BIBREF18"
}
],
"ref_spans": [],
"eq_spans": [],
"section": "Related work",
"sec_num": "6"
},
{
"text": "The most closely related work to ours is that recently presented by\u00d6zbal et al. (2011). Like us, Ozbal and colleagues use both a textual model and a visual model (as well as Google adjective-noun cooccurrence counts) to find the typical color of an object. However, their visual model works by analyzing pictures associated with an object, and determining the color of the object directly by image analysis. We attempt the more ambitious goal of separately associating a vector to nouns and adjectives, and defind no such difference (89% of adjective-noun pairs co-occur in at least one image in the literal set, 86% in the nonliteral set), because many of the relevant pairs describe concrete concepts that, while not necessarily of the \"right\" literal colour, are perfectly fit to be depicted in images (\"blue shark\", \"black boy\", \"white wine\"). termining the color of an object by the nearness of the noun denoting the object to the color term. In other words, we are trying to model the meaning of color terms and how they relate to other words, and not to directly extract the color of an object from pictures depicting them. Our second experiment is connected to the literature on the automated detection of figurative language (Shutova, 2010) . There is in particular some similarity with the tasks studied by Turney et al. (2011). Turney and colleagues try, among other things, to distinguish literal and metaphorical usages of adjectives when combined with nouns, including the highly visual adjective dark (dark hair vs. dark humour). Their method, based on automatically quantifying the degree of abstractness of the noun, is complementary to ours. Future work could combine our approach and theirs.",
"cite_spans": [
{
"start": 1234,
"end": 1249,
"text": "(Shutova, 2010)",
"ref_id": "BIBREF28"
}
],
"ref_spans": [],
"eq_spans": [],
"section": "Related work",
"sec_num": "6"
},
{
"text": "We have presented evidence that distributional semantic models based on text, while providing a good general semantic representation of word meaning, can be outperformed by models using visual information for semantic aspects of words where vision is relevant. More generally, this suggests that computer vision is mature enough to significantly contribute to perceptually grounded computational models of language. We have also shown that different types of visual features (LAB, SIFT) are appropriate for different tasks. Future research should investigate automated methods to discover which (if any) kind of visual information should be highlighted in which task, more sophisticated multimodal models, visual properties other than color, and larger color datasets, such as the one recently introduced by Mohammad (2011).",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Conclusion",
"sec_num": "7"
},
{
"text": "http://wacky.sslmit.unibo.it/ 2 LMI is obtained by multiplying raw counts by Pointwise Mutual Information, and it is a close approximation to the Log-Likelihood Ratio(Evert, 2005). It counteracts the tendency of PMI to favour extremely rare events.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "",
"sec_num": null
},
{
"text": "http://clic.cimec.unitn.it/dm 4 http://www.natcorp.ox.ac.uk/ 5 http://www.espgame.org",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "",
"sec_num": null
},
{
"text": "We use VLFeat (http://www.vlfeat.org/) for feature extraction(Vedaldi and Fulkerson, 2008).",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "",
"sec_num": null
},
{
"text": "https://github.com/s2m/FUSE",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "",
"sec_num": null
},
{
"text": "An updated version of MEN is available from http:// clic.cimec.unitn.it/\u02dcelia.bruni/MEN.html. The version used here contained 10 judgements per word pair.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "",
"sec_num": null
},
{
"text": "Dataset available from the second author's webpage, under resources.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "",
"sec_num": null
},
{
"text": "We also experimented with a model based on direct cooccurrence of adjectives and nouns, obtaining promising results in a preliminary version of Exp. 1. We abandoned this approach because such a model inherently lacks scalability, as it will not generalize behind cases where the training data contain direct examples of co-occurrences of the target pairs.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "",
"sec_num": null
},
{
"text": "Dataset available upon request to the second author.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "",
"sec_num": null
}
],
"back_matter": [
{
"text": "Acknowledgments E.B. and M.B. are partially supported by a Google Research Award.G.B. is partially supported by the Spanish Ministry of Science and Innovation (FFI2010-15006, TIN2009-14715-C04-04), the EU PASCAL2 Network of Excellence (FP7-ICT-216886) and the AGAUR (2010 BP-A 00070). The E2 evaluation set was created by G.B. with Louise McNally and Eva Maria Vecchi. Fig. 1 was adapted from a figure by Jasper Uijlings. G. B. thanks Margarita Torrent for taking care of her children while she worked hard to meet the Sunday deadline.",
"cite_spans": [],
"ref_spans": [
{
"start": 369,
"end": 375,
"text": "Fig. 1",
"ref_id": null
}
],
"eq_spans": [],
"section": "acknowledgement",
"sec_num": null
}
],
"bib_entries": {
"BIBREF0": {
"ref_id": "b0",
"title": "Concepts and properties in word spaces",
"authors": [
{
"first": "Marco",
"middle": [],
"last": "Baroni",
"suffix": ""
},
{
"first": "Alessandro",
"middle": [],
"last": "Lenci",
"suffix": ""
}
],
"year": 2008,
"venue": "Italian Journal of Linguistics",
"volume": "20",
"issue": "1",
"pages": "55--88",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Marco Baroni and Alessandro Lenci. 2008. Concepts and properties in word spaces. Italian Journal of Lin- guistics, 20(1):55-88.",
"links": null
},
"BIBREF1": {
"ref_id": "b1",
"title": "Distributional Memory: A general framework for corpus-based semantics",
"authors": [
{
"first": "Marco",
"middle": [],
"last": "Baroni",
"suffix": ""
},
{
"first": "Alessandro",
"middle": [],
"last": "Lenci",
"suffix": ""
}
],
"year": 2010,
"venue": "Computational Linguistics",
"volume": "36",
"issue": "4",
"pages": "673--721",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Marco Baroni and Alessandro Lenci. 2010. Dis- tributional Memory: A general framework for corpus-based semantics. Computational Linguistics, 36(4):673-721.",
"links": null
},
"BIBREF2": {
"ref_id": "b2",
"title": "Using visual information to predict lexical preference",
"authors": [
{
"first": "Shane",
"middle": [],
"last": "Bergsma",
"suffix": ""
},
{
"first": "Randy",
"middle": [],
"last": "Goebel",
"suffix": ""
}
],
"year": 2011,
"venue": "Proceedings of Recent Advances in Natural Language Processing",
"volume": "",
"issue": "",
"pages": "399--405",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Shane Bergsma and Randy Goebel. 2011. Using visual information to predict lexical preference. In Proceed- ings of Recent Advances in Natural Language Process- ing, pages 399-405, Hissar.",
"links": null
},
"BIBREF3": {
"ref_id": "b3",
"title": "Learning bilingual lexicons using the visual similarity of labeled web images",
"authors": [
{
"first": "Shane",
"middle": [],
"last": "Bergsma",
"suffix": ""
},
{
"first": "Benjamin",
"middle": [],
"last": "Van Durme",
"suffix": ""
}
],
"year": 2011,
"venue": "Proc. IJCAI",
"volume": "",
"issue": "",
"pages": "1764--1769",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Shane Bergsma and Benjamin Van Durme. 2011. Learn- ing bilingual lexicons using the visual similarity of la- beled web images. In Proc. IJCAI, pages 1764-1769, Barcelona, Spain, July.",
"links": null
},
"BIBREF4": {
"ref_id": "b4",
"title": "Basic Color Terms: Their Universality and Evolution",
"authors": [
{
"first": "Brent",
"middle": [],
"last": "Berlin",
"suffix": ""
},
{
"first": "Paul",
"middle": [],
"last": "Key",
"suffix": ""
}
],
"year": 1969,
"venue": "",
"volume": "",
"issue": "",
"pages": "",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Brent Berlin and Paul Key. 1969. Basic Color Terms: Their Universality and Evolution. University of Cali- fornia Press, Berkeley, CA.",
"links": null
},
"BIBREF5": {
"ref_id": "b5",
"title": "Image Classification using Random Forests and Ferns. In Computer Vision",
"authors": [
{
"first": "Anna",
"middle": [],
"last": "Bosch",
"suffix": ""
},
{
"first": "Andrew",
"middle": [],
"last": "Zisserman",
"suffix": ""
},
{
"first": "Xavier",
"middle": [],
"last": "Munoz",
"suffix": ""
}
],
"year": 2007,
"venue": "IEEE 11th International Conference on",
"volume": "",
"issue": "",
"pages": "1--8",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Anna Bosch, Andrew Zisserman, and Xavier Munoz. 2007. Image Classification using Random Forests and Ferns. In Computer Vision, 2007. ICCV 2007. IEEE 11th International Conference on, pages 1-8.",
"links": null
},
"BIBREF6": {
"ref_id": "b6",
"title": "Distributional semantics from text and images",
"authors": [
{
"first": "Elia",
"middle": [],
"last": "Bruni",
"suffix": ""
},
{
"first": "Marco",
"middle": [],
"last": "Giang Binh Tran",
"suffix": ""
},
{
"first": "",
"middle": [],
"last": "Baroni",
"suffix": ""
}
],
"year": 2011,
"venue": "Proceedings of the EMNLP GEMS Workshop",
"volume": "",
"issue": "",
"pages": "22--32",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Elia Bruni, Giang Binh Tran, and Marco Baroni. 2011. Distributional semantics from text and images. In Pro- ceedings of the EMNLP GEMS Workshop, pages 22- 32, Edinburgh.",
"links": null
},
"BIBREF7": {
"ref_id": "b7",
"title": "A computational approach to edge detection",
"authors": [
{
"first": "John",
"middle": [],
"last": "Canny",
"suffix": ""
}
],
"year": 1986,
"venue": "IEEE Trans. Pattern Anal. Mach. Intell",
"volume": "36",
"issue": "4",
"pages": "679--698",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "John Canny. 1986. A computational approach to edge detection. IEEE Trans. Pattern Anal. Mach. Intell, 36(4):679-698.",
"links": null
},
"BIBREF8": {
"ref_id": "b8",
"title": "Visual categorization with bags of keypoints",
"authors": [
{
"first": "Gabriella",
"middle": [],
"last": "Csurka",
"suffix": ""
},
{
"first": "Christopher",
"middle": [],
"last": "Dance",
"suffix": ""
},
{
"first": "Lixin",
"middle": [],
"last": "Fan",
"suffix": ""
},
{
"first": "Jutta",
"middle": [],
"last": "Willamowski",
"suffix": ""
},
{
"first": "C\u00e9dric",
"middle": [],
"last": "Bray",
"suffix": ""
}
],
"year": 2004,
"venue": "Workshop on Statistical Learning in Computer Vision, ECCV",
"volume": "",
"issue": "",
"pages": "1--22",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Gabriella Csurka, Christopher Dance, Lixin Fan, Jutta Willamowski, and C\u00e9dric Bray. 2004. Visual cate- gorization with bags of keypoints. In In Workshop on Statistical Learning in Computer Vision, ECCV, pages 1-22.",
"links": null
},
"BIBREF9": {
"ref_id": "b9",
"title": "The Statistics of Word Cooccurrences. Dissertation",
"authors": [
{
"first": "Stefan",
"middle": [
"Evert"
],
"last": "",
"suffix": ""
}
],
"year": 2005,
"venue": "",
"volume": "",
"issue": "",
"pages": "",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Stefan Evert. 2005. The Statistics of Word Cooccur- rences. Dissertation, Stuttgart University.",
"links": null
},
"BIBREF10": {
"ref_id": "b10",
"title": "Status of cie color appearance models",
"authors": [
{
"first": "D",
"middle": [],
"last": "Mark",
"suffix": ""
},
{
"first": "",
"middle": [],
"last": "Fairchild",
"suffix": ""
}
],
"year": 2005,
"venue": "",
"volume": "",
"issue": "",
"pages": "",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Mark D. Fairchild. 2005. Status of cie color appearance models.",
"links": null
},
"BIBREF11": {
"ref_id": "b11",
"title": "Every picture tells a story: Generating sentences from images",
"authors": [
{
"first": "A",
"middle": [],
"last": "Farhadi",
"suffix": ""
},
{
"first": "M",
"middle": [],
"last": "Hejrati",
"suffix": ""
},
{
"first": "M",
"middle": [],
"last": "Sadeghi",
"suffix": ""
},
{
"first": "P",
"middle": [],
"last": "Young",
"suffix": ""
},
{
"first": "C",
"middle": [],
"last": "Rashtchian",
"suffix": ""
},
{
"first": "J",
"middle": [],
"last": "Hockenmaier",
"suffix": ""
},
{
"first": "D",
"middle": [],
"last": "Forsyth",
"suffix": ""
}
],
"year": 2010,
"venue": "Proceedings of ECCV",
"volume": "",
"issue": "",
"pages": "",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "A. Farhadi, M. Hejrati, M. Sadeghi, P. Young, C. Rashtchian, J. Hockenmaier, and D. Forsyth. 2010. Every picture tells a story: Generating sentences from images. In Proceedings of ECCV.",
"links": null
},
"BIBREF12": {
"ref_id": "b12",
"title": "Visual information in semantic representation",
"authors": [
{
"first": "Yansong",
"middle": [],
"last": "Feng",
"suffix": ""
},
{
"first": "Mirella",
"middle": [],
"last": "Lapata",
"suffix": ""
}
],
"year": 2010,
"venue": "Proceedings of HLT-NAACL",
"volume": "",
"issue": "",
"pages": "91--99",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Yansong Feng and Mirella Lapata. 2010. Visual infor- mation in semantic representation. In Proceedings of HLT-NAACL, pages 91-99, Los Angeles, CA.",
"links": null
},
"BIBREF13": {
"ref_id": "b13",
"title": "Placing search in context: The concept revisited",
"authors": [
{
"first": "Lev",
"middle": [],
"last": "Finkelstein",
"suffix": ""
},
{
"first": "Evgeniy",
"middle": [],
"last": "Gabrilovich",
"suffix": ""
},
{
"first": "Yossi",
"middle": [],
"last": "Matias",
"suffix": ""
},
{
"first": "Ehud",
"middle": [],
"last": "Rivlin",
"suffix": ""
},
{
"first": "Zach",
"middle": [],
"last": "Solan",
"suffix": ""
},
{
"first": "Gadi",
"middle": [],
"last": "Wolfman",
"suffix": ""
},
{
"first": "Eytan",
"middle": [],
"last": "Ruppin",
"suffix": ""
}
],
"year": 2002,
"venue": "ACM Transactions on Information Systems",
"volume": "20",
"issue": "1",
"pages": "116--131",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Lev Finkelstein, Evgeniy Gabrilovich, Yossi Matias, Ehud Rivlin, Zach Solan, Gadi Wolfman, and Eytan Ruppin. 2002. Placing search in context: The concept revisited. ACM Transactions on Information Systems, 20(1):116-131.",
"links": null
},
"BIBREF14": {
"ref_id": "b14",
"title": "The pyramid match kernel: Discriminative classification with sets of image features",
"authors": [
{
"first": "Kristen",
"middle": [],
"last": "Grauman",
"suffix": ""
},
{
"first": "Trevor",
"middle": [],
"last": "Darrell",
"suffix": ""
}
],
"year": 2005,
"venue": "ICCV",
"volume": "",
"issue": "",
"pages": "1458--1465",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Kristen Grauman and Trevor Darrell. 2005. The pyramid match kernel: Discriminative classification with sets of image features. In In ICCV, pages 1458-1465.",
"links": null
},
"BIBREF15": {
"ref_id": "b15",
"title": "Baby talk: Understanding and generating simple image descriptions",
"authors": [
{
"first": "G",
"middle": [],
"last": "Kulkarni",
"suffix": ""
},
{
"first": "V",
"middle": [],
"last": "Premraj",
"suffix": ""
},
{
"first": "S",
"middle": [],
"last": "Dhar",
"suffix": ""
},
{
"first": "S",
"middle": [],
"last": "Li",
"suffix": ""
},
{
"first": "Y",
"middle": [],
"last": "Choi",
"suffix": ""
},
{
"first": "A",
"middle": [],
"last": "Berg",
"suffix": ""
},
{
"first": "T",
"middle": [],
"last": "Berg",
"suffix": ""
}
],
"year": 2011,
"venue": "Proceedings of CVPR",
"volume": "",
"issue": "",
"pages": "",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "G. Kulkarni, V. Premraj, S. Dhar, S. Li, Y. Choi, A. Berg, and T. Berg. 2011. Baby talk: Understanding and generating simple image descriptions. In Proceedings of CVPR.",
"links": null
},
"BIBREF16": {
"ref_id": "b16",
"title": "A solution to Plato's problem: The latent semantic analysis theory of acquisition, induction, and representation of knowledge",
"authors": [
{
"first": "Thomas",
"middle": [],
"last": "Landauer",
"suffix": ""
},
{
"first": "Susan",
"middle": [],
"last": "Dumais",
"suffix": ""
}
],
"year": 1997,
"venue": "Psychological Review",
"volume": "104",
"issue": "2",
"pages": "211--240",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Thomas Landauer and Susan Dumais. 1997. A solu- tion to Plato's problem: The latent semantic analysis theory of acquisition, induction, and representation of knowledge. Psychological Review, 104(2):211-240.",
"links": null
},
"BIBREF17": {
"ref_id": "b17",
"title": "Beyond bags of features: Spatial pyramid matching for recognizing natural scene categories",
"authors": [
{
"first": "Svetlana",
"middle": [],
"last": "Lazebnik",
"suffix": ""
},
{
"first": "Cordelia",
"middle": [],
"last": "Schmid",
"suffix": ""
},
{
"first": "Jean",
"middle": [],
"last": "Ponce",
"suffix": ""
}
],
"year": 2006,
"venue": "Proceedings of the 2006 IEEE Computer Society Conference on Computer Vision and Pattern Recognition",
"volume": "2",
"issue": "",
"pages": "2169--2178",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Svetlana Lazebnik, Cordelia Schmid, and Jean Ponce. 2006. Beyond bags of features: Spatial pyramid matching for recognizing natural scene categories. In Proceedings of the 2006 IEEE Computer Society Con- ference on Computer Vision and Pattern Recognition -Volume 2, CVPR 2006, pages 2169-2178, Washing- ton, DC, USA. IEEE Computer Society.",
"links": null
},
"BIBREF18": {
"ref_id": "b18",
"title": "Going beyond text: A hybrid image-text approach for measuring word relatedness",
"authors": [
{
"first": "Rada",
"middle": [],
"last": "Chee Wee Leong",
"suffix": ""
},
{
"first": "",
"middle": [],
"last": "Mihalcea",
"suffix": ""
}
],
"year": 2011,
"venue": "Proceedings of IJCNLP",
"volume": "",
"issue": "",
"pages": "1403--1407",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Chee Wee Leong and Rada Mihalcea. 2011. Going beyond text: A hybrid image-text approach for mea- suring word relatedness. In Proceedings of IJCNLP, pages 1403-1407, Chiang Mai, Thailand.",
"links": null
},
"BIBREF19": {
"ref_id": "b19",
"title": "Symbol interdependency in symbolic and embodied cognition",
"authors": [
{
"first": "Max",
"middle": [],
"last": "Louwerse",
"suffix": ""
}
],
"year": 2011,
"venue": "Topics in Cognitive Science",
"volume": "3",
"issue": "",
"pages": "273--302",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Max Louwerse. 2011. Symbol interdependency in sym- bolic and embodied cognition. Topics in Cognitive Science, 3:273-302.",
"links": null
},
"BIBREF20": {
"ref_id": "b20",
"title": "Object Recognition from Local Scale-Invariant Features. Computer Vision",
"authors": [
{
"first": "David",
"middle": [],
"last": "Lowe",
"suffix": ""
}
],
"year": 1999,
"venue": "IEEE International Conference on",
"volume": "2",
"issue": "",
"pages": "1150--1157",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "David Lowe. 1999. Object Recognition from Local Scale-Invariant Features. Computer Vision, IEEE In- ternational Conference on, 2:1150-1157 vol.2, Au- gust.",
"links": null
},
"BIBREF21": {
"ref_id": "b21",
"title": "Distinctive image features from scale-invariant keypoints",
"authors": [
{
"first": "David",
"middle": [],
"last": "Lowe",
"suffix": ""
}
],
"year": 2004,
"venue": "International Journal of Computer Vision",
"volume": "60",
"issue": "2",
"pages": "",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "David Lowe. 2004. Distinctive image features from scale-invariant keypoints. International Journal of Computer Vision, 60(2), November.",
"links": null
},
"BIBREF22": {
"ref_id": "b22",
"title": "Producing high-dimensional semantic spaces from lexical cooccurrence",
"authors": [
{
"first": "Kevin",
"middle": [],
"last": "Lund",
"suffix": ""
},
{
"first": "Curt",
"middle": [],
"last": "Burgess",
"suffix": ""
}
],
"year": 1996,
"venue": "Behavior Research Methods",
"volume": "28",
"issue": "",
"pages": "203--208",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Kevin Lund and Curt Burgess. 1996. Producing high-dimensional semantic spaces from lexical co- occurrence. Behavior Research Methods, 28:203-208.",
"links": null
},
"BIBREF23": {
"ref_id": "b23",
"title": "Colourful language: Measuring word-colour associations",
"authors": [
{
"first": "Saif",
"middle": [],
"last": "Mohammad",
"suffix": ""
}
],
"year": 2011,
"venue": "Proceedings of the 2nd Workshop on Cognitive Modeling and Computational Linguistics",
"volume": "",
"issue": "",
"pages": "97--106",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Saif Mohammad. 2011. Colourful language: Measuring word-colour associations. In Proceedings of the 2nd Workshop on Cognitive Modeling and Computational Linguistics, pages 97-106, Portland, Oregon.",
"links": null
},
"BIBREF24": {
"ref_id": "b24",
"title": "Learning to connect language and perception",
"authors": [
{
"first": "Raymond",
"middle": [
"J"
],
"last": "Mooney",
"suffix": ""
}
],
"year": 2008,
"venue": "",
"volume": "",
"issue": "",
"pages": "",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Raymond J. Mooney. 2008. Learning to connect lan- guage and perception.",
"links": null
},
"BIBREF25": {
"ref_id": "b25",
"title": "Scalable recognition with a vocabulary tree",
"authors": [
{
"first": "David",
"middle": [],
"last": "Nister",
"suffix": ""
},
{
"first": "Henrik",
"middle": [],
"last": "Stewenius",
"suffix": ""
}
],
"year": 2006,
"venue": "Proceedings of the 2006 IEEE Computer Society Conference on Computer Vision and Pattern Recognition",
"volume": "2",
"issue": "",
"pages": "2161--2168",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "David Nister and Henrik Stewenius. 2006. Scalable recognition with a vocabulary tree. In Proceedings of the 2006 IEEE Computer Society Conference on Computer Vision and Pattern Recognition -Volume 2, CVPR '06, pages 2161-2168.",
"links": null
},
"BIBREF26": {
"ref_id": "b26",
"title": "Modeling the shape of the scene: A holistic representation of the spatial envelope",
"authors": [
{
"first": "Aude",
"middle": [],
"last": "Oliva",
"suffix": ""
},
{
"first": "Antonio",
"middle": [],
"last": "Torralba",
"suffix": ""
}
],
"year": 2001,
"venue": "Int. J. Comput. Vision",
"volume": "42",
"issue": "",
"pages": "145--175",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Aude Oliva and Antonio Torralba. 2001. Modeling the shape of the scene: A holistic representation of the spatial envelope. Int. J. Comput. Vision, 42:145-175.",
"links": null
},
"BIBREF27": {
"ref_id": "b27",
"title": "A comparison of unsupervised methods to associate colors with words",
"authors": [
{
"first": "Carlo",
"middle": [],
"last": "G\u00f6zde\u00f6zbal",
"suffix": ""
},
{
"first": "Rada",
"middle": [],
"last": "Strapparava",
"suffix": ""
},
{
"first": "Daniele",
"middle": [],
"last": "Mihalcea",
"suffix": ""
},
{
"first": "",
"middle": [],
"last": "Pighin",
"suffix": ""
}
],
"year": 2011,
"venue": "Proceedings of ACII",
"volume": "",
"issue": "",
"pages": "42--51",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "G\u00f6zde\u00d6zbal, Carlo Strapparava, Rada Mihalcea, and Daniele Pighin. 2011. A comparison of unsupervised methods to associate colors with words. In Proceed- ings of ACII, pages 42-51, Memphis, TN.",
"links": null
},
"BIBREF28": {
"ref_id": "b28",
"title": "Models of metaphor in NLP",
"authors": [
{
"first": "Ekaterina",
"middle": [],
"last": "Shutova",
"suffix": ""
}
],
"year": 2010,
"venue": "Proceedings of ACL",
"volume": "",
"issue": "",
"pages": "688--697",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Ekaterina Shutova. 2010. Models of metaphor in NLP. In Proceedings of ACL, pages 688-697, Uppsala, Swe- den.",
"links": null
},
"BIBREF29": {
"ref_id": "b29",
"title": "Video Google: A text retrieval approach to object matching in videos",
"authors": [
{
"first": "Josef",
"middle": [],
"last": "Sivic",
"suffix": ""
},
{
"first": "Andrew",
"middle": [],
"last": "Zisserman",
"suffix": ""
}
],
"year": 2003,
"venue": "Proceedings of the International Conference on Computer Vision",
"volume": "2",
"issue": "",
"pages": "1470--1477",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Josef Sivic and Andrew Zisserman. 2003. Video Google: A text retrieval approach to object matching in videos. In Proceedings of the International Conference on Computer Vision, volume 2, pages 1470-1477, Octo- ber.",
"links": null
},
"BIBREF30": {
"ref_id": "b30",
"title": "Computer Vision : Algorithms and Applications",
"authors": [
{
"first": "Richard",
"middle": [],
"last": "Szeliski",
"suffix": ""
}
],
"year": 2010,
"venue": "",
"volume": "",
"issue": "",
"pages": "",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Richard Szeliski. 2010. Computer Vision : Algorithms and Applications. Springer-Verlag New York Inc.",
"links": null
},
"BIBREF31": {
"ref_id": "b31",
"title": "From frequency to meaning: Vector space models of semantics",
"authors": [
{
"first": "Peter",
"middle": [],
"last": "Turney",
"suffix": ""
},
{
"first": "Patrick",
"middle": [],
"last": "Pantel",
"suffix": ""
}
],
"year": 2010,
"venue": "Journal of Artificial Intelligence Research",
"volume": "37",
"issue": "",
"pages": "141--188",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Peter Turney and Patrick Pantel. 2010. From frequency to meaning: Vector space models of semantics. Jour- nal of Artificial Intelligence Research, 37:141-188.",
"links": null
},
"BIBREF32": {
"ref_id": "b32",
"title": "Literal and metaphorical sense identification through concrete and abstract context",
"authors": [
{
"first": "Peter",
"middle": [],
"last": "Turney",
"suffix": ""
},
{
"first": "Yair",
"middle": [],
"last": "Neuman",
"suffix": ""
},
{
"first": "Dan",
"middle": [],
"last": "Assaf",
"suffix": ""
},
{
"first": "Yohai",
"middle": [],
"last": "Cohen",
"suffix": ""
}
],
"year": 2011,
"venue": "Proceedings of EMNLP",
"volume": "",
"issue": "",
"pages": "680--690",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Peter Turney, Yair Neuman, Dan Assaf, and Yohai Co- hen. 2011. Literal and metaphorical sense identifi- cation through concrete and abstract context. In Pro- ceedings of EMNLP, pages 680-690, Edinburgh, UK.",
"links": null
},
"BIBREF33": {
"ref_id": "b33",
"title": "VLFeat: An open and portable library of computer vision algorithms",
"authors": [
{
"first": "Andrea",
"middle": [],
"last": "Vedaldi",
"suffix": ""
},
{
"first": "Brian",
"middle": [],
"last": "Fulkerson",
"suffix": ""
}
],
"year": 2008,
"venue": "",
"volume": "",
"issue": "",
"pages": "",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Andrea Vedaldi and Brian Fulkerson. 2008. VLFeat: An open and portable library of computer vision algo- rithms. http://www.vlfeat.org/.",
"links": null
},
"BIBREF34": {
"ref_id": "b34",
"title": "Labeling images with a computer game",
"authors": [
{
"first": "Laura",
"middle": [],
"last": "Luis Von Ahn",
"suffix": ""
},
{
"first": "",
"middle": [],
"last": "Dabbish",
"suffix": ""
}
],
"year": 2004,
"venue": "Proceedings of the SIGCHI Conference on Human Factors in Computing Systems",
"volume": "",
"issue": "",
"pages": "319--326",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Luis von Ahn and Laura Dabbish. 2004. Labeling im- ages with a computer game. In Proceedings of the SIGCHI Conference on Human Factors in Computing Systems, pages 319-326, Vienna, Austria.",
"links": null
},
"BIBREF35": {
"ref_id": "b35",
"title": "Evaluating bag-of-visualwords representations in scene classification",
"authors": [
{
"first": "Jun",
"middle": [],
"last": "Yang",
"suffix": ""
},
{
"first": "Yu-Gang",
"middle": [],
"last": "Jiang",
"suffix": ""
},
{
"first": "Alexander",
"middle": [
"G"
],
"last": "Hauptmann",
"suffix": ""
},
{
"first": "Chong-Wah",
"middle": [],
"last": "Ngo",
"suffix": ""
}
],
"year": 2007,
"venue": "Multimedia Information Retrieval",
"volume": "",
"issue": "",
"pages": "197--206",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Jun Yang, Yu-Gang Jiang, Alexander G. Hauptmann, and Chong-Wah Ngo. 2007. Evaluating bag-of-visual- words representations in scene classification. In Mul- timedia Information Retrieval, pages 197-206.",
"links": null
},
"BIBREF36": {
"ref_id": "b36",
"title": "What are textons?",
"authors": [
{
"first": "",
"middle": [],
"last": "Song Chun Zhu",
"suffix": ""
},
{
"first": "Ying",
"middle": [
"Nian"
],
"last": "Cheng En Guo",
"suffix": ""
},
{
"first": "Yizhou",
"middle": [],
"last": "Wu",
"suffix": ""
},
{
"first": "",
"middle": [],
"last": "Wang",
"suffix": ""
}
],
"year": 2002,
"venue": "Computer Vision -ECCV 2002, 7th European Conference on Computer Vision",
"volume": "",
"issue": "",
"pages": "793--807",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Song Chun Zhu, Cheng en Guo, Ying Nian Wu, and Yizhou Wang. 2002. What are textons? In Computer Vision -ECCV 2002, 7th European Conference on Computer Vision, Copenhagen, Denmark, May 28-31, 2002, Proceedings, Part IV, pages 793-807. Springer.",
"links": null
}
},
"ref_entries": {
"FIGREF0": {
"type_str": "figure",
"uris": null,
"num": null,
"text": "Procedure to build a visual representation for a word, exemplified with SIFT features."
},
"FIGREF1": {
"type_str": "figure",
"uris": null,
"num": null,
"text": "Discrimination of literal (L) vs. nonliteral (N) uses by the best visual and textual models."
}
}
}
} |