Search

Fine-tuning a pre-trained sequence-to-sequence-based language model has significantly advanced the field of abstractive summarization. However, the early models of abstractive summarization were limited by the gap between training and inference, and they did not fully utilize the potential of the language model. Recent studies have introduced a two-stage framework that allows the second-stage model to re-rank the candidate summary generated by the first-stage model, to resolve these limitations. In this study, we point out that the supervision method performed in the existing re-ranking model of the two-stage abstractive summarization framework cannot learn detailed and complex information of the data. In addition, we present the problem of positional bias in the existing encoder–decoder-based re-ranking model. To address these two limitations, this study proposes a hierarchical supervision method that jointly performs summary and sentence-level supervision. For sentence-level supervision, we designed two sentence-level loss functions: intra- and inter-intra-sentence ranking losses. Compared to the existing abstractive summarization model, the proposed method exhibited a performance improvement for both the CNN/DM and XSum datasets. The proposed model outperformed the baseline model under a few-shot setting. [ABSTRACT FROM AUTHOR]

Text summarization (TS) plays a crucial role in natural language processing (NLP) by automatically condensing and capturing key information from text documents. Its significance extends to diverse fields, including engineering, healthcare, and others, where it offers substantial time and resource savings. However, manual summarization is a laborious task, prompting the need for automated text summarization systems. In this paper, we propose a novel strategy for extractive summarization that leverages a generative adversarial network (GAN)-based method and Bidirectional Encoder Representations from Transformers (BERT) word embedding. BERT, a transformer-based architecture, processes sentence bidirectionally, considering both preceding and following words. This contextual understanding empowers BERT to generate word representations that carry a deeper meaning and accurately reflect their usage within specific contexts. Our method adopts a generator and discriminator within the GAN framework. The generator assesses the likelihood of each sentence in the summary while the discriminator evaluates the generated summary. To extract meaningful features in parallel, we introduce three dilated convolution layers in the generator and discriminator. Dilated convolution allows for capturing a larger context and incorporating long-range dependencies. By introducing gaps between filter weights, dilated convolution expands the receptive field, enabling the model to consider a broader context of words. To encourage the generator to explore diverse sentence combinations that lead to high-quality summaries, we introduce various noises to each document within our proposed GAN. This approach allows the generator to learn from a range of sentence permutations and select the most suitable ones. We evaluate the performance of our proposed model using the CNN/Daily Mail dataset. The results, measured using the ROUGE metric, demonstrate the superiority of our approach compared to other tested methods. This confirms the effectiveness of our GAN-based strategy, which integrates dilated convolution layers, BERT word embedding, and a generator-discriminator framework in achieving enhanced extractive summarization performance. [ABSTRACT FROM AUTHOR]

(1) Background: Information overload challenges decision-making in the Industry 4.0 era. While Natural Language Processing (NLP), especially Automatic Text Summarization (ATS), offers solutions, issues with factual accuracy persist. This research bridges cognitive neuroscience and NLP, aiming to improve model interpretability. (2) Methods: This research examined four fact extraction techniques: dependency relation, named entity recognition, part-of-speech tagging, and TF-IDF, in order to explore their correlation with human EEG signals. Representational Similarity Analysis (RSA) was applied to gauge the relationship between language models and brain activity. (3) Results: Named entity recognition showed the highest sensitivity to EEG signals, marking the most significant differentiation between factual and non-factual words with a score of −0.99. The dependency relation followed with −0.90, while part-of-speech tagging and TF-IDF resulted in 0.07 and −0.52, respectively. Deep language models such as GloVe, BERT, and GPT-2 exhibited noticeable influences on RSA scores, highlighting the nuanced interplay between brain activity and these models. (4) Conclusions: Our findings emphasize the crucial role of named entity recognition and dependency relations in fact extraction and demonstrate the independent effects of different models and TOIs on RSA scores. These insights aim to refine algorithms to reflect human text processing better, thereby enhancing ATS models' factual integrity. [ABSTRACT FROM AUTHOR]

The surge of digital legal documents has significantly expanded their usage. This has resulted in the sheer number of papers being used by various members of the judiciary, as well as by advocates and judicial officers. It can be incredibly challenging to keep up with all of them. Over four crore cases are still pending in Indian courts, and manually reviewing them can be a tedious and time-consuming process\cite{r1}. As machine learning has advanced, various text summarization models have been created to help legal professionals manage their documents. Due to the lack of publicly accessible datasets, it is difficult to fine-tune domain-independent models for Indian legal systems. The methodology proposed in this paper seeks to improve the overall performance of these models, and it also explores Indian legal documents' summarization techniques. In addition, this research also provides a study of the several summarization methods in-depth that have been on Indian legal documents, including PEGASUS, Bidirectional Auto- Regressive transformers (BART), TextRank, and Bidirectional Encoder Representations from Transformers (BERT). Through the process of extractive and abstract summarization, BART and PEGASUS will be able to gain a deeper understanding of the text normalization process. The outcomes of the text normalization process are evaluated by experts using the ROUGE metrics and multiple parameters. It shows that the proposed approach can work well in legal texts that have domain-independent frameworks. [ABSTRACT FROM AUTHOR]

The crucial aspect of extractive document summarization lies in understanding the interrelations between sentences. Documents inherently comprise a multitude of sentences, and sentence-level models frequently fail to consider the relationships between distantly-placed sentences, resulting in the omission of significant information in the summary. Moreover, information within documents tends to be distributed sparsely, challenging the efficacy of sentence-level models. In the realm of heterogeneous graph neural networks, it has been observed that semantic nodes with varying levels of granularity encapsulate distinct semantic connections. Initially, the incorporation of edge features into the computation of dynamic graph attention networks is performed to account for node relationships. Subsequently, given the multiplicity of topics in a document or a set of documents, a topic model is employed to extract topic-specific features and the probability distribution linking these topics with sentence nodes. Last but not least, the model defines nodes with different levels of granularity--ranging from documents and topics to sentences--and these various nodes necessitate different propagation widths and depths for capturing intricate relationships in the information being disseminated. Adaptive measures are taken to learn the importance and correlation between nodes of different granularities in terms of both width and depth. Experimental evidence from two benchmark datasets highlights the superior performance of the proposed model, as assessed by ROUGE metrics, in comparison to existing approaches, even in the absence of pre-trained language models. Additionally, an ablation study confirms the positive impact of each individual module on the model's ROUGE scores. [ABSTRACT FROM AUTHOR]

Automatic summarization is attracting increasing attention as one of the most promising research areas. This technology has been tried in various real-world applications in recent years and achieved a good response. However, the applicability of conventional evaluation metrics cannot keep up with rapidly evolving summarization task formats and ensuing indicator. After recent years of research, automatic summarization task requires not only readability and fluency, but also informativeness and consistency. Diversified application scenarios also bring new challenges both for generative language models and evaluation metrics. In this review, we analysis and specifically focus on the difference between the task format and the evaluation metrics. [ABSTRACT FROM AUTHOR]

The increasing amount of legal information available online is overwhelming for both citizens and legal professionals, making it difficult and time-consuming to find relevant information and keep up with the latest legal developments. Automatic text summarization techniques can be highly beneficial as they save time, reduce costs, and lessen the cognitive load of legal professionals. However, applying these techniques to legal documents poses several challenges due to the complexity of legal documents and the lack of needed resources, especially in linguistically under-resourced languages, such as the Greek language. In this paper, we address automatic summarization of Greek legal documents. A major challenge in this area is the lack of suitable datasets in the Greek language. In response, we developed a new metadata-rich dataset consisting of selected judgments from the Supreme Civil and Criminal Court of Greece, alongside their reference summaries and category tags, tailored for the purpose of automated legal document summarization. We also adopted several state-of-the-art methods for abstractive and extractive summarization and conducted a comprehensive evaluation of the methods using both human and automatic metrics. Our results: (i) revealed that, while extractive methods exhibit average performance, abstractive methods generate moderately fluent and coherent text, but they tend to receive low scores in relevance and consistency metrics; (ii) indicated the need for metrics that capture better a legal document summary's coherence, relevance, and consistency; (iii) demonstrated that fine-tuning BERT models on a specific upstream task can significantly improve the model's performance. [ABSTRACT FROM AUTHOR]

Abstractive dialogue summarization requires distilling and rephrasing key information from noisy multi-speaker documents. Combining pre-trained language models with input augmentation techniques has recently led to significant research progress. However, existing solutions still struggle to select relevant chat segments, primarily relying on open-domain and unsupervised annotators not tailored to the actual needs of the summarization task. In this paper, we propose DearWatson , a task-aware utterance-level annotation framework for improving the effectiveness and interpretability of pre-trained dialogue summarization models. Precisely, we learn relevant utterances in the source document and mark them with special tags, that then act as supporting evidence for the generated summary. Quantitative experiments are conducted on two datasets made up of real-life messenger conversations. The results show that DearWatson allows model attention to focus on salient tokens, achieving new state-of-the-art results in three evaluation metrics, including semantic and factuality measures. Human evaluation proves the superiority of our solution in semantic consistency and recall. Finally, extensive ablation studies confirm each module's importance, also exploring different annotation strategies and parameter-efficient fine-tuning of large generative language models. [ABSTRACT FROM AUTHOR]

With the development of pre-trained language models and large-scale datasets, automatic text summarization has attracted much attention from the community of natural language processing, but the progress of automatic summarization evaluation has stagnated. Although there have been efforts to improve automatic summarization evaluation, ROUGE has remained one of the most popular metrics for nearly 20 years due to its competitive evaluation performance. However, ROUGE is not perfect, there are studies have shown that it is suffering from inaccurate evaluation of abstractive summarization and limited diversity of generated summaries, both caused by lexical bias. To avoid the bias of lexical similarity, more and more meaningful embedding-based metrics have been proposed to evaluate summaries by measuring semantic similarity. Due to the challenge of accurately measuring semantic similarity, none of them can fully replace ROUGE as the default automatic evaluation toolkit for text summarization. To address the aforementioned problems, we propose a compromise evaluation framework (ROUGE-SEM) for improving ROUGE with semantic information, which compensates for the lack of semantic awareness through a semantic similarity module. According to the differences in semantic similarity and lexical similarity, summaries are classified into four categories for the first time, including good-summary, pearl-summary, glass-summary, and bad-summary. In particular, the back-translation technique is adopted to rewrite pearl-summary and glass-summary that are inaccurately evaluated by ROUGE to alleviate lexical bias. Through this pipeline framework, summaries are first classified by candidate summary classifier, then rewritten by categorized summary rewriter, and finally scored by rewritten summary scorer, which are efficiently evaluated in a manner consistent with human behavior. When measured using Pearson, Spearman, and Kendall rank coefficients, our proposal achieves comparable or higher correlations with human judgments than several state-of-the-art automatic summarization evaluation metrics in dimensions of coherence, consistency, fluency, and relevance. This also suggests that improving ROUGE with semantics is a promising direction for automatic summarization evaluation. [ABSTRACT FROM AUTHOR]