Retrieve_rerank_and_rewrite_Soft_template_based_neural_summarization.pdf

Transcript

1. Retrieve, Rerank and Rewrite: Soft Template Based Neural Summarization Ziqiang Cao, Wenjie Li, Furu Wei, Sujian Li Proceedings of the 56th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers), pp.152-161, 2018 Nagaoka University of Technology Takumi Maruyama Literature review:
2. 
   Ø template-based approachseq2seq Ø Informativeness
   , SOTA    Ø template ,   
3. 
    Ø “Seq2seq models tend to lose control sometimes” • '# ( • '# )" Ø Template based summarization • -%   • $+,template !*
   &  Template: [REGION] shares [open/close] [NUMBER] percent [lower/higher]
4.  Ø Soft-template based neural summarization • seq2seqtemplate  • 
         … •     Soft-templates
5. 
    Ø Modules • Retrieve: 
     • Rerank: Retrieve    • Rewrite: Rerank template  
6. 
    Ø Retrieve • Lucene (! ) • 
    
   (over 3M) ! • "30  https://lucene.apache.org/
7. 
    ØJointly Rerank and Rewrite • Rerank Bilinear networksaliency  • Rewrite x : , y*: , r : soft-templates (Retrieve
    ) Actual saliency : s*(r, y*) = ROUGE(r, y*) Saliency :
8. 
    ØJointly Rerank and Rewrite • Learning
9. 
   Ø Dataset Gigaword corpus
10. 
     Ø 
     Rerank -> Rewrite Rerank + Rewrite
11. 
     Ø Soft-template"& Rondom 
     ' First Retrieve ! $# Max Retrieve %!$# , ROUGE$# Optimal 
    %! ROUGE$# Rerank Rerank  !$#
12. 
     Ø Linguistic quality evaluation LEN_DIF !$& & &) LESS_3 3*(# & " COPY & &
     *( ' NEW_NE !$&&   +% "
13. 
     Ø template-based + seq2seq model  Ø Soft-templates  ,