An abstractive text summarization technique using transformer model with self-attention mechanism

Abstract

The realm of scientific text summarization has experienced remarkable progress due to the availability of annotated brief summaries and ample data. However, the utilization of multiple input modalities, such as videos and audio, has yet to be thoroughly explored. At present, scientific multimodal-input-based text summarization systems tend to employ longer target summaries like abstracts, leading to an underwhelming performance in the task of text summarization.In this paper, we deal with a novel task of extreme abstractive text summarization (aka TL;DR generation) by leveraging multiple input modalities. To this end, we introduce mTLDR, a first-of-itskind dataset for the aforementioned task, comprising videos, audio, and text, along with both author-composed summaries and expertannotated summaries. The mTLDR dataset accompanies a total of 4, 182 instances collected from various academic conference proceedings, such as ICLR, ACL, and CVPR. Subsequently, we present mTLDRgen, an encoder-decoder-based model that employs a novel dual-fused hyper-complex Transformer combined with a Wasserstein Riemannian Encoder Transformer, to dexterously capture the intricacies between different modalities in a hyper-complex latent geometric space. The hyper-complex Transformer captures the intrinsic properties between the modalities, while the Wasserstein Riemannian Encoder Transformer captures the latent structure of the modalities in the latent space geometry, thereby enabling the model to produce diverse sentences. mTLDRgen outperforms 20 baselines on mTLDR as well as another non-scientific dataset (How2) across three Rouge-based evaluation measures. Furthermore, based on the qualitative metrics, BERTScore and FEQA, and human evaluations, we demonstrate that the summaries generated by mTLDRgen are fluent and congruent to the original source material. CCS CONCEPTS• >Computing methodologies → Natural language processing.

Enhancing Text Summarization in a Limited Data and Long Sample Setting

“…Abstractive methods are also good at incorporating information that is not explicitly mentioned but implied or essential for understanding the context. Different studies and benchmarks reported neural abstractive summarization models' superior performance in capturing the essence of the text and producing more coherent and contextually relevant summaries ( [31], [32], [33], [34]).…”

Automatic Text Summarization for Indian Languages

“…In 2022, Sumanlata et al [2] proposed feature extraction-based extractive text summarization method using advanced optimization techniques.The proposed approach achieved better precision, recall, f-measure, and Rouge of 91.97%, 94.026%, 92.36%, and 77.5%, respectively for varying the training percentage. In 2023, Sandeep et al [3] proposed extractive text summarization using a transformer and self-attention model. They achieved 48.50% with minimized training loss.…”

Improving ROUGE‐1 by 6%: A novel multilingual transformer for abstractive news summarization

“…In this paper, we have proposed a fine-tuned model for abstractive news summarization on the Inshorts News English dataset. 36 The proposed model is fine-tuned on a Google-mT5 tokenizer. The fine-tuned framework employed for this endeavor comprises Transformers 4.32.1, Pytorch 2.1.0, Datasets 2.12.0, and Tokenizers 0.13.3, creating a robust and versatile environment for model development.…”