Text summarization using unsupervised deep learning

Yousefi-Azar, Mahmood; Hamey, Len

doi:10.1016/j.eswa.2016.10.017

Cited by 217 publications

(62 citation statements)

References 19 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…For extractive query-oriented single-document summarization, Yousefi-Azar and Hamey [27] used a deep autoencoder to compute a feature space from the term-frequency (tf) input. They developed a local word representation in which each vocabulary is designed to build the input representation for sentences in the document.…”

Section: Related Workmentioning

confidence: 99%

A Hierarchical Structured Self-Attentive Model for Extractive Document Summarization (HSSAS)

2018

View full text Add to dashboard Cite

The recent advance in neural network architecture and training algorithms has shown the effectiveness of representation learning. The neural-network-based models generate better representation than the traditional ones. They have the ability to automatically learn the distributed representation for sentences and documents. To this end, we proposed a novel model that addresses several issues that are not adequately modeled by the previously proposed models, such as the memory problem and incorporating the knowledge of document structure. Our model uses a hierarchical structured self-attention mechanism to create the sentence and document embeddings. This architecture mirrors the hierarchical structure of the document and in turn enables us to obtain better feature representation. The attention mechanism provides extra source of information to guide the summary extraction. The new model treated the summarization task as a classification problem in which the model computes the respective probabilities of sentence-summary membership. The model predictions are broken up by several features such as information content, salience, novelty, and positional representation. The proposed model was evaluated on two well-known datasets, the CNN/Daily Mail and DUC 2002. The experimental results show that our model outperforms the current extractive state of the art by a considerable margin.

show abstract

Section: Related Workmentioning

confidence: 99%

A Hierarchical Structured Self-Attentive Model for Extractive Document Summarization (HSSAS)

2018

View full text Add to dashboard Cite

show abstract

“…Deep neural network models have had great success in machine learning, particularly in various tasks of NLP. For example, automatic summarization [13], question answering [14], machine translation [15], words and phrases distributed representations [16], sentiment analysis [6] and other tasks. Kim [6] proposed a deep learning model for sentiment analysis using CNNs with different convolutional filter sizes.…”

Section: Background and Related Workmentioning

confidence: 99%

A Combined CNN and LSTM Model for Arabic Sentiment Analysis

Alayba

Palade

England

et al. 2018

Lecture Notes in Computer Science

141

View full text Add to dashboard Cite

Deep neural networks have shown good data modelling capabilities when dealing with challenging and large datasets from a wide range of application areas. Convolutional Neural Networks (CNNs) offer advantages in selecting good features and Long Short-Term Memory (LSTM) networks have proven good abilities of learning sequential data. Both approaches have been reported to provide improved results in areas such image processing, voice recognition, language translation and other Natural Language Processing (NLP) tasks. Sentiment classification for short text messages from Twitter is a challenging task, and the complexity increases for Arabic language sentiment classification tasks because Arabic is a rich language in morphology. In addition, the availability of accurate pre-processing tools for Arabic is another current limitation, along with limited research available in this area. In this paper, we investigate the benefits of integrating CNNs and LSTMs and report obtained improved accuracy for Arabic sentiment analysis on different datasets. Additionally, we seek to consider the morphological diversity of particular Arabic words by using different sentiment classification levels.

show abstract

“…In another study about NLP, stacked autoencoder applied on unsupervised extractive summarization of email with excellent performance. Summaries are highly informative and semantically similar to human abstracts [14]. On the robotic area, Ian Lenz, Honglak Lee and Ashutosh Saxena are using deep learning methods in order to solve the problem of detecting robotic grasps in an RGB-D view of a scene containing objects [15].…”

Section: Deep Learning Algorithmsmentioning

confidence: 99%

Stacked Autoencoder Method for Fabric Defect Detection

Şeker

Yüksek

2017

Cumhuriyet Science Journal

View full text Add to dashboard Cite

Abstract. The fabric defect detection has crucial importance in terms of sectoral quality. As fabric defection stage, accordingly the growing market volume and production capacity, detection via human vision has caused largely time-wasting and success rate decreasing until 60%. Due to a fabric has unique texture, there is necessity for it to work on separately from other images types while extracting its features. Features are vital material of computer vision especially classification problems. Hence, extracting right features is the most significant stage of error detection. This purpose in mind on this study, deep learning which distinguishes with multi-layer architectures and reveals high achievement on image and speech procession recent years by self-feature extraction is applied to fabric defect detection. Stacked autoencoder -a deep learning method-that aimed to represent input data via compression or decompression is tried to detect defect of fabrics and it gained acceptable success. The principal aim of this study is to increase achievement of feature extraction by tuning up the input value and hyper parameters autoencoder. Thanks to the fine tuning of hyper-parameters of deep model, we have 96% success rate on our own dataset. Keywords: Deep learning, fabric defect detection, autoencoder, hyper parameter Kumaş Hatası Tespiti için Yığınlanmış Oto-kodlayıcı YöntemiÖzet. Kumaş hatası tespiti sektörel kalite açısından önem arz etmektedir. Bu hataların tespitinde, gelişen pazar hacmi ve üretim kapasitelerinin büyüklüğü sebebiyle insan görüsü ile tespit, büyük oranda zaman kaybına ve hata tespit oranının %60 seviyelerine kadar düşmesine sebep olmaktadır. Bu bağlamda daha yüksek başarım elde edebilmek için görüntü işleme alanında bir çok yöntem denenmiştir. Kumaşın kendine has bir dokusunun olması sebebiyle, öznitelikleri çıkarılırken diğer görüntü türlerinden ayrı olarak incelenmesi gerektirmektedir. Öznitelikler bilgisayarlı görmede özellikle sınıflandırma problemlerinde ham madde olmaktadır. Bu yüzden doğru öznitelikleri çıkarmak, hata tespitinde en önemli aşamadır. Bu amaç doğrultusunda, çoklu-katman mimarisi ve kendi özniteliklerini çıkararak son yıllarda görüntü ve ses işleme alanında büyük başarılar getirmesi ile öne çıkan derin öğrenme kumaş hatası tespitine uygulanmıştır. Giriş verisini sıkıştırma ya da genişletme ile temsil eden yığınlı oto-kodlayıcılar -bir derin öğrenme yöntemi-kumaş hatası tespitinde denenmiş ve kabul edilebilir başarılar elde edilmiştir. Çalışmanın asıl amacı oto kodlayıcının hiper parametreleri ve giriş değeri ile oynamalar yaparak öznitelik çıkarımı başarısını artırmaktır. Derin modelin hiper parametrelerin ince ayarıyla, kendi veri setimizde %96'lık bir başarı oranı elde ettik.Anahtar Kelimeler: Derin öğrenme, kumaş hatası tespiti, oto-kodlayıcı, hiper parametre

show abstract

Text summarization using unsupervised deep learning

Cited by 217 publications

References 19 publications

A Hierarchical Structured Self-Attentive Model for Extractive Document Summarization (HSSAS)

A Hierarchical Structured Self-Attentive Model for Extractive Document Summarization (HSSAS)

A Combined CNN and LSTM Model for Arabic Sentiment Analysis

Stacked Autoencoder Method for Fabric Defect Detection

Contact Info

Product

Resources

About