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
|