| W13-3509 | approach was based on a naive | hard matching | between word lemmas . Below we |
| W09-0204 | In addi - tion , the problem of | hard matching | can be alleviated by processing |
| S13-1022 | method as the NLPM models . Here , | hard matching | was performed , where matching |
| P09-1106 | Another reason is that TER uses | hard matching | in computing edit distance . |
| C04-1078 | that is not matched by any of the | hard matching | pattern rules . Only those fields |
| C04-1078 | bootstrapping framework in which soft and | hard matching | pattern rules are combined in |
| D15-1210 | refer to its loss function as | hard matching | : AHM ( a1 , a2 ) = 1 − |
| P07-1026 | Tree Path feature is sensitive to | hard matching | between any two sub-trees without |
| P07-1026 | However , it only carries out | hard matching | , which may lead to over-fitting |
| C04-1078 | variations . The reason is that | hard matching | techniques result in relatively |
| D15-1210 | developing a soft version of the | hard matching | loss function because this will |
| P03-2040 | alignment , if we do more than | hard matching | of punctuations and take into |
| D15-1210 | BERKELEY BLEU . " HM " denotes the | hard matching | loss function , " SM " denotes |
| E95-1010 | of byte-length ratio measures , | hard matching | of numbers , string comparisons |
| P07-1026 | convolution tree kernel only carries out | hard matching | , so it fails to handle similar |
| C04-1078 | induction techniques based on | hard matching | ( i.e. , strict slot-by-slot |
| P06-2074 | Xiao et al. ( 2004 ) stated that | hard matching | techniques tend to have low recall |
| C04-1078 | <title> Cascading Use of Soft and | Hard Matching | Pattern Rules for Weakly Supervised |
