Sep 2026
Semantic-conditioned temporal structure distillation for audio-text retrieval
This work proposes Semantic-Conditioned Temporal Structure Distillation, SCTSD, which preserves the standard dual-encoder retrieval pipeline while learning temporal structural information induced by text conditions during training.
Wei-Jie Luo, Xiao-Hang Guan, Xi-Chang Cai et al.
· Multimedia Systems · 0 citations