858:, the speech signal and the corresponding textual representation can be mapped to each other in blocks in order. This is not always the case with the same text in two languages. For SMT, the machine translator can only manage small sequences of words, and word order has to be thought of by the program designer. Attempts at solutions have included re-ordering models, where a distribution of location changes for each item of translation is guessed from aligned bi-text. Different location changes can be ranked with the help of the language model and the best can be selected.
735:
850:
Word order in languages differ. Some classification can be done by naming the typical order of subject (S), verb (V) and object (O) in a sentence and one can talk, for instance, of SVO or VSO languages. There are also additional differences in word orders, for instance, where modifiers for nouns are
841:
This problem is connected with word alignment, as in very specific contexts the idiomatic expression aligned with words that resulted in an idiomatic expression of the same meaning in the target language. However, it is unlikely, as the alignment usually doesn't work in any other contexts. For that
801:
Function words that have no clear equivalent in the target language were another challenge for the statistical models. For example, when translating from
English to German, the sentence "John does not live here," the word "does" doesn't have a clear alignment in the translated sentence "John wohnt
866:
SMT systems typically store different word forms as separate symbols without any relation to each other and word forms or phrases that were not in the training data cannot be translated. This might be because of the lack of training data, changes in the human domain where the system is used, or
773:
In parallel corpora single sentences in one language can be found translated into several sentences in the other and vice versa. Long sentences may be broken up, short sentences may be merged. There are even some languages that use writing systems without clear indication of a sentence end (for
30:
approach, that superseded the previous, rule-based approach because it required explicit description of each and every linguistic rule, which was costly, and which often did not generalize to other languages. Since 2003, the statistical approach itself has been gradually superseded by the
830:
bilingual corpus primarily consists of parliamentary speech examples, where "Hear, Hear!" is frequently associated with "Bravo!" Using a model built on this corpus to translate ordinary speech in a conversational register would lead to incorrect translation of the word
656:
In phrase-based translation, the aim was to reduce the restrictions of word-based translation by translating whole sequences of words, where the lengths may differ. The sequences of words were called blocks or phrases, however, typically they were not linguistic
802:
hier nicht." Through logical reasoning, it may be aligned with the words "wohnt" (as in
English it contains grammatical information for the word "live") or "nicht" (as it only appears in the sentence because it is negated) or it may be unaligned.
672:
The chosen phrases were further mapped one-to-one based on a phrase translation table, and could be reordered. This table could be learnt based on word-alignment, or directly from a parallel corpus. The second model was trained using the
721:
rules, but the grammars could be constructed by an extension of methods for phrase-based translation without reference to linguistically motivated syntactic constituents. This idea was first introduced in Chiang's Hiero system (2005).
523:
567:
As the translation systems were not able to store all native strings and their translations, a document was typically translated sentence by sentence, but even this was not enough. Language models were typically approximated by
647:
The benefits obtained for translation between
Western European languages are not representative of results for other language pairs, owing to smaller training corpora and greater grammatical differences.
560:
that uses the foreign string, heuristics and other methods to limit the search space and at the same time keeping acceptable quality. This trade-off between quality and time usage can also be found in
810:
An example of such an anomaly was that "I took the train to Berlin" was mis-translated as "I took the train to Paris" due to the statistical abundance of "train to Paris" in the training set.
842:
reason, idioms could only be subjected to phrasal alignment, as they could not be decomposed further without losing their meaning. This problem was specific for word-based translation.
271:
1136:
Proceedings of the Joint
Conference on Human Language Technologies and the Annual Meeting of the North American Chapter of the Association of Computational Linguistics (HLT/NAACL)
665:
that were found using statistical methods from corpora. It has been shown that restricting the phrases to linguistic phrases (syntactically motivated groups of words, see
369:
576:, and similar approaches have been applied to translation models, but there was additional complexity due to different sentence lengths and word orders in the languages.
1320:
342:
is the probability of seeing that target language string. This decomposition is attractive as it splits the problem into two subproblems. Finding the best translation
308:
195:
115:
554:
340:
155:
135:
377:
1409:
1569:
698:
2163:
1313:
1547:
1306:
1958:
1402:
778:. Through this and other mathematical models efficient search and retrieval of the highest scoring sentence alignment is possible.
2127:
701:, the statistical counterpart of the old idea of syntax-based translation did not take off. Examples of this approach included
674:
1275:
1251:
1221:
569:
1868:
1559:
1395:
1341:
1241:
2122:
790:. To learn e.g. the translation model, however, we need to know which words align in a source-target sentence pair. The
1729:
1351:
746:
197:
has been approached in a number of ways. One approach which lends itself well to computer implementation is to apply
1883:
1714:
1361:
897:
717:
Hierarchical phrase-based translation combined the phrase-based and syntax-based approaches to translation. It used
1654:
1366:
718:
706:
644:
Statistical machine translation usually works less well for language pairs with significantly different word order.
590:
from
Stephan Vogel and Model 6 from Franz-Joseph Och), but significant advances were made with the introduction of
63:
1088:. In COLING ’96: The 16th International Conference on Computational Linguistics, pp. 836-841, Copenhagen, Denmark.
826:
might not receive a translation that accurately represents the original intent. For example, the popular
Canadian
529:
For a rigorous implementation of this one would have to perform an exhaustive search by going through all strings
2071:
1008:
965:
787:
775:
1719:
1464:
1346:
923:
795:
204:
1988:
1709:
2158:
1681:
606:
The most frequently cited benefits of statistical machine translation (SMT) over rule-based approach were:
2026:
2011:
1983:
1848:
1843:
1418:
1020:
973:
917:
823:
58:. Statistical machine translation was re-introduced in the late 1980s and early 1990s by researchers at
1763:
1734:
1512:
1376:
1371:
907:
36:
1606:
1459:
2132:
2056:
1788:
1744:
1629:
1527:
78:
345:
2036:
2006:
1673:
1507:
1156:
Proceedings of the 43rd Annual
Meeting of the Association for Computational Linguistics (ACL'05)
1893:
1586:
1564:
1554:
1522:
1497:
1753:
702:
693:
units, rather than single words or strings of words (as in phrase-based MT), i.e. (partial)
276:
163:
83:
2106:
1782:
1758:
1611:
912:
882:
791:
690:
678:
666:
532:
310:
is the probability that the source string is the translation of the target string, and the
786:
Sentence alignment is usually either provided by the corpus or obtained by aforementioned
316:
8:
2086:
2016:
1973:
1929:
1701:
1691:
1686:
1574:
1330:
587:
27:
518:{\displaystyle {\tilde {e}}=arg\max _{e\in e^{*}}p(e|f)=arg\max _{e\in e^{*}}p(f|e)p(e)}
2096:
1968:
1833:
1596:
1579:
1437:
855:
561:
140:
120:
74:
55:
1169:
2101:
1813:
1621:
1532:
1271:
1247:
1217:
1012:
969:
1298:
1978:
1863:
1838:
1639:
1542:
1292:— Includes links to freely available statistical machine translation software
1194:
1151:
1131:
1111:
1016:
902:
614:
1239:
2090:
2051:
2046:
1914:
1644:
1517:
1492:
1474:
1289:
928:
892:
1006:
963:
1798:
1778:
1502:
1116:
1099:
1053:
311:
51:
1387:
557:
2152:
2061:
1873:
1853:
1634:
198:
137:
in the target language (for example, English) is the translation of a string
47:
32:
2041:
1659:
1055:
977:
620:
Generally, SMT systems were not tailored to any specific pair of languages.
556:
in the native language. Performing the search efficiently is the work of a
1240:
Philip
Williams; Rico Sennrich; Matt Post; Philipp Koehn (1 August 2016).
1056:"The mathematics of statistical machine translation: parameter estimation"
1024:
1998:
1878:
1591:
1484:
1432:
1085:
851:
located, or where the same words are used as a question or a statement.
734:
1601:
694:
43:
765:
Problems that statistical machine translation did not solve included:
1469:
1152:
A Hierarchical Phrase-Based Model for
Statistical Machine Translation
641:
Results may have superficial fluency that masks translation problems.
1944:
1924:
1909:
1888:
1858:
1803:
1768:
1649:
1290:
Annotated list of statistical natural language processing resources
887:
662:
725:
371:
is done by picking up the one that gives the highest probability:
2081:
1939:
1919:
1793:
1537:
1452:
1100:"A Systematic Comparison of Various Statistical Alignment Models"
827:
697:
of sentences/utterances. Until the 1990s, with advent of strong
1447:
1442:
877:
774:
example, Thai). Sentence aligning can be performed through the
658:
595:
591:
1054:
P. Brown; S. Della Pietra; V. Della Pietra; R. Mercer (1993).
689:
Syntax-based translation was based on the idea of translating
2137:
1773:
819:
580:
73:
The idea behind statistical machine translation comes from
617:
in machine-readable format and even more monolingual data.
1934:
1170:"Has AI surpassed humans at translation? Not even close!"
625:
More fluent translations owing to use of a language model
584:
59:
712:
1235:
1233:
1328:
535:
380:
348:
319:
279:
207:
166:
160:
The problem of modeling the probability distribution
143:
123:
86:
1230:
1086:
HMM-based Word
Alignment in Statistical Translation
988:. Association for Computational Linguistics: 71–76
579:The statistical translation models were initially
548:
517:
363:
334:
302:
265:
189:
149:
129:
109:
2150:
1620:
978:"A statistical approach to language translation"
861:
461:
406:
1417:
1025:"A statistical approach to machine translation"
726:Challenges with statistical machine translation
610:More efficient use of human and data resources
1193:Philipp Koehn, Franz Josef Och, Daniel Marcu:
157:in the source language (for example, French).
1403:
1314:
1243:Syntax-based Statistical Machine Translation
1146:
1144:
638:Specific errors are hard to predict and fix.
77:. A document is translated according to the
1091:
684:
651:
1410:
1396:
1321:
1307:
950:W. Weaver (1955). Translation (1949). In:
818:Depending on the corpora used, the use of
1141:
1130:P. Koehn, F.J. Och, and D. Marcu (2003).
1115:
798:were attempts at solving this challenge.
50:in 1949, including the ideas of applying
1084:S. Vogel, H. Ney and C. Tillmann. 1996.
845:
805:
669:) decreased the quality of translation.
266:{\displaystyle p(e|f)\propto p(f|e)p(e)}
2164:Statistical natural language processing
1098:Och, Franz Josef; Ney, Hermann (2003).
938:
594:based models. Later work incorporated
46:machine translation were introduced by
2151:
1268:An Introduction to Machine Translation
1266:W. J. Hutchins and H. Somers. (1992).
1097:
1047:
1000:
957:
1391:
1302:
1211:
1207:
1205:
1203:
813:
768:
713:Hierarchical phrase-based translation
1869:Simple Knowledge Organization System
1246:. Morgan & Claypool Publishers.
1195:Statistical Phrase-Based Translation
1167:
1132:Statistical phrase based translation
1011:; S. Della Pietra; V. Della Pietra;
968:; S. Della Pietra; V. Della Pietra;
729:
13:
1200:
675:expectation maximization algorithm
14:
2175:
1884:Thesaurus (information retrieval)
1283:
898:Example-based machine translation
781:
707:synchronous context-free grammars
952:Machine Translation of Languages
733:
719:synchronous context-free grammar
64:Thomas J. Watson Research Center
1260:
1214:Statistical Machine Translation
1187:
788:Gale-Church alignment algorithm
776:Gale-Church alignment algorithm
629:
598:or quasi-syntactic structures.
20:Statistical machine translation
1465:Natural language understanding
1216:. Cambridge University Press.
1168:Zhou, Sharon (July 25, 2018).
1161:
1124:
1078:
944:
924:Rule-based machine translation
677:, similarly to the word-based
635:Corpus creation can be costly.
512:
506:
500:
493:
486:
445:
438:
431:
387:
355:
329:
323:
297:
290:
283:
273:, where the translation model
260:
254:
248:
241:
234:
225:
218:
211:
184:
177:
170:
104:
97:
90:
1:
1989:Optical character recognition
862:Out of vocabulary (OOV) words
1682:Multi-document summarization
364:{\displaystyle {\tilde {e}}}
16:Machine translation paradigm
7:
2012:Latent Dirichlet allocation
1984:Natural language generation
1849:Machine-readable dictionary
1844:Linguistic Linked Open Data
1419:Natural language processing
954:, MIT Press, Cambridge, MA.
918:Moses (machine translation)
870:
867:differences in morphology.
601:
558:machine translation decoder
10:
2180:
1764:Explicit semantic analysis
1513:Deep linguistic processing
1117:10.1162/089120103321337421
908:Hybrid machine translation
37:neural machine translation
2115:
2070:
2025:
1997:
1957:
1902:
1824:
1812:
1743:
1700:
1672:
1607:Word-sense disambiguation
1483:
1460:Computational linguistics
1425:
1337:
1104:Computational Linguistics
1060:Computational Linguistics
1029:Computational Linguistics
2133:Natural Language Toolkit
2057:Pronunciation assessment
1959:Automatic identification
1789:Latent semantic analysis
1745:Distributional semantics
1630:Compound-term processing
1528:Named-entity recognition
685:Syntax-based translation
652:Phrase-based translation
79:probability distribution
68:
2037:Automated essay scoring
2007:Document classification
1674:Automatic summarization
1212:Koehn, Philipp (2010).
1066:(2). MIT Press: 263–311
583:based (Models 1-5 from
1894:Universal Dependencies
1587:Terminology extraction
1570:Semantic decomposition
1565:Semantic role labeling
1555:Part-of-speech tagging
1523:Information extraction
1508:Coreference resolution
1498:Collocation extraction
550:
519:
365:
336:
304:
303:{\displaystyle p(f|e)}
267:
191:
190:{\displaystyle p(e|f)}
151:
131:
111:
110:{\displaystyle p(e|f)}
1655:Sentence segmentation
1035:(2). MIT Press: 79–85
1023:; P. Roossin (1990).
976:; P. Roossin (1988).
846:Different word orders
806:Statistical anomalies
551:
549:{\displaystyle e^{*}}
520:
366:
337:
305:
268:
192:
152:
132:
112:
2107:Voice user interface
1818:datasets and corpora
1759:Document-term matrix
1612:Word-sense induction
939:Notes and references
913:Microsoft Translator
883:Cache language model
705:-based MT and later
667:syntactic categories
533:
378:
346:
335:{\displaystyle p(e)}
317:
277:
205:
164:
141:
121:
84:
2159:Machine translation
2087:Interactive fiction
2017:Pachinko allocation
1974:Speech segmentation
1930:Google Ngram Viewer
1702:Machine translation
1692:Text simplification
1687:Sentence extraction
1575:Semantic similarity
1331:machine translation
933:Statistical parsing
929:SDL Language Weaver
824:linguistic register
588:Hidden Markov model
42:The first ideas of
28:machine translation
2097:Question answering
1969:Speech recognition
1834:Corpus linguistics
1814:Language resources
1597:Textual entailment
1580:Sentiment analysis
1150:D. Chiang (2005).
856:speech recognition
814:Idiom and register
769:Sentence alignment
745:. You can help by
699:stochastic parsers
562:speech recognition
546:
515:
482:
427:
361:
332:
300:
263:
187:
147:
127:
107:
75:information theory
56:information theory
2146:
2145:
2102:Virtual assistant
2027:Computer-assisted
1953:
1952:
1710:Computer-assisted
1668:
1667:
1660:Word segmentation
1622:Text segmentation
1560:Semantic analysis
1548:Syntactic parsing
1533:Ontology learning
1385:
1384:
1276:978-0-12-362830-5
1253:978-1-62705-502-4
1223:978-0-521-87415-1
1013:Frederick Jelinek
970:Frederick Jelinek
763:
762:
460:
405:
390:
358:
150:{\displaystyle f}
130:{\displaystyle e}
2171:
2123:Formal semantics
2072:Natural language
1979:Speech synthesis
1961:and data capture
1864:Semantic network
1839:Lexical resource
1822:
1821:
1640:Lexical analysis
1618:
1617:
1543:Semantic parsing
1412:
1405:
1398:
1389:
1388:
1342:Dictionary-based
1323:
1316:
1309:
1300:
1299:
1278:
1264:
1258:
1257:
1237:
1228:
1227:
1209:
1198:
1191:
1185:
1184:
1182:
1180:
1165:
1159:
1148:
1139:
1128:
1122:
1121:
1119:
1095:
1089:
1082:
1076:
1075:
1073:
1071:
1051:
1045:
1044:
1042:
1040:
1021:Robert L. Mercer
1017:John D. Lafferty
1004:
998:
997:
995:
993:
974:Robert L. Mercer
961:
955:
948:
903:Google Translate
758:
755:
737:
730:
615:parallel corpora
613:There were many
555:
553:
552:
547:
545:
544:
524:
522:
521:
516:
496:
481:
480:
479:
441:
426:
425:
424:
392:
391:
383:
370:
368:
367:
362:
360:
359:
351:
341:
339:
338:
333:
309:
307:
306:
301:
293:
272:
270:
269:
264:
244:
221:
196:
194:
193:
188:
180:
156:
154:
153:
148:
136:
134:
133:
128:
116:
114:
113:
108:
100:
2179:
2178:
2174:
2173:
2172:
2170:
2169:
2168:
2149:
2148:
2147:
2142:
2111:
2091:Syntax guessing
2073:
2066:
2052:Predictive text
2047:Grammar checker
2028:
2021:
1993:
1960:
1949:
1915:Bank of English
1898:
1826:
1817:
1808:
1739:
1696:
1664:
1616:
1518:Distant reading
1493:Argument mining
1479:
1475:Text processing
1421:
1416:
1386:
1381:
1333:
1327:
1296:
1286:
1281:
1265:
1261:
1254:
1238:
1231:
1224:
1210:
1201:
1192:
1188:
1178:
1176:
1166:
1162:
1149:
1142:
1129:
1125:
1096:
1092:
1083:
1079:
1069:
1067:
1052:
1048:
1038:
1036:
1005:
1001:
991:
989:
962:
958:
949:
945:
941:
936:
920:, free software
893:Europarl corpus
873:
864:
848:
816:
808:
784:
771:
759:
753:
750:
743:needs expansion
728:
715:
687:
654:
632:
604:
540:
536:
534:
531:
530:
492:
475:
471:
464:
437:
420:
416:
409:
382:
381:
379:
376:
375:
350:
349:
347:
344:
343:
318:
315:
314:
289:
278:
275:
274:
240:
217:
206:
203:
202:
176:
165:
162:
161:
142:
139:
138:
122:
119:
118:
96:
85:
82:
81:
71:
17:
12:
11:
5:
2177:
2167:
2166:
2161:
2144:
2143:
2141:
2140:
2135:
2130:
2125:
2119:
2117:
2113:
2112:
2110:
2109:
2104:
2099:
2094:
2084:
2078:
2076:
2074:user interface
2068:
2067:
2065:
2064:
2059:
2054:
2049:
2044:
2039:
2033:
2031:
2023:
2022:
2020:
2019:
2014:
2009:
2003:
2001:
1995:
1994:
1992:
1991:
1986:
1981:
1976:
1971:
1965:
1963:
1955:
1954:
1951:
1950:
1948:
1947:
1942:
1937:
1932:
1927:
1922:
1917:
1912:
1906:
1904:
1900:
1899:
1897:
1896:
1891:
1886:
1881:
1876:
1871:
1866:
1861:
1856:
1851:
1846:
1841:
1836:
1830:
1828:
1819:
1810:
1809:
1807:
1806:
1801:
1799:Word embedding
1796:
1791:
1786:
1779:Language model
1776:
1771:
1766:
1761:
1756:
1750:
1748:
1741:
1740:
1738:
1737:
1732:
1730:Transfer-based
1727:
1722:
1717:
1712:
1706:
1704:
1698:
1697:
1695:
1694:
1689:
1684:
1678:
1676:
1670:
1669:
1666:
1665:
1663:
1662:
1657:
1652:
1647:
1642:
1637:
1632:
1626:
1624:
1615:
1614:
1609:
1604:
1599:
1594:
1589:
1583:
1582:
1577:
1572:
1567:
1562:
1557:
1552:
1551:
1550:
1545:
1535:
1530:
1525:
1520:
1515:
1510:
1505:
1503:Concept mining
1500:
1495:
1489:
1487:
1481:
1480:
1478:
1477:
1472:
1467:
1462:
1457:
1456:
1455:
1450:
1440:
1435:
1429:
1427:
1423:
1422:
1415:
1414:
1407:
1400:
1392:
1383:
1382:
1380:
1379:
1374:
1369:
1364:
1359:
1354:
1352:Transfer-based
1349:
1344:
1338:
1335:
1334:
1329:Approaches to
1326:
1325:
1318:
1311:
1303:
1294:
1293:
1285:
1284:External links
1282:
1280:
1279:
1259:
1252:
1229:
1222:
1199:
1186:
1160:
1140:
1123:
1090:
1077:
1046:
999:
956:
942:
940:
937:
935:
934:
931:
926:
921:
915:
910:
905:
900:
895:
890:
885:
880:
874:
872:
869:
863:
860:
847:
844:
815:
812:
807:
804:
783:
782:Word alignment
780:
770:
767:
761:
760:
740:
738:
727:
724:
714:
711:
686:
683:
653:
650:
649:
648:
645:
642:
639:
636:
631:
628:
627:
626:
623:
622:
621:
618:
603:
600:
543:
539:
527:
526:
514:
511:
508:
505:
502:
499:
495:
491:
488:
485:
478:
474:
470:
467:
463:
459:
456:
453:
450:
447:
444:
440:
436:
433:
430:
423:
419:
415:
412:
408:
404:
401:
398:
395:
389:
386:
357:
354:
331:
328:
325:
322:
312:language model
299:
296:
292:
288:
285:
282:
262:
259:
256:
253:
250:
247:
243:
239:
236:
233:
230:
227:
224:
220:
216:
213:
210:
186:
183:
179:
175:
172:
169:
146:
126:
117:that a string
106:
103:
99:
95:
92:
89:
70:
67:
52:Claude Shannon
15:
9:
6:
4:
3:
2:
2176:
2165:
2162:
2160:
2157:
2156:
2154:
2139:
2136:
2134:
2131:
2129:
2128:Hallucination
2126:
2124:
2121:
2120:
2118:
2114:
2108:
2105:
2103:
2100:
2098:
2095:
2092:
2088:
2085:
2083:
2080:
2079:
2077:
2075:
2069:
2063:
2062:Spell checker
2060:
2058:
2055:
2053:
2050:
2048:
2045:
2043:
2040:
2038:
2035:
2034:
2032:
2030:
2024:
2018:
2015:
2013:
2010:
2008:
2005:
2004:
2002:
2000:
1996:
1990:
1987:
1985:
1982:
1980:
1977:
1975:
1972:
1970:
1967:
1966:
1964:
1962:
1956:
1946:
1943:
1941:
1938:
1936:
1933:
1931:
1928:
1926:
1923:
1921:
1918:
1916:
1913:
1911:
1908:
1907:
1905:
1901:
1895:
1892:
1890:
1887:
1885:
1882:
1880:
1877:
1875:
1874:Speech corpus
1872:
1870:
1867:
1865:
1862:
1860:
1857:
1855:
1854:Parallel text
1852:
1850:
1847:
1845:
1842:
1840:
1837:
1835:
1832:
1831:
1829:
1823:
1820:
1815:
1811:
1805:
1802:
1800:
1797:
1795:
1792:
1790:
1787:
1784:
1780:
1777:
1775:
1772:
1770:
1767:
1765:
1762:
1760:
1757:
1755:
1752:
1751:
1749:
1746:
1742:
1736:
1733:
1731:
1728:
1726:
1723:
1721:
1718:
1716:
1715:Example-based
1713:
1711:
1708:
1707:
1705:
1703:
1699:
1693:
1690:
1688:
1685:
1683:
1680:
1679:
1677:
1675:
1671:
1661:
1658:
1656:
1653:
1651:
1648:
1646:
1645:Text chunking
1643:
1641:
1638:
1636:
1635:Lemmatisation
1633:
1631:
1628:
1627:
1625:
1623:
1619:
1613:
1610:
1608:
1605:
1603:
1600:
1598:
1595:
1593:
1590:
1588:
1585:
1584:
1581:
1578:
1576:
1573:
1571:
1568:
1566:
1563:
1561:
1558:
1556:
1553:
1549:
1546:
1544:
1541:
1540:
1539:
1536:
1534:
1531:
1529:
1526:
1524:
1521:
1519:
1516:
1514:
1511:
1509:
1506:
1504:
1501:
1499:
1496:
1494:
1491:
1490:
1488:
1486:
1485:Text analysis
1482:
1476:
1473:
1471:
1468:
1466:
1463:
1461:
1458:
1454:
1451:
1449:
1446:
1445:
1444:
1441:
1439:
1436:
1434:
1431:
1430:
1428:
1426:General terms
1424:
1420:
1413:
1408:
1406:
1401:
1399:
1394:
1393:
1390:
1378:
1375:
1373:
1370:
1368:
1365:
1363:
1362:Example-based
1360:
1358:
1355:
1353:
1350:
1348:
1345:
1343:
1340:
1339:
1336:
1332:
1324:
1319:
1317:
1312:
1310:
1305:
1304:
1301:
1297:
1291:
1288:
1287:
1277:
1273:
1269:
1263:
1255:
1249:
1245:
1244:
1236:
1234:
1225:
1219:
1215:
1208:
1206:
1204:
1196:
1190:
1175:
1171:
1164:
1157:
1153:
1147:
1145:
1137:
1133:
1127:
1118:
1113:
1109:
1105:
1101:
1094:
1087:
1081:
1065:
1061:
1057:
1050:
1034:
1030:
1026:
1022:
1018:
1014:
1010:
1003:
987:
983:
979:
975:
971:
967:
960:
953:
947:
943:
932:
930:
927:
925:
922:
919:
916:
914:
911:
909:
906:
904:
901:
899:
896:
894:
891:
889:
886:
884:
881:
879:
876:
875:
868:
859:
857:
852:
843:
839:
838:
834:
829:
825:
821:
811:
803:
799:
797:
793:
789:
779:
777:
766:
757:
748:
744:
741:This section
739:
736:
732:
731:
723:
720:
710:
708:
704:
700:
696:
692:
682:
680:
676:
670:
668:
664:
660:
646:
643:
640:
637:
634:
633:
624:
619:
616:
612:
611:
609:
608:
607:
599:
597:
593:
589:
586:
582:
577:
575:
573:
565:
563:
559:
541:
537:
509:
503:
497:
489:
483:
476:
472:
468:
465:
457:
454:
451:
448:
442:
434:
428:
421:
417:
413:
410:
402:
399:
396:
393:
384:
374:
373:
372:
352:
326:
320:
313:
294:
286:
280:
257:
251:
245:
237:
231:
228:
222:
214:
208:
200:
199:Bayes Theorem
181:
173:
167:
158:
144:
124:
101:
93:
87:
80:
76:
66:
65:
61:
57:
53:
49:
48:Warren Weaver
45:
40:
38:
34:
33:deep learning
29:
25:
21:
2042:Concordancer
1724:
1438:Bag-of-words
1367:Interlingual
1356:
1295:
1270:, 18.3:322.
1267:
1262:
1242:
1213:
1189:
1177:. Retrieved
1174:Skynet Today
1173:
1163:
1155:
1135:
1126:
1107:
1103:
1093:
1080:
1068:. Retrieved
1063:
1059:
1049:
1037:. Retrieved
1032:
1028:
1002:
990:. Retrieved
985:
981:
959:
951:
946:
865:
853:
849:
840:
836:
832:
817:
809:
800:
796:HMM-approach
785:
772:
764:
751:
747:adding to it
742:
716:
688:
671:
655:
630:Shortcomings
605:
578:
574:-gram models
571:
566:
528:
159:
72:
41:
23:
19:
18:
1999:Topic model
1879:Text corpus
1725:Statistical
1592:Text mining
1433:AI-complete
1357:Statistical
695:parse trees
44:statistical
2153:Categories
1720:Rule-based
1602:Truecasing
1470:Stop words
1347:Rule-based
1009:John Cocke
1007:P. Brown;
966:John Cocke
964:P. Brown;
792:IBM-Models
201:, that is
2029:reviewing
1827:standards
1825:Types and
1110:: 19–51.
982:Coling'88
691:syntactic
679:IBM model
663:phrasemes
570:smoothed
542:∗
477:∗
469:∈
422:∗
414:∈
388:~
356:~
229:∝
1945:Wikidata
1925:FrameNet
1910:BabelNet
1889:Treebank
1859:PropBank
1804:Word2vec
1769:fastText
1650:Stemming
1179:2 August
1070:22 March
1039:22 March
992:22 March
888:Duolingo
871:See also
754:May 2012
602:Benefits
26:) was a
2116:Related
2082:Chatbot
1940:WordNet
1920:DBpedia
1794:Seq2seq
1538:Parsing
1453:Trigram
828:Hansard
794:or the
659:phrases
35:-based
2089:(c.f.
1747:models
1735:Neural
1448:Bigram
1443:n-gram
1377:Hybrid
1372:Neural
1274:
1250:
1220:
1197:(2003)
878:AppTek
837:Bravo!
661:, but
596:syntax
592:phrase
2138:spaCy
1783:large
1774:GloVe
1154:. In
1134:. In
820:idiom
69:Basis
1903:Data
1754:BERT
1272:ISBN
1248:ISBN
1218:ISBN
1181:2018
1072:2015
1041:2015
994:2015
833:hear
822:and
581:word
1935:UBY
1112:doi
854:In
835:as
749:.
703:DOP
585:IBM
462:max
407:max
62:'s
60:IBM
54:'s
24:SMT
2155::
1232:^
1202:^
1172:.
1143:^
1108:29
1106:.
1102:.
1064:19
1062:.
1058:.
1033:16
1031:.
1027:.
1019:;
1015:;
984:.
980:.
972:;
709:.
681:.
564:.
39:.
2093:)
1816:,
1785:)
1781:(
1411:e
1404:t
1397:v
1322:e
1315:t
1308:v
1256:.
1226:.
1183:.
1158:.
1138:.
1120:.
1114::
1074:.
1043:.
996:.
986:1
756:)
752:(
572:n
538:e
525:.
513:)
510:e
507:(
504:p
501:)
498:e
494:|
490:f
487:(
484:p
473:e
466:e
458:g
455:r
452:a
449:=
446:)
443:f
439:|
435:e
432:(
429:p
418:e
411:e
403:g
400:r
397:a
394:=
385:e
353:e
330:)
327:e
324:(
321:p
298:)
295:e
291:|
287:f
284:(
281:p
261:)
258:e
255:(
252:p
249:)
246:e
242:|
238:f
235:(
232:p
226:)
223:f
219:|
215:e
212:(
209:p
185:)
182:f
178:|
174:e
171:(
168:p
145:f
125:e
105:)
102:f
98:|
94:e
91:(
88:p
22:(
Text is available under the Creative Commons Attribution-ShareAlike License. Additional terms may apply.