GOOD WORK! Besides, in the Section 2.1.4, could you please clarify whether MinHash is implemented to deduplicate similar text across buckets?