Dynamic Semantic Channel Hashing (DSCH) replaces fixed-width SCH channels with continuous, similarity-dependent widths and positions, improving tie-aware mAP on most cross- and intra-modal retrieval tasks.
On the Stratification of Multi-label Data,
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.AI 1years
2026 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
DSCH-Loss: A Dynamic Semantic Channel Objective for Deep Semantic Hashing
Dynamic Semantic Channel Hashing (DSCH) replaces fixed-width SCH channels with continuous, similarity-dependent widths and positions, improving tie-aware mAP on most cross- and intra-modal retrieval tasks.