Topic
Multi-document summarization
About: Multi-document summarization is a research topic. Over the lifetime, 2270 publications have been published within this topic receiving 71850 citations.
Papers published on a yearly basis
Papers
More filters
TL;DR: The authors compared and assessed the effectiveness of two optimizers on a variety of datasets and compared their performance on various datasets as they are widely employed in text summarization, and concluded that the two utilized optimizers are adam and rmsprop.
TL;DR: The implementation of a general-purpose web application which performs automatic summarization on the text entered using the Text Rank Algorithm is discussed.
Abstract: Abstract: Automatic Text Summarization is one of the most trending research areas in the field of Natural Language Processing. The main aim of text summarization is to reduce the size of a text without losing any important information. Various techniques can be used for automatic summarization of text. In this paper we are going to focus on the automatic summarization of text using graph-based methods. In particular, we are going to discuss the implementation of a general-purpose web application which performs automatic summarization on the text entered using the Text Rank Algorithm. Summarization of text using graph-based approaches involves pre-processing and cleansing of text, tokenizing the sentences present in the text, representing the tokenized text in the form of numerical vectors, creating a similarity matrix which shows the semantic similarity between different sentences present in the text, representing the similarity matrix as a graph, scoring and ranking the sentences and extracting the summary. Keywords: Text Summarization, Unsupervised Learning, Text Rank, Page Rank, Web Application, Graph Based Summarization, Extractive Summarization
19 Jun 2017
TL;DR: This work presents a free Web API for single and multi-text summarization that integrates in a novel pipeline different text analysis techniques - ranging from keyword and entity extraction, to topic modelling and sentence clustering - and gives SoA competitive results.
Abstract: In this work we present a free Web API for single and multi-text summarization. The summarization algorithm follows an extractive approach, thus selecting the most relevant sentences from a single document or a document set. It integrates in a novel pipeline different text analysis techniques - ranging from keyword and entity extraction, to topic modelling and sentence clustering - and gives SoA competitive results. The application, written in Python, supports as input both plain texts and Web URLs. The API is publicly accessible for free using the specific conference token1 as described in the reference page2. The browser-based demo version, for summarization of single documents only, is publicly accessible at http://yonderlabs.com/demo.
Patent•
29 Feb 2012TL;DR: In this paper, an exclusive lock is acquired on a table storing scope information for the plurality of data summarization instances and the remaining tasks to be performed by the new summarization instance can be defined.
Abstract: Embodiments of the invention provide systems and methods for recovering a failed data summarization. According to one embodiment, recovering a failed instance can comprise processing existing summarization instances identified as instances for which a new data summarization instance needs to wait. Upon a completion or a timeout of each of the instances identified as instances for which the new data summarization instance needs to wait, an exclusive lock can be acquired on a table storing scope information for the plurality of data summarization instances. One or more existing data summarization instances that match the new data summarization instance or that have an overlapping scope with the new data summarization instance can be processed, remaining tasks to be performed by the new data summarization instance can be defined, the exclusive lock can be released, and the remaining tasks to be performed by the new data summarization instance can be performed.
TL;DR: The result obtained from the experiment shows promising result in summarization of Wolaita text, and the researcher justified the model performance with ROUGE evaluation metrics by comparing the system summaries with the expert summaries.
Abstract: Text summarization is the mechanism of summarizing a huge document comprising vast amount of information which is difficult to overcome and understand its message easily in any written documents for whatever languages without losing its entire message. A short and precise document which conveys intended information for the user in demand is expected in this information age. In addition to that, summarizing a document with vast amount of information is very difficult and time consuming specially for less resourced and technologically unfavored languages. Therefore in this study, the researcher proposed to address such problems for Wolaita by using graph based extractive text summarization approach. To attain the goal of this study the researcher prepared 92 documents for the study, explored extractive text summarization with graph-based approach to address the problems, performed text preprocessing tasks and finally developed text summarization model by using TextRank algorithms. The researcher used 92 documents, performed 92 various experiments, on documents and experimental results and findings were discussed in detail. To evaluate the model performance, three different expert summaries were collected for documents and computed system generated summaries with ROUGE evaluation metric. The researcher justified it with ROUGE evaluation metrics by comparing the system summaries with the expert summaries. The result obtained from the experiment shows promising result in summarization of Wolaita text. Finally, the experimental result of a 61.16% recall, 60.69% precision and 60.46% f-measures were obtained.