| P15-4009 | templates designed according to | Propbank SRL | labels . Propbank provides semantic |
| P15-2036 | for training ; the first uses | PropBank SRL | as guide features , and the second |
| W15-1012 | is based to a large degree on | PropBank SRL | , improving SRL alignment should |
| W06-1617 | accuracy as compared to that of the | PropBank SRL | task . 6.3 Integrating PropBank |
| D15-1169 | achieve state-of-the-art results on | PropBank SRL | . Our new A * parsing algorithm |
| N09-1017 | Gordon and Swanson ( 2007 ) for | PropBank SRL | and Pad ´ o et al. ( 2008 |
| S15-1027 | 5 , 2015 . mance of an English | PropBank SRL | system by 0.4 F1 points using |
| W06-1617 | previously shown to be effective in | PropBank SRL | are carefully selected and adapted |
| S15-1027 | other predicate type extensions of | PropBank SRL | . As our first attempt at automatically |
| P15-2036 | features ( 2.8 % F1 vs. 0.9 % F1 ) . | PropBank SRL | as guide features offers a small |
| N09-1017 | methodologies and representations used in | PropBank SRL | ( Pradhan et al. , 2005 ) can |
| D11-1116 | Hovy 's ( 2010 ) dataset . The | PropBank SRL | module achieves 89.5 F1 on predicate |
| P08-1063 | VerbNet roles or by using the | PropBank SRL | system and performing a posterior |
| P11-1023 | fine-grained measures . We adopted the | Propbank SRL | style predicate-argument framework |
| P14-1136 | previous-best single-parser systems on | PropBank SRL | . Unlike Das et al. ( 2014 ) |
| J14-1002 | Roth and Yih 2004 ) , as well as | PropBank SRL | ( Punyakanok et al. 2004 ) . |
