Hello, I found that in Eq. 5 of your paper mentioned that d should be truncated to $[-\mu, \mu]$. But it seems that it will cause $w_i$ of Eq.6 to always be one and in your code I think there's no truncation in eval_dist. It confuses me a little bit.
Thank you for your question! We found that typo a while ago. We have fixed it in the latest arXiv. Please check it and let me know if you have more questions.
Hello, I found that in Eq. 5 of your paper mentioned that d should be truncated to $[-\mu, \mu]$. But it seems that it will cause $w_i$ of Eq.6 to always be one and in your code I think there's no truncation in
eval_dist
. It confuses me a little bit.