Obviously the prerequisite for this would be that backpropogation of error actually occurs in the brain, but what do you think?
Your question uses the language of machine learning to speculate about human memory. That seems pretty thin.
I.e Why should I remember a bad ex girlfriend if there was nothing to learn from it?
Without retention of the memory of trauma, you won’t be forming new experiences.
Amygdala creates emotional responses and utilizes long term memories.
We know from ape studies and amphibian studies that amygdala evolved differently in humans than with other animals.
Isn't it also arguable that retention of trauma is inherently bad from an evolutionary perspective? Let's say someone retains a traumatic event, gets depressed, ends up committing suicide. In that case, if the brain optimized for survival, it would be beneficial to erase the memory.
I'd argue that the probability of death occurring as a result of severe trauma is greater than the probability of death occurring as a result of forgetting an experience
In evolutionary terms, individual organisms are optimized for reproducing and living long enough to give their offspring (more accurately, their genes) a good shot at survival. Richard Dawkins explains that in The Selfish Gene and other books. "Contentment," a human social construct, and classifying a memory as "traumatic" probably have little to do with what evolution has optimized us (or any organism) for. If we reproduce and pass on our genes, then slowly die a miserable and traumatic death, that makes no difference in terms of evolution.
Retention of traumatic memory is actually an evolutionary edge we have..if you don’t retain that memory, we will keep repeating the same fatal mistake again and again.
I would even go further and look into studies re amygdala and how it works with people who have substance abuse or addiction issues.
Suicidal tendencies have to do with brain chemical imbalances and not necessarily with the formation of long term and short term memory.