How do Buddhist logicians justify the reliability of inference if all entities are momentary and lack enduring substance?

In Nyāya epistemology, pramāṇas (means of valid knowledge) include perception, inference, comparison, and testimony. Since Buddhist thought often restricts ultimate reality to momentary dharmas (kṣaṇikavāda) and denies enduring substances, how do Buddhist logicians account for the reliability of inference across time, given that the subject and predicate cease to exist in the very next moment?