I guess a general answer will be opinion based (not a vote to close), but one general suggestion I would give is: ask your supervisor or a conference chair. (but last of them could be not responding, busy ppl. you know)
I had different supervisors during B.Sc. M.Sc.-Thesis and paper writings. Some of them were okay with headings and subheadings without "glue" text between, some really made a mass of it, that it would be the ugliest writing style they could imagine.
I have the habit to write those introducing glue paragraphs between headings (as @mwormser suggests too) just telling the reader what will come in the following subsections. But I would always ask a supervisor for what he/she likes more. In case of anonymous blind reviews you may chance 50/50 with the likes of the reviewers. Maybe the journal/conference/etc. offers a style guideline that makes suggestions to this question too, just have a look at the submission pages.
Anyway an introduction paragraph that has no "hard-fact / result / discussion" content - but is only a meta-text about what to come - should be short. Give the reader the chance to decide at this point if he/she is in need to deep-read the next (sub)sections or if he/she is familiar enough (eg. obligatory method describing intro chapters of a thesis) with their content.
Another solution I also saw was use of epigraphs (quotes, jokes etc.) to fill this gaps between headlines. This makes the glue paragraphs less repetitive, but I would be careful with this and (if that) only spread them very rarely.