D14-1144 research , we formulate plausible language transfer hypotheses supported by current
E14-4033 inherent challenge is that viable language transfer hypotheses are naturally difficult
D14-1144 weights . This yields two candidate language transfer feature lists per L1 . 4 Results
E14-4019 understand the SLA process and language transfer effects . They can enhance our
E14-4033 Swanson Eugene Charniak Abstract Language transfer , the preferential second language
D14-1144 output by formulating plausible language transfer hypotheses supported by current
E14-4033 that use NLI as a method to form language transfer hypotheses . 3 Methodology The
E14-4033 a fragment fits the profile of language transfer we adopt the expected per feature
E14-4019 data . Motivated by theories of language transfer , NLI is the task of identifying
C04-1005 language parser , source to target language transfer module , and target language
D14-1142 offer insights into two kinds of language transfer effects , namely word choice
E14-4033 F E F is capable of capturing language transfer phenomenon . The second is a
E14-4033 our confidence in a feature as a language transfer hypothesis . Swanson and Charniak
D11-1006 Projected Transfer Unlike most language transfer systems for parsers , the direct
D14-1144 current research . In addition to language transfer hypotheses , such systems could
C02-1163 source language paraphraser and a language transfer cooperates in translation by
D14-1142 some information about localized language transfer effects , since the features
D14-1144 Macquarie </title> NSW Sydney Abstract Language transfer , the characteristic second language
E14-4033 concise ranked list of candidate language transfer hypotheses and their usage statistics
E14-4033 refine C into a ranked list H of language transfer hypotheses , where H has also
hide detail