MIT Press

Table 5:

Previous work on issue-specific stance prediction (stance acc.).

Model	A	E	GM	GC	Avg
BERT (Devlin et al., 2019)	67.0	62.4	67.4	64.6	65.4
PSL (Sridhar et al., 2015b)	77.0	80.3	80.5	69.1	76.7
Struct. Rep. (Li et al., 2018)	86.5	82.2	87.6	83.1	84.9

DRaiLRelNets	89.2	82.4	90.1	83.1	86.2

Model	A	E	GM	GC	Avg
BERT (Devlin et al., 2019)	67.0	62.4	67.4	64.6	65.4
PSL (Sridhar et al., 2015b)	77.0	80.3	80.5	69.1	76.7
Struct. Rep. (Li et al., 2018)	86.5	82.2	87.6	83.1	84.9

DRaiLRelNets	89.2	82.4	90.1	83.1	86.2