Narrow Results

With an increasing number of new summarization systems proposed in recent years, an automatic text evaluation metric that can accurately and reliably rate the performance of summarization systems has been a pressing need. However, current automatic text evaluation metrics can only measure one or certain aspects of the quality between two summary texts and do not agree with human judgments consistently. In this paper, we show that combining multiple well-chosen evaluation metrics and training predictive models using human annotated datasets can lead to more reliable evaluation scores than using any individual automatic metric. Our predictive models trained on a human annotated subset of the CNN/DailyMail corpus demonstrate significant improvements (e.g. approximately 25% along coherence dimension) over selected individual metrics. Furthermore, a concise meta-evaluation on automatic metrics is provided along with an analysis of the performance of 12 predictive models. We also investigate the sensitivity of automatic metrics when mixed together for training these models. We have made the code, the instructions for experiment setup, and the trained models available as a tool for comparing and evaluating text summarization systems.^a

https://github.com/bzhao2718/ReliableSummEvalReg.

This paper presents some original methods for text summarization of a single source document by extraction. The methods are based on some of our own text segmentation algorithms. We denote them as logical segmentation because for all these methods (LTT, ArcInt and ArcReal) the score of a sentence is calculated starting from the number of sentences which are entailed by it. For a text (which is a sequence of sentences) the scores form a structure which indicates how the most important sentences alternate with less important ones and organizes the text according to its logical content. The second logical method, Pure Entailment also uses definition of the relation of entailment between two texts. At least to our knowledge, it is for the first time that the relation of Text Entailment between the sentences of a text is used for segmentation and summarization.

The third original method applies Dynamic Programming and centering theory to the sentences logically scored as above. The obtained ranked logical segments are used in the summarization. Our methods of segmentation and summarization are applied and evaluated against a manually realized segmentation and summarization of the same text by Donald Richie, "The Koan".

This paper addresses a novel sentence reduction algorithm based on a decision tree model in which semantic information is used to enhance the accuracy of sentence reduction. The proposed algorithm is able to deal with the changeable order problem in sentence reduction. Furthermore, the use of decision list to solve the fragment problem in sentence reduction based decision tree model is also discussed. Our experimental results show an improvement when compared with earlier methods.

Popularity of the Internet has contributed towards the explosive growth of online information, and it is especially useful to have tools which can help users digest information content. Text summarization addresses this need by taking a source text, selecting the most important portions of it, and presenting coherent summary to the user in a manner sensitive to the user's or application's needs. The goal of this paper is to show how these objectives can be achieved through an efficient use of lexical cohesion. The current work addresses both generic and query-based summaries in the context of single documents and sets of documents as in current news. We present an approach for identifying the most important portions of the text which are topically best suited to represent the source texts according to the author's views or in response to the user's interests. This identification must also take into consideration the degree of connectiveness among the chosen text portions so as to minimize the danger of producing summaries which contain poorly linked sentences. We present a system that handles these objectives, discuss its performance, and evaluate it and compare it to other systems in the context of Document Understanding Conference (DUC) evaluations.

Traditional text summarization systems have not used the category information of documents to be summarized. However, the estimated weights of each word can be often biased on small data such as a single document. Thus we proposed an effective feature-weighting method for document summarization that utilizes category information and solves the biased probability problem. The method uses a category-based smoothing method and a bootstrapping framework. As a result, in our experiments, our proposed summarization method achieves better performance than other statistical sentence-extraction methods.

A vast amount of textual information on the internet has amplified the importance of text summarization models. Abstractive summarization generates original words and sentences that may not exist in the source document to be summarized. Such abstractive models may suffer from shortcomings such as linguistic acceptability and hallucinations. Recall-Oriented Understudy for Gisting Evaluation (ROUGE) is a metric commonly used to evaluate abstractive summarization models. However, due to its n-gram-based approach, it ignores several critical linguistic aspects. In this work, we propose Similarity, Entailment, and Acceptability Score (SEAScore), an automatic evaluation metric for evaluating abstractive text summarization models using the power of state-of-the-art pre-trained language models. SEAScore comprises three language models (LMs) that extract meaningful linguistic features from candidate and reference summaries and a weighted sum aggregator that computes an evaluation score. Experimental results show that our LM-based SEAScore metric correlates better with human judgment than standard evaluation metrics such as ROUGE-N and BERTScore.

Large volumes of structured and semi-structured data are being generated every day. Processing this large amount of data and extracting important information is a challenging task. The goal of an automatic text summarization is to preserve the key information and the overall meaning of the article to be summarized. In this paper, a graph-based approach is followed to generate an extractive summary, where sentences of the article are considered as vertices, and weighted edges are introduced based on the cosine similarities among the vertices. A possible subset of maximal independent sets of vertices of the graph is identified with the assumption that adjacent vertices provide sentences with similar information. The degree centrality and clustering coefficient of the vertices are used to compute the score of each of the maximal independent sets. The set with the highest score provides the final summary of the article. The proposed method is evaluated using the benchmark BBC News data to demonstrate its effectiveness and is applied to the COVID-19 Twitter data to express its applicability in topic modeling. Both the application and comparative study with other methods illustrate the efficacy of the proposed methodology.

Recently, there has been an increasing need for artificial text summarizing algorithms that attempt to automatically condense a document into a shorter form due to the rapid and exponential growth of textual data. As a result, automated text summarization is utilized by an extensive wide range of companies to help people find the most significant information. But still, there is a challenge in creating a brief and short summary of the original material that contains the key concepts. This paper intends to propose an abstractive text summarization based on LSTM with transfer learning methods. Initially, the long text-based document summarization process includes two phases. In the first phase, the input text is preprocessed for stop word removal and stemming techniques to reduce the document size. Then features like improved bag of words, TF-IDF, aspect term extraction and average word length-based features are extracted from the preprocessed text. In the second phase, the knowledge extraction has been performed. In the knowledge extraction, tokens are generated using BERT tokenization and the co-occurrence is utilized using improved co-occurrence matrix generation. The proposed method achieves the highest recall in 80% of learning which is 1.75% 1.39%, 2.58%, 1.19%, 0.30%, and 2.28% better than the other models such as Bi-GRU, RNN, LSTM, SVM-LR, CNN, LSTM-GRU and GRU respectively. Further, the knowledge extraction and the extracted features are subjected to the improved LSTM network, where the features are trained as the pre-trained model via transfer learning. This is the phase, where the final summary is produced.

This paper introduces a mid-depth text understanding methodology. By mid-depth we mean that the methodology is: (1) based on shallow text understanding, (2) accomplished with currently available linguistic instruments, and (3) focused towards understanding texts as deeply as possible. To realize the methodology, we set up two new linguistic devices for sentence extraction — sentence abstraction and abductive chains. Sentence abstraction is processing for sentence revision with concept abstraction. Abductive chains are lexical chains with abduction. The chain consists of abductive links which connect coherent parts of abstracted sentences. It aims to identify the causal propensity described or implied in the text as well as to delineate its recurring concepts. With this methodology, a preliminary attempt at mid-depth text understanding is made to locate the topic sentences of a given text. We expect that this methodology will lead us to develop another practical text summarization system.

This paper proposes a new method of enhancing the accuracy of a decomposition task by using position checking and a semantic measure for each word within a summary document. The proposed model is an extension of the Hidden Markov Model for the human-written decomposition problem. Experimental results using DUC data and the Telecommunication Corpus show that the proposed method improves the accuracy of decomposition of human-written summary sentences.

In this paper, we present event-based summarization and investigate whether it can be enhanced by integrating temporal distribution information. We refer to events as event terms and the associated event elements. Event terms represent actions themselves and event elements are commonly those verb arguments. After anchoring events on the time line, we explore two statistical measures, i.e. tf*idf and x², for evaluating the importance of events on each day. Summary sentences are selected based on the weights of the events contained in them, in either sequential or round robin order. Experiments show that the combination of time-based tf*idf weighting scheme and sequential sentence selection strategy can improve the quality of summaries significantly. The improvement can be attributed to its capability of representing the trend of news topics based on event temporal distributions.

In this digital era, there is a tremendous increase in the volume of data, which adds difficulties to the person who utilizes particular applications, such as websites, email, and news. Text summarization targets to reduce the complexity of obtaining statistics from the websites as it compresses the textual document to a short summary without affecting the relevant information. The crucial step in multi-document summarization is obtaining a relationship between the cross-sentence. However, the conventional methods fail to determine the inter-sentence relationship, especially in long documents. This research develops a graph-based neural network to attain an inter-sentence relationship. The significant step in the proposed multi-document text summarization model is forming the weighted graph embedding features. Furthermore, the weighted graph embedding features are utilized to evaluate the relationship between the document’s words and sentences. Finally, the bidirectional long short-term memory (BiLSTM) classifier is utilized to summarize the multi-document text summarization. The experimental analysis uses the three standard datasets, the Daily Mail dataset, Document Understanding Conference (DUC) 2002, and Document Understanding Conference (DUC) 2004 dataset. The experimental outcome demonstrates that the proposed weighted graph embedding feature + BiLSTM model exceeds all the conventional methods with Precision, Recall, and F1 score of 0.5352, 0.6296, and 0.5429, respectively.

Text summarization is one of the most discussed topic in the field in information exchange and retrieval. Recently, the need for local language based text summarization methods are increasing. In this paper, a method for text summarization in Hindi language is plotted with help of extraction methods. The proposed approach is uses three major algorithms, fuzzy classifier, neural network and global search optimization (GSO). The fuzzy classifier and neural network are used for generating sentence score. The GSO algorithm is used with the neural network, in order to optimize the weights in the neural network. A hybrid score is generated from fuzzy method and neural network for each input sentences. Finally, based on the hybrid score from fuzzy classifier and neural network, the summary of the given input records are generated. An experimental analysis of the proposed approach will subjected based on the evaluation parameters precision, recall. Later on experimental analysis are conducted on the proposed approach in order to evaluate the performance. According to the experimental analysis, the proposed approach achieved an average precision rate 0.90 and average recall rate of 0.88 for compression rate 20%. The comparative analysis also provided reasonable results to prove the efficiency of the proposed approach.

Data availability is not a major issue at present times in view of the widespread use of Internet; however, information and knowledge availability are the issues. Due to data overload and time-critical nature of information need, automatic summarization of documents plays a significant role in information retrieval and text data mining. This paper discusses the design of a multi-document summarizer that uses Katz's K-mixture model for term distribution. The model helps in ranking the sentences by a modified term weight assignment. Highly ranked sentences are selected for the final summary. The sentences that are repetitive in nature are eliminated, and a tiled summary is produced. Our method avoids redundancy and produces a readable (even browsable) summary, which we refer to as an event-specific tiled summary. The system has been evaluated against the frequently occurring sentences in the summaries generated by a set of human subjects. Our system outperforms other auto-summarizers at different extraction levels of summarization with respect to the ideal summary, and is close to the ideal summary at 40% extraction level.

In recent years, deep learning has revolutionized natural language processing (NLP) by enabling the development of models that can learn complex representations of language data, leading to significant improvements in performance across a wide range of NLP tasks. Deep learning models for NLP typically use large amounts of data to train deep neural networks, allowing them to learn the patterns and relationships in language data. This is in contrast to traditional NLP approaches, which rely on hand-engineered features and rules to perform NLP tasks. The ability of deep neural networks to learn hierarchical representations of language data, handle variable-length input sequences, and perform well on large datasets makes them well-suited for NLP applications. Driven by the exponential growth of textual data and the increasing demand for condensed, coherent, and informative summaries, text summarization has been a critical research area in the field of NLP. Applying deep learning to text summarization refers to the use of deep neural networks to perform text summarization tasks. In this survey, we begin with a review of fashionable text summarization tasks in recent years, including extractive, abstractive, multi-document, and so on. Next, we discuss most deep learning-based models and their experimental results on these tasks. The paper also covers datasets and data representation for summarization tasks. Finally, we delve into the opportunities and challenges associated with summarization tasks and their corresponding methodologies, aiming to inspire future research efforts to advance the field further. A goal of our survey is to explain how these methods differ in their requirements as understanding them is essential for choosing a technique suited for a specific setting. This survey aims to provide a comprehensive review of existing techniques, evaluation methodologies, and practical applications of automatic text summarization.