File size: 186,307 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 2504 2505 2506 2507 2508 2509 2510 2511 2512 2513 2514 2515 2516 2517 2518 2519 2520 2521 2522 2523 2524 2525 2526 2527 2528 2529 2530 2531 2532 2533 2534 2535 2536 2537 2538 2539 2540 2541 2542 2543 2544 2545 2546 2547 2548 2549 2550 2551 2552 2553 2554 2555 2556 2557 2558 2559 2560 2561 2562 2563 2564 2565 2566 2567 2568 2569 2570 2571 2572 2573 2574 2575 2576 2577 2578 2579 2580 2581 2582 2583 2584 2585 2586 2587 2588 2589 2590 2591 2592 2593 2594 2595 2596 2597 2598 2599 2600 2601 2602 2603 2604 2605 2606 2607 2608 2609 2610 2611 2612 2613 2614 2615 2616 2617 2618 2619 2620 2621 2622 2623 2624 2625 2626 2627 2628 2629 2630 2631 2632 2633 2634 2635 2636 2637 2638 2639 2640 2641 2642 2643 2644 2645 2646 2647 2648 2649 2650 2651 2652 2653 2654 2655 2656 2657 2658 2659 2660 2661 2662 2663 2664 2665 2666 2667 2668 2669 2670 2671 2672 2673 2674 2675 2676 2677 2678 2679 2680 2681 2682 2683 2684 2685 2686 2687 2688 2689 2690 2691 2692 2693 2694 2695 2696 2697 2698 2699 2700 2701 2702 2703 2704 2705 2706 2707 2708 2709 2710 2711 2712 2713 2714 2715 2716 2717 2718 2719 2720 2721 2722 2723 2724 2725 2726 2727 2728 2729 2730 2731 2732 2733 2734 2735 2736 2737 2738 2739 2740 2741 2742 2743 2744 2745 2746 2747 2748 2749 2750 2751 2752 2753 2754 2755 2756 2757 2758 2759 2760 2761 2762 2763 2764 2765 2766 2767 2768 2769 2770 2771 2772 2773 2774 2775 2776 2777 2778 2779 2780 2781 2782 2783 2784 2785 2786 2787 2788 2789 2790 2791 2792 2793 2794 2795 2796 2797 2798 2799 2800 2801 2802 2803 2804 2805 2806 2807 2808 2809 2810 2811 2812 2813 2814 2815 2816 2817 2818 2819 2820 2821 2822 2823 2824 2825 2826 2827 2828 2829 2830 2831 2832 2833 2834 2835 2836 2837 2838 2839 2840 2841 2842 2843 2844 2845 2846 2847 2848 2849 2850 2851 2852 2853 2854 2855 2856 2857 2858 2859 2860 2861 2862 2863 2864 2865 2866 2867 2868 2869 2870 2871 2872 2873 2874 2875 2876 2877 2878 2879 2880 2881 2882 2883 2884 2885 2886 2887 2888 2889 2890 2891 2892 2893 2894 2895 2896 2897 2898 2899 2900 2901 2902 2903 2904 2905 2906 2907 2908 2909 2910 2911 2912 2913 2914 2915 2916 2917 2918 2919 2920 2921 2922 2923 2924 2925 2926 2927 2928 2929 2930 2931 2932 2933 2934 2935 2936 2937 2938 2939 2940 2941 2942 2943 2944 2945 2946 2947 2948 2949 2950 2951 2952 2953 2954 2955 2956 2957 2958 2959 2960 2961 2962 2963 2964 2965 2966 2967 2968 2969 2970 2971 2972 2973 2974 2975 2976 2977 2978 2979 2980 2981 2982 2983 2984 2985 2986 2987 2988 2989 2990 2991 2992 2993 2994 2995 2996 2997 2998 2999 3000 3001 3002 3003 3004 3005 3006 3007 3008 3009 3010 3011 3012 3013 3014 3015 3016 3017 3018 3019 3020 3021 3022 3023 3024 3025 3026 3027 3028 3029 3030 3031 3032 3033 3034 3035 3036 3037 3038 3039 3040 3041 3042 3043 3044 3045 3046 3047 3048 3049 3050 3051 3052 3053 3054 3055 3056 3057 3058 3059 3060 3061 3062 3063 3064 3065 3066 3067 3068 3069 3070 3071 3072 3073 3074 3075 3076 3077 3078 3079 3080 3081 3082 3083 3084 3085 3086 3087 3088 3089 3090 3091 3092 3093 3094 3095 3096 3097 3098 3099 3100 3101 3102 3103 3104 3105 3106 3107 3108 3109 3110 3111 3112 3113 3114 3115 3116 3117 3118 3119 3120 3121 3122 3123 3124 3125 3126 3127 3128 3129 3130 3131 3132 3133 3134 3135 3136 3137 3138 3139 3140 3141 3142 3143 3144 3145 3146 3147 3148 3149 3150 3151 3152 3153 3154 3155 3156 3157 3158 3159 3160 3161 3162 3163 3164 3165 3166 3167 3168 3169 3170 3171 3172 3173 3174 3175 3176 3177 3178 3179 3180 3181 3182 3183 3184 3185 3186 3187 3188 3189 3190 3191 3192 3193 3194 3195 3196 3197 3198 3199 3200 3201 3202 3203 3204 3205 3206 3207 3208 3209 3210 3211 3212 3213 3214 3215 3216 3217 3218 3219 3220 3221 3222 3223 3224 3225 3226 3227 3228 3229 3230 3231 3232 3233 3234 3235 3236 3237 3238 3239 3240 3241 3242 3243 3244 3245 3246 3247 3248 3249 3250 3251 3252 3253 3254 3255 3256 3257 3258 3259 3260 3261 3262 3263 3264 3265 3266 3267 3268 3269 3270 3271 3272 3273 3274 3275 3276 3277 3278 3279 3280 3281 3282 3283 3284 3285 3286 3287 3288 3289 3290 3291 3292 3293 3294 3295 3296 3297 3298 3299 3300 3301 3302 3303 3304 3305 3306 3307 3308 3309 3310 3311 3312 3313 3314 3315 3316 3317 3318 3319 3320 3321 3322 3323 3324 3325 3326 3327 3328 3329 3330 3331 3332 3333 3334 3335 3336 3337 3338 3339 3340 3341 3342 3343 3344 3345 3346 3347 3348 3349 3350 3351 3352 3353 3354 3355 3356 3357 3358 3359 3360 3361 3362 3363 3364 3365 3366 3367 3368 3369 3370 3371 3372 3373 3374 3375 3376 3377 3378 3379 3380 3381 3382 3383 3384 3385 3386 3387 3388 3389 3390 3391 3392 3393 3394 3395 3396 3397 3398 3399 3400 3401 3402 3403 3404 3405 3406 3407 3408 3409 3410 3411 3412 3413 3414 3415 3416 3417 3418 3419 3420 3421 3422 3423 3424 3425 3426 3427 3428 3429 3430 3431 3432 3433 3434 3435 3436 3437 3438 3439 3440 3441 3442 3443 3444 3445 3446 3447 3448 3449 3450 3451 3452 3453 3454 3455 3456 3457 3458 3459 3460 3461 3462 3463 3464 3465 3466 3467 3468 3469 3470 3471 3472 3473 3474 3475 3476 3477 3478 3479 3480 3481 3482 3483 3484 3485 3486 3487 3488 3489 3490 3491 3492 3493 3494 3495 3496 3497 3498 3499 3500 3501 3502 3503 3504 3505 3506 3507 3508 3509 3510 3511 3512 3513 3514 3515 3516 3517 3518 3519 3520 3521 3522 3523 3524 3525 3526 3527 3528 3529 3530 3531 3532 3533 3534 3535 3536 3537 3538 3539 3540 3541 3542 3543 3544 3545 3546 3547 3548 3549 3550 3551 3552 3553 3554 3555 3556 3557 3558 3559 3560 3561 3562 3563 3564 3565 3566 3567 3568 3569 3570 3571 3572 3573 3574 3575 3576 3577 3578 3579 3580 3581 3582 3583 3584 3585 3586 3587 3588 3589 3590 3591 3592 3593 3594 3595 3596 3597 3598 3599 3600 3601 3602 3603 3604 3605 3606 3607 3608 3609 3610 3611 3612 3613 3614 3615 3616 3617 3618 3619 3620 3621 3622 3623 3624 3625 3626 3627 3628 3629 3630 3631 3632 3633 3634 3635 3636 3637 3638 3639 3640 3641 3642 3643 3644 3645 3646 3647 3648 3649 3650 3651 3652 3653 3654 3655 3656 3657 3658 3659 3660 3661 3662 3663 3664 3665 3666 3667 3668 3669 3670 3671 3672 3673 3674 3675 3676 3677 3678 3679 3680 3681 3682 3683 3684 3685 3686 3687 3688 3689 3690 3691 3692 3693 3694 3695 3696 3697 3698 3699 3700 3701 3702 3703 3704 3705 3706 3707 3708 3709 3710 3711 3712 3713 3714 3715 3716 3717 3718 3719 3720 3721 3722 3723 3724 3725 3726 3727 3728 3729 3730 3731 3732 3733 3734 3735 3736 3737 3738 3739 3740 3741 3742 3743 3744 3745 3746 3747 3748 3749 3750 3751 3752 3753 3754 3755 3756 3757 3758 3759 3760 3761 3762 3763 3764 3765 3766 3767 3768 3769 3770 3771 3772 3773 3774 3775 3776 3777 3778 3779 3780 3781 3782 3783 3784 3785 3786 3787 3788 3789 3790 3791 3792 3793 3794 3795 3796 3797 3798 3799 3800 3801 3802 3803 3804 3805 3806 3807 3808 3809 3810 3811 3812 3813 3814 3815 3816 3817 3818 3819 3820 3821 3822 3823 3824 3825 3826 3827 3828 3829 3830 3831 3832 3833 3834 3835 3836 3837 3838 3839 3840 3841 3842 3843 3844 3845 3846 3847 3848 3849 3850 3851 3852 3853 3854 3855 3856 3857 3858 3859 3860 3861 3862 3863 3864 3865 3866 3867 3868 3869 3870 3871 3872 3873 3874 3875 3876 3877 3878 3879 3880 3881 3882 3883 3884 3885 3886 3887 3888 3889 3890 3891 3892 3893 3894 3895 3896 3897 3898 3899 3900 3901 3902 3903 3904 3905 3906 3907 3908 3909 3910 3911 3912 3913 3914 3915 3916 3917 3918 3919 3920 3921 3922 3923 3924 3925 3926 3927 3928 3929 3930 3931 3932 3933 3934 3935 3936 3937 3938 3939 3940 3941 3942 3943 3944 3945 3946 3947 3948 3949 3950 3951 3952 3953 3954 3955 3956 3957 3958 3959 3960 3961 3962 3963 3964 3965 3966 3967 3968 3969 3970 3971 3972 3973 3974 3975 3976 3977 3978 3979 3980 3981 3982 3983 3984 3985 3986 3987 3988 3989 3990 3991 3992 3993 3994 3995 3996 3997 3998 3999 4000 4001 4002 4003 4004 4005 4006 4007 4008 4009 4010 4011 4012 4013 4014 4015 4016 4017 4018 4019 4020 4021 4022 4023 4024 4025 4026 4027 4028 4029 4030 4031 4032 4033 4034 4035 4036 4037 4038 4039 | {
"paper_id": "2021",
"header": {
"generated_with": "S2ORC 1.0.0",
"date_generated": "2023-01-19T13:12:01.301454Z"
},
"title": "User-Friendly Automatic Transcription of Low-Resource Languages: Plugging ESPnet into Elpis",
"authors": [
{
"first": "Oliver",
"middle": [],
"last": "Adams",
"suffix": "",
"affiliation": {},
"email": "oliver.adams@gmail.com"
},
{
"first": "Benjamin",
"middle": [],
"last": "Galliot",
"suffix": "",
"affiliation": {
"laboratory": "Langues et Civilisations \u00e0 Tradition Orale (LACITO)",
"institution": "CNRS-Sorbonne Nouvelle",
"location": {
"country": "France"
}
},
"email": ""
},
{
"first": "Guillaume",
"middle": [],
"last": "Wisniewski",
"suffix": "",
"affiliation": {
"laboratory": "Laboratoire de Linguistique Formelle (LLF)",
"institution": "CNRS",
"location": {
"country": "France"
}
},
"email": "guillaume.wisniewski@u-paris.fr"
},
{
"first": "Nicholas",
"middle": [],
"last": "Lambourne D E",
"suffix": "",
"affiliation": {
"laboratory": "",
"institution": "The University of Queensland",
"location": {
"settlement": "Brisbane",
"country": "Australia"
}
},
"email": ""
},
{
"first": "Ben",
"middle": [],
"last": "Foley D E",
"suffix": "",
"affiliation": {
"laboratory": "",
"institution": "The University of Queensland",
"location": {
"settlement": "Brisbane",
"country": "Australia"
}
},
"email": ""
},
{
"first": "Rahasya",
"middle": [],
"last": "Sanders-Dwyer D E",
"suffix": "",
"affiliation": {
"laboratory": "",
"institution": "The University of Queensland",
"location": {
"settlement": "Brisbane",
"country": "Australia"
}
},
"email": ""
},
{
"first": "Janet",
"middle": [],
"last": "Wiles D E",
"suffix": "",
"affiliation": {
"laboratory": "",
"institution": "The University of Queensland",
"location": {
"settlement": "Brisbane",
"country": "Australia"
}
},
"email": ""
},
{
"first": "Alexis",
"middle": [],
"last": "Michaud",
"suffix": "",
"affiliation": {
"laboratory": "Langues et Civilisations \u00e0 Tradition Orale (LACITO)",
"institution": "CNRS-Sorbonne Nouvelle",
"location": {
"country": "France"
}
},
"email": ""
},
{
"first": "S\u00e9verine",
"middle": [],
"last": "Guillaume",
"suffix": "",
"affiliation": {
"laboratory": "Langues et Civilisations \u00e0 Tradition Orale (LACITO)",
"institution": "CNRS-Sorbonne Nouvelle",
"location": {
"country": "France"
}
},
"email": ""
},
{
"first": "Laurent",
"middle": [],
"last": "Besacier",
"suffix": "",
"affiliation": {
"laboratory": "Laboratoire d'Informatique de Grenoble (LIG)",
"institution": "Universit\u00e9 Grenoble Alpes",
"location": {
"country": "France"
}
},
"email": "laurent.besacier@univ-grenoble-alpes.fr"
},
{
"first": "Christopher",
"middle": [],
"last": "Cox",
"suffix": "",
"affiliation": {
"laboratory": "",
"institution": "University of Alberta",
"location": {
"country": "Canada"
}
},
"email": "cox.christopher@gmail.com"
},
{
"first": "Katya",
"middle": [],
"last": "Aplonova",
"suffix": "",
"affiliation": {
"laboratory": "Langues et Civilisation d'Afrique (LLACAN)",
"institution": "CNRS-INALCO",
"location": {
"settlement": "Langage",
"country": "France"
}
},
"email": ""
},
{
"first": "Guillaume",
"middle": [],
"last": "Jacques",
"suffix": "",
"affiliation": {
"laboratory": "Centre de Recherches Linguistiques sur l'Asie Orientale (CRLAO)",
"institution": "CNRS-EHESS",
"location": {
"country": "France"
}
},
"email": ""
},
{
"first": "Nathan",
"middle": [],
"last": "Hill",
"suffix": "",
"affiliation": {
"laboratory": "",
"institution": "University of London",
"location": {
"country": "United Kingdom"
}
},
"email": ""
}
],
"year": "",
"venue": null,
"identifiers": {},
"abstract": "This paper reports on progress integrating the speech recognition toolkit ESPnet into Elpis, a web front-end originally designed to provide access to the Kaldi automatic speech recognition toolkit. The goal of this work is to make end-to-end speech recognition models available to language workers via a user-friendly graphical interface. Encouraging results are reported on (i) development of an ESPnet recipe for use in Elpis, with preliminary results on data sets previously used for training acoustic models with the Persephone toolkit along with a new data set that had not previously been used in speech recognition, and (ii) incorporating ESPnet into Elpis along with UI enhancements and a CUDA-supported Dockerfile.",
"pdf_parse": {
"paper_id": "2021",
"_pdf_hash": "",
"abstract": [
{
"text": "This paper reports on progress integrating the speech recognition toolkit ESPnet into Elpis, a web front-end originally designed to provide access to the Kaldi automatic speech recognition toolkit. The goal of this work is to make end-to-end speech recognition models available to language workers via a user-friendly graphical interface. Encouraging results are reported on (i) development of an ESPnet recipe for use in Elpis, with preliminary results on data sets previously used for training acoustic models with the Persephone toolkit along with a new data set that had not previously been used in speech recognition, and (ii) incorporating ESPnet into Elpis along with UI enhancements and a CUDA-supported Dockerfile.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Abstract",
"sec_num": null
}
],
"body_text": [
{
"text": "Transcription of speech is an important part of language documentation, and yet speech recognition technology has not been widely harnessed to aid linguists. Despite revolutionary progress in the performance of speech recognition systems in the past decade (Hinton et al., 2012; Hannun et al., 2014; Zeyer et al., 2018; Hadian et al., 2018; Ravanelli et al., 2019; Zhou et al., 2020) , including its application to low-resource languages (Besacier et al., 2014; Blokland et al., 2015; van Esch et al., 2019; Hjortnaes et al., 2020) , these advances are yet to play a common role in language documentation workflows. Speech recognition software often requires effective command line skills and a reasonably detailed understanding of the underlying modeling. People involved in language documentation, language description, and language revitalization projects (this includes, but is not limited to, linguists who carry out fieldwork) seldom have such knowledge. Thus, the tools are largely inaccessible by many people who would benefit from their use.",
"cite_spans": [
{
"start": 257,
"end": 278,
"text": "(Hinton et al., 2012;",
"ref_id": "BIBREF25"
},
{
"start": 279,
"end": 299,
"text": "Hannun et al., 2014;",
"ref_id": "BIBREF21"
},
{
"start": 300,
"end": 319,
"text": "Zeyer et al., 2018;",
"ref_id": "BIBREF65"
},
{
"start": 320,
"end": 340,
"text": "Hadian et al., 2018;",
"ref_id": "BIBREF20"
},
{
"start": 341,
"end": 364,
"text": "Ravanelli et al., 2019;",
"ref_id": "BIBREF47"
},
{
"start": 365,
"end": 383,
"text": "Zhou et al., 2020)",
"ref_id": "BIBREF66"
},
{
"start": 438,
"end": 461,
"text": "(Besacier et al., 2014;",
"ref_id": "BIBREF4"
},
{
"start": 462,
"end": 484,
"text": "Blokland et al., 2015;",
"ref_id": "BIBREF5"
},
{
"start": 485,
"end": 507,
"text": "van Esch et al., 2019;",
"ref_id": "BIBREF14"
},
{
"start": 508,
"end": 531,
"text": "Hjortnaes et al., 2020)",
"ref_id": "BIBREF26"
}
],
"ref_spans": [],
"eq_spans": [],
"section": "Introduction",
"sec_num": "1"
},
{
"text": "Elpis 1 is a tool created to allow language workers with minimal computational experience to build their own speech recognition models and automatically transcribe audio (Foley et al., 2018 . Elpis uses the Kaldi 2 automatic speech recognition (ASR) toolkit (Povey et al., 2011) as its backend. Kaldi is a mature, widely used and well-supported speech recognition toolkit which supports a range of hidden Markov model based speech recognition models.",
"cite_spans": [
{
"start": 170,
"end": 189,
"text": "(Foley et al., 2018",
"ref_id": "BIBREF15"
},
{
"start": 258,
"end": 278,
"text": "(Povey et al., 2011)",
"ref_id": "BIBREF46"
}
],
"ref_spans": [],
"eq_spans": [],
"section": "Introduction",
"sec_num": "1"
},
{
"text": "In this paper we report on the ongoing integration of ESPnet 3 into Elpis as an alternative to the current Kaldi system. We opted to integrate ESPnet as it is a widely used and actively developed tool with state-of-the-art end-to-end neural network models. By supporting ESPnet in Elpis, we aim to bring a wider range of advances in speech recognition to a broad group of users, and provide alternative model options that may better suit some data circumstances, such as an absence of a pronunciation lexicon.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Introduction",
"sec_num": "1"
},
{
"text": "In the rest of this paper, we describe changes to the Elpis toolkit to support the new backend, and preliminary experiments applying our ESPnet recipe to several datasets from a language documentation context. Finally, we discuss plans going forward with this project.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Introduction",
"sec_num": "1"
},
{
"text": "Automatic phonetic/phonemic transcription in language documentation As a subset of speech recognition research, work has been done in applying speech recognition systems to the very low-resource phonemic data scenarios typical in the language documentation context. Encouraging results capitalizing on the advances in speech recognition technology for automatic phonemic transcription in a language documentation context were reported by . Their work used a neural network architecture with connectionist temporal classification (Graves et al., 2006) for phonemic (including tonal) transcription. A command line toolkit was released called Persephone. To assess the reproducibility of the results on other languages, experiments were extended beyond the Chatino, Na and Tsuut'ina data sets, to a sample of languages from the Pangloss Collection, an online archive of under-resourced languages (Michailovsky et al., 2014) . The results confirmed that end-to-end models for automatic phonemic transcription deliver promising performance, and also suggested that preprocessing tasks can to a large extent be automated, thereby increasing the attractiveness of the tool for language documentation workflows (Wisniewski et al., 2020) . Another effort in this space is Allosaurus (Li et al., 2020) , which leverages multilingual models for phonetic transcription and jointly models language independent phones and language-dependent phonemes. This stands as a 3 https://github.com/espnet/espnet promising step towards effective universal phonetic recognition, which would be of great value in the language documentation process.",
"cite_spans": [
{
"start": 529,
"end": 550,
"text": "(Graves et al., 2006)",
"ref_id": "BIBREF18"
},
{
"start": 893,
"end": 920,
"text": "(Michailovsky et al., 2014)",
"ref_id": "BIBREF36"
},
{
"start": 1203,
"end": 1228,
"text": "(Wisniewski et al., 2020)",
"ref_id": "BIBREF61"
},
{
"start": 1274,
"end": 1291,
"text": "(Li et al., 2020)",
"ref_id": "BIBREF33"
}
],
"ref_spans": [],
"eq_spans": [],
"section": "Related Work",
"sec_num": "2"
},
{
"text": "Since such research tools do not have user friendly interfaces, efforts have been put into making these tools accessible to wider audience of users. The authors of Allosaurus provide a web interface online. 4 To integrate Persephone into the language documentation workflow, a plugin, Persephone-ELAN, 5 was developed for ELAN, 6 a piece of software that is widely used for annotation in language documentation (Cox, 2019) .",
"cite_spans": [
{
"start": 207,
"end": 208,
"text": "4",
"ref_id": null
},
{
"start": 411,
"end": 422,
"text": "(Cox, 2019)",
"ref_id": "BIBREF8"
}
],
"ref_spans": [],
"eq_spans": [],
"section": "User-friendly speech recognition interfaces",
"sec_num": null
},
{
"text": "Meanwhile, Elpis is a toolkit that provides a user-friendly front-end to the Kaldi speech recognition system. The interface steps the user through the process of preparing language recordings using existing ELAN transcription files, training a model and applying the model to obtain a hypothesis orthographic transcription for untranscribed speech recordings.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "User-friendly speech recognition interfaces",
"sec_num": null
},
{
"text": "ESPnet is an end-to-end neural network-based speech recognition toolkit. Developed with Pytorch (Paszke et al., 2019) in a research context, the tool satisfies three desiderata for our purposes: (a) it is easy to modify training recipes, which consist of collections of scripts and configuration files that make it easy to perform training and decoding by calling a wrapper script. These recipes describe a wide range of the hyperparameters and architecture choices of the model; (b) it is actively developed, with frequent integration of the latest advances in end-to-end speech recognition; and (c) it supports Kaldi-style data formatting, which makes it a natural end-to-end counterpart to Kaldi backend that was already supported in Elpis. These points make it a more appealing candidate backend than Persephone, primarily due to ESPnet's larger developer base.",
"cite_spans": [
{
"start": 96,
"end": 117,
"text": "(Paszke et al., 2019)",
"ref_id": "BIBREF42"
}
],
"ref_spans": [],
"eq_spans": [],
"section": "Bringing ESPnet to Elpis",
"sec_num": "3"
},
{
"text": "One goal of the integration is to create a default ESPnet recipe for Elpis to use, that performs well across a variety of languages and with the small amount and type of data typically available in a language documentation context. To get a sense of how easy it would be using ESPnet to get similar performance as previously attained we applied it to the single-speaker Na and Chatino datasets as used in Adams et al. (2018) (see Table 1 , which includes other details of the datasets used, including the amount of training data). We report character error rate (CER) rather than phoneme error rate (PER) because it is general, does not require a subsequent languagespecific post-processing step, and also captures characters that a linguist might want transcribed that aren't strictly phonemic. Because of minor differences in the training sets, their preprocessing, and metrics used, these numbers are not intended to be directly comparable with previous work. While these results are not directly comparable to the results they reported, the performance was good enough to confirm that integrating ES-Pnet was preferable to Persephone. We do no language-specific preprocessing, though the Elpis interface allows the user to define a character set for which instances of those characters will be removed from the text. For the Na data and the Japhug data in \u00a74, the Pangloss XML format is converted to ELAN XML using a XSLT-based tool, Pangloss-Elpis. 7 7 gitlab.com/lacito/pangloss-elpis While we did not aggressively tune hyperparameters and architecture details, they do have a substantial impact on performance and computational requirements. Owing to the small datasets and limited computational resources of many of the machines that Elpis may run on, we used a relatively small neural network. In the future we aim to grow a representative suite of evaluation languages from a language documentation setting for further tuning to determine what hyperparameters and architecture best suit different scenarios. Though we aim for a recipe that does well across a range of possible language documentation data circumstances, the best architecture and hyperparameters will vary depending on the characteristics of the input dataset. Rather than have the user fiddle with such hyperparameters directly, which would undermine the user-friendliness of the tool, there is potential to automatically adjust the hyperparameters of the model on the basis of the data supplied to the model. For example, the parameters could be automatically set depending on the number of speakers in the ELAN file and the total amount of speech.",
"cite_spans": [],
"ref_spans": [
{
"start": 430,
"end": 437,
"text": "Table 1",
"ref_id": "TABREF0"
}
],
"eq_spans": [],
"section": "Development of an ESPnet recipe for Elpis",
"sec_num": "3.1"
},
{
"text": "The architecture we used for these experiments is a hybrid CTC-attention model (Watanabe et al., 2017b ) with a 3-layer BiLSTM encoder and a single layer decoder. We use a hidden size of 320 and use an equal weighting between the CTC and ",
"cite_spans": [
{
"start": 79,
"end": 102,
"text": "(Watanabe et al., 2017b",
"ref_id": "BIBREF60"
}
],
"ref_spans": [],
"eq_spans": [],
"section": "Development of an ESPnet recipe for Elpis",
"sec_num": "3.1"
},
{
"text": "Beyond integration of ESPnet into Elpis, several other noteworthy enhancements have been made to Elpis.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Elpis enhancements",
"sec_num": "3.2"
},
{
"text": "Detailed training feedback Prior to the work reported in this paper, the progress of training and transcribing stages was shown as a spinning icon with no other feedback. Due to the amount of time it takes to train even small speech recognition models, the lack of detailed feedback may cause a user to wonder what stage the training was at, or whether a fault had caused the system to fail. During training and transcription, the backend processes' logs are now output to the screen (see Figure 1) . Although the information in these logs may be more detailed than what the intended audience of the tool needs to understand, it does serve to give any user feedback on how training is going, and reassure them that it is still running (or notify them if a process has failed). The logs can also provide useful contextual information when debugging an experiment in collaborations between language workers and software engineers.",
"cite_spans": [],
"ref_spans": [
{
"start": 489,
"end": 498,
"text": "Figure 1)",
"ref_id": "FIGREF0"
}
],
"eq_spans": [],
"section": "Elpis enhancements",
"sec_num": "3.2"
},
{
"text": "CUDA-supported Docker image The type of Kaldi model which Elpis trains (using a Gaussian mixture model as the acoustic model) was originally selected to be computationally efficient, and able to run on the type of computers commonly used by language researchers, which often don't have a GPU (graphics processing unit). With the addition of ESPnet, the benefit of using a GPU will be felt through vastly reduced training times for the neural network. To this end, Elpis has been 8 https://github.com/ persephone-tools/espnet/commit/ 1c529eab738cc8e68617aebbae520f7c9c919081 adapted to include Compute Unified Device Architecture (CUDA) support, which is essential in order to leverage a GPU when training ESPnet on a machine that has one available.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Elpis enhancements",
"sec_num": "3.2"
},
{
"text": "The point of this work is to provide a tool that can be used by linguists in their limited-data scenarios. To this end we aim to experiment with diverse datasets that reflect the breadth of language documentation contexts. Going forward, this will be useful in getting a sense of what sort of model performance users can expect given the characteristics of dataset. In this section we report on further application of the model underpinning the Elpis-ESPnet integration to another data set.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Application to a new data set: Japhug",
"sec_num": "4"
},
{
"text": "Japhug is a Sino-Tibetan language with a rich system of consonant clusters, as well as flamboyant morphology. In Japhug, syllables can have initial clusters containing at most three consonants, and at most one coda (Jacques, 2019) . Japhug does not have lexical tones. The language's phonological profile is thus very different from Na (about which see Michaud, 2017) and Chatino (Cruz, 2011; Cruz and Woodbury, 2014; Cavar et al., 2016) .",
"cite_spans": [
{
"start": 215,
"end": 230,
"text": "(Jacques, 2019)",
"ref_id": null
},
{
"start": 353,
"end": 367,
"text": "Michaud, 2017)",
"ref_id": "BIBREF37"
},
{
"start": 380,
"end": 392,
"text": "(Cruz, 2011;",
"ref_id": "BIBREF9"
},
{
"start": 393,
"end": 417,
"text": "Cruz and Woodbury, 2014;",
"ref_id": "BIBREF10"
},
{
"start": 418,
"end": 437,
"text": "Cavar et al., 2016)",
"ref_id": "BIBREF6"
}
],
"ref_spans": [],
"eq_spans": [],
"section": "Application to a new data set: Japhug",
"sec_num": "4"
},
{
"text": "The data set comprises a total of about 30 hours of transcribed recordings of narratives, timealigned at the level of the sentence (Macaire, 2020) , which is a huge amount in a language documentation context. The recordings were made in the course of field trips from the first years of the century until now, in a quiet environment, and almost all of a single speaker. Our tests on various data sets so far suggest that these settings (one speaker -hence no speaker overlap -and clean audio) are those in which performance is most likely to be good when one happens to be training an acoustic model from scratch.",
"cite_spans": [
{
"start": 131,
"end": 146,
"text": "(Macaire, 2020)",
"ref_id": "BIBREF35"
}
],
"ref_spans": [],
"eq_spans": [],
"section": "Application to a new data set: Japhug",
"sec_num": "4"
},
{
"text": "The full data set is openly accessible online from the Pangloss Collection, under a Creative Commons license, allowing visitors to browse the texts, and computer scientists to try their hand at the data set. 9 The data collector's generous approach to data sharing sets an impressive example, putting into practice some principles which gather increasing support, but which are not yet systematically translated into institutional and editorial policies (Garellek et al., 2020) .",
"cite_spans": [
{
"start": 208,
"end": 209,
"text": "9",
"ref_id": null
},
{
"start": 454,
"end": 477,
"text": "(Garellek et al., 2020)",
"ref_id": "BIBREF17"
}
],
"ref_spans": [],
"eq_spans": [],
"section": "Application to a new data set: Japhug",
"sec_num": "4"
},
{
"text": "The dataset can be downloaded by sending a request to the Cocoon data repository, which hosts the Pangloss Collection. A script, retriever.py, 10 retrieves resources with a certain language name. Data sets can then be created in various ways, such as sorting by speaker (tests suggest that singlespeaker models are a good way to start) and by genre, e.g. excluding materials such as songs, which are a very different kettle of fish from ordinary speech and complicate model training. Figure 2 shows how the phoneme error rate decreases as the amount of training data increases up to 170 minutes. Tests are currently being conducted to verify whether performance stagnates when the amount of data is increased beyond 170 minutes. As with the other experiments the recipe described in \u00a73.1 was used. For each amount of training data, the model was trained for 20 epochs for each of these training runs, with the smaller sets always as a subset of all larger sets. Figure 3 shows the training profile for a given training run using 170 minutes of data. ",
"cite_spans": [],
"ref_spans": [
{
"start": 484,
"end": 492,
"text": "Figure 2",
"ref_id": null
},
{
"start": 962,
"end": 970,
"text": "Figure 3",
"ref_id": null
}
],
"eq_spans": [],
"section": "Application to a new data set: Japhug",
"sec_num": "4"
},
{
"text": "Devoting a section to reflections about adoption of automatic speech recognition tools in language documentation may seem superfluous here. The audience of a conference on the use of computational methods in the study of endangered languages is highly knowledgeable about the difficulties and the rewards of interdisciplinary projects, as a matter of course. But it seemed useful to include a few general thoughts on this topic nonetheless, for the attention of the broader readership which we hope will probe into the Proceedings of the ComputEL-4 conference: colleagues who may consider joining international efforts for wider adoption of natural language processing tools in language documentation workflows. We briefly address a few types of doubts and misgivings.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Challenges concerning adoption of automatic speech recognition tools in language documentation",
"sec_num": "5"
},
{
"text": "A first concern is that automatic speech recognition software is simply too complex for language workers. But it should be recalled that new technologies that seem inaccessible to language workers can be game-changers in linguistics. For instance, the L A T E X software is the typesetting backend used by the journal Glossa (Rooryck, 2016) and by the publishing house Language Science Press (Nordhoff, 2018) , which publish research in linguistics, offering high-quality open-access venues with no author fees or reader fees. Thus, L A T E X, a piece of software which is notorious for its complexity, is used on a large scale in linguistics publishing: Glossa publishes more than 100 articles a year, and Language Science Press about 30 books a year. Key to this success is an organizational setup whereby linguists receive not only a set of stylesheets and instructions, but also hands-on support from a L A T E X expert all along the typesetting process. Undeniably complex software is only accessible to people with no prior knowledge of it if support is available. Automatic speech recognition software should be equally accessible for language workers, given the right organization and setup. Accordingly, special emphasis is placed on user design in the Elpis project. This aspect of the work falls outside of the scope of the present paper, but we wanted to reassure potential users that it is clear to Elpis developers that the goal is to make the technology available to people who do not use the command line. If users can operate software such as ELAN then they will be more than equipped for the skills of uploading ELAN files to Elpis and clicking the Train button.",
"cite_spans": [
{
"start": 325,
"end": 340,
"text": "(Rooryck, 2016)",
"ref_id": null
},
{
"start": 392,
"end": 408,
"text": "(Nordhoff, 2018)",
"ref_id": "BIBREF41"
}
],
"ref_spans": [],
"eq_spans": [],
"section": "Is automatic speech recognition software too complex for language workers?",
"sec_num": "5.1"
},
{
"text": "A second concern among language workers is whether the technology can deliver on its promise, or whether transcription acceleration projects are a case of \"digital innovation fetishism\" (Ampuja, 2020) . Some language workers have reported a feeling that integration of automatic transcription into the language documentation workflow (as described in ) feels out of reach for them. There is no denying that natural language processing tools such as ESPnet and Kaldi are very complex, and that currently, the help of specialists is still needed to make use of this technology in language documentation. However, progress is clearly being made, and a motivated interdisciplinary community is growing at the intersection of language documentation and computer science, comprising linguists who are interested in investing time to learn about natural language processing and computer scientists who want to achieve \"great things with small languages\" (Thieberger and Nordlinger, 2006) . It seems well worth investing in computational methods to assist in the urgent task of documenting the world's languages.",
"cite_spans": [
{
"start": 186,
"end": 200,
"text": "(Ampuja, 2020)",
"ref_id": "BIBREF2"
},
{
"start": 947,
"end": 980,
"text": "(Thieberger and Nordlinger, 2006)",
"ref_id": "BIBREF53"
}
],
"ref_spans": [],
"eq_spans": [],
"section": "Will the technology deliver on its promise?",
"sec_num": "5.2"
},
{
"text": "5.3 Keeping up with the state of the art vs. stabilizing the tool",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Will the technology deliver on its promise?",
"sec_num": "5.2"
},
{
"text": "Finally, a concern among linguists is that the state of the art in computer science is evolving so rapidly that the tool cannot be stabilized, and hence cannot be proposed to language workers for enduring integration into the language documentation workflow. In cases where significant, high-frequency updates are required to keep up with changes in speech recognition software, the investment could be too much for the relatively small communities of programmers involved in transcription acceleration projects. Our optimistic answer is that state-of-the-art code, or code close to the state of the art, need not be difficult to integrate, use or maintain. For example, the developers of Huggingface's Transformers 11 do an impressive job of wrapping the latest and greatest in natural language processing into an easy-to-use interface (Wolf et al., 2019) . They have shown an ability to integrate new models quickly after their initial publication. Usability and stability of the interface is dictated by the quality of the code that is written by the authors of the backend library. If this is done well then the state of the art can be integrated with minimal coding effort by users of the library. For this reason, we are not so concerned about the shifting sands of the underlying building blocks, but the choice of quality backend library does count here. While it is true that there will have to be some modest effort to keep up to date with ESPnet (as would be the case using any other tool), in using ESPnet we are optimistic that the models supported by Elpis can remain up to date with the state of the art without too much hassle.",
"cite_spans": [
{
"start": 837,
"end": 856,
"text": "(Wolf et al., 2019)",
"ref_id": "BIBREF62"
}
],
"ref_spans": [],
"eq_spans": [],
"section": "Will the technology deliver on its promise?",
"sec_num": "5.2"
},
{
"text": "The broader context to the work reported here is a rapidly evolving field in which various initiatives aim to package natural language processing toolkits in intuitive interfaces so as to allow a wider audience to leverage the power of these toolkits. Directions for new developments in Elpis include (i) refining the ESPnet recipe, (ii) refining the user interface through user design processes, (iii) preparing pre-trained models that can be adapted to a small amount of data in a target language, and (iv) providing Elpis as a web service.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Further improvements",
"sec_num": "6"
},
{
"text": "Refinement of the ESPnet recipe that is used in the Elpis pipeline, such that it works as well as possible given the type of data found in language documentation contexts, is a top priority. This work focuses on achieving lower error rates across data sets, starting with refining hyperparameters for model training and extends to other project objectives including providing pre-trained models (see \u00a76.3). This work is of a more experimental nature and can be done largely independently of the Elpis front-end.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Refining the ESPnet recipe",
"sec_num": "6.1"
},
{
"text": "In parallel with the technical integration of ESPnet with Elpis, a user-design process has been investigating how users expect to use these new features. In a series of sessions, linguists and language workers discussed their diverse needs with a designer. The feedback from this process informed the building of a prototype interface based on the latest version of Elpis at the time. The test interface was then used in individual testing sessions to discover points of confusion and uncertainty in the interface. Results of the design process will guide an update to the interface and further work on writing supporting documentation and user guides. The details of this process are beyond the scope of this paper and will be reported separately in future.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Refining the interface",
"sec_num": "6.2"
},
{
"text": "Adapting a trained model to a new language has a long history in speech recognition, having been used both for Hidden Markov Model based systems (Schultz and Waibel, 2001; Le and Besacier, 2005; Stolcke et al., 2006; T\u00f3th et al., 2008; Plahl et al., 2011; Thomas et al., 2012; Imseng et al., 2014; Do et al., 2014; Heigold et al., 2013; Scharenborg et al., 2017) and end-to-end neural systems (Toshniwal et al., 2017; Chiu et al., 2018; M\u00fcller et al., 2017; Dalmia et al., 2018; Watanabe et al., 2017a; Inaguma et al., 2018; Yi et al., 2018; Adams et al., 2019) . In scenarios where data in the target domain or language is limited, leveraging models trained on a number of speakers in different languages often can result in a better performance. The model can learn to cope with acoustic and phonetic characteristics that are common between languages, such as building robustness to channel variability due to different recording conditions, as well as learning common features of phones and sequences of phones between languages.",
"cite_spans": [
{
"start": 145,
"end": 171,
"text": "(Schultz and Waibel, 2001;",
"ref_id": "BIBREF51"
},
{
"start": 172,
"end": 194,
"text": "Le and Besacier, 2005;",
"ref_id": "BIBREF32"
},
{
"start": 195,
"end": 216,
"text": "Stolcke et al., 2006;",
"ref_id": "BIBREF52"
},
{
"start": 217,
"end": 235,
"text": "T\u00f3th et al., 2008;",
"ref_id": "BIBREF56"
},
{
"start": 236,
"end": 255,
"text": "Plahl et al., 2011;",
"ref_id": "BIBREF45"
},
{
"start": 256,
"end": 276,
"text": "Thomas et al., 2012;",
"ref_id": "BIBREF54"
},
{
"start": 277,
"end": 297,
"text": "Imseng et al., 2014;",
"ref_id": "BIBREF28"
},
{
"start": 298,
"end": 314,
"text": "Do et al., 2014;",
"ref_id": "BIBREF13"
},
{
"start": 315,
"end": 336,
"text": "Heigold et al., 2013;",
"ref_id": null
},
{
"start": 337,
"end": 362,
"text": "Scharenborg et al., 2017)",
"ref_id": "BIBREF50"
},
{
"start": 393,
"end": 417,
"text": "(Toshniwal et al., 2017;",
"ref_id": "BIBREF55"
},
{
"start": 418,
"end": 436,
"text": "Chiu et al., 2018;",
"ref_id": "BIBREF7"
},
{
"start": 437,
"end": 457,
"text": "M\u00fcller et al., 2017;",
"ref_id": "BIBREF40"
},
{
"start": 458,
"end": 478,
"text": "Dalmia et al., 2018;",
"ref_id": "BIBREF11"
},
{
"start": 479,
"end": 502,
"text": "Watanabe et al., 2017a;",
"ref_id": "BIBREF58"
},
{
"start": 503,
"end": 524,
"text": "Inaguma et al., 2018;",
"ref_id": "BIBREF29"
},
{
"start": 525,
"end": 541,
"text": "Yi et al., 2018;",
"ref_id": "BIBREF63"
},
{
"start": 542,
"end": 561,
"text": "Adams et al., 2019)",
"ref_id": "BIBREF1"
}
],
"ref_spans": [],
"eq_spans": [],
"section": "Pre-trained models and transfer learning",
"sec_num": "6.3"
},
{
"text": "In recent years pre-training of models on large amounts of unannotated data has led to breakthrough results in text-based natural language processing, initially gaining widespread popularity with the context-independent embeddings of word2vec (Mikolov et al., 2013) and GloVe (Pennington et al., 2014) , before the recent contextual word embedding revolution (Peters et al., 2018; Devlin et al., 2019; Liu et al., 2020) that has harnessed the transformer architecture (Vaswani et al., 2017) . It is now the case that the best approaches in natural language processing are typically characterized by pre-training of a model on a large amount of unannotated data using the cloze task (a.k.a masked language model training) before fine-tuning to a target task. Models pre-trained in this way best make use of available data since the amount of unannotated data far outweighs annotated data and such pre-training is advantageous to downstream learning, whether a small or large amount of data is available in the target task (Gururangan et al., 2020) . Despite the established nature of pre-training in natural language processing, it is less well established in speech recognition, though there has been recent work (Rivi\u00e8re et al., 2020; Baevski et al., 2020) .",
"cite_spans": [
{
"start": 243,
"end": 265,
"text": "(Mikolov et al., 2013)",
"ref_id": "BIBREF39"
},
{
"start": 276,
"end": 301,
"text": "(Pennington et al., 2014)",
"ref_id": "BIBREF43"
},
{
"start": 359,
"end": 380,
"text": "(Peters et al., 2018;",
"ref_id": "BIBREF44"
},
{
"start": 381,
"end": 401,
"text": "Devlin et al., 2019;",
"ref_id": "BIBREF12"
},
{
"start": 402,
"end": 419,
"text": "Liu et al., 2020)",
"ref_id": null
},
{
"start": 468,
"end": 490,
"text": "(Vaswani et al., 2017)",
"ref_id": "BIBREF57"
},
{
"start": 1021,
"end": 1046,
"text": "(Gururangan et al., 2020)",
"ref_id": "BIBREF19"
},
{
"start": 1213,
"end": 1235,
"text": "(Rivi\u00e8re et al., 2020;",
"ref_id": "BIBREF48"
},
{
"start": 1236,
"end": 1257,
"text": "Baevski et al., 2020)",
"ref_id": null
}
],
"ref_spans": [],
"eq_spans": [],
"section": "Pre-trained models and transfer learning",
"sec_num": "6.3"
},
{
"text": "The language documentation scenario, where annotated data is very limited is a scenario that we argue stands most to gain from such pretraining (both supervised and self-supervised outof-domain); followed by model adaptation to limited target language data. One of the features Elpis could provide is to include pre-trained models in its distribution or via an online service. Such models may be pre-trained in a self-supervised manner on lots of untranscribed speech, trained in a supervised manner on transcribed speech in other languages, or use a combination of both pre-training tasks. In cases where the pre-trained model was trained in a supervised manner, there is scope to deploy techniques to reconcile the differences in acoustic realization between phonemes of different languages via methods such as that of Allosaurus (Li et al., 2020) which uses a joint model of language-independent phones and language-dependent phonemes. Providing a variety of pretrained models would be valuable, since the best seed model for adaptation may vary on the basis of the data in the target language (Adams et al., 2019) .",
"cite_spans": [
{
"start": 832,
"end": 849,
"text": "(Li et al., 2020)",
"ref_id": "BIBREF33"
},
{
"start": 1097,
"end": 1117,
"text": "(Adams et al., 2019)",
"ref_id": "BIBREF1"
}
],
"ref_spans": [],
"eq_spans": [],
"section": "Pre-trained models and transfer learning",
"sec_num": "6.3"
},
{
"text": "A recognized problem in language documentation is that, owing to the transcription bottleneck, a large amount of unannotated and untranscribed data ends up in data graveyards (Himmelmann, 2006) : archived recordings that go unused in linguistic research. It is frequently the case that the vast majority of speech collected by field linguists is untranscribed. Here too, self-supervised pretraining in the target language is likely a promising avenue to pursue, perhaps in tandem with supervised pre-training regiments. For this reason, we are optimistic that automatic transcription will have a role to play in almost all data scenarios found in the language documentation contexteven when training data is extremely limited -and is not just reserved for certain single-speaker corpora with consistently high quality audio and clean alignments with text. In the past one could plausibly argue that the limited amount of transcribed speech as training data is an insurmountable hurdle in a language documentation context, but that will likely not remain the case.",
"cite_spans": [
{
"start": 175,
"end": 193,
"text": "(Himmelmann, 2006)",
"ref_id": "BIBREF24"
}
],
"ref_spans": [],
"eq_spans": [],
"section": "Pre-trained models and transfer learning",
"sec_num": "6.3"
},
{
"text": "One of the next steps planned for Elpis is to allow for acoustic models to be exported and loaded. Beyond the immediate benefit of saving the trouble of training models anew each time, having a library of acoustic models available in an online repository would facilitate further research on adaptation of acoustic models to (i) more speakers, and (ii) more language varieties. Building universal phone recognition systems is an active area of research (Li et al., 2020) ; these developments could benefit from the availability of acoustic models on a range of languages. Hosting acoustic models in an online repository, and using them for transfer learning, appear as promising perspectives.",
"cite_spans": [
{
"start": 453,
"end": 470,
"text": "(Li et al., 2020)",
"ref_id": "BIBREF33"
}
],
"ref_spans": [],
"eq_spans": [],
"section": "Pre-trained models and transfer learning",
"sec_num": "6.3"
},
{
"text": "Training models requires a lot of computing power. Elpis now supports high-speed parallel processing in situations where the user's operating system has compatible GPUs (see Section 3.2). However, many users don't have this technology in the computers they have ready access to, so we also plan to investigate possibilities for host-ing Elpis on a high-capacity server for end-user access. Providing language technologies via web services appears to be a successful method of making tools widely available, with examples including the WebMAUS forced-alignment tool. 12 The suite of tools provided by the Bavarian Speech Archive (Kisler et al., 2017) have successfully processed more than ten million media files since their introduction in 2012. For users who want to avoid sending data to a server, there are other possibilities: Kaldi can be compiled to Web Assembly so it can do decoding in a browser (Hu et al., 2020) . But for the type of user scenarios considered here, hosting on a server would have major advantages, and transfer over secure connection is a strong protection against data theft (for those data sets that must not be made public, to follow the consultants' wishes or protect the data collectors' exclusive access rights to the data so that they will not be scooped in research and placed at a disadvantage in job applications).",
"cite_spans": [
{
"start": 566,
"end": 568,
"text": "12",
"ref_id": null
},
{
"start": 628,
"end": 649,
"text": "(Kisler et al., 2017)",
"ref_id": "BIBREF31"
},
{
"start": 904,
"end": 921,
"text": "(Hu et al., 2020)",
"ref_id": "BIBREF27"
}
],
"ref_spans": [],
"eq_spans": [],
"section": "Providing Elpis as a web service",
"sec_num": "6.4"
},
{
"text": "This context suggests that it would be highly desirable to design web hosting for Elpis. It would facilitate conducting broad sets of tests training acoustic models, and would also facilitate the transcription of untranscribed recordings.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Providing Elpis as a web service",
"sec_num": "6.4"
},
{
"text": "In this paper we have reported on integrating ES-Pnet, an end-to-end neural network speech recognition system, into Elpis, the user-friendly speech recognition interface. We described changes that have been made to the front-end, the addition of a CUDA supported Elpis Dockerfile, and the creation of an ESPnet recipe for Elpis. We reported preliminary results on several languages and articulated plans going forward.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Conclusion",
"sec_num": "7"
},
{
"text": "Religion, region, language and the state\" ) and Agence Nationale de la Recherche (as part of two projects, \"Computational Language Documentation by 2025\" [ANR-19-CE38-0015-04] and \"Empirical Foundations of Linguistics\" [ANR-10-LABX-0083]).",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Conclusion",
"sec_num": "7"
},
{
"text": "Linguistic resources used in the present study were collected as part of projects funded by the European Research Council (\"Discourse reporting in African storytelling\" ) and by Agence Nationale de la Recherche (\"Parallel corpora in languages of the Greater Himalayan area\" [ANR-12-CORP-0006]).",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Conclusion",
"sec_num": "7"
},
{
"text": "https://github.com/CoEDL/elpis 2 https://github.com/kaldi-asr/kaldi",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "",
"sec_num": null
},
{
"text": "https://www.dictate.app 5 https://github.com/coxchristopher/ persephone-elan 6 https://archive.mpi.nl/tla/elan",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "",
"sec_num": null
},
{
"text": "Each text has a Digital Object Identifier, allowing for one-click access. Readers are invited to take a look: https: //doi.org/10.24397/pangloss-000336010 https://gitlab.com/lacito/panglosselpis",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "",
"sec_num": null
},
{
"text": "https://github.com/huggingface/ transformers",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "",
"sec_num": null
},
{
"text": "https://clarin.phonetik.uni-muenchen. de/BASWebServices/interface",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "",
"sec_num": null
}
],
"back_matter": [
{
"text": "Many thanks to the three reviewers for comments and suggestions.We are grateful for financial support to the Elpis project from the Australian Research Council Centre of Excellence for the Dynamics of Language, the University of Queensland, the Institut des langues rares (ILARA) at \u00c9cole Pratique des Hautes \u00c9tudes, the European Research Council (as part of the project \"Beyond boundaries:",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Acknowledgments",
"sec_num": null
}
],
"bib_entries": {
"BIBREF0": {
"ref_id": "b0",
"title": "Evaluating phonemic transcription of low-resource tonal languages for language documentation",
"authors": [
{
"first": "Oliver",
"middle": [],
"last": "Adams",
"suffix": ""
},
{
"first": "Trevor",
"middle": [],
"last": "Cohn",
"suffix": ""
},
{
"first": "Graham",
"middle": [],
"last": "Neubig",
"suffix": ""
},
{
"first": "Hilaria",
"middle": [],
"last": "Cruz",
"suffix": ""
},
{
"first": "Steven",
"middle": [],
"last": "Bird",
"suffix": ""
},
{
"first": "Alexis",
"middle": [],
"last": "Michaud",
"suffix": ""
}
],
"year": 2018,
"venue": "Proceedings of LREC 2018 (Language Resources and Evaluation Conference)",
"volume": "",
"issue": "",
"pages": "3356--3365",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Oliver Adams, Trevor Cohn, Graham Neubig, Hi- laria Cruz, Steven Bird, and Alexis Michaud. 2018. Evaluating phonemic transcription of low-resource tonal languages for language documentation. In Proceedings of LREC 2018 (Language Resources and Evaluation Conference), pages 3356-3365, Miyazaki. https://halshs.archives- ouvertes.fr/halshs-01709648.",
"links": null
},
"BIBREF1": {
"ref_id": "b1",
"title": "Massively multilingual adversarial speech recognition",
"authors": [
{
"first": "Oliver",
"middle": [],
"last": "Adams",
"suffix": ""
},
{
"first": "Matthew",
"middle": [],
"last": "Wiesner",
"suffix": ""
},
{
"first": "Shinji",
"middle": [],
"last": "Watanabe",
"suffix": ""
},
{
"first": "David",
"middle": [],
"last": "Yarowsky",
"suffix": ""
}
],
"year": 2019,
"venue": "Proceedings of the 2019 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies",
"volume": "",
"issue": "",
"pages": "96--108",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Oliver Adams, Matthew Wiesner, Shinji Watanabe, and David Yarowsky. 2019. Massively multilin- gual adversarial speech recognition. In Proceed- ings of the 2019 Conference of the North Ameri- can Chapter of the Association for Computational Linguistics: Human Language Technologies, Vol- ume 1 (Long and Short Papers), pages 96-108, Minneapolis, Minnesota. Association for Compu- tational Linguistics. https://www.aclweb. org/anthology/N19-1009.",
"links": null
},
"BIBREF2": {
"ref_id": "b2",
"title": "The blind spots of digital innovation fetishism",
"authors": [
{
"first": "Marko",
"middle": [],
"last": "Ampuja",
"suffix": ""
}
],
"year": 2020,
"venue": "The digital age and its discontents: Critical reflections in education",
"volume": "",
"issue": "",
"pages": "31--54",
"other_ids": {
"DOI": [
"10.33134/HUP-4-2"
]
},
"num": null,
"urls": [],
"raw_text": "Marko Ampuja. 2020. The blind spots of digital in- novation fetishism. In Matteo Stocchetti, editor, The digital age and its discontents: Critical reflec- tions in education, pages 31-54. Helsinki Univer- sity Press, Helsinki. https://doi.org/10. 33134/HUP-4-2.",
"links": null
},
"BIBREF3": {
"ref_id": "b3",
"title": "2020. wav2vec 2.0: A framework for self-supervised learning of speech representations",
"authors": [
{
"first": "Alexei",
"middle": [],
"last": "Baevski",
"suffix": ""
},
{
"first": "Henry",
"middle": [],
"last": "Zhou",
"suffix": ""
},
{
"first": "Abdelrahman",
"middle": [],
"last": "Mohamed",
"suffix": ""
},
{
"first": "Michael",
"middle": [],
"last": "Auli",
"suffix": ""
}
],
"year": null,
"venue": "",
"volume": "",
"issue": "",
"pages": "",
"other_ids": {
"arXiv": [
"arXiv:2006.11477"
]
},
"num": null,
"urls": [],
"raw_text": "Alexei Baevski, Henry Zhou, Abdelrahman Mohamed, and Michael Auli. 2020. wav2vec 2.0: A framework for self-supervised learning of speech representa- tions. arXiv preprint arXiv:2006.11477. https: //arxiv.org/abs/2006.11477.",
"links": null
},
"BIBREF4": {
"ref_id": "b4",
"title": "Automatic speech recognition for under-resourced languages: A survey",
"authors": [
{
"first": "Laurent",
"middle": [],
"last": "Besacier",
"suffix": ""
},
{
"first": "Etienne",
"middle": [],
"last": "Barnard",
"suffix": ""
},
{
"first": "Alexey",
"middle": [],
"last": "Karpov",
"suffix": ""
},
{
"first": "Tanja",
"middle": [],
"last": "Schultz",
"suffix": ""
}
],
"year": 2014,
"venue": "Speech Communication",
"volume": "56",
"issue": "",
"pages": "85--100",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Laurent Besacier, Etienne Barnard, Alexey Karpov, and Tanja Schultz. 2014. Automatic speech recog- nition for under-resourced languages: A survey. Speech Communication, 56:85-100.",
"links": null
},
"BIBREF5": {
"ref_id": "b5",
"title": "Language documentation meets language technology",
"authors": [
{
"first": "Rogier",
"middle": [],
"last": "Blokland",
"suffix": ""
},
{
"first": "Marina",
"middle": [],
"last": "Fedina",
"suffix": ""
},
{
"first": "Ciprian",
"middle": [],
"last": "Gerstenberger",
"suffix": ""
},
{
"first": "Niko",
"middle": [],
"last": "Partanen",
"suffix": ""
},
{
"first": "Michael",
"middle": [],
"last": "Rie\u00dfler",
"suffix": ""
},
{
"first": "Joshua",
"middle": [],
"last": "Wilbur",
"suffix": ""
}
],
"year": 2015,
"venue": "Proceedings of the First International Workshop on Computational Linguistics for Uralic Languages -Septentrio Conference Series",
"volume": "",
"issue": "",
"pages": "8--18",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Rogier Blokland, Marina Fedina, Ciprian Gersten- berger, Niko Partanen, Michael Rie\u00dfler, and Joshua Wilbur. 2015. Language documen- tation meets language technology. In Pro- ceedings of the First International Workshop on Computational Linguistics for Uralic Lan- guages -Septentrio Conference Series, pages 8- 18. http://septentrio.uit.no/index. php/SCS/article/view/3457/3386.",
"links": null
},
"BIBREF6": {
"ref_id": "b6",
"title": "Endangered language documentation: bootstrapping a Chatino speech corpus, forced aligner, ASR",
"authors": [
{
"first": "Ma\u0142gorzata",
"middle": [],
"last": "Cavar",
"suffix": ""
},
{
"first": "Damir",
"middle": [],
"last": "Cavar",
"suffix": ""
},
{
"first": "Hilaria",
"middle": [],
"last": "Cruz",
"suffix": ""
}
],
"year": 2016,
"venue": "Proceedings of LREC 2016, Tenth International Conference on Language Resources and Evaluation",
"volume": "",
"issue": "",
"pages": "4004--4011",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Ma\u0142gorzata Cavar, Damir Cavar, and Hilaria Cruz. 2016. Endangered language documentation: boot- strapping a Chatino speech corpus, forced aligner, ASR. In Proceedings of LREC 2016, Tenth Interna- tional Conference on Language Resources and Eval- uation, pages 4004-4011, Portoro\u017e, Slovenia.",
"links": null
},
"BIBREF7": {
"ref_id": "b7",
"title": "State-of-the-art speech recognition with sequence-to-sequence models",
"authors": [
{
"first": "Chung-Cheng",
"middle": [],
"last": "Chiu",
"suffix": ""
},
{
"first": "Tara",
"middle": [
"N"
],
"last": "Sainath",
"suffix": ""
},
{
"first": "Yonghui",
"middle": [],
"last": "Wu",
"suffix": ""
},
{
"first": "Rohit",
"middle": [],
"last": "Prabhavalkar",
"suffix": ""
},
{
"first": "Patrick",
"middle": [],
"last": "Nguyen",
"suffix": ""
},
{
"first": "Zhifeng",
"middle": [],
"last": "Chen",
"suffix": ""
},
{
"first": "Anjuli",
"middle": [],
"last": "Kannan",
"suffix": ""
},
{
"first": "Ron",
"middle": [
"J"
],
"last": "Weiss",
"suffix": ""
},
{
"first": "Kanishka",
"middle": [],
"last": "Rao",
"suffix": ""
},
{
"first": "Katya",
"middle": [],
"last": "Gonina",
"suffix": ""
}
],
"year": 2018,
"venue": "ICASSP",
"volume": "",
"issue": "",
"pages": "4774--4778",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Chung-Cheng Chiu, Tara N Sainath, Yonghui Wu, Ro- hit Prabhavalkar, Patrick Nguyen, Zhifeng Chen, Anjuli Kannan, Ron J Weiss, Kanishka Rao, Katya Gonina, et al. 2018. State-of-the-art speech recogni- tion with sequence-to-sequence models. In ICASSP, pages 4774-4778. https://arxiv.org/abs/ 1712.01769.",
"links": null
},
"BIBREF8": {
"ref_id": "b8",
"title": "Persephone-ELAN (software)",
"authors": [
{
"first": "Christopher",
"middle": [],
"last": "Cox",
"suffix": ""
}
],
"year": 2019,
"venue": "",
"volume": "",
"issue": "",
"pages": "",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Christopher Cox. 2019. Persephone-ELAN (software). https://github.com/coxchristopher/persephone-elan. https://github.com/coxchristopher/ persephone-elan.",
"links": null
},
"BIBREF9": {
"ref_id": "b9",
"title": "Phonology, tone and the functions of tone in San Juan Quiahije Chatino",
"authors": [
{
"first": "Emiliana",
"middle": [],
"last": "Cruz",
"suffix": ""
}
],
"year": 2011,
"venue": "Ph.D",
"volume": "",
"issue": "",
"pages": "",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Emiliana Cruz. 2011. Phonology, tone and the functions of tone in San Juan Quiahije Chatino. Ph.D., University of Texas at Austin, Austin. http://hdl.handle.net/2152/ ETD-UT-2011-08-4280.",
"links": null
},
"BIBREF10": {
"ref_id": "b10",
"title": "Finding a way into a family of tone languages: The story and methods of the Chatino Language Documentation Project. Language Documentation and Conservation",
"authors": [
{
"first": "Emiliana",
"middle": [],
"last": "Cruz",
"suffix": ""
},
{
"first": "Tony",
"middle": [],
"last": "Woodbury",
"suffix": ""
}
],
"year": 2014,
"venue": "",
"volume": "8",
"issue": "",
"pages": "490--524",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Emiliana Cruz and Tony Woodbury. 2014. Finding a way into a family of tone languages: The story and methods of the Chatino Language Documentation Project. Language Documentation and Conserva- tion, 8:490-524. http://hdl.handle.net/ 10125/24615.",
"links": null
},
"BIBREF11": {
"ref_id": "b11",
"title": "Sequence-based multilingual low resource speech recognition",
"authors": [
{
"first": "Siddharth",
"middle": [],
"last": "Dalmia",
"suffix": ""
},
{
"first": "Ramon",
"middle": [],
"last": "Sanabria",
"suffix": ""
},
{
"first": "Florian",
"middle": [],
"last": "Metze",
"suffix": ""
},
{
"first": "Alan",
"middle": [
"W"
],
"last": "Black",
"suffix": ""
}
],
"year": 2018,
"venue": "ICASSP",
"volume": "",
"issue": "",
"pages": "",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Siddharth Dalmia, Ramon Sanabria, Florian Metze, and Alan W Black. 2018. Sequence-based multi- lingual low resource speech recognition. In ICASSP. https://arxiv.org/abs/1802.07420.",
"links": null
},
"BIBREF12": {
"ref_id": "b12",
"title": "BERT: Pre-training of deep bidirectional transformers for language understanding",
"authors": [
{
"first": "Jacob",
"middle": [],
"last": "Devlin",
"suffix": ""
},
{
"first": "Ming-Wei",
"middle": [],
"last": "Chang",
"suffix": ""
},
{
"first": "Kenton",
"middle": [],
"last": "Lee",
"suffix": ""
},
{
"first": "Kristina",
"middle": [],
"last": "Toutanova",
"suffix": ""
}
],
"year": 2019,
"venue": "Proceedings of the 2019 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies",
"volume": "1",
"issue": "",
"pages": "4171--4186",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Jacob Devlin, Ming-Wei Chang, Kenton Lee, and Kristina Toutanova. 2019. BERT: Pre-training of deep bidirectional transformers for language under- standing. In Proceedings of the 2019 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies, Volume 1 (Long and Short Papers), pages 4171-4186, Minneapolis, Minnesota. Asso- ciation for Computational Linguistics. https:// www.aclweb.org/anthology/N19-1423.",
"links": null
},
"BIBREF13": {
"ref_id": "b13",
"title": "Cross-lingual phone mapping for large vocabulary speech recognition of underresourced languages",
"authors": [
{
"first": "Xiong",
"middle": [],
"last": "Van Hai Do",
"suffix": ""
},
{
"first": "",
"middle": [],
"last": "Xiao",
"suffix": ""
},
{
"first": "Haizhou",
"middle": [],
"last": "Eng Siong Chng",
"suffix": ""
},
{
"first": "",
"middle": [],
"last": "Li",
"suffix": ""
}
],
"year": 2014,
"venue": "IEICE Transactions on Information and Systems",
"volume": "",
"issue": "2",
"pages": "285--295",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Van Hai Do, Xiong Xiao, Eng Siong Chng, and Haizhou Li. 2014. Cross-lingual phone mapping for large vocabulary speech recognition of under- resourced languages. IEICE Transactions on Infor- mation and Systems, E97-D(2):285-295.",
"links": null
},
"BIBREF14": {
"ref_id": "b14",
"title": "Future directions in technological support for language documentation",
"authors": [
{
"first": "Ben",
"middle": [],
"last": "Daan Van Esch",
"suffix": ""
},
{
"first": "Nay",
"middle": [],
"last": "Foley",
"suffix": ""
},
{
"first": "",
"middle": [],
"last": "San",
"suffix": ""
}
],
"year": 2019,
"venue": "Proceedings of the Workshop on Computational Methods for Endangered Languages",
"volume": "1",
"issue": "",
"pages": "",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Daan van Esch, Ben Foley, and Nay San. 2019. Future directions in technological support for language documentation. In Proceedings of the Workshop on Computational Methods for Endan- gered Languages, volume 1, Honolulu, Hawai'i.",
"links": null
},
"BIBREF15": {
"ref_id": "b15",
"title": "Building speech recognition systems for language documentation: the CoEDL Endangered Language Pipeline and Inference System (ELPIS)",
"authors": [
{
"first": "Ben",
"middle": [],
"last": "Foley",
"suffix": ""
},
{
"first": "Josh",
"middle": [],
"last": "Arnold",
"suffix": ""
},
{
"first": "Rolando",
"middle": [],
"last": "Coto-Solano",
"suffix": ""
},
{
"first": "Gautier",
"middle": [],
"last": "Durantin",
"suffix": ""
},
{
"first": "T",
"middle": [
"Mark"
],
"last": "Ellison",
"suffix": ""
}
],
"year": 2018,
"venue": "Proceedings of the 6th Intl. Workshop on Spoken Language Technologies for Under-Resourced Languages (SLTU)",
"volume": "",
"issue": "",
"pages": "200--204",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Ben Foley, Josh Arnold, Rolando Coto-Solano, Gau- tier Durantin, and T. Mark Ellison. 2018. Building speech recognition systems for language documen- tation: the CoEDL Endangered Language Pipeline and Inference System (ELPIS). In Proceedings of the 6th Intl. Workshop on Spoken Language Tech- nologies for Under-Resourced Languages (SLTU), 29-31 August 2018, pages 200-204, Gurugram, India. ISCA. https://www.isca-speech. org/archive/SLTU_2018/pdfs/Ben.pdf.",
"links": null
},
"BIBREF16": {
"ref_id": "b16",
"title": "Elpis, an accessible speech-to-text tool",
"authors": [
{
"first": "Ben",
"middle": [],
"last": "Foley",
"suffix": ""
},
{
"first": "Alina",
"middle": [],
"last": "Rakhi",
"suffix": ""
},
{
"first": "Nicholas",
"middle": [],
"last": "Lambourne",
"suffix": ""
},
{
"first": "Nicholas",
"middle": [],
"last": "Buckeridge",
"suffix": ""
},
{
"first": "Janet",
"middle": [],
"last": "Wiles",
"suffix": ""
}
],
"year": 2019,
"venue": "Proceedings of Interspeech 2019",
"volume": "",
"issue": "",
"pages": "306--310",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Ben Foley, Alina Rakhi, Nicholas Lambourne, Nicholas Buckeridge, and Janet Wiles. 2019. Elpis, an accessible speech-to-text tool. In Proceedings of Interspeech 2019, pages 306-310, Graz. https: //www.isca-speech.org/archive/ Interspeech_2019/pdfs/8006.pdf.",
"links": null
},
"BIBREF17": {
"ref_id": "b17",
"title": "Toward open data policies in phonetics: What we can gain and how we can avoid pitfalls",
"authors": [
{
"first": "Marc",
"middle": [],
"last": "Garellek",
"suffix": ""
},
{
"first": "Matthew",
"middle": [],
"last": "Gordon",
"suffix": ""
},
{
"first": "James",
"middle": [],
"last": "Kirby",
"suffix": ""
},
{
"first": "Wai-Sum",
"middle": [],
"last": "Lee",
"suffix": ""
},
{
"first": "Alexis",
"middle": [],
"last": "Michaud",
"suffix": ""
},
{
"first": "Christine",
"middle": [],
"last": "Mooshammer",
"suffix": ""
},
{
"first": "Oliver",
"middle": [],
"last": "Niebuhr",
"suffix": ""
},
{
"first": "Daniel",
"middle": [],
"last": "Recasens",
"suffix": ""
},
{
"first": "Timo",
"middle": [],
"last": "Roettger",
"suffix": ""
},
{
"first": "Adrian",
"middle": [],
"last": "Simpson",
"suffix": ""
},
{
"first": "Kristine",
"middle": [
"M"
],
"last": "Yu",
"suffix": ""
}
],
"year": 2020,
"venue": "Journal of Speech Science",
"volume": "9",
"issue": "1",
"pages": "",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Marc Garellek, Matthew Gordon, James Kirby, Wai- Sum Lee, Alexis Michaud, Christine Mooshammer, Oliver Niebuhr, Daniel Recasens, Timo Roettger, Adrian Simpson, and Kristine M. Yu. 2020. Toward open data policies in phonetics: What we can gain and how we can avoid pitfalls. Journal of Speech Science, 9(1). https://halshs.archives- ouvertes.fr/halshs-02894375.",
"links": null
},
"BIBREF18": {
"ref_id": "b18",
"title": "Connectionist Temporal Classification : Labelling Unsegmented Sequence Data with Recurrent Neural Networks",
"authors": [
{
"first": "Alex",
"middle": [],
"last": "Graves",
"suffix": ""
},
{
"first": "Santiago",
"middle": [],
"last": "Fernandez",
"suffix": ""
},
{
"first": "Faustino",
"middle": [],
"last": "Gomez",
"suffix": ""
},
{
"first": "Jurgen",
"middle": [],
"last": "Schmidhuber",
"suffix": ""
}
],
"year": 2006,
"venue": "Proceedings of the 23rd international conference on Machine Learning",
"volume": "",
"issue": "",
"pages": "369--376",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Alex Graves, Santiago Fernandez, Faustino Gomez, and Jurgen Schmidhuber. 2006. Connection- ist Temporal Classification : Labelling Unseg- mented Sequence Data with Recurrent Neural Networks. Proceedings of the 23rd interna- tional conference on Machine Learning, pages 369-376. http://www.cs.utoronto.ca/ graves/icml_2006.pdf.",
"links": null
},
"BIBREF19": {
"ref_id": "b19",
"title": "Don't stop pretraining: Adapt language models to domains and tasks",
"authors": [
{
"first": "Ana",
"middle": [],
"last": "Suchin Gururangan",
"suffix": ""
},
{
"first": "Swabha",
"middle": [],
"last": "Marasovi\u0107",
"suffix": ""
},
{
"first": "Kyle",
"middle": [],
"last": "Swayamdipta",
"suffix": ""
},
{
"first": "Iz",
"middle": [],
"last": "Lo",
"suffix": ""
},
{
"first": "Doug",
"middle": [],
"last": "Beltagy",
"suffix": ""
},
{
"first": "Noah",
"middle": [
"A"
],
"last": "Downey",
"suffix": ""
},
{
"first": "",
"middle": [],
"last": "Smith",
"suffix": ""
}
],
"year": 2020,
"venue": "Proceedings of the 58th Annual Meeting of the Association for Computational Linguistics",
"volume": "",
"issue": "",
"pages": "8342--8360",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Suchin Gururangan, Ana Marasovi\u0107, Swabha Swayamdipta, Kyle Lo, Iz Beltagy, Doug Downey, and Noah A. Smith. 2020. Don't stop pretraining: Adapt language models to domains and tasks. In Proceedings of the 58th Annual Meeting of the Asso- ciation for Computational Linguistics, pages 8342- 8360. Association for Computational Linguistics. https://arxiv.org/abs/2004.10964.",
"links": null
},
"BIBREF20": {
"ref_id": "b20",
"title": "End-to-end speech recognition using lattice-free MMI",
"authors": [
{
"first": "Hossein",
"middle": [],
"last": "Hadian",
"suffix": ""
},
{
"first": "Hossein",
"middle": [],
"last": "Sameti",
"suffix": ""
},
{
"first": "Daniel",
"middle": [],
"last": "Povey",
"suffix": ""
},
{
"first": "Sanjeev",
"middle": [],
"last": "Khudanpur",
"suffix": ""
}
],
"year": 2018,
"venue": "Interspeech",
"volume": "",
"issue": "",
"pages": "12--16",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Hossein Hadian, Hossein Sameti, Daniel Povey, and Sanjeev Khudanpur. 2018. End-to-end speech recognition using lattice-free MMI. In Interspeech, pages 12-16. https://danielpovey.com/ files/2018_interspeech_end2end.pdf.",
"links": null
},
"BIBREF21": {
"ref_id": "b21",
"title": "Deep speech: Scaling up end-to-end speech recognition",
"authors": [
{
"first": "Awni",
"middle": [],
"last": "Hannun",
"suffix": ""
},
{
"first": "Carl",
"middle": [],
"last": "Case",
"suffix": ""
},
{
"first": "Jared",
"middle": [],
"last": "Casper",
"suffix": ""
},
{
"first": "Bryan",
"middle": [],
"last": "Catanzaro",
"suffix": ""
},
{
"first": "Greg",
"middle": [],
"last": "Diamos",
"suffix": ""
},
{
"first": "Erich",
"middle": [],
"last": "Elsen",
"suffix": ""
},
{
"first": "Ryan",
"middle": [],
"last": "Prenger",
"suffix": ""
},
{
"first": "Sanjeev",
"middle": [],
"last": "Satheesh",
"suffix": ""
},
{
"first": "Shubho",
"middle": [],
"last": "Sengupta",
"suffix": ""
},
{
"first": "Adam",
"middle": [],
"last": "Coates",
"suffix": ""
}
],
"year": 2014,
"venue": "",
"volume": "",
"issue": "",
"pages": "",
"other_ids": {
"arXiv": [
"arXiv:1412.5567"
]
},
"num": null,
"urls": [],
"raw_text": "Awni Hannun, Carl Case, Jared Casper, Bryan Catanzaro, Greg Diamos, Erich Elsen, Ryan Prenger, Sanjeev Satheesh, Shubho Sengupta, Adam Coates, et al. 2014. Deep speech: Scaling up end-to-end speech recognition. arXiv preprint arXiv:1412.5567. https://arxiv.org/abs/ 1412.5567.",
"links": null
},
"BIBREF23": {
"ref_id": "b23",
"title": "Multilingual acoustic models using distributed deep neural networks",
"authors": [
{
"first": "Devin",
"middle": [],
"last": "",
"suffix": ""
},
{
"first": "Jeffrey",
"middle": [],
"last": "Dean",
"suffix": ""
}
],
"year": 2013,
"venue": "Proceedings of ICASSP",
"volume": "",
"issue": "",
"pages": "8619--8623",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Devin, and Jeffrey Dean. 2013. Multilingual acous- tic models using distributed deep neural networks. In Proceedings of ICASSP, pages 8619-8623.",
"links": null
},
"BIBREF24": {
"ref_id": "b24",
"title": "Language documentation: what is it and what is it good for?",
"authors": [
{
"first": "Nikolaus",
"middle": [],
"last": "Himmelmann",
"suffix": ""
}
],
"year": 2006,
"venue": "Josh Gippert, Nikolaus Himmelmann, and Ulrike Mosel",
"volume": "",
"issue": "",
"pages": "1--30",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Nikolaus Himmelmann. 2006. Language documenta- tion: what is it and what is it good for? In Josh Gip- pert, Nikolaus Himmelmann, and Ulrike Mosel, ed- itors, Essentials of language documentation, pages 1-30. De Gruyter, Berlin/New York.",
"links": null
},
"BIBREF25": {
"ref_id": "b25",
"title": "Deep neural networks for acoustic modeling in speech recognition: The shared views of four research groups",
"authors": [
{
"first": "Geoffrey",
"middle": [],
"last": "Hinton",
"suffix": ""
},
{
"first": "Li",
"middle": [],
"last": "Deng",
"suffix": ""
},
{
"first": "Dong",
"middle": [],
"last": "Yu",
"suffix": ""
},
{
"first": "George",
"middle": [
"E"
],
"last": "Dahl",
"suffix": ""
},
{
"first": "Abdel-Rahman",
"middle": [],
"last": "Mohamed",
"suffix": ""
},
{
"first": "Navdeep",
"middle": [],
"last": "Jaitly",
"suffix": ""
},
{
"first": "Andrew",
"middle": [],
"last": "Senior",
"suffix": ""
},
{
"first": "Vincent",
"middle": [],
"last": "Vanhoucke",
"suffix": ""
},
{
"first": "Patrick",
"middle": [],
"last": "Nguyen",
"suffix": ""
},
{
"first": "Tara",
"middle": [
"N"
],
"last": "Sainath",
"suffix": ""
},
{
"first": "Others",
"middle": [],
"last": "",
"suffix": ""
}
],
"year": 2012,
"venue": "Signal Processing Magazine",
"volume": "29",
"issue": "6",
"pages": "82--97",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Geoffrey Hinton, Li Deng, Dong Yu, George E Dahl, Abdel-rahman Mohamed, Navdeep Jaitly, Andrew Senior, Vincent Vanhoucke, Patrick Nguyen, Tara N Sainath, and Others. 2012. Deep neural networks for acoustic modeling in speech recognition: The shared views of four research groups. Signal Processing Magazine, IEEE, 29(6):82-97.",
"links": null
},
"BIBREF26": {
"ref_id": "b26",
"title": "Towards a speech recognizer for Komi, an endangered and low-resource Uralic language",
"authors": [
{
"first": "Nils",
"middle": [],
"last": "Hjortnaes",
"suffix": ""
},
{
"first": "Niko",
"middle": [],
"last": "Partanen",
"suffix": ""
},
{
"first": "Michael",
"middle": [],
"last": "Rie\u00dfler",
"suffix": ""
},
{
"first": "Francis",
"middle": [
"M"
],
"last": "Tyers",
"suffix": ""
}
],
"year": 2020,
"venue": "Proceedings of the Sixth International Workshop on Computational Linguistics of Uralic Languages",
"volume": "",
"issue": "",
"pages": "31--37",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Nils Hjortnaes, Niko Partanen, Michael Rie\u00dfler, and Francis M. Tyers. 2020. Towards a speech recognizer for Komi, an endangered and low-resource Uralic language. In Proceed- ings of the Sixth International Workshop on Computational Linguistics of Uralic Languages, pages 31-37, Wien. Association for Computa- tional Linguistics. https://www.aclweb.",
"links": null
},
"BIBREF27": {
"ref_id": "b27",
"title": "Kaldi-web: An installationfree, on-device speech recognition system",
"authors": [
{
"first": "Mathieu",
"middle": [],
"last": "Hu",
"suffix": ""
},
{
"first": "Laurent",
"middle": [],
"last": "Pierron",
"suffix": ""
},
{
"first": "Emmanuel",
"middle": [],
"last": "Vincent",
"suffix": ""
},
{
"first": "Denis",
"middle": [],
"last": "Jouvet",
"suffix": ""
}
],
"year": 2020,
"venue": "Proceedings of Interspeech 2020 Show & Tell, Shanghai",
"volume": "",
"issue": "",
"pages": "",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Mathieu Hu, Laurent Pierron, Emmanuel Vincent, and Denis Jouvet. 2020. Kaldi-web: An installation- free, on-device speech recognition system. In Pro- ceedings of Interspeech 2020 Show & Tell, Shang- hai. https://hal.archives-ouvertes. fr/hal-02910876.",
"links": null
},
"BIBREF28": {
"ref_id": "b28",
"title": "Using out-of-language data to improve an under-resourced speech recognizer",
"authors": [
{
"first": "David",
"middle": [],
"last": "Imseng",
"suffix": ""
},
{
"first": "Petr",
"middle": [],
"last": "Motlicek",
"suffix": ""
},
{
"first": "Herv\u00e9",
"middle": [],
"last": "Bourlard",
"suffix": ""
},
{
"first": "Philip N",
"middle": [],
"last": "Garner",
"suffix": ""
}
],
"year": 2014,
"venue": "Speech Communication",
"volume": "56",
"issue": "",
"pages": "142--151",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "David Imseng, Petr Motlicek, Herv\u00e9 Bourlard, and Philip N Garner. 2014. Using out-of-language data to improve an under-resourced speech recognizer. Speech Communication, 56:142-151.",
"links": null
},
"BIBREF29": {
"ref_id": "b29",
"title": "Transfer learning of languageindependent end-to-end ASR with language model fusion",
"authors": [
{
"first": "Hirofumi",
"middle": [],
"last": "Inaguma",
"suffix": ""
},
{
"first": "Jaejin",
"middle": [],
"last": "Cho",
"suffix": ""
},
{
"first": "Murali",
"middle": [
"Karthick"
],
"last": "Baskar",
"suffix": ""
},
{
"first": "Tatsuya",
"middle": [],
"last": "Kawahara",
"suffix": ""
},
{
"first": "Shinji",
"middle": [],
"last": "Watanabe",
"suffix": ""
}
],
"year": 2018,
"venue": "",
"volume": "",
"issue": "",
"pages": "",
"other_ids": {
"arXiv": [
"arXiv:1811.02134"
]
},
"num": null,
"urls": [],
"raw_text": "Hirofumi Inaguma, Jaejin Cho, Murali Karthick Baskar, Tatsuya Kawahara, and Shinji Watan- abe. 2018. Transfer learning of language- independent end-to-end ASR with language model fusion. arXiv:1811.02134. https://arxiv. org/abs/1811.02134.",
"links": null
},
"BIBREF31": {
"ref_id": "b31",
"title": "Multilingual processing of speech via web services",
"authors": [
{
"first": "Thomas",
"middle": [],
"last": "Kisler",
"suffix": ""
},
{
"first": "Uwe",
"middle": [],
"last": "Reichel",
"suffix": ""
},
{
"first": "Florian",
"middle": [],
"last": "Schiel",
"suffix": ""
}
],
"year": 2017,
"venue": "Computer Speech & Language",
"volume": "45",
"issue": "",
"pages": "885--2308",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Thomas Kisler, Uwe Reichel, and Florian Schiel. 2017. Multilingual processing of speech via web services. Computer Speech & Language, 45:326-347. ISBN: 0885-2308 Publisher: Elsevier.",
"links": null
},
"BIBREF32": {
"ref_id": "b32",
"title": "First steps in fast acoustic modeling for a new target language: application to Vietnamese",
"authors": [
{
"first": "Bac",
"middle": [],
"last": "Viet",
"suffix": ""
},
{
"first": "Laurent",
"middle": [],
"last": "Le",
"suffix": ""
},
{
"first": "",
"middle": [],
"last": "Besacier",
"suffix": ""
}
],
"year": 2005,
"venue": "ICASSP",
"volume": "",
"issue": "",
"pages": "",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Viet Bac Le and Laurent Besacier. 2005. First steps in fast acoustic modeling for a new target language: application to Vietnamese. In ICASSP.",
"links": null
},
"BIBREF33": {
"ref_id": "b33",
"title": "Universal phone recognition with a multilingual allophone system",
"authors": [
{
"first": "Xinjian",
"middle": [],
"last": "Li",
"suffix": ""
},
{
"first": "Siddharth",
"middle": [],
"last": "Dalmia",
"suffix": ""
},
{
"first": "Juncheng",
"middle": [],
"last": "Li",
"suffix": ""
},
{
"first": "Matthew",
"middle": [],
"last": "Lee",
"suffix": ""
},
{
"first": "Patrick",
"middle": [],
"last": "Littell",
"suffix": ""
},
{
"first": "Jiali",
"middle": [],
"last": "Yao",
"suffix": ""
},
{
"first": "Antonios",
"middle": [],
"last": "Anastasopoulos",
"suffix": ""
},
{
"first": "David",
"middle": [
"R"
],
"last": "Mortensen",
"suffix": ""
},
{
"first": "Graham",
"middle": [],
"last": "Neubig",
"suffix": ""
},
{
"first": "Alan",
"middle": [
"W"
],
"last": "Black",
"suffix": ""
}
],
"year": 2020,
"venue": "ICASSP 2020-2020 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP)",
"volume": "",
"issue": "",
"pages": "8249--8253",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Xinjian Li, Siddharth Dalmia, Juncheng Li, Matthew Lee, Patrick Littell, Jiali Yao, Antonios Anas- tasopoulos, David R. Mortensen, Graham Neu- big, and Alan W. Black. 2020. Universal phone recognition with a multilingual allophone system. In ICASSP 2020-2020 IEEE International Confer- ence on Acoustics, Speech and Signal Processing (ICASSP), pages 8249-8253. IEEE. https:// arxiv.org/abs/2002.11800.",
"links": null
},
"BIBREF34": {
"ref_id": "b34",
"title": "2020. A survey on contextual embeddings",
"authors": [
{
"first": "Qi",
"middle": [],
"last": "Liu",
"suffix": ""
},
{
"first": "J",
"middle": [],
"last": "Matt",
"suffix": ""
},
{
"first": "Phil",
"middle": [],
"last": "Kusner",
"suffix": ""
},
{
"first": "",
"middle": [],
"last": "Blunsom",
"suffix": ""
}
],
"year": null,
"venue": "",
"volume": "",
"issue": "",
"pages": "",
"other_ids": {
"arXiv": [
"arXiv:2003.07278"
]
},
"num": null,
"urls": [],
"raw_text": "Qi Liu, Matt J Kusner, and Phil Blunsom. 2020. A survey on contextual embeddings. arXiv preprint arXiv:2003.07278. https://arxiv. org/abs/2003.07278.",
"links": null
},
"BIBREF35": {
"ref_id": "b35",
"title": "Alignement temporel entre transcriptions et audio de donn\u00e9es de langue japhug",
"authors": [
{
"first": "C\u00e9cile",
"middle": [],
"last": "Macaire",
"suffix": ""
}
],
"year": 2020,
"venue": "Actes des Journ\u00e9es scientifiques du Groupement de Recherche \"Linguistique informatique, formelle et de terrain\" (LIFT)",
"volume": "",
"issue": "",
"pages": "",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "C\u00e9cile Macaire. 2020. Alignement temporel entre tran- scriptions et audio de donn\u00e9es de langue japhug. In Actes des Journ\u00e9es scientifiques du Groupement de Recherche \"Linguistique informatique, formelle et de terrain\" (LIFT), Paris. https://hal. archives-ouvertes.fr/hal-03047146.",
"links": null
},
"BIBREF36": {
"ref_id": "b36",
"title": "Documenting and researching endangered languages: the Pangloss Collection. Language Documentation and Conservation",
"authors": [
{
"first": "Boyd",
"middle": [],
"last": "Michailovsky",
"suffix": ""
},
{
"first": "Martine",
"middle": [],
"last": "Mazaudon",
"suffix": ""
},
{
"first": "Alexis",
"middle": [],
"last": "Michaud",
"suffix": ""
},
{
"first": "S\u00e9verine",
"middle": [],
"last": "Guillaume",
"suffix": ""
}
],
"year": 2014,
"venue": "",
"volume": "8",
"issue": "",
"pages": "119--135",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Boyd Michailovsky, Martine Mazaudon, Alexis Michaud, S\u00e9verine Guillaume, Alexandre Fran\u00e7ois, and Evangelia Adamou. 2014. Doc- umenting and researching endangered lan- guages: the Pangloss Collection. Language Documentation and Conservation, 8:119- 135. https://halshs.archives- ouvertes.fr/halshs-01003734.",
"links": null
},
"BIBREF37": {
"ref_id": "b37",
"title": "Tone in Yongning Na: lexical tones and morphotonology",
"authors": [
{
"first": "Alexis",
"middle": [],
"last": "Michaud",
"suffix": ""
}
],
"year": 2017,
"venue": "Number 13 in Studies in Diversity Linguistics",
"volume": "",
"issue": "",
"pages": "",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Alexis Michaud. 2017. Tone in Yongning Na: lex- ical tones and morphotonology. Number 13 in Studies in Diversity Linguistics. Language Science Press, Berlin. http://langsci-press.org/ catalog/book/109.",
"links": null
},
"BIBREF38": {
"ref_id": "b38",
"title": "Integrating automatic transcription into the language documentation workflow: experiments with Na data and the Persephone toolkit. Language Documentation and Conservation",
"authors": [
{
"first": "Alexis",
"middle": [],
"last": "Michaud",
"suffix": ""
},
{
"first": "Oliver",
"middle": [],
"last": "Adams",
"suffix": ""
},
{
"first": "Trevor",
"middle": [],
"last": "Cohn",
"suffix": ""
},
{
"first": "Graham",
"middle": [],
"last": "Neubig",
"suffix": ""
},
{
"first": "S\u00e9verine",
"middle": [],
"last": "Guillaume",
"suffix": ""
}
],
"year": 2018,
"venue": "",
"volume": "12",
"issue": "",
"pages": "393--429",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Alexis Michaud, Oliver Adams, Trevor Cohn, Graham Neubig, and S\u00e9verine Guillaume. 2018. Integrat- ing automatic transcription into the language docu- mentation workflow: experiments with Na data and the Persephone toolkit. Language Documentation and Conservation, 12:393-429. http://hdl. handle.net/10125/24793.",
"links": null
},
"BIBREF39": {
"ref_id": "b39",
"title": "Distributed representations of words and phrases and their compositionality",
"authors": [
{
"first": "Tom\u00e1\u0161",
"middle": [],
"last": "Mikolov",
"suffix": ""
},
{
"first": "Ilya",
"middle": [],
"last": "Sutskever",
"suffix": ""
},
{
"first": "Kai",
"middle": [],
"last": "Chen",
"suffix": ""
},
{
"first": "Greg",
"middle": [
"S"
],
"last": "Corrado",
"suffix": ""
},
{
"first": "Jeff",
"middle": [],
"last": "Dean",
"suffix": ""
}
],
"year": 2013,
"venue": "Advances in neural information processing systems",
"volume": "",
"issue": "",
"pages": "3111--3119",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Tom\u00e1\u0161 Mikolov, Ilya Sutskever, Kai Chen, Greg S Cor- rado, and Jeff Dean. 2013. Distributed representa- tions of words and phrases and their compositional- ity. In Advances in neural information processing systems, pages 3111-3119.",
"links": null
},
"BIBREF40": {
"ref_id": "b40",
"title": "Phonemic and graphemic multilingual CTC based speech recognition",
"authors": [
{
"first": "Markus",
"middle": [],
"last": "M\u00fcller",
"suffix": ""
},
{
"first": "Sebastian",
"middle": [],
"last": "St\u00fcker",
"suffix": ""
},
{
"first": "Alex",
"middle": [],
"last": "Waibel",
"suffix": ""
}
],
"year": 2017,
"venue": "",
"volume": "",
"issue": "",
"pages": "",
"other_ids": {
"arXiv": [
"arXiv:1711.04564"
]
},
"num": null,
"urls": [],
"raw_text": "Markus M\u00fcller, Sebastian St\u00fcker, and Alex Waibel. 2017. Phonemic and graphemic multilingual CTC based speech recognition. arXiv:1711.04564. https://arxiv.org/abs/1711.04564.",
"links": null
},
"BIBREF41": {
"ref_id": "b41",
"title": "Language Science Press business model: Evaluated version of the 2015 model",
"authors": [
{
"first": "Sebastian",
"middle": [],
"last": "Nordhoff",
"suffix": ""
}
],
"year": 2018,
"venue": "",
"volume": "",
"issue": "",
"pages": "",
"other_ids": {
"DOI": [
"10.5281/zenodo.1286972"
]
},
"num": null,
"urls": [],
"raw_text": "Sebastian Nordhoff. 2018. Language Science Press business model: Evaluated version of the 2015 model. Language Science Press, Berlin. https: //doi.org/10.5281/zenodo.1286972.",
"links": null
},
"BIBREF42": {
"ref_id": "b42",
"title": "Pytorch: An imperative style, high-performance deep learning library",
"authors": [
{
"first": "Adam",
"middle": [],
"last": "Paszke",
"suffix": ""
},
{
"first": "Sam",
"middle": [],
"last": "Gross",
"suffix": ""
},
{
"first": "Francisco",
"middle": [],
"last": "Massa",
"suffix": ""
},
{
"first": "Adam",
"middle": [],
"last": "Lerer",
"suffix": ""
},
{
"first": "James",
"middle": [],
"last": "Bradbury",
"suffix": ""
},
{
"first": "Gregory",
"middle": [],
"last": "Chanan",
"suffix": ""
},
{
"first": "Trevor",
"middle": [],
"last": "Killeen",
"suffix": ""
},
{
"first": "Zeming",
"middle": [],
"last": "Lin",
"suffix": ""
},
{
"first": "Natalia",
"middle": [],
"last": "Gimelshein",
"suffix": ""
},
{
"first": "Luca",
"middle": [],
"last": "Antiga",
"suffix": ""
}
],
"year": 2019,
"venue": "Advances in neural information processing systems. Proceedings of the 33rd Conference on Neural Information Processing Systems",
"volume": "",
"issue": "",
"pages": "8026--8037",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Adam Paszke, Sam Gross, Francisco Massa, Adam Lerer, James Bradbury, Gregory Chanan, Trevor Killeen, Zeming Lin, Natalia Gimelshein, and Luca Antiga. 2019. Pytorch: An imperative style, high-performance deep learning library. In Advances in neural information processing systems. Proceedings of the 33rd Conference on Neural Information Processing Systems (NeurIPS 2019), pages 8026-8037, Vancouver, Canada. https://papers.nips.cc/paper/9015- pytorch-an-imperative-style- high-performance-deep-learning- library.pdf.",
"links": null
},
"BIBREF43": {
"ref_id": "b43",
"title": "Glove: global vectors for word representation",
"authors": [
{
"first": "Jeffrey",
"middle": [],
"last": "Pennington",
"suffix": ""
},
{
"first": "Richard",
"middle": [],
"last": "Socher",
"suffix": ""
},
{
"first": "Christopher D",
"middle": [],
"last": "Manning",
"suffix": ""
}
],
"year": 2014,
"venue": "Proceedings of the Empiricial Methods in Natural Language Processing",
"volume": "12",
"issue": "",
"pages": "1532--1543",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Jeffrey Pennington, Richard Socher, and Christopher D Manning. 2014. Glove: global vectors for word rep- resentation. Proceedings of the Empiricial Methods in Natural Language Processing (EMNLP 2014), 12:1532-1543.",
"links": null
},
"BIBREF44": {
"ref_id": "b44",
"title": "Deep contextualized word representations",
"authors": [
{
"first": "E",
"middle": [],
"last": "Matthew",
"suffix": ""
},
{
"first": "Mark",
"middle": [],
"last": "Peters",
"suffix": ""
},
{
"first": "Mohit",
"middle": [],
"last": "Neumann",
"suffix": ""
},
{
"first": "Matt",
"middle": [],
"last": "Iyyer",
"suffix": ""
},
{
"first": "Christopher",
"middle": [],
"last": "Gardner",
"suffix": ""
},
{
"first": "Kenton",
"middle": [],
"last": "Clark",
"suffix": ""
},
{
"first": "Luke",
"middle": [],
"last": "Lee",
"suffix": ""
},
{
"first": "",
"middle": [],
"last": "Zettlemoyer",
"suffix": ""
}
],
"year": 2018,
"venue": "",
"volume": "",
"issue": "",
"pages": "",
"other_ids": {
"arXiv": [
"arXiv:1802.05365"
]
},
"num": null,
"urls": [],
"raw_text": "Matthew E Peters, Mark Neumann, Mohit Iyyer, Matt Gardner, Christopher Clark, Kenton Lee, and Luke Zettlemoyer. 2018. Deep contextualized word rep- resentations. arXiv preprint arXiv:1802.05365. https://arxiv.org/abs/1802.05365.",
"links": null
},
"BIBREF45": {
"ref_id": "b45",
"title": "Cross-lingual portability of Chinese and English neural network features for French and German LVCSR",
"authors": [
{
"first": "Christian",
"middle": [],
"last": "Plahl",
"suffix": ""
},
{
"first": "Ralf",
"middle": [],
"last": "Schl\u00fcter",
"suffix": ""
},
{
"first": "Hermann",
"middle": [],
"last": "Ney",
"suffix": ""
}
],
"year": 2011,
"venue": "IEEE Workshop on Automatic Speech Recognition and Understanding (ASRU)",
"volume": "",
"issue": "",
"pages": "371--376",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Christian Plahl, Ralf Schl\u00fcter, and Hermann Ney. 2011. Cross-lingual portability of Chinese and En- glish neural network features for French and Ger- man LVCSR. In IEEE Workshop on Automatic Speech Recognition and Understanding (ASRU), pages 371-376.",
"links": null
},
"BIBREF46": {
"ref_id": "b46",
"title": "The Kaldi speech recognition toolkit",
"authors": [
{
"first": "Daniel",
"middle": [],
"last": "Povey",
"suffix": ""
},
{
"first": "Arnab",
"middle": [],
"last": "Ghoshal",
"suffix": ""
},
{
"first": "Gilles",
"middle": [],
"last": "Boulianne",
"suffix": ""
},
{
"first": "Lukas",
"middle": [],
"last": "Burget",
"suffix": ""
},
{
"first": "Ondrej",
"middle": [],
"last": "Glembek",
"suffix": ""
},
{
"first": "Nagendra",
"middle": [],
"last": "Goel",
"suffix": ""
},
{
"first": "Mirko",
"middle": [],
"last": "Hannemann",
"suffix": ""
},
{
"first": "Petr",
"middle": [],
"last": "Motlicek",
"suffix": ""
},
{
"first": "Yanmin",
"middle": [],
"last": "Qian",
"suffix": ""
},
{
"first": "Petr",
"middle": [],
"last": "Schwarz",
"suffix": ""
},
{
"first": "Jan",
"middle": [],
"last": "Silovsky",
"suffix": ""
},
{
"first": "Georg",
"middle": [],
"last": "Stemmer",
"suffix": ""
},
{
"first": "Karel",
"middle": [],
"last": "Vesely",
"suffix": ""
}
],
"year": 2011,
"venue": "IEEE 2011 Workshop on Automatic Speech Recognition and Understanding",
"volume": "",
"issue": "",
"pages": "",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Daniel Povey, Arnab Ghoshal, Gilles Boulianne, Lukas Burget, Ondrej Glembek, Nagendra Goel, Mirko Hannemann, Petr Motlicek, Yanmin Qian, Petr Schwarz, Jan Silovsky, Georg Stemmer, and Karel Vesely. 2011. The Kaldi speech recognition toolkit. In IEEE 2011 Workshop on Automatic Speech Recognition and Understanding. IEEE Signal Processing Society. https://infoscience. epfl.ch/record/192584/files/Povey_ ASRU2011_2011.pdf.",
"links": null
},
"BIBREF47": {
"ref_id": "b47",
"title": "The pytorch-kaldi speech recognition toolkit",
"authors": [
{
"first": "Mirco",
"middle": [],
"last": "Ravanelli",
"suffix": ""
},
{
"first": "Titouan",
"middle": [],
"last": "Parcollet",
"suffix": ""
},
{
"first": "Yoshua",
"middle": [],
"last": "Bengio",
"suffix": ""
}
],
"year": 2019,
"venue": "ICASSP 2019-2019 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP)",
"volume": "",
"issue": "",
"pages": "6465--6469",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Mirco Ravanelli, Titouan Parcollet, and Yoshua Ben- gio. 2019. The pytorch-kaldi speech recogni- tion toolkit. In ICASSP 2019-2019 IEEE Interna- tional Conference on Acoustics, Speech and Sig- nal Processing (ICASSP), pages 6465-6469. IEEE. https://arxiv.org/abs/1811.07453.",
"links": null
},
"BIBREF48": {
"ref_id": "b48",
"title": "Unsupervised pretraining transfers well across languages",
"authors": [
{
"first": "Morgane",
"middle": [],
"last": "Rivi\u00e8re",
"suffix": ""
},
{
"first": "Armand",
"middle": [],
"last": "Joulin",
"suffix": ""
},
{
"first": "Pierre-Emmanuel",
"middle": [],
"last": "Mazar\u00e9",
"suffix": ""
},
{
"first": "Emmanuel",
"middle": [],
"last": "Dupoux",
"suffix": ""
}
],
"year": 2020,
"venue": "ICASSP 2020 -2020 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP)",
"volume": "",
"issue": "",
"pages": "7414--7418",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Morgane Rivi\u00e8re, Armand Joulin, Pierre-Emmanuel Mazar\u00e9, and Emmanuel Dupoux. 2020. Unsuper- vised pretraining transfers well across languages. In ICASSP 2020 -2020 IEEE International Confer- ence on Acoustics, Speech and Signal Processing (ICASSP), pages 7414-7418. https://arxiv. org/abs/2002.02848.",
"links": null
},
"BIBREF50": {
"ref_id": "b50",
"title": "Building an ASR system for a low-research language through the adaptation of a high-resource language ASR system: preliminary results",
"authors": [
{
"first": "Odette",
"middle": [],
"last": "Scharenborg",
"suffix": ""
},
{
"first": "Francesco",
"middle": [],
"last": "Ciannella",
"suffix": ""
},
{
"first": "Shruti",
"middle": [],
"last": "Palaskar",
"suffix": ""
},
{
"first": "Alan",
"middle": [],
"last": "Black",
"suffix": ""
},
{
"first": "Florian",
"middle": [],
"last": "Metze",
"suffix": ""
},
{
"first": "Lucas",
"middle": [],
"last": "Ondel",
"suffix": ""
},
{
"first": "Mark",
"middle": [],
"last": "Hasegawa-Johnson",
"suffix": ""
}
],
"year": 2017,
"venue": "International Conference on Natural Language, Signal and Speech Processing",
"volume": "",
"issue": "",
"pages": "",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Odette Scharenborg, Francesco Ciannella, Shruti Palaskar, Alan Black, Florian Metze, Lucas Ondel, and Mark Hasegawa-Johnson. 2017. Building an ASR system for a low-research language through the adaptation of a high-resource language ASR system: preliminary results. In International Conference on Natural Language, Signal and Speech Processing (ICNLSSP).",
"links": null
},
"BIBREF51": {
"ref_id": "b51",
"title": "Experiments on cross-language acoustic modeling",
"authors": [
{
"first": "Tanja",
"middle": [],
"last": "Schultz",
"suffix": ""
},
{
"first": "Alex",
"middle": [],
"last": "Waibel",
"suffix": ""
}
],
"year": 2001,
"venue": "",
"volume": "",
"issue": "",
"pages": "2721--2724",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Tanja Schultz and Alex Waibel. 2001. Experi- ments on cross-language acoustic modeling. EU- ROSPEECH'01, pages 2721-2724.",
"links": null
},
"BIBREF52": {
"ref_id": "b52",
"title": "Cross-domain and cross-language portability of acoustic features estimated by multilayer perceptrons",
"authors": [
{
"first": "Andreas",
"middle": [],
"last": "Stolcke",
"suffix": ""
},
{
"first": "Frantisek",
"middle": [],
"last": "Grezl",
"suffix": ""
},
{
"first": "Mei-Yuh",
"middle": [],
"last": "Hwang",
"suffix": ""
},
{
"first": "Xin",
"middle": [],
"last": "Lei",
"suffix": ""
},
{
"first": "Nelson",
"middle": [],
"last": "Morgan",
"suffix": ""
},
{
"first": "Dimitra",
"middle": [],
"last": "Vergyri",
"suffix": ""
}
],
"year": 2006,
"venue": "ICASSP",
"volume": "",
"issue": "",
"pages": "",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Andreas Stolcke, Frantisek Grezl, Mei-Yuh Hwang, Xin Lei, Nelson Morgan, and Dimitra Vergyri. 2006. Cross-domain and cross-language portabil- ity of acoustic features estimated by multilayer per- ceptrons. In ICASSP. http://ieeexplore. ieee.org/document/1660022/.",
"links": null
},
"BIBREF53": {
"ref_id": "b53",
"title": "Doing great things with small languages (Australian Research Council grant DP0984419",
"authors": [
{
"first": "Nick",
"middle": [],
"last": "Thieberger",
"suffix": ""
},
{
"first": "Rachel",
"middle": [],
"last": "Nordlinger",
"suffix": ""
}
],
"year": 2006,
"venue": "",
"volume": "",
"issue": "",
"pages": "",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Nick Thieberger and Rachel Nordlinger. 2006. Doing great things with small languages (Australian Research Council grant DP0984419). https: //arts.unimelb.edu.au/school-of- languages-and-linguistics/our- research/past-research-projects/ great-things-small-languages.",
"links": null
},
"BIBREF54": {
"ref_id": "b54",
"title": "Multilingual MLP features for low-resource LVCSR systems",
"authors": [
{
"first": "Samuel",
"middle": [],
"last": "Thomas",
"suffix": ""
},
{
"first": "Sriram",
"middle": [],
"last": "Ganapathy",
"suffix": ""
}
],
"year": 2012,
"venue": "ICASSP",
"volume": "",
"issue": "",
"pages": "4269--4272",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Samuel Thomas, Sriram Ganapathy, Hynek Herman- sky, and Speech Processing. 2012. Multilingual MLP features for low-resource LVCSR systems. In ICASSP, pages 4269-4272.",
"links": null
},
"BIBREF55": {
"ref_id": "b55",
"title": "Multilingual speech recognition with a single end-to-end model",
"authors": [
{
"first": "Shubham",
"middle": [],
"last": "Toshniwal",
"suffix": ""
},
{
"first": "Tara",
"middle": [
"N"
],
"last": "Sainath",
"suffix": ""
},
{
"first": "Ron",
"middle": [
"J"
],
"last": "Weiss",
"suffix": ""
},
{
"first": "Bo",
"middle": [],
"last": "Li",
"suffix": ""
},
{
"first": "Pedro",
"middle": [],
"last": "Moreno",
"suffix": ""
},
{
"first": "Eugene",
"middle": [],
"last": "Weinstein",
"suffix": ""
},
{
"first": "Kanishka",
"middle": [],
"last": "Rao",
"suffix": ""
}
],
"year": 2017,
"venue": "ICASSP",
"volume": "",
"issue": "",
"pages": "",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Shubham Toshniwal, Tara N. Sainath, Ron J. Weiss, Bo Li, Pedro Moreno, Eugene Weinstein, and Kan- ishka Rao. 2017. Multilingual speech recognition with a single end-to-end model. In ICASSP. http: //arxiv.org/abs/1711.01694.",
"links": null
},
"BIBREF56": {
"ref_id": "b56",
"title": "Cross-lingual portability of MLP-based tandem features -a case study for English and Hungarian",
"authors": [
{
"first": "L\u00e1szl\u00f3",
"middle": [],
"last": "T\u00f3th",
"suffix": ""
},
{
"first": "Joe",
"middle": [],
"last": "Frankel",
"suffix": ""
},
{
"first": "G\u00e1bor",
"middle": [],
"last": "Gosztolya",
"suffix": ""
},
{
"first": "Simon",
"middle": [],
"last": "King",
"suffix": ""
}
],
"year": 2008,
"venue": "",
"volume": "",
"issue": "",
"pages": "",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "L\u00e1szl\u00f3 T\u00f3th, Joe Frankel, G\u00e1bor Gosztolya, and Simon King. 2008. Cross-lingual portability of MLP-based tandem features -a case study for English and Hun- garian. INTERSPEECH.",
"links": null
},
"BIBREF57": {
"ref_id": "b57",
"title": "Attention is all you need",
"authors": [
{
"first": "Ashish",
"middle": [],
"last": "Vaswani",
"suffix": ""
},
{
"first": "Noam",
"middle": [],
"last": "Shazeer",
"suffix": ""
},
{
"first": "Niki",
"middle": [],
"last": "Parmar",
"suffix": ""
},
{
"first": "Jakob",
"middle": [],
"last": "Uszkoreit",
"suffix": ""
},
{
"first": "Llion",
"middle": [],
"last": "Jones",
"suffix": ""
},
{
"first": "Aidan",
"middle": [
"N"
],
"last": "Gomez",
"suffix": ""
},
{
"first": "\u0141ukasz",
"middle": [],
"last": "Kaiser",
"suffix": ""
},
{
"first": "Illia",
"middle": [],
"last": "Polosukhin",
"suffix": ""
}
],
"year": 2017,
"venue": "Advances in neural information processing systems",
"volume": "",
"issue": "",
"pages": "5998--6008",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Ashish Vaswani, Noam Shazeer, Niki Parmar, Jakob Uszkoreit, Llion Jones, Aidan N Gomez, \u0141ukasz Kaiser, and Illia Polosukhin. 2017. Attention is all you need. In Advances in neural information pro- cessing systems, pages 5998-6008.",
"links": null
},
"BIBREF58": {
"ref_id": "b58",
"title": "Language independent end-to-end architecture for joint language identification and speech recognition",
"authors": [
{
"first": "Shinji",
"middle": [],
"last": "Watanabe",
"suffix": ""
},
{
"first": "Takaaki",
"middle": [],
"last": "Hori",
"suffix": ""
},
{
"first": "John",
"middle": [
"R"
],
"last": "Hershey",
"suffix": ""
}
],
"year": 2017,
"venue": "IEEE Workshop on Automatic Speech Recognition and Understanding Workshop (ASRU)",
"volume": "",
"issue": "",
"pages": "265--271",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Shinji Watanabe, Takaaki Hori, and John R. Her- shey. 2017a. Language independent end-to-end architecture for joint language identification and speech recognition. In IEEE Workshop on Automatic Speech Recognition and Under- standing Workshop (ASRU), pages 265-271. https://www.merl.com/publications/ docs/TR2017-182.pdf.",
"links": null
},
"BIBREF59": {
"ref_id": "b59",
"title": "ESPnet: End-to-end speech processing toolkit",
"authors": [
{
"first": "Shinji",
"middle": [],
"last": "Watanabe",
"suffix": ""
},
{
"first": "Takaaki",
"middle": [],
"last": "Hori",
"suffix": ""
},
{
"first": "Shigeki",
"middle": [],
"last": "Karita",
"suffix": ""
},
{
"first": "Tomoki",
"middle": [],
"last": "Hayashi",
"suffix": ""
},
{
"first": "Jiro",
"middle": [],
"last": "Nishitoba",
"suffix": ""
},
{
"first": "Yuya",
"middle": [],
"last": "Unno",
"suffix": ""
},
{
"first": "Nelson",
"middle": [
"Enrique"
],
"last": "",
"suffix": ""
},
{
"first": "Yalta",
"middle": [],
"last": "Soplin",
"suffix": ""
},
{
"first": "Jahn",
"middle": [],
"last": "Heymann",
"suffix": ""
},
{
"first": "Matthew",
"middle": [],
"last": "Wiesner",
"suffix": ""
},
{
"first": "Nanxin",
"middle": [],
"last": "Chen",
"suffix": ""
}
],
"year": 2018,
"venue": "",
"volume": "",
"issue": "",
"pages": "",
"other_ids": {
"arXiv": [
"arXiv:1804.00015"
]
},
"num": null,
"urls": [],
"raw_text": "Shinji Watanabe, Takaaki Hori, Shigeki Karita, Tomoki Hayashi, Jiro Nishitoba, Yuya Unno, Nelson Enrique Yalta Soplin, Jahn Heymann, Matthew Wiesner, and Nanxin Chen. 2018. ESPnet: End-to-end speech processing toolkit. arXiv preprint arXiv:1804.00015. https://arxiv. org/abs/1804.00015.",
"links": null
},
"BIBREF60": {
"ref_id": "b60",
"title": "Hybrid CTC/attention architecture for end-to-end speech recognition",
"authors": [
{
"first": "Shinji",
"middle": [],
"last": "Watanabe",
"suffix": ""
},
{
"first": "Takaaki",
"middle": [],
"last": "Hori",
"suffix": ""
},
{
"first": "Suyoun",
"middle": [],
"last": "Kim",
"suffix": ""
},
{
"first": "R",
"middle": [],
"last": "John",
"suffix": ""
},
{
"first": "Tomoki",
"middle": [],
"last": "Hershey",
"suffix": ""
},
{
"first": "",
"middle": [],
"last": "Hayashi",
"suffix": ""
}
],
"year": 2017,
"venue": "IEEE Journal of Selected Topics in Signal Processing",
"volume": "11",
"issue": "8",
"pages": "1240--1253",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Shinji Watanabe, Takaaki Hori, Suyoun Kim, John R Hershey, and Tomoki Hayashi. 2017b. Hy- brid CTC/attention architecture for end-to-end speech recognition. IEEE Journal of Selected Topics in Signal Processing, 11(8):1240-1253. https://www.merl.com/publications/ docs/TR2017-190.pdf.",
"links": null
},
"BIBREF61": {
"ref_id": "b61",
"title": "Phonemic transcription of low-resource languages: To what extent can preprocessing be automated?",
"authors": [
{
"first": "Guillaume",
"middle": [],
"last": "Wisniewski",
"suffix": ""
},
{
"first": "S\u00e9verine",
"middle": [],
"last": "Guillaume",
"suffix": ""
},
{
"first": "Alexis",
"middle": [],
"last": "Michaud",
"suffix": ""
}
],
"year": 2020,
"venue": "Proceedings of the 1st Joint SLTU (Spoken Language Technologies for Under-resourced languages) and CCURL (Collaboration and Computing for Under-Resourced Languages) Workshop",
"volume": "",
"issue": "",
"pages": "306--315",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Guillaume Wisniewski, S\u00e9verine Guillaume, and Alexis Michaud. 2020. Phonemic transcription of low-resource languages: To what extent can pre- processing be automated? In Proceedings of the 1st Joint SLTU (Spoken Language Technologies for Under-resourced languages) and CCURL (Col- laboration and Computing for Under-Resourced Languages) Workshop, pages 306-315, Marseille, France. European Language Resources Associa- tion (ELRA). https://halshs.archives- ouvertes.fr/hal-02513914.",
"links": null
},
"BIBREF62": {
"ref_id": "b62",
"title": "Huggingface's transformers: State-of-the-art natural language processing",
"authors": [
{
"first": "Thomas",
"middle": [],
"last": "Wolf",
"suffix": ""
},
{
"first": "Lysandre",
"middle": [],
"last": "Debut",
"suffix": ""
},
{
"first": "Victor",
"middle": [],
"last": "Sanh",
"suffix": ""
},
{
"first": "Julien",
"middle": [],
"last": "Chaumond",
"suffix": ""
},
{
"first": "Clement",
"middle": [],
"last": "Delangue",
"suffix": ""
},
{
"first": "Anthony",
"middle": [],
"last": "Moi",
"suffix": ""
},
{
"first": "Pierric",
"middle": [],
"last": "Cistac",
"suffix": ""
},
{
"first": "Tim",
"middle": [],
"last": "Rault",
"suffix": ""
},
{
"first": "R\u00e9mi",
"middle": [],
"last": "Louf",
"suffix": ""
},
{
"first": "Morgan",
"middle": [],
"last": "Funtowicz",
"suffix": ""
},
{
"first": "Joe",
"middle": [],
"last": "Davison",
"suffix": ""
},
{
"first": "Sam",
"middle": [],
"last": "Shleifer",
"suffix": ""
},
{
"first": "Clara",
"middle": [],
"last": "Patrick Von Platen",
"suffix": ""
},
{
"first": "Yacine",
"middle": [],
"last": "Ma",
"suffix": ""
},
{
"first": "Julien",
"middle": [],
"last": "Jernite",
"suffix": ""
},
{
"first": "Canwen",
"middle": [],
"last": "Plu",
"suffix": ""
},
{
"first": "Teven",
"middle": [
"Le"
],
"last": "Xu",
"suffix": ""
},
{
"first": "Sylvain",
"middle": [],
"last": "Scao",
"suffix": ""
},
{
"first": "Mariama",
"middle": [],
"last": "Gugger",
"suffix": ""
},
{
"first": "Quentin",
"middle": [],
"last": "Drame",
"suffix": ""
},
{
"first": "Alexander",
"middle": [
"M"
],
"last": "Lhoest",
"suffix": ""
},
{
"first": "",
"middle": [],
"last": "Rush",
"suffix": ""
}
],
"year": 2019,
"venue": "ArXiv",
"volume": "",
"issue": "",
"pages": "",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Thomas Wolf, Lysandre Debut, Victor Sanh, Julien Chaumond, Clement Delangue, Anthony Moi, Pier- ric Cistac, Tim Rault, R\u00e9mi Louf, Morgan Funtow- icz, Joe Davison, Sam Shleifer, Patrick von Platen, Clara Ma, Yacine Jernite, Julien Plu, Canwen Xu, Teven Le Scao, Sylvain Gugger, Mariama Drame, Quentin Lhoest, and Alexander M. Rush. 2019. Huggingface's transformers: State-of-the-art natu- ral language processing. ArXiv, abs/1910.03771. https://arxiv.org/abs/1910.03771.",
"links": null
},
"BIBREF63": {
"ref_id": "b63",
"title": "Adversarial multilingual training for lowresource speech recognition",
"authors": [
{
"first": "Jiangyan",
"middle": [],
"last": "Yi",
"suffix": ""
},
{
"first": "Jianhua",
"middle": [],
"last": "Tao",
"suffix": ""
},
{
"first": "Zhengqi",
"middle": [],
"last": "Wen",
"suffix": ""
},
{
"first": "Ye",
"middle": [],
"last": "Bai",
"suffix": ""
}
],
"year": 2018,
"venue": "ICASSP",
"volume": "",
"issue": "",
"pages": "4899--4903",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Jiangyan Yi, Jianhua Tao, Zhengqi Wen, and Ye Bai. 2018. Adversarial multilingual training for low- resource speech recognition. ICASSP, pages 4899- 4903.",
"links": null
},
"BIBREF64": {
"ref_id": "b64",
"title": "Adadelta: an adaptive learning rate method",
"authors": [
{
"first": "D",
"middle": [],
"last": "Matthew",
"suffix": ""
},
{
"first": "",
"middle": [],
"last": "Zeiler",
"suffix": ""
}
],
"year": 2012,
"venue": "",
"volume": "",
"issue": "",
"pages": "",
"other_ids": {
"arXiv": [
"arXiv:1212.5701"
]
},
"num": null,
"urls": [],
"raw_text": "Matthew D. Zeiler. 2012. Adadelta: an adaptive learn- ing rate method. arXiv preprint arXiv:1212.5701. https://arxiv.org/abs/1212.5701.",
"links": null
},
"BIBREF65": {
"ref_id": "b65",
"title": "Improved training of end-to-end attention models for speech recognition",
"authors": [
{
"first": "Albert",
"middle": [],
"last": "Zeyer",
"suffix": ""
},
{
"first": "Kazuki",
"middle": [],
"last": "Irie",
"suffix": ""
},
{
"first": "Ralf",
"middle": [],
"last": "Schl\u00fcter",
"suffix": ""
},
{
"first": "Hermann",
"middle": [],
"last": "Ney",
"suffix": ""
}
],
"year": 2018,
"venue": "",
"volume": "",
"issue": "",
"pages": "",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Albert Zeyer, Kazuki Irie, Ralf Schl\u00fcter, and Hermann Ney. 2018. Improved training of end-to-end at- tention models for speech recognition. https: //arxiv.org/abs/1805.03294.",
"links": null
},
"BIBREF66": {
"ref_id": "b66",
"title": "The RWTH ASR system for TED-LIUM Release 2: Improving Hybrid HMM with SpecAugment",
"authors": [
{
"first": "Wei",
"middle": [],
"last": "Zhou",
"suffix": ""
},
{
"first": "Wilfried",
"middle": [],
"last": "Michel",
"suffix": ""
},
{
"first": "Kazuki",
"middle": [],
"last": "Irie",
"suffix": ""
},
{
"first": "Markus",
"middle": [],
"last": "Kitza",
"suffix": ""
},
{
"first": "Ralf",
"middle": [],
"last": "Schl\u00fcter",
"suffix": ""
},
{
"first": "Hermann",
"middle": [],
"last": "Ney",
"suffix": ""
}
],
"year": 2020,
"venue": "",
"volume": "",
"issue": "",
"pages": "",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Wei Zhou, Wilfried Michel, Kazuki Irie, Markus Kitza, Ralf Schl\u00fcter, and Hermann Ney. 2020. The RWTH ASR system for TED-LIUM Release 2: Improv- ing Hybrid HMM with SpecAugment. https: //arxiv.org/abs/2004.00960.",
"links": null
}
},
"ref_entries": {
"FIGREF0": {
"num": null,
"text": "Training stages of the Elpis interface. Notice the choice of backend in the upper right-hand corner.",
"uris": null,
"type_str": "figure"
},
"FIGREF1": {
"num": null,
"text": "Character error rate for Japhug as a function of the amount of training data, using the ESPnet recipe included in Elpis. Character error rate on the training set (blue) and validation set (orange) for Japhug as training progresses (up to 20 epochs), using the ESPnet recipe included in Elpis.",
"uris": null,
"type_str": "figure"
},
"TABREF0": {
"html": null,
"num": null,
"type_str": "table",
"text": "Information on the evaluation datasets used and the character error rate performance of the current recipe.",
"content": "<table><tr><td colspan=\"3\">Language Num speakers Type</td><td colspan=\"2\">Train (minutes) CER (%)</td></tr><tr><td>Na</td><td>1</td><td>Spontaneous narratives</td><td>273</td><td>14.5</td></tr><tr><td>Na</td><td>1</td><td colspan=\"2\">Elicited words & phrases 188</td><td>4.7</td></tr><tr><td>Chatino</td><td>1</td><td>Read speech</td><td>81</td><td>23.5</td></tr><tr><td>Japhug</td><td>1</td><td>Spontaneous narratives</td><td>170</td><td>12.8</td></tr><tr><td colspan=\"3\">attention objectives. For optimization we use a</td><td/><td/></tr><tr><td colspan=\"3\">batch length of 30 and the Adadelta gradient de-</td><td/><td/></tr><tr><td colspan=\"3\">scent algorithm (Zeiler, 2012). For more details,</td><td/><td/></tr><tr><td colspan=\"2\">we include a link to the recipe. 8</td><td/><td/><td/></tr></table>"
}
}
}
} |