The higher order model can't be ensembled, can it? Each trained model would give different antecedents and spans aren't necessarily gonna be the same after pruning. Is there a way to ensemble it?
The ensembling approach from your first paper is not gonna work, is it?
Ensembling is performed for both the span pruning and antecedent decisions. At test time, we first average the mention scores sm(i) over each model before pruning the spans. Given the same pruned spans, each model then computes the antecedent scores sa(i, j) separately, and they are averaged to produce the final scores
The higher order model can't be ensembled, can it? Each trained model would give different antecedents and spans aren't necessarily gonna be the same after pruning. Is there a way to ensemble it? The ensembling approach from your first paper is not gonna work, is it?