Skip to Main Content
Table 2 

Comparisons of decoding results.


iteration
model
pruning
#rules
dev
test
time (s)
Baseline  heuristic Hiero 3.59M 25.5 25.1 809 
No-singleton  heuristic Hiero 1.09M 24.7 24.2 638 
 
Sampled 0th (init) Pitman-Yor scope < 3 212K 19.9 19.1 489 
Sampled 100th Pitman-Yor scope < 3 313K 23.9 23.3 1,214 
Sampled averaged (0 to 70) Pitman-Yor scope < 3 885K 26.2 24.5 1,488 
Sampled averaged (0 to 70) Pitman-Yor Hiero 785K 25.6 25.1 532 
 
Sampled averaged (0 to 70) Dirichlet scope < 3 774K 24.6 23.8 930 

iteration
model
pruning
#rules
dev
test
time (s)
Baseline  heuristic Hiero 3.59M 25.5 25.1 809 
No-singleton  heuristic Hiero 1.09M 24.7 24.2 638 
 
Sampled 0th (init) Pitman-Yor scope < 3 212K 19.9 19.1 489 
Sampled 100th Pitman-Yor scope < 3 313K 23.9 23.3 1,214 
Sampled averaged (0 to 70) Pitman-Yor scope < 3 885K 26.2 24.5 1,488 
Sampled averaged (0 to 70) Pitman-Yor Hiero 785K 25.6 25.1 532 
 
Sampled averaged (0 to 70) Dirichlet scope < 3 774K 24.6 23.8 930 
Close Modal

or Create an Account

Close Modal
Close Modal