Some genetic changes occur too fast for selection to take place.
Can you please define "genetic change"? Selection can only change the frequency of alleles in a population. It cannot make a new mutation. If by, "genetic change", you mean "mutation", then selection just can't do it. If by "genetic change", you mean change in allele frequency, then the change resulting selection is not necessarily much fast than if genetic drift alone is at play.
Often scientists presume that selection has taken place for an observed genetic change. Is this assumption still valid, based on the current knowledge of genetic changes?
No, really not! Whatever you mean by "genetic change", there is no false assumption made here.
For example genetic changes caused by viruses and transposons seems to have the capacity to establish fast in populations.
Again, do you mean "mutation" or "change in allele frequency" by "genetic change"? Assuming you meant "change in allele frequency", can you please cite a peer-reviewed paper as an example but the claim is really not that obvious to me.
Can a genetic change established in a population be disadvantageous?
Yes! Most mutations are deleterious and genetic drift can lead to some drift load. No one has ever assumed that genetic drift could not happen. Even Darwin knew about it!
Can it be advantageous to make a pseudo gene functional again?
Sure! I have no idea how that relate to the rest of the question that being said.