Align and Attend: Multimodal Summarization with Dual Contrastive Losses

Benchmark Model Rank Results
extractive-document-summarization-on-cnnA2Summ#2ROUGE-1: 44.11ROUGE-2: 20.31ROUGE-L: 35.92
supervised-video-summarization-on-summeA2Summ#3F1-score (Canonical): 55.0Kendall's Tau: 0.108Spearman's Rho: 0.129
supervised-video-summarization-on-tvsumA2Summ#5F1-score (Canonical): 63.4Kendall's Tau: 0.137Spearman's Rho: 0.165