Contrastive summarization is the problem of jointly generating summaries for two entities in order to highlight their differences. In this paper we present an investigation into contrastive summarization through an implementation and evaluation of a contrastive opinion summarizer in the consumer reviews domain.

Our team submitted runs for the first running of the TREC Temporal Summarization track. We focused on the Sequential Update Summarization task. This task involves simulating processing a temporally ordered stream of over 1 billion documents to identify sentences that are relevant to a specific breaking news stories which contain new content. In this paper, we describe our approach and evaluatio...

In this paper we present our participation at TAC 2008 opinion summarization task. We make use of a statistical model for opinion extraction and the subsequent summarization of the extracted opinions. We also present its performance as per the official TAC 2008 evaluation results.

This paper presents our extractive summarization systems at the update summarization track of TAC 2008. We proposed two novel methods, one is based on the information distance theory, and the other is based on the sentence centrality which derives from the centrality concept in the graph theory. The evaluation results show that the two submitted runs are very competitive to generate extractive ...

This paper presents our multi-document summarization system ICTGSP-S at DUC 2007. We propose a new method for representing and summarizing documents by integrating subtopics partition with graph representation. The method starts from the assumption that capturing subtopic structure of document collection is essential for summarization. The evaluation results show the benefit of this approach.

This paper presents our work on queryfocused multi-document summarization with the enhanced IS_SUM system. We focus on improving its lexical chain algorithm for efficiency enhancement, applying the WordNet for similarity calculation and adapting it to query-focused multi-document summarization. We present its performance in terms of its official DUC2007 evaluation results together with some oth...

We develop a Recurrent Neural Network (RNN) Language Model to extract sentences from Yelp Review Data for the purpose of automatic summarization. We compare these extracted sentences against user-generated tips in the Yelp Academic Dataset using ROUGE and BLEU metrics for summarization evaluation. The performance of a uni-directional RNN is compared against word-vector averaging.

This paper describes the architecture of the summarization system IS_SUM from Institute of Software, Chinese Academy of Sciences for DUC2006. The improvements on lexical chain algorithm are given in detail in order to enhance its efficiency and adapt it to query based summarization. We conclude our paper with the different evaluation results and the very primary analysis.

The CCNU summarization system, PUSMS (Proceeding to Using Semantic Method for Summarization), join in TAC (formerly DUC) for the first time. For the update summarization tasks, we used syntacticbased anaphora resolution and sentence compression algorithms in our system. Term significance was then obtained by frequency-related topic significance and query-related significance by obtaining cooccu...

NeATS is a multi-document summarization system that attempts to extract relevant or interesting portions from a set of documents about some topic and present them in coherent order. NeATS is among the best performers in the large scale summarization evaluation DUC-01.