| P07-1024 | the corpus using the unlabeled | DLA | . 2 . For each combination of |
| P07-1024 | realistic case of a " la - beled " | DLA | , which is required to have syntactically |
| P07-1024 | ordering . Once we find the optimal | DLA | , two questions can be asked |
| P07-1024 | We begin with an " unlabeled " | DLA | , which simply minimizes dependency |
| P07-1024 | of finding the optimal labeled | DLA | . If we model a DLA as a set |
| P07-1024 | each set as the order in the new | DLA | . In the first step we used the |
| P07-1024 | correctly ordered . optimal labeled | DLA | can found using the following |
| C88-2157 | Dependency Localization Analysis ( | DLA | ) is used . This identifies the |
| N10-1002 | construction type we target for | DLA | in this paper is English Verb |
| P07-1024 | Secondly , how similar is the optimal | DLA | to English in terms of the actual |
| P07-1024 | English to that of this optimal | DLA | ? Secondly , how similar is the |
| P07-1024 | opposed to the " unla - beled " | DLA | presented above . 4 Labeled DLAs |
| N10-1002 | dataset to perform three distinct | DLA | tasks , as detailed in Table |
| P07-1024 | Optimized Labeled DLA While the | DLA | presented above is a good deal |
| P07-1024 | rules . We call this a " labeled " | DLA | , as opposed to the " unla - |
| P07-1024 | the problem down is to model the | DLA | as a set of weights for each |
| N10-1002 | to a deep lexical acquisition ( | DLA | ) task , using a maximum entropy |
| C88-2157 | rules . In the extractkm step , | DLA | is used . This process first |
| P07-1024 | side . Searching over all such | DLAs | would be exponentially expensive |
| P07-1024 | significantly better than a random | DLA | , indicating that dependency |
