Concept-Based Classification For Multi-Document Summarization

  • Asli Celikyilmaz
  • Dilek Hakkani-Tür

IEEE Intl. Conf. on Acoustics, Speech and Signal Processing - ICASSP |

Documents often contain inherently many concepts reflecting specific and generic aspects. To automatically generate a short summary text of documents on similar topics, it is imperative that we discover general aspects in documents be cause summaries usually contain general rather than specific concepts. This paper presents a semi-supervised extractive summarization model based upon latent concept classification that can differentiate between the two types of aspects as hidden concepts being mentioned in documents. A classifier is trained on hidden concepts discovered from documents and their corresponding human-generated summaries using a probabilistic Bayesian model: the summary-focused topic model. Experimental results based on ROUGE evaluations indicate that ranking sentences to be included in summary text based on the latent summary concept classification has improvements on the quality of the generated summaries.