{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2025,8,2]],"date-time":"2025-08-02T17:27:35Z","timestamp":1754155655128,"version":"3.41.2"},"reference-count":65,"publisher":"Emerald","issue":"4","license":[{"start":{"date-parts":[[2025,3,6]],"date-time":"2025-03-06T00:00:00Z","timestamp":1741219200000},"content-version":"tdm","delay-in-days":0,"URL":"https:\/\/www.emerald.com\/insight\/site-policies"}],"content-domain":{"domain":[],"crossmark-restriction":false},"short-container-title":["IMDS"],"published-print":{"date-parts":[[2025,3,17]]},"abstract":"<jats:sec><jats:title content-type=\"abstract-subheading\">Purpose<\/jats:title><jats:p>Recently, due to the practicability in several domains, generative adversarial network (GAN) has successfully been adopted in the field of natural language generation (NLG). The purpose of this paper focuses on improving the quality of text and generating sequences similar to human writing for several real applications.<\/jats:p><\/jats:sec><jats:sec><jats:title content-type=\"abstract-subheading\">Design\/methodology\/approach<\/jats:title><jats:p>A novel model, GAN<jats:sup>2<\/jats:sup>, is developed based on a GAN with dual adversarial architecture. We train the generator by an internal discriminator with a beam search technique to improve the quality of generated sequences. Then, we enhance the generator with an external discriminator to optimize and strengthen the learning process of sequence generation.<\/jats:p><\/jats:sec><jats:sec><jats:title content-type=\"abstract-subheading\">Findings<\/jats:title><jats:p>The proposed GAN<jats:sup>2<\/jats:sup> model could be utilized in widespread applications, such as chatbots, machine translation and image description. By the proposed dual adversarial structure, we significantly improve the quality of the generated text. The average and top-1 metrics, such as NLL, BLEU and ROUGE, are used to measure the generated sentences from the GAN<jats:sup>2<\/jats:sup> model over all baselines. Several experiments are conducted to demonstrate the performance and superiority of the proposed model compared with the state-of-the-art methods on numerous evaluation metrics.<\/jats:p><\/jats:sec><jats:sec><jats:title content-type=\"abstract-subheading\">Originality\/value<\/jats:title><jats:p>Generally, reward sparsity and mode collapse are two main challenging issues when adopt GAN to real NLG applications. In this study, GAN<jats:sup>2<\/jats:sup> exploits a dual adversarial architecture which facilitates the learning process in the early training stage for solving the problem of reward sparsity. The occurrence of mode collapse also could be reduced in the later training stage with the introduced comparative discriminator by avoiding high rewards for training in a specific mode. Furthermore, the proposed model is applied to several synthetic and real datasets to show the practicability and exhibit great generalization with all discussed metrics.<\/jats:p><\/jats:sec>","DOI":"10.1108\/imds-05-2024-0435","type":"journal-article","created":{"date-parts":[[2025,3,5]],"date-time":"2025-03-05T06:34:43Z","timestamp":1741156483000},"page":"1279-1305","source":"Crossref","is-referenced-by-count":0,"title":["A dual adversarial structure of\u00a0generative adversarial network for\u00a0nature language generation"],"prefix":"10.1108","volume":"125","author":[{"given":"Kuen-Liang","family":"Sue","sequence":"first","affiliation":[]},{"ORCID":"https:\/\/orcid.org\/0000-0001-7153-0410","authenticated-orcid":false,"given":"Yi-Cheng","family":"Chen","sequence":"additional","affiliation":[]}],"member":"140","published-online":{"date-parts":[[2025,3,6]]},"reference":[{"key":"key2025031505573908000_ref001","doi-asserted-by":"publisher","first-page":"17043","DOI":"10.1109\/access.2022.3146405","article-title":"Adversarial machine learning in text processing: a literature survey","volume":"10","year":"2022","journal-title":"IEEE Access"},{"first-page":"214","article-title":"Wasserstein generative adversarial networks","year":"2017","key":"key2025031505573908000_ref002"},{"issue":"4","key":"key2025031505573908000_ref003","doi-asserted-by":"publisher","first-page":"431","DOI":"10.1017\/s1351324907004664","article-title":"Automatic generation of weather forecast texts using comprehensive probabilistic generation-space models","volume":"14","year":"2008","journal-title":"Natural Language Engineering"},{"key":"key2025031505573908000_ref004","first-page":"1877","article-title":"Language models are few-shot learners","volume":"33","year":"2020","journal-title":"Advances in Neural Information Processing Systems"},{"key":"key2025031505573908000_ref005","doi-asserted-by":"publisher","first-page":"17","DOI":"10.3115\/1610163.1610168","article-title":"Stochastic realisation ranking for a free word order language","year":"2007"},{"year":"2021","key":"key2025031505573908000_ref006","article-title":"Evaluation of text generation: a survey"},{"issue":"1","key":"key2025031505573908000_ref007","doi-asserted-by":"publisher","DOI":"10.1088\/2632-2153\/ad1f77","article-title":"Ten years of generative adversarial nets (GANs): a survey of the state-of-the-art","volume":"5","year":"2024","journal-title":"Machine Learning: Science and Technology"},{"year":"2017","key":"key2025031505573908000_ref008","article-title":"Maximum-likelihood augmented discrete generative adversarial networks"},{"key":"key2025031505573908000_ref009","unstructured":"Chen, X., Duan, Y., Houthooft, R., Schulman, J., Sutskever, I. and Abbeel, P. (2016), \u201cInfoGAN: interpretable representation learning by information maximizing generative adversarial nets\u201d, in Advances in Neural Information Processing Systems, Vol.\u00a029."},{"key":"key2025031505573908000_ref010","doi-asserted-by":"publisher","first-page":"72716","DOI":"10.1109\/access.2019.2919974","article-title":"Keyphrase enhanced diverse beam search: a content-introducing approach to neural text generation","volume":"7","year":"2019","journal-title":"IEEE Access"},{"key":"key2025031505573908000_ref011","doi-asserted-by":"publisher","first-page":"125","DOI":"10.1016\/j.neucom.2018.12.092","article-title":"Customizable text generation via conditional text generative adversarial network","volume":"416","year":"2020","journal-title":"Neurocomputing"},{"first-page":"1724","article-title":"Learning phrase representations using RNN encoder\u2013decoder for statistical machine translation","year":"2014","key":"key2025031505573908000_ref012"},{"first-page":"4171","article-title":"BERT: pre-training of deep bidirectional transformers for language understanding","year":"2019","key":"key2025031505573908000_ref013"},{"first-page":"13063","article-title":"Unified Language model pre-training for Natural Language understanding and generation","year":"2019","key":"key2025031505573908000_ref014"},{"year":"2017","key":"key2025031505573908000_ref015","article-title":"Real-valued (medical) time series generation with recurrent conditional GANs"},{"year":"2018","key":"key2025031505573908000_ref016","article-title":"MaskGAN: better text generation via filling in the _______"},{"key":"key2025031505573908000_ref017","doi-asserted-by":"publisher","first-page":"1","DOI":"10.18653\/v1\/2023.inlg-main.1","article-title":"Guided beam search to improve generalization in low-resource data-to-text generation","year":"2023"},{"key":"key2025031505573908000_ref018","doi-asserted-by":"publisher","first-page":"65","DOI":"10.1613\/jair.5477","article-title":"Survey of the state of the art in Natural Language Generation: core tasks, applications and evaluation","volume":"61","year":"2018","journal-title":"Journal of Artificial Intelligence Research"},{"first-page":"2672","article-title":"Generative adversarial nets","year":"2014","key":"key2025031505573908000_ref019"},{"year":"2014","key":"key2025031505573908000_ref020","article-title":"Generating sequences with recurrent neural networks"},{"key":"key2025031505573908000_ref021","doi-asserted-by":"publisher","first-page":"5141","DOI":"10.1609\/aaai.v32i1.11957","article-title":"Long text generation via adversarial training with leaked information","year":"2018"},{"key":"key2025031505573908000_ref022","doi-asserted-by":"publisher","first-page":"5087","DOI":"10.18653\/v1\/2021.emnlp-main.415","article-title":"Exposure bias versus self-recovery: are distortions really incremental for autoregressive text generation?","year":"2021"},{"issue":"8","key":"key2025031505573908000_ref023","doi-asserted-by":"publisher","first-page":"1735","DOI":"10.1162\/neco.1997.9.8.1735","article-title":"Long short-term memory","volume":"9","year":"1997","journal-title":"Neural Computation"},{"key":"key2025031505573908000_ref024","doi-asserted-by":"publisher","first-page":"2525","DOI":"10.1145\/3503161.3547854","article-title":"SingGAN: generative adversarial network for high-fidelity singing voice generation","year":"2022"},{"year":"2015","key":"key2025031505573908000_ref025","article-title":"How (not) to train your generative model: scheduled sampling, likelihood Adversary?"},{"key":"key2025031505573908000_ref026","doi-asserted-by":"publisher","first-page":"5967","DOI":"10.1109\/cvpr.2017.632","article-title":"Image-to-Image translation with conditional adversarial networks","year":"2017"},{"issue":"12","key":"key2025031505573908000_ref027","doi-asserted-by":"publisher","first-page":"1","DOI":"10.1145\/3571730","article-title":"Survey of hallucination in Natural Language Generation","volume":"55","year":"2023","journal-title":"ACM Computing Surveys"},{"first-page":"3675","article-title":"Hierarchical deep reinforcement learning: integrating temporal abstraction and intrinsic motivation","year":"2016","key":"key2025031505573908000_ref028"},{"year":"2016","key":"key2025031505573908000_ref029","article-title":"GANS for sequences of discrete elements with the gumbel-softmax distribution"},{"first-page":"11891","article-title":"Generative cooperative networks for natural language generation","year":"2022","key":"key2025031505573908000_ref030"},{"first-page":"170","article-title":"Forest-based statistical sentence generation","year":"2000","key":"key2025031505573908000_ref031"},{"key":"key2025031505573908000_ref032","doi-asserted-by":"publisher","first-page":"133","DOI":"10.1007\/978-3-031-09342-5_13","article-title":"TTS-GAN: a transformer-based time-series generative adversarial network","year":"2022"},{"key":"key2025031505573908000_ref033","unstructured":"Lin, C. (2004), \u201cROUGE: a package for automatic evaluation of summaries\u201d, in Text Summarization Branches Out, pp.\u00a074-81."},{"key":"key2025031505573908000_ref034","doi-asserted-by":"publisher","first-page":"740","DOI":"10.1007\/978-3-319-10602-1_48","article-title":"Microsoft COCO: common objects in context","year":"2014"},{"first-page":"3155","article-title":"Adversarial ranking for Language Generation","year":"2017","key":"key2025031505573908000_ref035"},{"year":"2019","key":"key2025031505573908000_ref036","article-title":"LSTM vs GRU vs bidirectional RNN for script generation"},{"key":"key2025031505573908000_ref037","article-title":"Introducing Meta Llama 3: the most capable openly available LLM to date","author":"Meta","year":"2024","journal-title":"Meta post"},{"year":"2014","key":"key2025031505573908000_ref038","article-title":"Conditional generative adversarial nets"},{"key":"key2025031505573908000_ref039","doi-asserted-by":"publisher","first-page":"364","DOI":"10.1016\/j.neucom.2021.03.097","article-title":"DGSAN: discrete generative self-adversarial network","volume":"448","year":"2021","journal-title":"Neurocomputing"},{"year":"2019","key":"key2025031505573908000_ref040","article-title":"RelGAN: relational generative adversarial networks for text generation"},{"key":"key2025031505573908000_ref041","doi-asserted-by":"publisher","first-page":"131","DOI":"10.1007\/978-3-031-43412-9_8","article-title":"Adversarial imitation learning with controllable rewards for text generation","year":"2023"},{"article-title":"GPT-4 technical report","year":"2023","author":"OpenAI","key":"key2025031505573908000_ref042"},{"key":"key2025031505573908000_ref043","doi-asserted-by":"publisher","first-page":"311","DOI":"10.3115\/1073083.1073135","article-title":"Bleu: a method for automatic evaluation of machine translation","year":"2002"},{"issue":"8","key":"key2025031505573908000_ref044","first-page":"9","article-title":"Language models are unsupervised multitask learners","volume":"1","year":"2019","journal-title":"OpenAI blog"},{"volume-title":"Building Natural Language Generation Systems","year":"2009","key":"key2025031505573908000_ref045"},{"issue":"6","key":"key2025031505573908000_ref046","doi-asserted-by":"publisher","first-page":"1","DOI":"10.1016\/j.ipm.2020.102217","article-title":"Adversarial text generation with context adapted global knowledge and a self-attentive discriminator","volume":"57","year":"2020","journal-title":"Information Processing and Management"},{"first-page":"8555","article-title":"Discriminative adversarial search for abstractive summarization","year":"2020","key":"key2025031505573908000_ref047"},{"key":"key2025031505573908000_ref048","first-page":"26585","article-title":"To beam or not to beam: that is a question of cooperation for language gans","volume":"34","year":"2021","journal-title":"Advances in Neural Information Processing Systems"},{"key":"key2025031505573908000_ref049","doi-asserted-by":"publisher","first-page":"4361","DOI":"10.24963\/ijcai.2018\/606","article-title":"Toward diverse text generation with inverse reinforcement learning","year":"2018"},{"first-page":"3104","article-title":"Sequence to sequence learning with neural networks","year":"2014","key":"key2025031505573908000_ref050"},{"first-page":"5998","article-title":"Attention is all you need","year":"2017","key":"key2025031505573908000_ref051"},{"key":"key2025031505573908000_ref052","doi-asserted-by":"publisher","first-page":"11880","DOI":"10.1109\/access.2020.2966291","article-title":"Text generation service model based on truth-guided SeqGAN","volume":"8","year":"2020","journal-title":"IEEE Access"},{"key":"key2025031505573908000_ref053","doi-asserted-by":"publisher","first-page":"14067","DOI":"10.1609\/aaai.v35i16.17656","article-title":"TextGAIL: generative adversarial imitation learning for text generation","year":"2021"},{"first-page":"14067","article-title":"Textgail: generative adversarial imitation learning for text generation","year":"2021","key":"key2025031505573908000_ref054"},{"key":"key2025031505573908000_ref055","doi-asserted-by":"publisher","first-page":"2721","DOI":"10.1145\/3580305.3599370","article-title":"Graph contrastive learning with generative adversarial network","year":"2023"},{"issue":"3","key":"key2025031505573908000_ref056","doi-asserted-by":"publisher","first-page":"1443","DOI":"10.1109\/tcsvt.2021.3074032","article-title":"Local and global perception generative adversarial network for facial expression synthesis","volume":"32","year":"2022","journal-title":"IEEE Transactions on Circuits and Systems for Video Technology"},{"first-page":"5753","article-title":"XLNet: generalized autoregressive pretraining for language understanding","year":"2019","key":"key2025031505573908000_ref057"},{"key":"key2025031505573908000_ref058","doi-asserted-by":"publisher","first-page":"105217","DOI":"10.1109\/access.2020.2993928","article-title":"FGGAN: feature-guiding generative adversarial networks for text generation","volume":"8","year":"2020","journal-title":"IEEE Access"},{"issue":"1","key":"key2025031505573908000_ref059","doi-asserted-by":"publisher","first-page":"A292","DOI":"10.1137\/18m1225409","article-title":"Physics-informed generative adversarial networks for stochastic differential equations","volume":"42","year":"2020","journal-title":"SIAM Journal on Scientific Computing"},{"key":"key2025031505573908000_ref060","doi-asserted-by":"publisher","first-page":"9466","DOI":"10.1609\/aaai.v34i05.6490","article-title":"Meta-CoTGAN: a Meta cooperative training paradigm for improving adversarial text generation","year":"2020"},{"key":"key2025031505573908000_ref061","doi-asserted-by":"publisher","first-page":"2852","DOI":"10.1609\/aaai.v31i1.10804","article-title":"SeqGAN: sequence generative adversarial nets with policy gradient","year":"2017"},{"first-page":"5907","article-title":"Stackgan: text to photo-realistic image synthesis with stacked generative adversarial networks","year":"2017","key":"key2025031505573908000_ref062"},{"year":"2020","key":"key2025031505573908000_ref063","article-title":"Self-adversarial learning with comparative discrimination for text generation"},{"first-page":"1097","article-title":"Texygen: a benchmarking platform for text generation models","year":"2018","key":"key2025031505573908000_ref065"},{"key":"key2025031505573908000_ref064","first-page":"2223","article-title":"Unpaired image-to-image translation using cycle-consistent adversarial networks","volume-title":"Proceedings of the IEEE international conference on computer vision (CVPR 2017)","year":"2017"}],"container-title":["Industrial Management &amp; Data Systems"],"original-title":[],"language":"en","link":[{"URL":"https:\/\/www.emerald.com\/insight\/content\/doi\/10.1108\/IMDS-05-2024-0435\/full\/xml","content-type":"application\/xml","content-version":"vor","intended-application":"text-mining"},{"URL":"https:\/\/www.emerald.com\/insight\/content\/doi\/10.1108\/IMDS-05-2024-0435\/full\/html","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2025,7,24]],"date-time":"2025-07-24T21:52:18Z","timestamp":1753393938000},"score":1,"resource":{"primary":{"URL":"http:\/\/www.emerald.com\/imds\/article\/125\/4\/1279-1305\/1245545"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2025,3,6]]},"references-count":65,"journal-issue":{"issue":"4","published-online":{"date-parts":[[2025,3,6]]},"published-print":{"date-parts":[[2025,3,17]]}},"alternative-id":["10.1108\/IMDS-05-2024-0435"],"URL":"https:\/\/doi.org\/10.1108\/imds-05-2024-0435","relation":{},"ISSN":["0263-5577","1758-5783"],"issn-type":[{"type":"print","value":"0263-5577"},{"type":"electronic","value":"1758-5783"}],"subject":[],"published":{"date-parts":[[2025,3,6]]}}}