File size: 148,676 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 | {
"paper_id": "P17-1003",
"header": {
"generated_with": "S2ORC 1.0.0",
"date_generated": "2023-01-19T08:18:28.994921Z"
},
"title": "Neural Symbolic Machines: Learning Semantic Parsers on Freebase with Weak Supervision",
"authors": [
{
"first": "Chen",
"middle": [],
"last": "Liang",
"suffix": "",
"affiliation": {},
"email": "chenliang2013@u.northwestern.edu"
},
{
"first": "Jonathan",
"middle": [],
"last": "Berant",
"suffix": "",
"affiliation": {},
"email": "joberant@cs.tau.ac.il"
},
{
"first": "Quoc",
"middle": [],
"last": "Le",
"suffix": "",
"affiliation": {},
"email": ""
},
{
"first": "Kenneth",
"middle": [
"D"
],
"last": "Forbus",
"suffix": "",
"affiliation": {},
"email": "forbus@u.northwestern.edu"
},
{
"first": "Ni",
"middle": [],
"last": "Lao",
"suffix": "",
"affiliation": {},
"email": "nlao@google.com"
}
],
"year": "",
"venue": null,
"identifiers": {},
"abstract": "Harnessing the statistical power of neural networks to perform language understanding and symbolic reasoning is difficult, when it requires executing efficient discrete operations against a large knowledge-base. In this work, we introduce a Neural Symbolic Machine (NSM), which contains (a) a neural \"programmer\", i.e., a sequence-to-sequence model that maps language utterances to programs and utilizes a key-variable memory to handle compositionality (b) a symbolic \"computer\", i.e., a Lisp interpreter that performs program execution, and helps find good programs by pruning the search space. We apply REINFORCE to directly optimize the task reward of this structured prediction problem. To train with weak supervision and improve the stability of REINFORCE we augment it with an iterative maximum-likelihood training process. NSM outperforms the state-of-theart on the WEBQUESTIONSSP dataset when trained from question-answer pairs only, without requiring any feature engineering or domain-specific knowledge.",
"pdf_parse": {
"paper_id": "P17-1003",
"_pdf_hash": "",
"abstract": [
{
"text": "Harnessing the statistical power of neural networks to perform language understanding and symbolic reasoning is difficult, when it requires executing efficient discrete operations against a large knowledge-base. In this work, we introduce a Neural Symbolic Machine (NSM), which contains (a) a neural \"programmer\", i.e., a sequence-to-sequence model that maps language utterances to programs and utilizes a key-variable memory to handle compositionality (b) a symbolic \"computer\", i.e., a Lisp interpreter that performs program execution, and helps find good programs by pruning the search space. We apply REINFORCE to directly optimize the task reward of this structured prediction problem. To train with weak supervision and improve the stability of REINFORCE we augment it with an iterative maximum-likelihood training process. NSM outperforms the state-of-theart on the WEBQUESTIONSSP dataset when trained from question-answer pairs only, without requiring any feature engineering or domain-specific knowledge.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Abstract",
"sec_num": null
}
],
"body_text": [
{
"text": "Deep neural networks have achieved impressive performance in supervised classification and structured prediction tasks such as speech recognition (Hinton et al., 2012) , machine translation and more. However, training neural networks for semantic parsing (Zelle and Mooney, 1996; Zettlemoyer and Collins, 2005; Liang et al., 2011) or program induction, where language is mapped to a sym-\u21e4 Work done while the author was interning at Google \u2020 Work done while the author was visiting Google",
"cite_spans": [
{
"start": 146,
"end": 167,
"text": "(Hinton et al., 2012)",
"ref_id": "BIBREF11"
},
{
"start": 255,
"end": 279,
"text": "(Zelle and Mooney, 1996;",
"ref_id": "BIBREF42"
},
{
"start": 280,
"end": 310,
"text": "Zettlemoyer and Collins, 2005;",
"ref_id": "BIBREF43"
},
{
"start": 311,
"end": 330,
"text": "Liang et al., 2011)",
"ref_id": "BIBREF18"
}
],
"ref_spans": [],
"eq_spans": [],
"section": "Introduction",
"sec_num": "1"
},
{
"text": "x: Largest city in the US \u21d2 y: NYC Figure 1 : The main challenges of training a semantic parser from weak supervision: (a) compositionality: we use variables (v0, v1, v2) to store execution results of intermediate generated programs. (b) search: we prune the search space and augment REINFORCE with pseudo-gold programs.",
"cite_spans": [
{
"start": 158,
"end": 170,
"text": "(v0, v1, v2)",
"ref_id": null
}
],
"ref_spans": [
{
"start": 35,
"end": 43,
"text": "Figure 1",
"ref_id": null
}
],
"eq_spans": [],
"section": "Introduction",
"sec_num": "1"
},
{
"text": "bolic representation that is executed by an executor, through weak supervision remains challenging. This is because the model must interact with a symbolic executor through non-differentiable operations to search over a large program space. In semantic parsing, recent work handled this (Dong and Lapata, 2016; Jia and Liang, 2016) by training from manually annotated programs and avoiding program execution at training time. However, annotating programs is known to be expensive and scales poorly. In program induction, attempts to address this problem (Graves et al., 2014; Reed and de Freitas, 2016; Kaiser and Sutskever, 2015; Graves et al., 2016b; Andreas et al., 2016) either utilized low-level memory (Zaremba and Sutskever, 2015) , or required memory to be differentiable (Neelakantan et al., 2015; Yin et al., 2015 ) so that the model can be trained with backpropagation. This makes it difficult to use the efficient discrete operations and memory of a traditional computer, and limited the application to synthetic or small knowledge bases.",
"cite_spans": [
{
"start": 287,
"end": 310,
"text": "(Dong and Lapata, 2016;",
"ref_id": "BIBREF7"
},
{
"start": 311,
"end": 331,
"text": "Jia and Liang, 2016)",
"ref_id": "BIBREF12"
},
{
"start": 554,
"end": 575,
"text": "(Graves et al., 2014;",
"ref_id": "BIBREF8"
},
{
"start": 576,
"end": 602,
"text": "Reed and de Freitas, 2016;",
"ref_id": "BIBREF24"
},
{
"start": 603,
"end": 630,
"text": "Kaiser and Sutskever, 2015;",
"ref_id": "BIBREF14"
},
{
"start": 631,
"end": 652,
"text": "Graves et al., 2016b;",
"ref_id": "BIBREF10"
},
{
"start": 653,
"end": 674,
"text": "Andreas et al., 2016)",
"ref_id": "BIBREF0"
},
{
"start": 708,
"end": 737,
"text": "(Zaremba and Sutskever, 2015)",
"ref_id": "BIBREF41"
},
{
"start": 780,
"end": 806,
"text": "(Neelakantan et al., 2015;",
"ref_id": "BIBREF19"
},
{
"start": 807,
"end": 823,
"text": "Yin et al., 2015",
"ref_id": "BIBREF39"
}
],
"ref_spans": [],
"eq_spans": [],
"section": "Introduction",
"sec_num": "1"
},
{
"text": "In this paper, we propose to utilize the memory and discrete operations of a traditional com-puter in a novel Manager-Programmer-Computer (MPC) framework for neural program induction, which integrates three components: 1. A \"manager\" that provides weak supervision (e.g., 'NYC' in Figure 1 ) through a reward indicating how well a task is accomplished. Unlike full supervision, weak supervision is easy to obtain at scale (Section 3.1). 2. A \"programmer\" that takes natural language as input and generates a program that is a sequence of tokens ( Figure 2 ). The programmer learns from the reward and must overcome the hard search problem of finding correct programs (Section 2.2). 3. A \"computer\" that executes programs in a high level programming language. Its nondifferentiable memory enables abstract, scalable and precise operations, but makes training more challenging (Section 2.3). To help the \"programmer\" prune the search space, it provides a friendly neural computer interface, which detects and eliminates invalid choices (Section 2.1). Within this framework, we introduce the Neural Symbolic Machine (NSM) and apply it to semantic parsing. NSM contains a neural sequenceto-sequence (seq2seq) \"programmer\" (Sutskever et al., 2014 ) and a symbolic non-differentiable Lisp interpreter (\"computer\") that executes programs against a large knowledge-base (KB).",
"cite_spans": [
{
"start": 1218,
"end": 1241,
"text": "(Sutskever et al., 2014",
"ref_id": "BIBREF29"
}
],
"ref_spans": [
{
"start": 281,
"end": 289,
"text": "Figure 1",
"ref_id": null
},
{
"start": 547,
"end": 555,
"text": "Figure 2",
"ref_id": null
}
],
"eq_spans": [],
"section": "Introduction",
"sec_num": "1"
},
{
"text": "Our technical contribution in this work is threefold. First, to support language compositionality, we augment the standard seq2seq model with a key-variable memory to save and reuse intermediate execution results ( Figure 1 ). This is a novel application of pointer networks (Vinyals et al., 2015) to compositional semantics.",
"cite_spans": [
{
"start": 275,
"end": 297,
"text": "(Vinyals et al., 2015)",
"ref_id": "BIBREF31"
}
],
"ref_spans": [
{
"start": 215,
"end": 223,
"text": "Figure 1",
"ref_id": null
}
],
"eq_spans": [],
"section": "Introduction",
"sec_num": "1"
},
{
"text": "Second, to alleviate the search problem of finding correct programs when training from questionanswer pairs,we use the computer to execute partial programs and prune the programmer's search space by checking the syntax and semantics of generated programs. This generalizes the weakly supervised semantic parsing framework (Liang et al., 2011; Berant et al., 2013) by leveraging semantic denotations during structural search.",
"cite_spans": [
{
"start": 322,
"end": 342,
"text": "(Liang et al., 2011;",
"ref_id": "BIBREF18"
},
{
"start": 343,
"end": 363,
"text": "Berant et al., 2013)",
"ref_id": "BIBREF2"
}
],
"ref_spans": [],
"eq_spans": [],
"section": "Introduction",
"sec_num": "1"
},
{
"text": "Third, to train from weak supervision and directly maximize the expected reward we turn to the REINFORCE (Williams, 1992) algorithm. Since learning from scratch is difficult for RE-INFORCE, we combine it with an iterative max-imum likelihood (ML) training process, where beam search is used to find pseudo-gold programs, which are then used to augment the objective of REINFORCE.",
"cite_spans": [
{
"start": 105,
"end": 121,
"text": "(Williams, 1992)",
"ref_id": "BIBREF34"
}
],
"ref_spans": [],
"eq_spans": [],
"section": "Introduction",
"sec_num": "1"
},
{
"text": "On the WEBQUESTIONSSP dataset (Yih et al., 2016) , NSM achieves new state-of-the-art results with weak supervision, significantly closing the gap between weak and full supervision for this task. Unlike prior works, it is trained end-toend, and does not require feature engineering or domain-specific knowledge.",
"cite_spans": [
{
"start": 30,
"end": 48,
"text": "(Yih et al., 2016)",
"ref_id": "BIBREF38"
}
],
"ref_spans": [],
"eq_spans": [],
"section": "Introduction",
"sec_num": "1"
},
{
"text": "We now introduce NSM by first describing the \"computer\", a non-differentiable Lisp interpreter that executes programs against a large KB and provides code assistance (Section 2.1). We then propose a seq2seq model (\"programmer\") that supports compositionality using a key-variable memory to save and reuse intermediate results (Section 2.2). Finally, we describe a training procedure that is based on REINFORCE, but is augmented with pseudo-gold programs found by an iterative ML training procedure (Section 2.3).",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Neural Symbolic Machines",
"sec_num": "2"
},
{
"text": "Before diving into details, we define the semantic parsing task: given a knowledge base K, and a question x = (w 1 , w 2 , ..., w m ), produce a program or logical form z that when executed against K generates the right answer y. Let E denote a set of entities (e.g., ABELINCOLN), 1 and let P denote a set of properties (e.g., PLACEOFBIRTH). A knowledge base K is a set of assertions or triples (e 1 , p, e 2 ) 2 E \u21e5 P \u21e5 E, such as (ABELINCOLN, PLACEOFBIRTH, HODGENVILLE).",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Neural Symbolic Machines",
"sec_num": "2"
},
{
"text": "Semantic parsing typically requires using a set of operations to query the knowledge base and process the results. Operations learned with neural networks such as addition and sorting do not perfectly generalize to inputs that are larger than the ones observed in the training data (Graves et al., 2014; Reed and de Freitas, 2016) . In contrast, operations implemented in high level programming languages are abstract, scalable, and precise, thus generalizes perfectly to inputs of arbitrary size. Based on this observation, we implement operations necessary for semantic parsing with an or-dinary programming language instead of trying to learn them with a neural network.",
"cite_spans": [
{
"start": 282,
"end": 303,
"text": "(Graves et al., 2014;",
"ref_id": "BIBREF8"
},
{
"start": 304,
"end": 330,
"text": "Reed and de Freitas, 2016)",
"ref_id": "BIBREF24"
}
],
"ref_spans": [],
"eq_spans": [],
"section": "Computer: Lisp Interpreter with Code Assistance",
"sec_num": "2.1"
},
{
"text": "We adopt a Lisp interpreter as the \"computer\". A program C is a list of expressions (c 1 ...c N ), where each expression is either a special token \"Return\" indicating the end of the program, or a list of tokens enclosed by parentheses \"(F A 1 ...A K )\". F is a function, which takes as input K arguments of specific types. Table 1 defines the semantics of each function and the types of its arguments (either a property p or a variable r). When a function is executed, it returns an entity list that is the expression's denotation in K, and save it to a new variable.",
"cite_spans": [],
"ref_spans": [
{
"start": 323,
"end": 330,
"text": "Table 1",
"ref_id": "TABREF0"
}
],
"eq_spans": [],
"section": "Computer: Lisp Interpreter with Code Assistance",
"sec_num": "2.1"
},
{
"text": "By introducing variables that save the intermediate results of execution, the program naturally models language compositionality and describes from left to right a bottom-up derivation of the full meaning of the natural language input, which is convenient in a seq2seq model (Figure 1 ). This is reminiscent of the floating parser (Wang et al., 2015; Pasupat and Liang, 2015) , where a derivation tree that is not grounded in the input is incrementally constructed.",
"cite_spans": [
{
"start": 331,
"end": 350,
"text": "(Wang et al., 2015;",
"ref_id": "BIBREF32"
},
{
"start": 351,
"end": 375,
"text": "Pasupat and Liang, 2015)",
"ref_id": "BIBREF21"
}
],
"ref_spans": [
{
"start": 275,
"end": 284,
"text": "(Figure 1",
"ref_id": null
}
],
"eq_spans": [],
"section": "Computer: Lisp Interpreter with Code Assistance",
"sec_num": "2.1"
},
{
"text": "The set of programs defined by our functions is equivalent to the subset of -calculus presented in (Yih et al., 2015). We did not use full Lisp programming language here, because constructs like control flow and loops are unnecessary for most current semantic parsing tasks, and it is simple to add more functions to the model when necessary.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Computer: Lisp Interpreter with Code Assistance",
"sec_num": "2.1"
},
{
"text": "To create a friendly neural computer interface, the interpreter provides code assistance to the programmer by producing a list of valid tokens at each step. First, a valid token should not cause a syntax error: e.g., if the previous token is \"(\", the next token must be a function name, and if the previous token is \"Hop\", the next token must be a variable. More importantly, a valid token should not cause a semantic (run-time) error: this is detected using the denotation saved in the variables. For example, if the previously generated tokens were \"( Hop r\", the next available token is restricted to properties {p | 9e, e 0 : e 2 r, (e, p, e 0 ) 2 K} that are reachable from entities in r in the KB. These checks are enabled by the variables and can be derived from the definition of the functions in Table 1 . The interpreter prunes the \"programmer\"'s search space by orders of magnitude, and enables learning from weak supervision on a large KB.",
"cite_spans": [],
"ref_spans": [
{
"start": 805,
"end": 812,
"text": "Table 1",
"ref_id": "TABREF0"
}
],
"eq_spans": [],
"section": "Computer: Lisp Interpreter with Code Assistance",
"sec_num": "2.1"
},
{
"text": "Given the \"computer\", the \"programmer\" needs to map natural language into a program, which is a sequence of tokens that reference operations and values in the \"computer\". We base our programmer on a standard seq2seq model with attention, but extend it with a key-variable memory that allows the model to learn to represent and refer to program variables ( Figure 2 ). Sequence-to-sequence models consist of two RNNs, an encoder and a decoder. We used a 1-layer GRU for both the encoder and decoder. Given a sequence of words w 1 , w 2 ...w m , each word w t is mapped to an embedding q t (embedding details are in Section 3). Then, the encoder reads these embeddings and updates its hidden state step by step using h t+1 = GRU (h t , q t , \u2713 Encoder ), where \u2713 Encoder are the GRU parameters. The decoder updates its hidden states u t by u t+1 = GRU (u t , c t 1 , \u2713 Decoder ), where c t 1 is the embedding of last step's output token a t 1 , and \u2713 Decoder are the GRU parameters. The last hidden state of the encoder h T is used as the decoder's initial state. We also adopt a dot-product attention similar to Dong and Lapata (2016) . The tokens of the program a 1 , a 2 ...a n are generated one by one using a softmax over the vocabulary of valid tokens at each step, as provided by the \"computer\" (Section 2.1).",
"cite_spans": [
{
"start": 1111,
"end": 1133,
"text": "Dong and Lapata (2016)",
"ref_id": "BIBREF7"
}
],
"ref_spans": [
{
"start": 356,
"end": 364,
"text": "Figure 2",
"ref_id": null
}
],
"eq_spans": [],
"section": "Programmer: Seq2seq Model with Key-Variable Memory",
"sec_num": "2.2"
},
{
"text": "To achieve compositionality, the decoder must learn to represent and refer to intermediate variables whose value was saved in the \"computer\" after execution. Therefore, we augment the model with a key-variable memory, where each entry has two components: a continuous embedding key v i , and a corresponding variable token R i referencing the value in the \"computer\" (see Figure 2 ). During encoding, we use an entity linker to link text spans (e.g., \"US\") to KB entities. For each linked entity we add a memory entry where the key is the average of GRU hidden states over the entity span, and the variable token (R 1 ) is the name of a variable in the computer holding the linked entity (m.USA) as its value. During decoding, when a full expression is generated (i.e., the decoder generates \")\"), it gets executed, and the result is stored as the value of a new variable in the \"computer\". This variable is keyed by the GRU hidden state at that step. When a new variable R 1 with key embedding v 1 is added into the key-variable memory, ( Hop r p ) ) {e 2 |e 1 2 r, (e 1 , p, e 2 ) 2 K} ( ArgMax r p ) ) {e 1 |e 1 2 r, 9e 2 2 E : (e 1 , p, e 2 ) 2 K, 8e : (e 1 , p, e) 2 K, e 2 e} ( ArgMin r p ) ) {e 1 |e 1 2 r, 9e 2 2 E : (e 1 , p, e 2 ) 2 K, 8e : (e 1 , p, e) 2 K, e 2 \uf8ff e} ( Filter r 1 r 2 p ) ) {e 1 |e 1 2 r 1 , 9e 2 2 r 2 : (e 1 , p, e 2 ) 2 K} Figure 2 : Semantic Parsing with NSM. The key embeddings of the key-variable memory are the output of the sequence model at certain encoding or decoding steps. For illustration purposes, we also show the values of the variables in parentheses, but the sequence model never sees these values, and only references them with the name of the variable (\"R1\"). A special token \"GO\" indicates the start of decoding, and \"Return\" indicates the end of decoding.",
"cite_spans": [],
"ref_spans": [
{
"start": 372,
"end": 380,
"text": "Figure 2",
"ref_id": null
},
{
"start": 1353,
"end": 1361,
"text": "Figure 2",
"ref_id": null
}
],
"eq_spans": [],
"section": "Programmer: Seq2seq Model with Key-Variable Memory",
"sec_num": "2.2"
},
{
"text": "the token R 1 is added into the decoder vocabulary with v 1 as its embedding. The final answer returned by the \"programmer\" is the value of the last computed variable.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Programmer: Seq2seq Model with Key-Variable Memory",
"sec_num": "2.2"
},
{
"text": "Similar to pointer networks (Vinyals et al., 2015) , the key embeddings for variables are dynamically generated for each example. During training, the model learns to represent variables by backpropagating gradients from a time step where a variable is selected by the decoder, through the key-variable memory, to an earlier time step when the key embedding was computed. Thus, the encoder/decoder learns to generate representations for variables such that they can be used at the right time to construct the correct program.",
"cite_spans": [
{
"start": 28,
"end": 50,
"text": "(Vinyals et al., 2015)",
"ref_id": "BIBREF31"
}
],
"ref_spans": [],
"eq_spans": [],
"section": "Programmer: Seq2seq Model with Key-Variable Memory",
"sec_num": "2.2"
},
{
"text": "While the key embeddings are differentiable, the values referenced by the variables (lists of entities), stored in the \"computer\", are symbolic and non-differentiable. This distinguishes the keyvariable memory from other memory-augmented neural networks that use continuous differentiable embeddings as the values of memory entries (Weston et al., 2014; Graves et al., 2016a).",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Programmer: Seq2seq Model with Key-Variable Memory",
"sec_num": "2.2"
},
{
"text": "NSM executes non-differentiable operations against a KB, and thus end-to-end backpropagation is not possible. Therefore, we base our training procedure on REINFORCE (Williams, 1992; . When the reward signal is sparse and the search space is large, it is common to utilize some full supervision to pre-train REINFORCE .",
"cite_spans": [
{
"start": 165,
"end": 181,
"text": "(Williams, 1992;",
"ref_id": "BIBREF34"
}
],
"ref_spans": [],
"eq_spans": [],
"section": "Training NSM with Weak Supervision",
"sec_num": "2.3"
},
{
"text": "To train from weak supervision, we suggest an iterative ML procedure for finding pseudo-gold programs that will bootstrap REINFORCE.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Training NSM with Weak Supervision",
"sec_num": "2.3"
},
{
"text": "REINFORCE We can formulate training as a reinforcement learning problem: given a question x, the state, action and reward at each time step t 2 {0, 1, ..., T } are (s t , a t , r t ). Since the environment is deterministic, the state is defined by the question x and the action sequence: s t = (x, a 0:t 1 ), where a 0:t 1 = (a 0 , ..., a t 1 ) is the history of actions at time t. A valid action at time t is a t 2 A(s t ), where A(s t ) is the set of valid tokens given by the \"computer\". Since each action corresponds to a token, the full history a 0:T corresponds to a program. The reward r",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Training NSM with Weak Supervision",
"sec_num": "2.3"
},
{
"text": "t = I[t = T ] \u2022 F 1 (x, a 0:T )",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Training NSM with Weak Supervision",
"sec_num": "2.3"
},
{
"text": "is non-zero only at the last step of decoding, and is the F 1 score computed comparing the gold answer and the answer generated by executing the program a 0:T . Thus, the cumulative reward of a program a 0:T is",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Training NSM with Weak Supervision",
"sec_num": "2.3"
},
{
"text": "R(x, a 0:T ) = X t r t = F 1 (x, a 0:T ).",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Training NSM with Weak Supervision",
"sec_num": "2.3"
},
{
"text": "The agent's decision making procedure at each time is defined by a policy, \u21e1 \u2713 (s, a) = P \u2713 (a t = a|x, a 0:t 1 ), where \u2713 are the model parameters. Since the environment is deterministic, the probability of generating a program a 0:T is",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Training NSM with Weak Supervision",
"sec_num": "2.3"
},
{
"text": "P \u2713 (a 0:T |x) = Y t P \u2713 (a t | x, a 0:t 1 ).",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Training NSM with Weak Supervision",
"sec_num": "2.3"
},
{
"text": "We can define our objective to be the expected cumulative reward and use policy gradient meth-ods such as REINFORCE for training. The objective and gradient are:",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Training NSM with Weak Supervision",
"sec_num": "2.3"
},
{
"text": "J RL (\u2713) = X x E P \u2713 (a 0:T |x) [R(x, a 0:T )], r \u2713 J RL (\u2713) = X x X a 0:T P \u2713 (a 0:T | x) \u2022 [R(x, a 0:T ) B(x)] \u2022 r \u2713 log P \u2713 (a 0:T | x),",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Training NSM with Weak Supervision",
"sec_num": "2.3"
},
{
"text": "where B(x) = P a 0:T P \u2713 (a 0:T | x)R(x, a 0:T ) is a baseline that reduces the variance of the gradient estimation without introducing bias. Having a separate network to predict the baseline is an interesting future direction.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Training NSM with Weak Supervision",
"sec_num": "2.3"
},
{
"text": "While REINFORCE assumes a stochastic policy, we use beam search for gradient estimation. Thus, in contrast with common practice of approximating the gradient by sampling from the model, we use the top-k action sequences (programs) in the beam with normalized probabilities. This allows training to focus on sequences with high probability, which are on the decision boundaries, and reduces the variance of the gradient.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Training NSM with Weak Supervision",
"sec_num": "2.3"
},
{
"text": "Empirically (and in line with prior work), RE-INFORCE converged slowly and often got stuck in local optima (see Section 3). The difficulty of training resulted from the sparse reward signal in the large search space, which caused model probabilities for programs with non-zero reward to be very small at the beginning. If the beam size k is small, good programs fall off the beam, leading to zero gradients for all programs in the beam. If the beam size k is large, training is very slow, and the normalized probabilities of good programs when the model is untrained are still very small, leading to (1) near zero baselines, thus near zero gradients on \"bad\" programs (2) near zero gradients on good programs due to the low probability P \u2713 (a 0:T | x).",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Training NSM with Weak Supervision",
"sec_num": "2.3"
},
{
"text": "To combat this, we present an alternative training strategy based on maximum-likelihood.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Training NSM with Weak Supervision",
"sec_num": "2.3"
},
{
"text": "Iterative ML If we had gold programs, we could directly optimize their likelihood. Since we do not have gold programs, we can perform an iterative procedure (similar to hard Expectation-Maximization (EM)), where we search for good programs given fixed parameters, and then optimize the probability of the best program found so far. We do decoding on an example with a large beam size and declare a best 0:T (x) to be the pseudogold program, which achieved highest reward with shortest length among the programs decoded on x in all previous iterations. Then, we can optimize the ML objective:",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Training NSM with Weak Supervision",
"sec_num": "2.3"
},
{
"text": "EQUATION",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [
{
"start": 0,
"end": 8,
"text": "EQUATION",
"ref_id": "EQREF",
"raw_str": "J ML (\u2713) = X x log P \u2713 (a best 0:T (x) | x)",
"eq_num": "(1)"
}
],
"section": "Training NSM with Weak Supervision",
"sec_num": "2.3"
},
{
"text": "A question x is not included if we did not find any program with positive reward.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Training NSM with Weak Supervision",
"sec_num": "2.3"
},
{
"text": "Training with iterative ML is fast because there is at most one program per example and the gradient is not weighted by model probability. while decoding with a large beam size is slow, we could train for multiple epochs after each decoding. This iterative process has a bootstrapping effect that a better model leads to a better program a best 0:T (x) through decoding, and a better program a best 0:T (x) leads to a better model through training.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Training NSM with Weak Supervision",
"sec_num": "2.3"
},
{
"text": "Even with a large beam size, some programs are hard to find because of the large search space. A common solution to this problem is to use curriculum learning (Zaremba and Sutskever, 2015; Reed and de Freitas, 2016) . The size of the search space is controlled by both the set of functions used in the program and the program length. We apply curriculum learning by gradually increasing both these quantities (see details in Section 3) when performing iterative ML.",
"cite_spans": [
{
"start": 159,
"end": 188,
"text": "(Zaremba and Sutskever, 2015;",
"ref_id": "BIBREF41"
},
{
"start": 189,
"end": 215,
"text": "Reed and de Freitas, 2016)",
"ref_id": "BIBREF24"
}
],
"ref_spans": [],
"eq_spans": [],
"section": "Training NSM with Weak Supervision",
"sec_num": "2.3"
},
{
"text": "Nevertheless, iterative ML uses only pseudogold programs and does not directly optimize the objective we truly care about. This has two adverse effects: (1) The best program a best 0:T (x) could be a spurious program that accidentally produces the correct answer (e.g., using the property PLACEOFBIRTH instead of PLACEOFDEATH when the two places are the same), and thus does not generalize to other questions. (2) Because training does not observe full negative programs, the model often fails to distinguish between tokens that are related to one another. For example, differentiating PARENTSOF vs. SIBLINGSOF vs. CHILDRENOF can be challenging. We now present learning where we combine iterative ML with REINFORCE.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Training NSM with Weak Supervision",
"sec_num": "2.3"
},
{
"text": "Augmented REINFORCE To bootstrap REIN-FORCE, we can use iterative ML to find pseudogold programs, and then add these programs to the beam with a reasonably large probability. This is similar to methods from imitation learning (Ross et al., 2011; Jiang et al., 2012) that define a proposal distribution by linearly interpolating the model distribution and an oracle. ",
"cite_spans": [
{
"start": 226,
"end": 245,
"text": "(Ross et al., 2011;",
"ref_id": "BIBREF25"
},
{
"start": 246,
"end": 265,
"text": "Jiang et al., 2012)",
"ref_id": "BIBREF13"
}
],
"ref_spans": [],
"eq_spans": [],
"section": "Training NSM with Weak Supervision",
"sec_num": "2.3"
},
{
"text": "(x, y) in D do C Decode BRL programs from x for j in 1...|C| do if Rx,y(Cj) > Rx,y(C \u21e4 x ) then C \u21e4 x Cj C C [ {C \u21e4 x } for j in 1...|C| d\u00f4 pj (1 \u21b5)\u2022 p j P j 0 p j 0 where pj = P \u2713 (Cj | x) if Cj = C \u21e4 x thenpj pj + \u21b5 DRL DRL [ {(x, Cj,pj)} \u2713 REINFORCE training with DRL",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Training NSM with Weak Supervision",
"sec_num": "2.3"
},
{
"text": "Algorithm 1 describes our overall training procedure. We first run iterative ML for N ML iterations and record the best program found for every example x i . Then, we run REINFORCE, where we normalize the probabilities of the programs in beam to sum to (1 \u21b5) and add \u21b5 to the probability of the best found program C \u21e4 (x i ). Consequently, the model always puts a reasonable amount of probability on a program with high reward during training. Note that we randomly initialized the parameters for REINFORCE, since initializing from the final ML parameters seems to get stuck in a local optimum and produced worse results.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Training NSM with Weak Supervision",
"sec_num": "2.3"
},
{
"text": "On top of imitation learning, our approach is related to the common practice in reinforcement learning (Schaul et al., 2016) to replay rare successful experiences to reduce the training variance and improve training efficiency. This is also similar to recent developments in machine translation, where ML and RL objectives are linearly combined, because anchoring the model to some high-reward outputs stabilizes training.",
"cite_spans": [
{
"start": 103,
"end": 124,
"text": "(Schaul et al., 2016)",
"ref_id": "BIBREF26"
}
],
"ref_spans": [],
"eq_spans": [],
"section": "Training NSM with Weak Supervision",
"sec_num": "2.3"
},
{
"text": "We now empirically show that NSM can learn a semantic parser from weak supervision over a large KB. We evaluate on WEBQUESTIONSSP, a challenging semantic parsing dataset with strong baselines. Experiments show that NSM achieves new state-of-the-art performance on WEBQUES-TIONSSP with weak supervision, and significantly closes the gap between weak and full supervisions for this task.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Experiments and Analysis",
"sec_num": "3"
},
{
"text": "The WEBQUESTIONSSP dataset (Yih et al., 2016) contains full semantic parses for a subset of the questions from WEBQUESTIONS (Berant et al., 2013) , because 18.5% of the original dataset were found to be \"not answerable\". It consists of 3,098 question-answer pairs for training and 1,639 for testing, which were collected using Google Suggest API, and the answers were originally obtained using Amazon Mechanical Turk workers. They were updated in (Yih et al., 2016) by annotators who were familiar with the design of Freebase and added semantic parses. We further separated out 620 questions from the training set as a validation set. For query pre-processing we used an in-house named entity linking system to find the entities in a question. The quality of the entity linker is similar to that of (Yih et al., 2015) at 94% of the gold root entities being included. Similar to Dong and Lapata (2016), we replaced named entity tokens with a special token \"ENT\". For example, the question \"who plays meg in family guy\" is changed to \"who plays ENT in ENT ENT\". This helps reduce overfitting, because instead of memorizing the correct program for a specific entity, the model has to focus on other context words in the sentence, which improves generalization.",
"cite_spans": [
{
"start": 27,
"end": 45,
"text": "(Yih et al., 2016)",
"ref_id": "BIBREF38"
},
{
"start": 124,
"end": 145,
"text": "(Berant et al., 2013)",
"ref_id": "BIBREF2"
},
{
"start": 447,
"end": 465,
"text": "(Yih et al., 2016)",
"ref_id": "BIBREF38"
}
],
"ref_spans": [],
"eq_spans": [],
"section": "The WEBQUESTIONSSP dataset",
"sec_num": "3.1"
},
{
"text": "Following (Yih et al., 2015) we used the last publicly available snapshot of Freebase (Bollacker et al., 2008) . Since NSM training requires random access to Freebase during decoding, we preprocessed Freebase by removing predicates that are not related to world knowledge (starting with \"/common/ \", \"/type/ \", \"/freebase/ \"), 2 and removing all text valued predicates, which are rarely the answer. Out of all 27K relations, 434 relations are removed during preprocessing. This results in a graph that fits in memory with 23K relations, 82M nodes, and 417M edges.",
"cite_spans": [
{
"start": 10,
"end": 28,
"text": "(Yih et al., 2015)",
"ref_id": "BIBREF37"
},
{
"start": 86,
"end": 110,
"text": "(Bollacker et al., 2008)",
"ref_id": "BIBREF4"
}
],
"ref_spans": [],
"eq_spans": [],
"section": "The WEBQUESTIONSSP dataset",
"sec_num": "3.1"
},
{
"text": "For pre-trained word embeddings, we used the 300 dimension GloVe word embeddings trained on 840B tokens (Pennington et al., 2014) . On the encoder side, we added a projection matrix to transform the embeddings into 50 dimensions. On the decoder side, we used the same GloVe embeddings to construct an embedding for each property using its Freebase id, and also added a projection matrix to transform this embedding to 50 dimensions. A Freebase id contains three parts: domain, type, and property. For example, the Freebase id for PARENTSOF is \"/people/person/parents\". \"people\" is the domain, \"person\" is the type and \"parents\" is the property. The embedding is constructed by concatenating the average of word embeddings in the domain and type name to the average of word embeddings in the property name. For example, if the embedding dimension is 300, the embedding dimension for \"/people/person/parents\" will be 600. The first 300 dimensions will be the average of the embeddings for \"people\" and \"person\", and the second 300 dimensions will be the embedding for \"parents\".",
"cite_spans": [
{
"start": 104,
"end": 129,
"text": "(Pennington et al., 2014)",
"ref_id": "BIBREF22"
}
],
"ref_spans": [],
"eq_spans": [],
"section": "Model Details",
"sec_num": "3.2"
},
{
"text": "The dimension of encoder hidden state, decoder hidden state and key embeddings are all 50. The embeddings for the functions and special tokens (e.g., \"UNK\", \"GO\") are randomly initialized by a truncated normal distribution with mean=0.0 and stddev=0.1. All the weight matrices are initialized with a uniform distribution in [",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Model Details",
"sec_num": "3.2"
},
{
"text": "p 3 d , p 3 d ]",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Model Details",
"sec_num": "3.2"
},
{
"text": "where d is the input dimension. Dropout rate is set to 0.5, and we see a clear tendency for larger dropout rate to produce better performance, indicating overfitting is a major problem for learning.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Model Details",
"sec_num": "3.2"
},
{
"text": "In iterative ML training, the decoder uses a beam of size k = 100 to update the pseudo-gold programs and the model is trained for 20 epochs after each decoding step. We use the Adam optimizer (Kingma and Ba, 2014) with initial learning rate 0.001. In our experiment, this process usually converges after a few (5-8) iterations.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Training Details",
"sec_num": "3.3"
},
{
"text": "For REINFORCE training, the best hyperparameters are chosen using the validation set. We use a beam of size k = 5 for decoding, and \u21b5 is set to 0.1. Because the dataset is small and some relations are only used once in the whole training set, we train the model on the entire training set for 200 iterations with the best hyperparameters. Then we train the model with learning rate decay until convergence. Learning rate is decayed as",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Training Details",
"sec_num": "3.3"
},
{
"text": "g t = g 0 \u21e5 max(0,t ts) m",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Training Details",
"sec_num": "3.3"
},
{
"text": ", where g 0 = 0.001, = 0.5 m = 1000, and t s is the number of training steps at the end of iteration 200.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Training Details",
"sec_num": "3.3"
},
{
"text": "Since decoding needs to query the knowledge base (KB) constantly, the speed bottleneck for training is decoding. We address this problem in our implementation by partitioning the dataset, and using multiple decoders in parallel to handle each partition. We use 100 decoders, which queries 50 KG servers, and one trainer. The neural network model is implemented in TensorFlow. Since the model is small, we didn't see a significant speedup by using GPU, so all the decoders and the trainer are using CPU only.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Training Details",
"sec_num": "3.3"
},
{
"text": "Inspired by the staged generation process in Yih et al. 2015, curriculum learning includes two steps. We first run iterative ML for 10 iterations with programs constrained to only use the \"Hop\" function and the maximum number of expressions is 2. Then, we run iterative ML again, but use both \"Hop\" and \"Filter\". The maximum number of expressions is 3, and the relations used by \"Hop\" are restricted to those that appeared in a best 0:T (q) in the first step.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Training Details",
"sec_num": "3.3"
},
{
"text": "We evaluate performance using the offical evaluation script for WEBQUESTIONSSP. Because the answer to a question may contain multiple entities or values, precision, recall and F1 are computed based on the output of each individual question, and average F1 is reported as the main evaluation metric. Accuracy measures the proportion of questions that are answered exactly.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Results and discussion",
"sec_num": "3.4"
},
{
"text": "A comparison to STAGG, the previous state-ofthe-art model (Yih et al., 2016 (Yih et al., , 2015 , is shown in Table 2 . Our model beats STAGG with weak supervision by a significant margin on all metrics, while relying on no feature engineering or handcrafted rules. When STAGG is trained with strong supervision it obtains an F1 of 71.7, and thus NSM closes half the gap between training with weak and full supervision.",
"cite_spans": [
{
"start": 58,
"end": 75,
"text": "(Yih et al., 2016",
"ref_id": "BIBREF38"
},
{
"start": 76,
"end": 95,
"text": "(Yih et al., , 2015",
"ref_id": "BIBREF37"
}
],
"ref_spans": [
{
"start": 110,
"end": 117,
"text": "Table 2",
"ref_id": "TABREF2"
}
],
"eq_spans": [],
"section": "Results and discussion",
"sec_num": "3.4"
},
{
"text": "Prec. Rec. F1 Acc. STAGG 67.3 73.1 66.8 58.8 NSM 70.8 76.0 69.0 59.5 Four key ingredients lead to the final performance of NSM. The first one is the neural computer interface that provides code assistance by checking for syntax and semantic errors. We find that semantic checks are very effective for opendomain KBs with a large number of properties. For our task, the average number of choices is reduced from 23K per step (all properties) to less than 100 (the average number of properties connected to an entity).",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Model",
"sec_num": null
},
{
"text": "The second ingredient is augmented REIN-FORCE training. Table 3 compares augmented REINFORCE, REINFORCE, and iterative ML on the validation set. REINFORCE gets stuck in local optimum and performs poorly. Iterative ML training is not directly optimizing the F1 measure, and achieves sub-optimal results. In contrast, augmented REINFORCE is able to bootstrap using pseudo-gold programs found by iterative ML and achieves the best performance on both the training and validation set.",
"cite_spans": [],
"ref_spans": [
{
"start": 56,
"end": 63,
"text": "Table 3",
"ref_id": "TABREF4"
}
],
"eq_spans": [],
"section": "Model",
"sec_num": null
},
{
"text": "Train The third ingredient is curriculum learning during iterative ML. We compare the performance of the best programs found with and without curriculum learning in Table 4 . We find that the best programs found with curriculum learning are substantially better than those found without curriculum learning by a large margin on every metric.",
"cite_spans": [],
"ref_spans": [
{
"start": 165,
"end": 172,
"text": "Table 4",
"ref_id": null
}
],
"eq_spans": [],
"section": "Settings",
"sec_num": null
},
{
"text": "Prec. Rec. F1 Acc. No curriculum 79.1 91.1 78.5 67.2 Curriculum 88.6 96.1 89.5 79.8 Table 4 : Evaluation of the programs with the highest F1 score in the beam (a best 0:t ) with and without curriculum learning.",
"cite_spans": [],
"ref_spans": [
{
"start": 84,
"end": 91,
"text": "Table 4",
"ref_id": null
}
],
"eq_spans": [],
"section": "Settings",
"sec_num": null
},
{
"text": "The last important ingredient is reducing overfitting. Given the small size of the dataset, overfitting is a major problem for training neural network models. We show the contributions of different techniques for controlling overfitting in Table 5. Note that after all the techniques have been applied, the model is still overfitting with training F1@1=83.0% and validation F1@1=67.2%.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Settings",
"sec_num": null
},
{
"text": "Among the programs generated by the model, a significant portion (36.7%) uses more than one expression. From Table 6 , we can see that the performance doesn't decrease much as the composi-Settings F1@1 Pretrained word embeddings 5.5 Pretrained property embeddings 2.7 Dropout on GRU input and output 2.4 Dropout on softmax 1.1 Anonymize entity tokens 2.0 tional depth increases, indicating that the model is effective at capturing compositionality. We observe that programs with three expressions use a more limited set of properties, mainly focusing on answering a few types of questions such as \"who plays meg in family guy\", \"what college did jeff corwin go to\" and \"which countries does russia border\". In contrast, programs with two expressions use a more diverse set of properties, which could explain the lower performance compared to programs with three expressions.",
"cite_spans": [],
"ref_spans": [
{
"start": 109,
"end": 116,
"text": "Table 6",
"ref_id": "TABREF6"
}
],
"eq_spans": [],
"section": "Settings",
"sec_num": null
},
{
"text": "Error analysis Error analysis on the validation set shows two main sources of errors: 1. Search failure: Programs with high reward are not found during search for pseudo-gold programs, either because the beam size is not large enough, or because the set of functions implemented by the interpreter is insufficient. The 89.5% F1 score in Table 4 indicates that at least 10% of the questions are of this kind. 2. Ranking failure: Programs with high reward exist in the beam, but are not ranked at the top during decoding. Because the training error is low, this is largely due to overfitting or spurious programs. The 67.2% F1 score in Table 3 indicates that about 20% of the questions are of this kind.",
"cite_spans": [],
"ref_spans": [
{
"start": 337,
"end": 344,
"text": "Table 4",
"ref_id": null
},
{
"start": 634,
"end": 641,
"text": "Table 3",
"ref_id": "TABREF4"
}
],
"eq_spans": [],
"section": "Settings",
"sec_num": null
},
{
"text": "model. Therefore, both models rely on REIN-FORCE for training. The main difference between the two is the abstraction level of the programming language. RL-NTM uses lower level operations such as memory address manipulation and byte reading/writing, while NSM uses a high level programming language over a large knowledge base that includes operations such as following properties from entities, or sorting based on a property, which is more suitable for representing semantics. Earlier works such as OOPS (Schmidhuber, 2004) has desirable characteristics, for example, the ability to define new functions. These remain to be future improvements for NSM.",
"cite_spans": [
{
"start": 506,
"end": 525,
"text": "(Schmidhuber, 2004)",
"ref_id": "BIBREF27"
}
],
"ref_spans": [],
"eq_spans": [],
"section": "Settings",
"sec_num": null
},
{
"text": "We formulate NSM training as an instance of reinforcement learning (Sutton and Barto, 1998) in order to directly optimize the task reward of the structured prediction problem Li et al., 2016; Yu et al., 2017) . Compared to imitation learning methods (Daume et al., 2009; Ross et al., 2011 ) that interpolate a model distribution with an oracle, NSM needs to solve a challenging search problem of training from weak supervisions in a large search space. Our solution employs two techniques (a) a symbolic \"computer\" helps find good programs by pruning the search space (b) an iterative ML training process, where beam search is used to find pseudogold programs. Wiseman and Rush (Wiseman and Rush, 2016) proposed a max-margin approach to train a sequence-to-sequence scorer. However, their training procedure is more involved, and we did not implement it in this work. MIXER (Ranzato et al., 2015) also proposed to combine ML training and REINFORCE, but they only considered tasks with full supervisions. applied imitation learning to semantic parsing, but still requires hand crafted grammars and features.",
"cite_spans": [
{
"start": 67,
"end": 91,
"text": "(Sutton and Barto, 1998)",
"ref_id": "BIBREF30"
},
{
"start": 175,
"end": 191,
"text": "Li et al., 2016;",
"ref_id": "BIBREF7"
},
{
"start": 192,
"end": 208,
"text": "Yu et al., 2017)",
"ref_id": "BIBREF40"
},
{
"start": 250,
"end": 270,
"text": "(Daume et al., 2009;",
"ref_id": "BIBREF6"
},
{
"start": 271,
"end": 288,
"text": "Ross et al., 2011",
"ref_id": "BIBREF25"
},
{
"start": 661,
"end": 702,
"text": "Wiseman and Rush (Wiseman and Rush, 2016)",
"ref_id": "BIBREF35"
}
],
"ref_spans": [],
"eq_spans": [],
"section": "Settings",
"sec_num": null
},
{
"text": "NSM is similar to Neural Programmer (Neelakantan et al., 2015) and Dynamic Neural Module Network (Andreas et al., 2016) in that they all solve the problem of semantic parsing from structured data, and generate programs using similar semantics. The main difference between these approaches is how an intermediate result (the memory) is represented. Neural Programmer and Dynamic-NMN chose to represent results as vectors of weights (row selectors and attention vectors), which enables backpropagation and search through all possible programs in parallel. How-ever, their strategy is not applicable to a large KB such as Freebase, which contains about 100M entities, and more than 20k properties. Instead, NSM chooses a more scalable approach, where the \"computer\" saves intermediate results, and the neural network only refers to them with variable names (e.g., \"R 1 \" for all cities in the US).",
"cite_spans": [
{
"start": 36,
"end": 62,
"text": "(Neelakantan et al., 2015)",
"ref_id": "BIBREF19"
},
{
"start": 97,
"end": 119,
"text": "(Andreas et al., 2016)",
"ref_id": "BIBREF0"
}
],
"ref_spans": [],
"eq_spans": [],
"section": "Settings",
"sec_num": null
},
{
"text": "NSM is similar to the Path Ranking Algorithm (PRA) (Lao et al., 2011) in that semantics is encoded as a sequence of actions, and denotations are used to prune the search space during learning. NSM is more powerful than PRA by 1) allowing more complex semantics to be composed through the use of a key-variable memory; 2) controlling the search procedure with a trained neural network, while PRA only samples actions uniformly; 3) allowing input questions to express complex relations, and then dynamically generating action sequences. PRA can combine multiple semantic representations to produce the final prediction, which remains to be future work for NSM.",
"cite_spans": [
{
"start": 51,
"end": 69,
"text": "(Lao et al., 2011)",
"ref_id": "BIBREF16"
}
],
"ref_spans": [],
"eq_spans": [],
"section": "Settings",
"sec_num": null
},
{
"text": "We propose the Manager-Programmer-Computer framework for neural program induction. It integrates neural networks with a symbolic nondifferentiable computer to support abstract, scalable and precise operations through a friendly neural computer interface. Within this framework, we introduce the Neural Symbolic Machine, which integrates a neural sequence-to-sequence \"programmer\" with key-variable memory, and a symbolic Lisp interpreter with code assistance. Because the interpreter is non-differentiable and to directly optimize the task reward, we apply REIN-FORCE and use pseudo-gold programs found by an iterative ML training process to bootstrap training. NSM achieves new state-of-the-art results on a challenging semantic parsing dataset with weak supervision, and significantly closes the gap between weak and full supervision. It is trained endto-end, and does not require any feature engineering or domain-specific knowledge.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Conclusion",
"sec_num": "5"
},
{
"text": "We also consider numbers (e.g., \"1.33\") and date-times (e.g., \"1999-1-1\") as entities.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "",
"sec_num": null
},
{
"text": "We kept \"/common/topic/notable types\".",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "",
"sec_num": null
},
{
"text": "Related workAmong deep learning models for program induction, Reinforcement Learning Neural Turing Machines (RL-NTMs)(Zaremba and Sutskever, 2015) are the most similar to NSM, as a nondifferentiable machine is controlled by a sequence",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "",
"sec_num": null
}
],
"back_matter": [
{
"text": "We thank for discussions and help from Arvind Neelakantan, Mohammad Norouzi, Tom Kwiatkowski, Eugene Brevdo, Lukasz Kaizer, Thomas Strohmann, Yonghui Wu, Zhifeng Chen, Alexandre Lacoste, and John Blitzer. The second author is partially supported by the Israel Science Foundation, grant 942/16.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Acknowledgements",
"sec_num": null
}
],
"bib_entries": {
"BIBREF0": {
"ref_id": "b0",
"title": "Learning to compose neural networks for question answering",
"authors": [
{
"first": "Jacob",
"middle": [],
"last": "Andreas",
"suffix": ""
},
{
"first": "Marcus",
"middle": [],
"last": "Rohrbach",
"suffix": ""
},
{
"first": "Trevor",
"middle": [],
"last": "Darrell",
"suffix": ""
},
{
"first": "Dan",
"middle": [],
"last": "Klein",
"suffix": ""
}
],
"year": 2016,
"venue": "",
"volume": "",
"issue": "",
"pages": "",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Jacob Andreas, Marcus Rohrbach, Trevor Darrell, and Dan Klein. 2016. Learning to compose neural networks for question answering. CoRR abs/1601.01705.",
"links": null
},
"BIBREF1": {
"ref_id": "b1",
"title": "Neural machine translation by jointly learning to align and translate",
"authors": [
{
"first": "Dzmitry",
"middle": [],
"last": "Bahdanau",
"suffix": ""
},
{
"first": "Kyunghyun",
"middle": [],
"last": "Cho",
"suffix": ""
},
{
"first": "Yoshua",
"middle": [],
"last": "Bengio",
"suffix": ""
}
],
"year": 2014,
"venue": "",
"volume": "",
"issue": "",
"pages": "",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Dzmitry Bahdanau, Kyunghyun Cho, and Yoshua Bengio. 2014. Neural machine translation by jointly learning to align and translate. CoRR abs/1409.0473. http://arxiv.org/abs/1409.0473.",
"links": null
},
"BIBREF2": {
"ref_id": "b2",
"title": "Semantic parsing on freebase from question-answer pairs",
"authors": [
{
"first": "Jonathan",
"middle": [],
"last": "Berant",
"suffix": ""
},
{
"first": "Andrew",
"middle": [],
"last": "Chou",
"suffix": ""
},
{
"first": "Roy",
"middle": [],
"last": "Frostig",
"suffix": ""
},
{
"first": "Percy",
"middle": [],
"last": "Liang",
"suffix": ""
}
],
"year": 2013,
"venue": "EMNLP",
"volume": "2",
"issue": "",
"pages": "",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Jonathan Berant, Andrew Chou, Roy Frostig, and Percy Liang. 2013. Semantic parsing on freebase from question-answer pairs. In EMNLP. volume 2, page 6.",
"links": null
},
"BIBREF3": {
"ref_id": "b3",
"title": "Imitation learning of agenda-based semantic parsers",
"authors": [
{
"first": "Jonathan",
"middle": [],
"last": "Berant",
"suffix": ""
},
{
"first": "Percy",
"middle": [],
"last": "Liang",
"suffix": ""
}
],
"year": 2015,
"venue": "TACL",
"volume": "3",
"issue": "",
"pages": "545--558",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Jonathan Berant and Percy Liang. 2015. Imitation learning of agenda-based semantic parsers. TACL 3:545-558.",
"links": null
},
"BIBREF4": {
"ref_id": "b4",
"title": "Freebase: a collaboratively created graph database for structuring human knowledge",
"authors": [
{
"first": "K",
"middle": [],
"last": "Bollacker",
"suffix": ""
},
{
"first": "C",
"middle": [],
"last": "Evans",
"suffix": ""
},
{
"first": "P",
"middle": [],
"last": "Paritosh",
"suffix": ""
},
{
"first": "T",
"middle": [],
"last": "Sturge",
"suffix": ""
},
{
"first": "J",
"middle": [],
"last": "Taylor",
"suffix": ""
}
],
"year": 2008,
"venue": "International Conference on Management of Data (SIGMOD)",
"volume": "",
"issue": "",
"pages": "1247--1250",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "K. Bollacker, C. Evans, P. Paritosh, T. Sturge, and J. Taylor. 2008. Freebase: a collaboratively created graph database for structuring human knowledge. In International Conference on Management of Data (SIGMOD). pages 1247-1250.",
"links": null
},
"BIBREF5": {
"ref_id": "b5",
"title": "Learning phrase representations using rnn encoder-decoder for statistical machine translation",
"authors": [
{
"first": "Kyunghyun",
"middle": [],
"last": "Cho",
"suffix": ""
},
{
"first": "Bart",
"middle": [],
"last": "Van Merrienboer",
"suffix": ""
},
{
"first": "Caglar",
"middle": [],
"last": "Gulcehre",
"suffix": ""
},
{
"first": "Dzmitry",
"middle": [],
"last": "Bahdanau",
"suffix": ""
},
{
"first": "Fethi",
"middle": [],
"last": "Bougares",
"suffix": ""
},
{
"first": "Holger",
"middle": [],
"last": "Schwenk",
"suffix": ""
},
{
"first": "Yoshua",
"middle": [],
"last": "Bengio",
"suffix": ""
}
],
"year": 2014,
"venue": "Proceedings of the 2014 Conference on Empirical Methods in Natural Language Processing (EMNLP)",
"volume": "",
"issue": "",
"pages": "1724--1734",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Kyunghyun Cho, Bart van Merrienboer, Caglar Gul- cehre, Dzmitry Bahdanau, Fethi Bougares, Holger Schwenk, and Yoshua Bengio. 2014. Learning phrase representations using rnn encoder-decoder for statistical machine translation. In Proceedings of the 2014 Conference on Empirical Methods in Nat- ural Language Processing (EMNLP). Association for Computational Linguistics, Doha, Qatar, pages 1724-1734. http://www.aclweb.org/anthology/D14- 1179.",
"links": null
},
"BIBREF6": {
"ref_id": "b6",
"title": "Searchbased structured prediction",
"authors": [
{
"first": "H",
"middle": [],
"last": "Daume",
"suffix": ""
},
{
"first": "J",
"middle": [],
"last": "Langford",
"suffix": ""
},
{
"first": "D",
"middle": [],
"last": "Marcu",
"suffix": ""
}
],
"year": 2009,
"venue": "Machine Learning",
"volume": "75",
"issue": "",
"pages": "297--325",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "H. Daume, J. Langford, and D. Marcu. 2009. Search- based structured prediction. Machine Learning 75:297-325.",
"links": null
},
"BIBREF7": {
"ref_id": "b7",
"title": "Language to logical form with neural attention",
"authors": [
{
"first": "Li",
"middle": [],
"last": "Dong",
"suffix": ""
},
{
"first": "Mirella",
"middle": [],
"last": "Lapata",
"suffix": ""
}
],
"year": 2016,
"venue": "Association for Computational Linguistics (ACL)",
"volume": "",
"issue": "",
"pages": "",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Li Dong and Mirella Lapata. 2016. Language to log- ical form with neural attention. In Association for Computational Linguistics (ACL).",
"links": null
},
"BIBREF8": {
"ref_id": "b8",
"title": "Neural turing machines",
"authors": [
{
"first": "Alex",
"middle": [],
"last": "Graves",
"suffix": ""
},
{
"first": "Greg",
"middle": [],
"last": "Wayne",
"suffix": ""
},
{
"first": "Ivo",
"middle": [],
"last": "Danihelka",
"suffix": ""
}
],
"year": 2014,
"venue": "",
"volume": "",
"issue": "",
"pages": "",
"other_ids": {
"arXiv": [
"arXiv:1410.5401"
]
},
"num": null,
"urls": [],
"raw_text": "Alex Graves, Greg Wayne, and Ivo Danihelka. 2014. Neural turing machines. arXiv preprint arXiv:1410.5401 .",
"links": null
},
"BIBREF9": {
"ref_id": "b9",
"title": "Hybrid computing using a neural network with dynamic external memory",
"authors": [
{
"first": "Alex",
"middle": [],
"last": "Graves",
"suffix": ""
},
{
"first": "Greg",
"middle": [],
"last": "Wayne",
"suffix": ""
},
{
"first": "Malcolm",
"middle": [],
"last": "Reynolds",
"suffix": ""
},
{
"first": "Tim",
"middle": [],
"last": "Harley",
"suffix": ""
},
{
"first": "Ivo",
"middle": [],
"last": "Danihelka",
"suffix": ""
},
{
"first": "Agnieszka",
"middle": [],
"last": "Grabska-Barwinska",
"suffix": ""
},
{
"first": "Sergio",
"middle": [
"G"
],
"last": "Colmenarejo",
"suffix": ""
},
{
"first": "Edward",
"middle": [],
"last": "Grefenstette",
"suffix": ""
},
{
"first": "Tiago",
"middle": [],
"last": "Ramalho",
"suffix": ""
},
{
"first": "John",
"middle": [],
"last": "Agapiou",
"suffix": ""
},
{
"first": "Adri\u00e3",
"middle": [
"P"
],
"last": "Badia",
"suffix": ""
},
{
"first": "Karl",
"middle": [
"M"
],
"last": "Hermann",
"suffix": ""
},
{
"first": "Yori",
"middle": [],
"last": "Zwols",
"suffix": ""
},
{
"first": "Georg",
"middle": [],
"last": "Ostrovski",
"suffix": ""
}
],
"year": 2016,
"venue": "",
"volume": "",
"issue": "",
"pages": "",
"other_ids": {
"DOI": [
"10.1038/nature20101"
]
},
"num": null,
"urls": [],
"raw_text": "Alex Graves, Greg Wayne, Malcolm Reynolds, Tim Harley, Ivo Danihelka, Agnieszka Grabska- Barwinska, Sergio G. Colmenarejo, Edward Grefen- stette, Tiago Ramalho, John Agapiou, Adri\u00c3 P. Badia, Karl M. Hermann, Yori Zwols, Georg Ostrovski, Adam Cain, Helen King, Christopher Summerfield, Phil Blunsom, Koray Kavukcuoglu, and Demis Hassabis. 2016a. Hybrid comput- ing using a neural network with dynamic exter- nal memory. Nature advance online publication. https://doi.org/10.1038/nature20101.",
"links": null
},
"BIBREF10": {
"ref_id": "b10",
"title": "Hybrid computing using a neural network with dynamic external memory",
"authors": [
{
"first": "Alex",
"middle": [],
"last": "Graves",
"suffix": ""
},
{
"first": "Greg",
"middle": [],
"last": "Wayne",
"suffix": ""
},
{
"first": "Malcolm",
"middle": [],
"last": "Reynolds",
"suffix": ""
},
{
"first": "Tim",
"middle": [],
"last": "Harley",
"suffix": ""
},
{
"first": "Ivo",
"middle": [],
"last": "Danihelka",
"suffix": ""
},
{
"first": "Agnieszka",
"middle": [],
"last": "Grabska-Barwi\u0144ska",
"suffix": ""
},
{
"first": "Sergio",
"middle": [
"G\u00f3mez"
],
"last": "Colmenarejo",
"suffix": ""
},
{
"first": "Edward",
"middle": [],
"last": "Grefenstette",
"suffix": ""
},
{
"first": "Tiago",
"middle": [],
"last": "Ramalho",
"suffix": ""
},
{
"first": "John",
"middle": [],
"last": "Agapiou",
"suffix": ""
}
],
"year": 2016,
"venue": "Nature",
"volume": "",
"issue": "",
"pages": "",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Alex Graves, Greg Wayne, Malcolm Reynolds, Tim Harley, Ivo Danihelka, Agnieszka Grabska- Barwi\u0144ska, Sergio G\u00f3mez Colmenarejo, Edward Grefenstette, Tiago Ramalho, John Agapiou, et al. 2016b. Hybrid computing using a neural network with dynamic external memory. Nature .",
"links": null
},
"BIBREF11": {
"ref_id": "b11",
"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": ""
}
],
"year": 2012,
"venue": "IEEE 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, et al. 2012. Deep neural networks for acoustic modeling in speech recognition: The shared views of four research groups. IEEE Signal Process- ing Magazine 29(6):82-97.",
"links": null
},
"BIBREF12": {
"ref_id": "b12",
"title": "Data recombination for neural semantic parsing",
"authors": [
{
"first": "Robin",
"middle": [],
"last": "Jia",
"suffix": ""
},
{
"first": "Percy",
"middle": [],
"last": "Liang",
"suffix": ""
}
],
"year": 2016,
"venue": "Association for Computational Linguistics (ACL)",
"volume": "",
"issue": "",
"pages": "",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Robin Jia and Percy Liang. 2016. Data recombination for neural semantic parsing. In Association for Com- putational Linguistics (ACL).",
"links": null
},
"BIBREF13": {
"ref_id": "b13",
"title": "Learned prioritization for trading off accuracy and speed",
"authors": [
{
"first": "J",
"middle": [],
"last": "Jiang",
"suffix": ""
},
{
"first": "A",
"middle": [],
"last": "Teichert",
"suffix": ""
},
{
"first": "J",
"middle": [],
"last": "Eisner",
"suffix": ""
},
{
"first": "H",
"middle": [],
"last": "Daume",
"suffix": ""
}
],
"year": 2012,
"venue": "Advances in Neural Information Processing Systems (NIPS)",
"volume": "",
"issue": "",
"pages": "",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "J. Jiang, A. Teichert, J. Eisner, and H. Daume. 2012. Learned prioritization for trading off accuracy and speed. In Advances in Neural Information Process- ing Systems (NIPS).",
"links": null
},
"BIBREF14": {
"ref_id": "b14",
"title": "Neural gpus learn algorithms",
"authors": [
{
"first": "\u0141ukasz",
"middle": [],
"last": "Kaiser",
"suffix": ""
},
{
"first": "Ilya",
"middle": [],
"last": "Sutskever",
"suffix": ""
}
],
"year": 2015,
"venue": "",
"volume": "",
"issue": "",
"pages": "",
"other_ids": {
"arXiv": [
"arXiv:1511.08228"
]
},
"num": null,
"urls": [],
"raw_text": "\u0141ukasz Kaiser and Ilya Sutskever. 2015. Neural gpus learn algorithms. arXiv preprint arXiv:1511.08228 .",
"links": null
},
"BIBREF15": {
"ref_id": "b15",
"title": "Adam: A method for stochastic optimization",
"authors": [
{
"first": "P",
"middle": [],
"last": "Diederik",
"suffix": ""
},
{
"first": "Jimmy",
"middle": [],
"last": "Kingma",
"suffix": ""
},
{
"first": "",
"middle": [],
"last": "Ba",
"suffix": ""
}
],
"year": 2014,
"venue": "",
"volume": "",
"issue": "",
"pages": "",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Diederik P. Kingma and Jimmy Ba. 2014. Adam: A method for stochastic optimization. CoRR abs/1412.6980. http://arxiv.org/abs/1412.6980.",
"links": null
},
"BIBREF16": {
"ref_id": "b16",
"title": "Random walk inference and learning in a large scale knowledge base",
"authors": [
{
"first": "Ni",
"middle": [],
"last": "Lao",
"suffix": ""
},
{
"first": "Tom",
"middle": [],
"last": "Mitchell",
"suffix": ""
},
{
"first": "William",
"middle": [
"W"
],
"last": "Cohen",
"suffix": ""
}
],
"year": 2011,
"venue": "Proceedings of the Conference on Empirical Methods in Natural Language Processing",
"volume": "",
"issue": "",
"pages": "529--539",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Ni Lao, Tom Mitchell, and William W Cohen. 2011. Random walk inference and learning in a large scale knowledge base. In Proceedings of the Conference on Empirical Methods in Natural Language Pro- cessing. Association for Computational Linguistics, pages 529-539.",
"links": null
},
"BIBREF17": {
"ref_id": "b17",
"title": "Deep reinforcement learning for dialogue generation",
"authors": [
{
"first": "Jiwei",
"middle": [],
"last": "Li",
"suffix": ""
},
{
"first": "Will",
"middle": [],
"last": "Monroe",
"suffix": ""
},
{
"first": "Alan",
"middle": [],
"last": "Ritter",
"suffix": ""
},
{
"first": "Michel",
"middle": [],
"last": "Galley",
"suffix": ""
},
{
"first": "Jianfeng",
"middle": [],
"last": "Gao",
"suffix": ""
},
{
"first": "Dan",
"middle": [],
"last": "Jurafsky",
"suffix": ""
}
],
"year": 2016,
"venue": "",
"volume": "",
"issue": "",
"pages": "",
"other_ids": {
"arXiv": [
"arXiv:1606.01541"
]
},
"num": null,
"urls": [],
"raw_text": "Jiwei Li, Will Monroe, Alan Ritter, Michel Galley, Jianfeng Gao, and Dan Jurafsky. 2016. Deep rein- forcement learning for dialogue generation. arXiv preprint arXiv:1606.01541 .",
"links": null
},
"BIBREF18": {
"ref_id": "b18",
"title": "Learning dependency-based compositional semantics",
"authors": [
{
"first": "P",
"middle": [],
"last": "Liang",
"suffix": ""
},
{
"first": "M",
"middle": [
"I"
],
"last": "Jordan",
"suffix": ""
},
{
"first": "D",
"middle": [],
"last": "Klein",
"suffix": ""
}
],
"year": 2011,
"venue": "Association for Computational Linguistics (ACL)",
"volume": "",
"issue": "",
"pages": "590--599",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "P. Liang, M. I. Jordan, and D. Klein. 2011. Learn- ing dependency-based compositional semantics. In Association for Computational Linguistics (ACL). pages 590-599.",
"links": null
},
"BIBREF19": {
"ref_id": "b19",
"title": "Neural programmer: Inducing latent programs with gradient descent",
"authors": [
{
"first": "Arvind",
"middle": [],
"last": "Neelakantan",
"suffix": ""
},
{
"first": "V",
"middle": [],
"last": "Quoc",
"suffix": ""
},
{
"first": "Ilya",
"middle": [],
"last": "Le",
"suffix": ""
},
{
"first": "",
"middle": [],
"last": "Sutskever",
"suffix": ""
}
],
"year": 2015,
"venue": "",
"volume": "",
"issue": "",
"pages": "",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Arvind Neelakantan, Quoc V. Le, and Ilya Sutskever. 2015. Neural programmer: Inducing la- tent programs with gradient descent. CoRR abs/1511.04834.",
"links": null
},
"BIBREF20": {
"ref_id": "b20",
"title": "Reward augmented maximum likelihood for neural structured prediction",
"authors": [
{
"first": "Mohammad",
"middle": [],
"last": "Norouzi",
"suffix": ""
},
{
"first": "Samy",
"middle": [],
"last": "Bengio",
"suffix": ""
},
{
"first": "Navdeep",
"middle": [],
"last": "Zhifeng Chen",
"suffix": ""
},
{
"first": "Mike",
"middle": [],
"last": "Jaitly",
"suffix": ""
},
{
"first": "Yonghui",
"middle": [],
"last": "Schuster",
"suffix": ""
},
{
"first": "Dale",
"middle": [],
"last": "Wu",
"suffix": ""
},
{
"first": "",
"middle": [],
"last": "Schuurmans",
"suffix": ""
}
],
"year": 2016,
"venue": "Advances in Neural Information Processing Systems (NIPS)",
"volume": "",
"issue": "",
"pages": "",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Mohammad Norouzi, Samy Bengio, zhifeng Chen, Navdeep Jaitly, Mike Schuster, Yonghui Wu, and Dale Schuurmans. 2016. Reward augmented max- imum likelihood for neural structured prediction. In Advances in Neural Information Processing Systems (NIPS).",
"links": null
},
"BIBREF21": {
"ref_id": "b21",
"title": "Compositional semantic parsing on semi-structured tables",
"authors": [
{
"first": "Panupong",
"middle": [],
"last": "Pasupat",
"suffix": ""
},
{
"first": "Percy",
"middle": [],
"last": "Liang",
"suffix": ""
}
],
"year": 2015,
"venue": "ACL",
"volume": "",
"issue": "",
"pages": "",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Panupong Pasupat and Percy Liang. 2015. Composi- tional semantic parsing on semi-structured tables. In ACL.",
"links": null
},
"BIBREF22": {
"ref_id": "b22",
"title": "Glove: Global vectors for word representation",
"authors": [
{
"first": "Jeffrey",
"middle": [],
"last": "Pennington",
"suffix": ""
},
{
"first": "Richard",
"middle": [],
"last": "Socher",
"suffix": ""
},
{
"first": "Christopher",
"middle": [
"D"
],
"last": "Manning",
"suffix": ""
}
],
"year": 2014,
"venue": "EMNLP",
"volume": "",
"issue": "",
"pages": "",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Jeffrey Pennington, Richard Socher, and Christo- pher D. Manning. 2014. Glove: Global vectors for word representation. In EMNLP.",
"links": null
},
"BIBREF23": {
"ref_id": "b23",
"title": "Sequence level training with recurrent neural networks",
"authors": [
{
"first": "Aurelio",
"middle": [],
"last": "Marc",
"suffix": ""
},
{
"first": "Sumit",
"middle": [],
"last": "Ranzato",
"suffix": ""
},
{
"first": "Michael",
"middle": [],
"last": "Chopra",
"suffix": ""
},
{
"first": "Wojciech",
"middle": [],
"last": "Auli",
"suffix": ""
},
{
"first": "",
"middle": [],
"last": "Zaremba",
"suffix": ""
}
],
"year": 2015,
"venue": "",
"volume": "",
"issue": "",
"pages": "",
"other_ids": {
"arXiv": [
"arXiv:1511.06732"
]
},
"num": null,
"urls": [],
"raw_text": "Marc'Aurelio Ranzato, Sumit Chopra, Michael Auli, and Wojciech Zaremba. 2015. Sequence level train- ing with recurrent neural networks. arXiv preprint arXiv:1511.06732 .",
"links": null
},
"BIBREF24": {
"ref_id": "b24",
"title": "Neural programmer-interpreters",
"authors": [
{
"first": "Scott",
"middle": [],
"last": "Reed",
"suffix": ""
},
{
"first": "",
"middle": [],
"last": "Nando De Freitas",
"suffix": ""
}
],
"year": 2016,
"venue": "ICLR",
"volume": "",
"issue": "",
"pages": "",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Scott Reed and Nando de Freitas. 2016. Neural programmer-interpreters. In ICLR.",
"links": null
},
"BIBREF25": {
"ref_id": "b25",
"title": "A reduction of imitation learning and structured prediction to noregret online learning",
"authors": [
{
"first": "S",
"middle": [],
"last": "Ross",
"suffix": ""
},
{
"first": "G",
"middle": [],
"last": "Gordon",
"suffix": ""
},
{
"first": "A",
"middle": [],
"last": "Bagnell",
"suffix": ""
}
],
"year": 2011,
"venue": "Artificial Intelligence and Statistics (AISTATS)",
"volume": "",
"issue": "",
"pages": "",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "S. Ross, G. Gordon, and A. Bagnell. 2011. A reduction of imitation learning and structured prediction to no- regret online learning. In Artificial Intelligence and Statistics (AISTATS).",
"links": null
},
"BIBREF26": {
"ref_id": "b26",
"title": "Prioritized experience replay",
"authors": [
{
"first": "Tom",
"middle": [],
"last": "Schaul",
"suffix": ""
},
{
"first": "John",
"middle": [],
"last": "Quan",
"suffix": ""
},
{
"first": "Ioannis",
"middle": [],
"last": "Antonoglou",
"suffix": ""
},
{
"first": "David",
"middle": [],
"last": "Silver",
"suffix": ""
}
],
"year": 2016,
"venue": "International Conference on Learning Representations",
"volume": "",
"issue": "",
"pages": "",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Tom Schaul, John Quan, Ioannis Antonoglou, and David Silver. 2016. Prioritized experience replay. In International Conference on Learning Represen- tations. Puerto Rico.",
"links": null
},
"BIBREF27": {
"ref_id": "b27",
"title": "Optimal ordered problem solver",
"authors": [
{
"first": "J\u00fcrgen",
"middle": [],
"last": "Schmidhuber",
"suffix": ""
}
],
"year": 2004,
"venue": "Machine Learning",
"volume": "54",
"issue": "3",
"pages": "211--254",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "J\u00fcrgen Schmidhuber. 2004. Optimal ordered problem solver. Machine Learning 54(3):211-254.",
"links": null
},
"BIBREF28": {
"ref_id": "b28",
"title": "Mastering the game of go with deep neural networks and tree search",
"authors": [
{
"first": "David",
"middle": [],
"last": "Silver",
"suffix": ""
},
{
"first": "Aja",
"middle": [],
"last": "Huang",
"suffix": ""
},
{
"first": "Chris",
"middle": [
"J"
],
"last": "Maddison",
"suffix": ""
},
{
"first": "Arthur",
"middle": [],
"last": "Guez",
"suffix": ""
},
{
"first": "Laurent",
"middle": [],
"last": "Sifre",
"suffix": ""
},
{
"first": "George",
"middle": [],
"last": "Van Den",
"suffix": ""
},
{
"first": "Julian",
"middle": [],
"last": "Driessche",
"suffix": ""
},
{
"first": "Ioannis",
"middle": [],
"last": "Schrittwieser",
"suffix": ""
},
{
"first": "Veda",
"middle": [],
"last": "Antonoglou",
"suffix": ""
},
{
"first": "Marc",
"middle": [],
"last": "Panneershelvam",
"suffix": ""
},
{
"first": "",
"middle": [],
"last": "Lanctot",
"suffix": ""
}
],
"year": 2016,
"venue": "Nature",
"volume": "529",
"issue": "7587",
"pages": "484--489",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "David Silver, Aja Huang, Chris J Maddison, Arthur Guez, Laurent Sifre, George Van Den Driessche, Ju- lian Schrittwieser, Ioannis Antonoglou, Veda Pan- neershelvam, Marc Lanctot, et al. 2016. Mastering the game of go with deep neural networks and tree search. Nature 529(7587):484-489.",
"links": null
},
"BIBREF29": {
"ref_id": "b29",
"title": "Sequence to sequence learning with neural networks",
"authors": [
{
"first": "Ilya",
"middle": [],
"last": "Sutskever",
"suffix": ""
},
{
"first": "Oriol",
"middle": [],
"last": "Vinyals",
"suffix": ""
},
{
"first": "Quoc V",
"middle": [],
"last": "Le",
"suffix": ""
}
],
"year": 2014,
"venue": "Advances in neural information processing systems",
"volume": "",
"issue": "",
"pages": "3104--3112",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Ilya Sutskever, Oriol Vinyals, and Quoc V Le. 2014. Sequence to sequence learning with neural net- works. In Advances in neural information process- ing systems. pages 3104-3112.",
"links": null
},
"BIBREF30": {
"ref_id": "b30",
"title": "Reinforcement Learning: An Introduction",
"authors": [
{
"first": "Richard",
"middle": [
"S"
],
"last": "Sutton",
"suffix": ""
},
{
"first": "Andrew",
"middle": [
"G"
],
"last": "Barto",
"suffix": ""
}
],
"year": 1998,
"venue": "",
"volume": "",
"issue": "",
"pages": "",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Richard S. Sutton and Andrew G. Barto. 1998. Rein- forcement Learning: An Introduction. MIT Press, Cambridge, MA.",
"links": null
},
"BIBREF31": {
"ref_id": "b31",
"title": "Pointer networks",
"authors": [
{
"first": "Oriol",
"middle": [],
"last": "Vinyals",
"suffix": ""
},
{
"first": "Meire",
"middle": [],
"last": "Fortunato",
"suffix": ""
},
{
"first": "Navdeep",
"middle": [],
"last": "Jaitly",
"suffix": ""
}
],
"year": 2015,
"venue": "Advances in Neural Information Processing Systems (NIPS)",
"volume": "",
"issue": "",
"pages": "",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Oriol Vinyals, Meire Fortunato, and Navdeep Jaitly. 2015. Pointer networks. In Advances in Neural In- formation Processing Systems (NIPS).",
"links": null
},
"BIBREF32": {
"ref_id": "b32",
"title": "Building a semantic parser overnight",
"authors": [
{
"first": "Yushi",
"middle": [],
"last": "Wang",
"suffix": ""
},
{
"first": "Jonathan",
"middle": [],
"last": "Berant",
"suffix": ""
},
{
"first": "Percy",
"middle": [],
"last": "Liang",
"suffix": ""
}
],
"year": 2015,
"venue": "Association for Computational Linguistics (ACL)",
"volume": "",
"issue": "",
"pages": "",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Yushi Wang, Jonathan Berant, and Percy Liang. 2015. Building a semantic parser overnight. In Associa- tion for Computational Linguistics (ACL).",
"links": null
},
"BIBREF34": {
"ref_id": "b34",
"title": "Simple statistical gradientfollowing algorithms for connectionist reinforcement learning",
"authors": [
{
"first": "Ronald",
"middle": [
"J"
],
"last": "Williams",
"suffix": ""
}
],
"year": 1992,
"venue": "Machine Learning",
"volume": "",
"issue": "",
"pages": "229--256",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Ronald J. Williams. 1992. Simple statistical gradient- following algorithms for connectionist reinforce- ment learning. In Machine Learning. pages 229- 256.",
"links": null
},
"BIBREF35": {
"ref_id": "b35",
"title": "Sequence-to-sequence learning as beamsearch optimization",
"authors": [
{
"first": "Sam",
"middle": [],
"last": "Wiseman",
"suffix": ""
},
{
"first": "Alexander",
"middle": [
"M"
],
"last": "Rush",
"suffix": ""
}
],
"year": 2016,
"venue": "",
"volume": "",
"issue": "",
"pages": "",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Sam Wiseman and Alexander M. Rush. 2016. Sequence-to-sequence learning as beam- search optimization. CoRR abs/1606.02960.",
"links": null
},
"BIBREF36": {
"ref_id": "b36",
"title": "Google's neural machine translation system: Bridging the gap between human and machine translation",
"authors": [
{
"first": "Yonghui",
"middle": [],
"last": "Wu",
"suffix": ""
},
{
"first": "Mike",
"middle": [],
"last": "Schuster",
"suffix": ""
},
{
"first": "Zhifeng",
"middle": [],
"last": "Chen",
"suffix": ""
},
{
"first": "Quoc",
"middle": [
"V"
],
"last": "Le",
"suffix": ""
},
{
"first": "Mohammad",
"middle": [],
"last": "Norouzi",
"suffix": ""
},
{
"first": "Wolfgang",
"middle": [],
"last": "Macherey",
"suffix": ""
},
{
"first": "Maxim",
"middle": [],
"last": "Krikun",
"suffix": ""
},
{
"first": "Yuan",
"middle": [],
"last": "Cao",
"suffix": ""
},
{
"first": "Qin",
"middle": [],
"last": "Gao",
"suffix": ""
},
{
"first": "Klaus",
"middle": [],
"last": "Macherey",
"suffix": ""
},
{
"first": "Jeff",
"middle": [],
"last": "Klingner",
"suffix": ""
},
{
"first": "Apurva",
"middle": [],
"last": "Shah",
"suffix": ""
},
{
"first": "Melvin",
"middle": [],
"last": "Johnson",
"suffix": ""
},
{
"first": "Xiaobing",
"middle": [],
"last": "Liu",
"suffix": ""
},
{
"first": "Lukasz",
"middle": [],
"last": "Kaiser",
"suffix": ""
},
{
"first": "Stephan",
"middle": [],
"last": "Gouws",
"suffix": ""
},
{
"first": "Yoshikiyo",
"middle": [],
"last": "Kato",
"suffix": ""
},
{
"first": "Taku",
"middle": [],
"last": "Kudo",
"suffix": ""
},
{
"first": "Hideto",
"middle": [],
"last": "Kazawa",
"suffix": ""
},
{
"first": "Keith",
"middle": [],
"last": "Stevens",
"suffix": ""
},
{
"first": "George",
"middle": [],
"last": "Kurian",
"suffix": ""
},
{
"first": "Nishant",
"middle": [],
"last": "Patil",
"suffix": ""
},
{
"first": "Wei",
"middle": [],
"last": "Wang",
"suffix": ""
}
],
"year": 2016,
"venue": "Oriol Vinyals",
"volume": "",
"issue": "",
"pages": "",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Yonghui Wu, Mike Schuster, Zhifeng Chen, Quoc V. Le, Mohammad Norouzi, Wolfgang Macherey, Maxim Krikun, Yuan Cao, Qin Gao, Klaus Macherey, Jeff Klingner, Apurva Shah, Melvin Johnson, Xiaobing Liu, Lukasz Kaiser, Stephan Gouws, Yoshikiyo Kato, Taku Kudo, Hideto Kazawa, Keith Stevens, George Kurian, Nishant Patil, Wei Wang, Cliff Young, Jason Smith, Jason Riesa, Alex Rudnick, Oriol Vinyals, Greg Corrado, Macduff Hughes, and Jeffrey Dean. 2016. Google's neural machine translation system: Bridging the gap between human and machine translation. CoRR abs/1609.08144. http://arxiv.org/abs/1609.08144.",
"links": null
},
"BIBREF37": {
"ref_id": "b37",
"title": "Semantic parsing via staged query graph generation: Question answering with knowledge base",
"authors": [
{
"first": "Ming-Wei",
"middle": [],
"last": "Wen-Tau Yih",
"suffix": ""
},
{
"first": "Xiaodong",
"middle": [],
"last": "Chang",
"suffix": ""
},
{
"first": "Jianfeng",
"middle": [],
"last": "He",
"suffix": ""
},
{
"first": "",
"middle": [],
"last": "Gao",
"suffix": ""
}
],
"year": 2015,
"venue": "Association for Computational Linguistics (ACL)",
"volume": "",
"issue": "",
"pages": "",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Wen-tau Yih, Ming-Wei Chang, Xiaodong He, and Jianfeng Gao. 2015. Semantic parsing via staged query graph generation: Question answering with knowledge base. In Association for Computational Linguistics (ACL).",
"links": null
},
"BIBREF38": {
"ref_id": "b38",
"title": "The value of semantic parse labeling for knowledge base question answering",
"authors": [
{
"first": "Matthew",
"middle": [],
"last": "Wen-Tau Yih",
"suffix": ""
},
{
"first": "Chris",
"middle": [],
"last": "Richardson",
"suffix": ""
},
{
"first": "Ming-Wei",
"middle": [],
"last": "Meek",
"suffix": ""
},
{
"first": "Jina",
"middle": [],
"last": "Chang",
"suffix": ""
},
{
"first": "",
"middle": [],
"last": "Suh",
"suffix": ""
}
],
"year": 2016,
"venue": "Association for Computational Linguistics (ACL)",
"volume": "",
"issue": "",
"pages": "",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Wen-tau Yih, Matthew Richardson, Chris Meek, Ming- Wei Chang, and Jina Suh. 2016. The value of se- mantic parse labeling for knowledge base question answering. In Association for Computational Lin- guistics (ACL).",
"links": null
},
"BIBREF39": {
"ref_id": "b39",
"title": "Neural enquirer: Learning to query tables",
"authors": [
{
"first": "Pengcheng",
"middle": [],
"last": "Yin",
"suffix": ""
},
{
"first": "Zhengdong",
"middle": [],
"last": "Lu",
"suffix": ""
},
{
"first": "Hang",
"middle": [],
"last": "Li",
"suffix": ""
},
{
"first": "Ben",
"middle": [],
"last": "Kao",
"suffix": ""
}
],
"year": 2015,
"venue": "",
"volume": "",
"issue": "",
"pages": "",
"other_ids": {
"arXiv": [
"arXiv:1512.00965"
]
},
"num": null,
"urls": [],
"raw_text": "Pengcheng Yin, Zhengdong Lu, Hang Li, and Ben Kao. 2015. Neural enquirer: Learning to query tables. arXiv preprint arXiv:1512.00965 .",
"links": null
},
"BIBREF40": {
"ref_id": "b40",
"title": "Learning to skim text",
"authors": [
{
"first": "Adam",
"middle": [],
"last": "Yu",
"suffix": ""
},
{
"first": "Hongrae",
"middle": [],
"last": "Lee",
"suffix": ""
},
{
"first": "Quoc",
"middle": [],
"last": "Le",
"suffix": ""
}
],
"year": 2017,
"venue": "ACL",
"volume": "",
"issue": "",
"pages": "",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Adam Yu, Hongrae Lee, and Quoc Le. 2017. Learning to skim text. In ACL.",
"links": null
},
"BIBREF41": {
"ref_id": "b41",
"title": "Reinforcement learning neural turing machines",
"authors": [
{
"first": "Wojciech",
"middle": [],
"last": "Zaremba",
"suffix": ""
},
{
"first": "Ilya",
"middle": [],
"last": "Sutskever",
"suffix": ""
}
],
"year": 2015,
"venue": "",
"volume": "",
"issue": "",
"pages": "",
"other_ids": {
"arXiv": [
"arXiv:1505.00521"
]
},
"num": null,
"urls": [],
"raw_text": "Wojciech Zaremba and Ilya Sutskever. 2015. Rein- forcement learning neural turing machines. arXiv preprint arXiv:1505.00521 .",
"links": null
},
"BIBREF42": {
"ref_id": "b42",
"title": "Learning to parse database queries using inductive logic programming",
"authors": [
{
"first": "M",
"middle": [],
"last": "Zelle",
"suffix": ""
},
{
"first": "R",
"middle": [
"J"
],
"last": "Mooney",
"suffix": ""
}
],
"year": 1996,
"venue": "Association for the Advancement of Artificial Intelligence (AAAI)",
"volume": "",
"issue": "",
"pages": "1050--1055",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "M. Zelle and R. J. Mooney. 1996. Learning to parse database queries using inductive logic program- ming. In Association for the Advancement of Arti- ficial Intelligence (AAAI). pages 1050-1055.",
"links": null
},
"BIBREF43": {
"ref_id": "b43",
"title": "Learning to map sentences to logical form: Structured classification with probabilistic categorial grammars",
"authors": [
{
"first": "L",
"middle": [
"S"
],
"last": "Zettlemoyer",
"suffix": ""
},
{
"first": "M",
"middle": [],
"last": "Collins",
"suffix": ""
}
],
"year": 2005,
"venue": "Uncertainty in Artificial Intelligence (UAI)",
"volume": "",
"issue": "",
"pages": "658--666",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "L. S. Zettlemoyer and M. Collins. 2005. Learning to map sentences to logical form: Structured classifica- tion with probabilistic categorial grammars. In Un- certainty in Artificial Intelligence (UAI). pages 658- 666.",
"links": null
}
},
"ref_entries": {
"TABREF0": {
"content": "<table><tr><td>Key</td><td/><td>Variable</td><td/><td/><td>Key</td><td>Variable</td><td/><td/><td/><td/><td>Key</td><td colspan=\"2\">Variable</td><td colspan=\"2\">m.NYC</td></tr><tr><td>v 1</td><td/><td>R1(m.USA)</td><td colspan=\"2\">Execute ( Hop R1 !CityIn )</td><td>v 1</td><td colspan=\"2\">R1(m.USA)</td><td/><td colspan=\"2\">Execute ( Argmax R2 Population )</td><td>...</td><td>...</td><td colspan=\"2\">Execute Return</td></tr><tr><td colspan=\"3\">Entity Resolver</td><td/><td/><td>v 2</td><td colspan=\"3\">R2(list of US cities)</td><td/><td/><td>v 3</td><td colspan=\"2\">R3(m.NYC)</td><td/></tr><tr><td/><td/><td/><td/><td>(</td><td colspan=\"2\">Hop</td><td>R1</td><td>!CityIn</td><td>)</td><td>(</td><td>Argmax</td><td>R2</td><td>Population</td><td>)</td><td>Return</td></tr><tr><td>Largest</td><td>city</td><td>in</td><td>US</td><td>GO</td><td>(</td><td colspan=\"2\">Hop</td><td>R1</td><td>!CityIn</td><td>)</td><td>(</td><td>Argmax</td><td>R2</td><td>Population</td><td>)</td></tr></table>",
"type_str": "table",
"html": null,
"text": "Interpreter functions. r represents a variable, p a property in Freebase. and \uf8ff are defined on numbers and dates.",
"num": null
},
"TABREF1": {
"content": "<table><tr><td>Iterative ML Decode BML programs given x for n = 1 to NML do for (x, y) in D do C for j in 1...|C| do if Rx,y(Cj) > Rx,y(C \u21e4 x ) then C \u21e4 x Cj</td></tr><tr><td>\u2713 Initialize model \u2713 randomly ML training with DML = {(x, C \u21e4 x )} . REINFORCE for n = 1 to NRL do DRL ; is the RL training set for</td></tr></table>",
"type_str": "table",
"html": null,
"text": "Algorithm 1 IML-REINFORCEInput: question-answer pairs D = {(xi, yi)}, mix ratio \u21b5, reward function R(\u2022), training iterations NML, NRL, and beam sizes BML, BRL. Procedure: Initialize C \u21e4 x = ; the best program so far for x Initialize model \u2713 randomly .",
"num": null
},
"TABREF2": {
"content": "<table/>",
"type_str": "table",
"html": null,
"text": "Results on the test set. Average F1 is the main evaluation metric and NSM outperforms STAGG with no domainspecific knowledge or feature engineering.",
"num": null
},
"TABREF4": {
"content": "<table/>",
"type_str": "table",
"html": null,
"text": "Average F1 on the validation set for augmented RE-INFORCE, REINFORCE, and iterative ML.",
"num": null
},
"TABREF5": {
"content": "<table><tr><td>#Expressions</td><td>0</td><td>1</td><td>2</td><td>3</td></tr><tr><td>Percentage</td><td colspan=\"4\">0.4% 62.9% 29.8% 6.9%</td></tr><tr><td>F1</td><td>0.0</td><td>73.5</td><td>59.9</td><td>70.3</td></tr></table>",
"type_str": "table",
"html": null,
"text": "Contributions of different overfitting techniques on the validation set.",
"num": null
},
"TABREF6": {
"content": "<table/>",
"type_str": "table",
"html": null,
"text": "Percentage and performance of model generated programs with different complexity (number of expressions).",
"num": null
}
}
}
} |