SOUGATA MUKHERJEE

Hi,

In Section II of the paper (https://ieeexplore.ieee.org/stamp/stamp.jsp?arnumber=10479171), it is mentioned that the SYSPIN corpus contains several types of errors.

Could you please elaborate on the data preparation pipeline used here? Specifically, I would like to understand what measures or quality control steps were taken to detect and mitigate such errors in the SYSPIN corpus to ensure better generation quality.

Thanks!