We highlight the importance of transfer learning as a solution to mitigate this issue. Our methodology involves pretraining a classification model on data from alternative sources, such as Synthetic ...