Results in paper
Opened this issue · 1 comments
BeerTai commented
Hi,how were the results in the paper obtained?
The maximum value of the sum of each batch or the maximum value of each dataset?
YoungXiyuan commented
Sorry for my late reply.
First, we select results of the best epoch during each run. And the best epoch means the DCA framework acquires good enough value (might not the maximum value) in every dataset.
Second, we run the DCA framework five times and get the average of their own selected results as the final published results.