Search

Extractive text summarization selects the most important sentences from a document, preserves their original meaning, and produces an objective and fact-based summary. It is faster and less computationally intensive than abstract summarization techniques. Learning cross-sentence relationships is crucial for extractive text summarization. However, most of the language models currently in use process text data sequentially, which makes it difficult to capture such inter-sentence relations, especially in long documents. This paper proposes an extractive summarization model based on the graph neural network (GNN) to address this problem. The model effectively represents cross-sentence relationships using a graph-structured document representation. In addition to sentence nodes, we introduce two nodes with different granularity in the graph structure, words and topics, which bring different levels of semantic information. The node representations are updated by the graph attention network (GAT). The final summary is obtained using the binary classification of the sentence nodes. Our text summarization method was demonstrated to be highly effective, as supported by the results of our experiments on the CNN/DM and NYT datasets. To be specific, our approach outperformed baseline models of the same type in terms of ROUGE scores on both datasets, indicating the potential of our proposed model for enhancing text summarization tasks. [ABSTRACT FROM AUTHOR]

Deep learning has developed rapidly in recent years, attracting the attention of numerous researchers. Since a wide range of topics are covered in this field, we are wondering what topics researchers have concerned about. However, after investigation, we find that very few researchers have paid attention to this demand. In this paper, we conduct a large-scale study to analyze the questions faced by deep learning developers. We use Stack Overflow, one of the largest question&answer sites, as our data source, and extract 32969 posts about deep learning as our studied dataset. After filtering, augmenting and pre-processing the post datasets from Stack Overflow, we use the latent Dirichlet allocation (LDA) topic model to summarize 30 topics based on their text content. In addition, we measure the difficulty and popularity of each topic, compare the different issues faced by different deep learning frameworks, and analyze the development trend of each topic. Our main results are as follows: (1) developers ask a broad spectrum of questions about deep-learning, ranging from Data Shape to Object Detection; (2) Gradient Propagation is the most popular among all the topics and (3) Object Detection is the most difficult; (4) issues of Package Installation, Code Understanding and Method Introduction are common in the current different deep learning frameworks; (5) there are three trends in these topics, e.g., a significant rising trend is found in the number of discussion on Data Shape. Finally, based on our research findings, we make some targeted and valuable suggestions for developers, researchers, educators, and framework providers of deep learning. [ABSTRACT FROM AUTHOR]