It depends on the conference or workshop, of course. In my field (Computer Science), it is uncommon but not unheard-of for workshops to have what is called "post-proceedings". That is, authors submit an initial version of their manuscript, present at the workshop, and then have a chance to revise for the actual published paper. In this case it is possible, maybe expected, that you actually improve your paper based on the comments you received (although usually no additional review happens at this point).
However, most venues don't have that. The more usual case is that you submit a paper and the proceedings are already published when the conference starts. In this case, you obviously have no chance to revise your paper, because by the time the conference happens the paper is already available on archival sources. The advantage, of course, is that the paper is available for people to read during, or even before, the conference.
Aren't workshops/conference mechanisms to get feedback?
For conferences in CS, this is most certainly not what conferences are for. In this field, conferences are a means to communicate the result of finished studies, not to get input on how to do them. Sure, you can get ideas for future work, but it will pretty much by definition be too late to change what you have already done. For workshops, things are a bit muddier, but in practice authors usually use the feedback they receive at workshops to improve the larger study that they are usually working on (which will be submitted to a conference or journal), not so much to improve the workshop paper itself.