In this paper, we introduce Iterative Text Summarization (ITS), an iteration-based model for supervised extractive text summarization, inspired by the observation that it is often necessary for a human to read an article multiple times in order to fully understand and summarize its contents. Current summarization approaches read through a document only once to generate a document representation, resulting in a sub-optimal representation. To address this issue we introduce a model which iteratively polishes the document representation on many passes through the document. As part of our model, we also introduce a selective reading mechanism that decides more accurately the extent to which each sentence in the model should be updated. Experimental results on the CNN/DailyMail and DUC2002 datasets demonstrate that our model significantly outperforms state-of-the-art extractive systems when evaluated by machines and by humans.
This paper introduces a sentence to vector encoding framework suitable for advanced natural language processing. Our latent representation is shown to encode sentences with common semantic information with similar vector representations. The vector representation is extracted from an encoder-decoder model which is trained on sentence paraphrase pairs. We demonstrate the application of the sentence representations for two different tasks -- sentence paraphrasing and paragraph summarization, making it attractive for commonly used recurrent frameworks that process text. Experimental results help gain insight how vector representations are suitable for advanced language embedding.
Extractive compression is a challenging natural language processing problem. This work contributes by formulating neural extractive compression as a parse tree transduction problem, rather than a sequence transduction task. Motivated by this, we introduce a deep neural model for learning structure-to-substructure tree transductions by extending the standard Long Short-Term Memory, considering the parent-child relationships in the structural recursion. The proposed model can achieve state of the art performance on sentence compression benchmarks, both in terms of accuracy and compression rate.
In Automatic Text Summarization, preprocessing is an important phase to reduce the space of textual representation. Classically, stemming and lemmatization have been widely used for normalizing words. However, even using normalization on large texts, the curse of dimensionality can disturb the performance of summarizers. This paper describes a new method for normalization of words to further reduce the space of representation. We propose to reduce each word to its initial letters, as a form of Ultra-stemming. The results show that Ultra-stemming not only preserve the content of summaries produced by this representation, but often the performances of the systems can be dramatically improved. Summaries on trilingual corpora were evaluated automatically with Fresa. Results confirm an increase in the performance, regardless of summarizer system used.
In sentence compression, the task of shortening sentences while retaining the original meaning, models tend to be trained on large corpora containing pairs of verbose and compressed sentences. To remove the need for paired corpora, we emulate a summarization task and add noise to extend sentences and train a denoising auto-encoder to recover the original, constructing an end-to-end training regime without the need for any examples of compressed sentences. We conduct a human evaluation of our model on a standard text summarization dataset and show that it performs comparably to a supervised baseline based on grammatical correctness and retention of meaning. Despite being exposed to no target data, our unsupervised models learn to generate imperfect but reasonably readable sentence summaries. Although we underperform supervised models based on ROUGE scores, our models are competitive with a supervised baseline based on human evaluation for grammatical correctness and retention of meaning.
Neural machine translation usually adopts autoregressive models and suffers from exposure bias as well as the consequent error propagation problem. Many previous works have discussed the relationship between error propagation and the \emph{accuracy drop} (i.e., the left part of the translated sentence is often better than its right part in left-to-right decoding models) problem. In this paper, we conduct a series of analyses to deeply understand this problem and get several interesting findings. (1) The role of error propagation on accuracy drop is overstated in the literature, although it indeed contributes to the accuracy drop problem. (2) Characteristics of a language play a more important role in causing the accuracy drop: the left part of the translation result in a right-branching language (e.g., English) is more likely to be more accurate than its right part, while the right part is more accurate for a left-branching language (e.g., Japanese). Our discoveries are confirmed on different model structures including Transformer and RNN, and in other sequence generation tasks such as text summarization.
In this paper we propose a new approach to evaluate the informativeness of transcriptions coming from Automatic Speech Recognition systems. This approach, based in the notion of informativeness, is focused on the framework of Automatic Text Summarization performed over these transcriptions. At a first glance we estimate the informative content of the various automatic transcriptions, then we explore the capacity of Automatic Text Summarization to overcome the informative loss. To do this we use an automatic summary evaluation protocol without reference (based on the informative content), which computes the divergence between probability distributions of different textual representations: manual and automatic transcriptions and their summaries. After a set of evaluations this analysis allowed us to judge both the quality of the transcriptions in terms of informativeness and to assess the ability of automatic text summarization to compensate the problems raised during the transcription phase.
We introduce extreme summarization, a new single-document summarization task which does not favor extractive strategies and calls for an abstractive modeling approach. The idea is to create a short, one-sentence news summary answering the question "What is the article about?". We collect a real-world, large-scale dataset for this task by harvesting online articles from the British Broadcasting Corporation (BBC). We propose a novel abstractive model which is conditioned on the article's topics and based entirely on convolutional neural networks. We demonstrate experimentally that this architecture captures long-range dependencies in a document and recognizes pertinent content, outperforming an oracle extractive system and state-of-the-art abstractive approaches when evaluated automatically and by humans.
Abstractive text summarization aims to shorten long text documents into a human readable form that contains the most important facts from the original document. However, the level of actual abstraction as measured by novel phrases that do not appear in the source document remains low in existing approaches. We propose two techniques to improve the level of abstraction of generated summaries. First, we decompose the decoder into a contextual network that retrieves relevant parts of the source document, and a pretrained language model that incorporates prior knowledge about language generation. Second, we propose a novelty metric that is optimized directly through policy learning to encourage the generation of novel phrases. Our model achieves results comparable to state-of-the-art models, as determined by ROUGE scores and human evaluations, while achieving a significantly higher level of abstraction as measured by n-gram overlap with the source document.
Sequence to sequence (Seq2Seq) learning has recently been used for abstractive and extractive summarization. In current study, Seq2Seq models have been used for eBay product description summarization. We propose a novel Document-Context based Seq2Seq models using RNNs for abstractive and extractive summarizations. Intuitively, this is similar to humans reading the title, abstract or any other contextual information before reading the document. This gives humans a high-level idea of what the document is about. We use this idea and propose that Seq2Seq models should be started with contextual information at the first time-step of the input to obtain better summaries. In this manner, the output summaries are more document centric, than being generic, overcoming one of the major hurdles of using generative models. We generate document-context from user-behavior and seller provided information. We train and evaluate our models on human-extracted-golden-summaries. The document-contextual Seq2Seq models outperform standard Seq2Seq models. Moreover, generating human extracted summaries is prohibitively expensive to scale, we therefore propose a semi-supervised technique for extracting approximate summaries and using it for training Seq2Seq models at scale. Semi-supervised models are evaluated against human extracted summaries and are found to be of similar efficacy. We provide side by side comparison for abstractive and extractive summarizers (contextual and non-contextual) on same evaluation dataset. Overall, we provide methodologies to use and evaluate the proposed techniques for large document summarization. Furthermore, we found these techniques to be highly effective, which is not the case with existing techniques.
With rise of digital age, there is an explosion of information in the form of news, articles, social media, and so on. Much of this data lies in unstructured form and manually managing and effectively making use of it is tedious, boring and labor intensive. This explosion of information and need for more sophisticated and efficient information handling tools gives rise to Information Extraction(IE) and Information Retrieval(IR) technology. Information Extraction systems takes natural language text as input and produces structured information specified by certain criteria, that is relevant to a particular application. Various sub-tasks of IE such as Named Entity Recognition, Coreference Resolution, Named Entity Linking, Relation Extraction, Knowledge Base reasoning forms the building blocks of various high end Natural Language Processing (NLP) tasks such as Machine Translation, Question-Answering System, Natural Language Understanding, Text Summarization and Digital Assistants like Siri, Cortana and Google Now. This paper introduces Information Extraction technology, its various sub-tasks, highlights state-of-the-art research in various IE subtasks, current challenges and future research directions.
Automatic Text Summarization strategies have been successfully employed to digest text collections and extract its essential content. Usually, summaries are generated using textual corpora that belongs to the same domain area where the summary will be used. Nonetheless, there are special cases where it is not found enough textual sources, and one possible alternative is to generate a summary from a different domain. One manner to summarize texts consists of using a graph model. This model allows giving more importance to words corresponding to the main concepts from the target domain found in the summarized text. This gives the reader an overview of the main text concepts as well as their relationships. However, this kind of summarization presents a significant number of repeated terms when compared to human-generated summaries. In this paper, we present an approach to produce graph-model extractive summaries of texts, meeting the target domain exigences and treating the terms repetition problem. To evaluate the proposition, we performed a series of experiments showing that the proposed approach statistically improves the performance of a model based on Graph Centrality, achieving better coverage, accuracy, and recall.
So far and trying to reach human capabilities, research in automatic summarization has been based on hypothesis that are both enabling and limiting. Some of these limitations are: how to take into account and reflect (in the generated summary) the implicit information conveyed in the text, the author intention, the reader intention, the context influence, the general world knowledge. Thus, if we want machines to mimic human abilities, then they will need access to this same large variety of knowledge. The implicit is affecting the orientation and the argumentation of the text and consequently its summary. Most of Text Summarizers (TS) are processing as compressing the initial data and they necessarily suffer from information loss. TS are focusing on features of the text only, not on what the author intended or why the reader is reading the text. In this paper, we address this problem and we present a system focusing on acquiring knowledge that is implicit. We principally spotlight the implicit information conveyed by the argumentative connectives such as: but, even, yet and their effect on the summary.
Keyphrases are the phrases, consisting of one or more words, representing the important concepts in the articles. Keyphrases are useful for a variety of tasks such as text summarization, automatic indexing, clustering/classification, text mining etc. This paper presents a hybrid approach to keyphrase extraction from medical documents. The keyphrase extraction approach presented in this paper is an amalgamation of two methods: the first one assigns weights to candidate keyphrases based on an effective combination of features such as position, term frequency, inverse document frequency and the second one assign weights to candidate keyphrases using some knowledge about their similarities to the structure and characteristics of keyphrases available in the memory (stored list of keyphrases). An efficient candidate keyphrase identification method as the first component of the proposed keyphrase extraction system has also been introduced in this paper. The experimental results show that the proposed hybrid approach performs better than some state-of-the art keyphrase extraction approaches.
Matching texts in highly inflected languages such as Arabic by simple stemming strategy is unlikely to perform well. In this paper, we present a strategy for automatic text matching technique for for inflectional languages, using Arabic as the test case. The system is an extension of ROUGE test in which texts are matched on token's lemma level. The experimental results show an enhancement of detecting similarities between different sentences having same semantics but written in different lexical forms..
In this paper we present REG, a graph-based approach for study a fundamental problem of Natural Language Processing (NLP): the automatic text summarization. The algorithm maps a document as a graph, then it computes the weight of their sentences. We have applied this approach to summarize documents in three languages.
Text segmentation task is an essential processing task for many of Natural Language Processing (NLP) such as text summarization, text translation, dialogue language understanding, among others. Turns segmentation considered the key player in dialogue understanding task for building automatic Human-Computer systems. In this paper, we introduce a novel approach to turn segmentation into utterances for Egyptian spontaneous dialogues and Instance Messages (IM) using Machine Learning (ML) approach as a part of automatic understanding Egyptian spontaneous dialogues and IM task. Due to the lack of Egyptian dialect dialogue corpus the system evaluated by our corpus includes 3001 turns, which are collected, segmented, and annotated manually from Egyptian call-centers. The system achieves F1 scores of 90.74% and accuracy of 95.98%.
The development of methods to deal with the informative contents of the text units in the matching process is a major challenge in automatic summary evaluation systems that use fixed n-gram matching. The limitation causes inaccurate matching between units in a peer and reference summaries. The present study introduces a new Keyphrase based Summary Evaluator KpEval for evaluating automatic summaries. The KpEval relies on the keyphrases since they convey the most important concepts of a text. In the evaluation process, the keyphrases are used in their lemma form as the matching text unit. The system was applied to evaluate different summaries of Arabic multi-document data set presented at TAC2011. The results showed that the new evaluation technique correlates well with the known evaluation systems: Rouge1, Rouge2, RougeSU4, and AutoSummENG MeMoG. KpEval has the strongest correlation with AutoSummENG MeMoG, Pearson and spearman correlation coefficient measures are 0.8840, 0.9667 respectively.
Summarization is one of the key features of human intelligence. It plays an important role in understanding and representation. With rapid and continual expansion of texts, pictures and videos in cyberspace, automatic summarization becomes more and more desirable. Text summarization has been studied for over half century, but it is still hard to automatically generate a satisfied summary. Traditional methods process texts empirically and neglect the fundamental characteristics and principles of language use and understanding. This paper summarizes previous text summarization approaches in a multi-dimensional classification space, introduces a multi-dimensional methodology for research and development, unveils the basic characteristics and principles of language use and understanding, investigates some fundamental mechanisms of summarization, studies the dimensions and forms of representations, and proposes a multi-dimensional evaluation mechanisms. Investigation extends to the incorporation of pictures into summary and to the summarization of videos, graphs and pictures, and then reaches a general summarization framework.
ROUGE is a widely adopted, automatic evaluation measure for text summarization. While it has been shown to correlate well with human judgements, it is biased towards surface lexical similarities. This makes it unsuitable for the evaluation of abstractive summarization, or summaries with substantial paraphrasing. We study the effectiveness of word embeddings to overcome this disadvantage of ROUGE. Specifically, instead of measuring lexical overlaps, word embeddings are used to compute the semantic similarity of the words used in summaries instead. Our experimental results show that our proposal is able to achieve better correlations with human judgements when measured with the Spearman and Kendall rank coefficients.