The cases we have discussed so far talk about improvements in target tasks based on knowledge transfer from the source task. There are cases when transfer learning can lead to a drop in performance. Negative transfer refers to scenarios where the transfer of knowledge from the source to the target does not lead to any improvements, but rather causes a drop in the overall performance of the target task. There can be various reasons for negative transfer, such as cases when the source task is not sufficiently related to the target task or if the transfer method could not leverage the relationship between the source and target tasks very well. Avoiding negative transfer is very important and requires careful investigation. In their work, Rosenstien and their co-authors present empirically how brute-force transfer degrades performance in target tasks when the source and target are too dissimilar. Bayesian approaches by Bakker and their co-authors, along with other techniques exploring clustering-based solutions to identify relatedness, are being researched to avoid negative transfers.