Skip to Main Content
Table 6: 

Results of abstractiveness-controlled models with different specified bins on CNN/DM dataset. Our CMDP framework significantly improves the BERTScore and MoverScore (p < 0.003, approx. randomization test) over all the bins. It also significantly improves the bin % for bin 2 and 3 (p < 0.00001, approx. randomization test).

bin 1bin 2bin 3
MethodBERTS.MoverS.Bin %BERTS.MoverS.Bin %BERTS.MoverS.Bin %
ControlSum 38.55 23.56 99.94 39.47 23.23 1.09 37.51 20.58 0.03 
 
PG 35.42 20.00 99.85 34.91 18.67 1.42 32.53 15.87 0.21 
PG+CMDP 41.77 25.96 100.00 40.79 23.54 75.10 34.22 17.71 48.62 
 
D.GPT2 39.02 23.82 99.90 39.58 23.30 1.93 38.15 21.23 0.01 
D.GPT2+CMDP 43.23 26.65 99.56 44.07 26.39 62.60 42.03 24.71 1.94 
 
D.GPT2+MDP 42.56 26.43 99.77 43.59 26.23 55.67 41.44 24.42 2.09 
bin 1bin 2bin 3
MethodBERTS.MoverS.Bin %BERTS.MoverS.Bin %BERTS.MoverS.Bin %
ControlSum 38.55 23.56 99.94 39.47 23.23 1.09 37.51 20.58 0.03 
 
PG 35.42 20.00 99.85 34.91 18.67 1.42 32.53 15.87 0.21 
PG+CMDP 41.77 25.96 100.00 40.79 23.54 75.10 34.22 17.71 48.62 
 
D.GPT2 39.02 23.82 99.90 39.58 23.30 1.93 38.15 21.23 0.01 
D.GPT2+CMDP 43.23 26.65 99.56 44.07 26.39 62.60 42.03 24.71 1.94 
 
D.GPT2+MDP 42.56 26.43 99.77 43.59 26.23 55.67 41.44 24.42 2.09 
Close Modal

or Create an Account

Close Modal
Close Modal