{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,7,18]],"date-time":"2026-07-18T19:43:36Z","timestamp":1784403816378,"version":"3.55.0"},"reference-count":46,"publisher":"Institution of Engineering and Technology (IET)","issue":"1","license":[{"start":{"date-parts":[[2025,11,10]],"date-time":"2025-11-10T00:00:00Z","timestamp":1762732800000},"content-version":"vor","delay-in-days":313,"URL":"http:\/\/creativecommons.org\/licenses\/by\/4.0\/"},{"start":{"date-parts":[[2025,1,1]],"date-time":"2025-01-01T00:00:00Z","timestamp":1735689600000},"content-version":"tdm","delay-in-days":0,"URL":"http:\/\/doi.wiley.com\/10.1002\/tdm_license_1.1"}],"funder":[{"DOI":"10.13039\/501100015401","name":"Key Research and Development Projects of Shaanxi Province","doi-asserted-by":"publisher","award":["2022ZDLSF07-07"],"award-info":[{"award-number":["2022ZDLSF07-07"]}],"id":[{"id":"10.13039\/501100015401","id-type":"DOI","asserted-by":"publisher"}]},{"DOI":"10.13039\/501100001809","name":"National Natural Science Foundation of China","doi-asserted-by":"publisher","award":["62272378"],"award-info":[{"award-number":["62272378"]}],"id":[{"id":"10.13039\/501100001809","id-type":"DOI","asserted-by":"publisher"}]}],"content-domain":{"domain":["ietresearch.onlinelibrary.wiley.com"],"crossmark-restriction":true},"short-container-title":["IET Information Security"],"published-print":{"date-parts":[[2025,1]]},"abstract":"<jats:p>Backdoor attacks have significantly threatened the models of natural language processing (NLP). However, most textual backdoor attacks exhibit low levels of stealthiness, making them susceptible to detection and removal by defense strategies. In order to improve the performance and stealthiness of such backdoor attacks, this article introduces a novel backdoor attack named Bad Padding (BPad) based on steganography. BPad employs a word\u2010substitution steganographic method to hide triggers in sentences, thereby generating poisoned data. To ensure a high level of stealthiness for these poisoned samples, BPad developed a word substitution strategy that enhances both the diversity of the substituted words and the contextual coherence of the sentences. BPad also modifies the preprocessing stage by extracting triggers from the sentences and padding them as tokens at the end, effectively amplifying the impact of the trigger and making it easier for the model to learn the shortcut from the trigger to the target label, thereby achieving the injection of a backdoor. This article uses various metrics to present experimental measures of the attack performance and stealthiness of BPad. The results find that BPad achieved competitive results compared to baseline methods in non\u2010defense scenarios and outperforms baseline methods under both training and inference defense. Besides that, the attack samples generated by BPad demonstrate strong stealthiness in terms of semantic coherence, perplexity, and grammaticality.<\/jats:p>","DOI":"10.1049\/ise2\/8880733","type":"journal-article","created":{"date-parts":[[2025,11,11]],"date-time":"2025-11-11T06:31:34Z","timestamp":1762842694000},"update-policy":"https:\/\/doi.org\/10.1002\/crossmark_policy","source":"Crossref","is-referenced-by-count":1,"title":["Bad Padding: A Highly Stealthy Backdoor Attack Using Steganography at the Padding Stage"],"prefix":"10.1049","volume":"2025","author":[{"ORCID":"https:\/\/orcid.org\/0009-0003-3632-5288","authenticated-orcid":false,"given":"Zhuowei","family":"Niu","sequence":"first","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0003-2019-7886","authenticated-orcid":false,"given":"Qindong","family":"Sun","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0002-8983-4049","authenticated-orcid":false,"given":"Kai","family":"Lin","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0002-4391-3913","authenticated-orcid":false,"given":"Mingkai","family":"Ding","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]}],"member":"265","published-online":{"date-parts":[[2025,11,10]]},"reference":[{"key":"e_1_2_14_1_2","doi-asserted-by":"publisher","DOI":"10.18653\/v1\/W17-1101"},{"key":"e_1_2_14_2_2","doi-asserted-by":"publisher","DOI":"10.1016\/j.asoc.2023.110993"},{"key":"e_1_2_14_3_2","doi-asserted-by":"publisher","DOI":"10.1109\/ACCESS.2019.2941376"},{"key":"e_1_2_14_4_2","doi-asserted-by":"publisher","DOI":"10.1145\/3485832.3485837"},{"key":"e_1_2_14_5_2","first-page":"984","volume-title":"Proceedings of the 29th International Conference on Computational LinguisticsInternational Committee on Computational Linguistics","author":"Hy Lu","year":"2022"},{"key":"e_1_2_14_6_2","doi-asserted-by":"publisher","DOI":"10.18653\/v1\/2021.acl-long.431"},{"key":"e_1_2_14_7_2","doi-asserted-by":"crossref","unstructured":"IrtizaS. KhanL. andHamlenK. W. SentMod: Hidden Backdoor Attack on Unstructured Textual Data 2022 IEEE 8th Intl Conference on Big Data Security on Cloud (BigDataSecurity) IEEE Intl Conference on High Performance and Smart Computing (HPSC) and IEEE Intl Conference on Intelligent Data and Security (IDS) 2022 Jinan China IEEE 224\u2013231 https:\/\/doi.org\/10.1109\/BigDataSecurityHPSCIDS54978.2022.00050.","DOI":"10.1109\/BigDataSecurityHPSCIDS54978.2022.00050"},{"key":"e_1_2_14_8_2","unstructured":"AntypasD. UshioA. Camacho-ColladosJ. NevesL. SilvaV. andBarbieriF. Twitter Topic Classification Proceedings of the 29th International Conference on Computational Linguistics 2022 Gyeongju Republic of Korea International Committee on Computational Linguistics 3386\u20133400."},{"key":"e_1_2_14_9_2","doi-asserted-by":"publisher","DOI":"10.1007\/978-981-15-0694-9_41"},{"key":"e_1_2_14_10_2","doi-asserted-by":"publisher","DOI":"10.1007\/s10994-023-06370-5"},{"key":"e_1_2_14_11_2","unstructured":"SunL. Natural Backdoor Attack on Text Data 2021."},{"key":"e_1_2_14_12_2","doi-asserted-by":"publisher","DOI":"10.3390\/math9212829"},{"key":"e_1_2_14_13_2","doi-asserted-by":"publisher","DOI":"10.1016\/j.ins.2022.07.120"},{"key":"e_1_2_14_14_2","doi-asserted-by":"publisher","DOI":"10.18653\/v1\/2021.naacl-main.433"},{"key":"e_1_2_14_15_2","doi-asserted-by":"publisher","DOI":"10.18653\/v1\/2020.acl-main.249"},{"key":"e_1_2_14_16_2","doi-asserted-by":"publisher","DOI":"10.18653\/v1\/2021.emnlp-main.752"},{"key":"e_1_2_14_17_2","doi-asserted-by":"publisher","DOI":"10.1016\/j.cose.2021.102433"},{"key":"e_1_2_14_18_2","doi-asserted-by":"publisher","DOI":"10.18653\/v1\/2021.acl-long.37"},{"key":"e_1_2_14_19_2","doi-asserted-by":"publisher","DOI":"10.18653\/v1\/2021.emnlp-main.374"},{"key":"e_1_2_14_20_2","doi-asserted-by":"publisher","DOI":"10.18653\/v1\/2021.acl-long.377"},{"key":"e_1_2_14_21_2","doi-asserted-by":"crossref","unstructured":"ZhangX. ZhangZ. JiS. andWangT. Trojaning Language Models for Fun and Profit 2021 IEEE European Symposium on Security and Privacy (EuroS&P) 2020 Vienna Austria IEEE 179\u2013197 https:\/\/doi.org\/10.1109\/EuroSP51992.2021.00022.","DOI":"10.1109\/EuroSP51992.2021.00022"},{"key":"e_1_2_14_22_2","doi-asserted-by":"crossref","unstructured":"DuW. YuanT. X. ZhaoH. D. andLiuG. S. NWS: Natural Textual Backdoor Attacks Via Word Substitution ICASSP 2024 - 2024 IEEE International Conference on Acoustics Speech and Signal Processing (ICASSP) 2024 Seoul Korea Republic of IEEE 4680\u20134684 https:\/\/doi.org\/10.1109\/ICASSP48485.2024.10447968.","DOI":"10.1109\/ICASSP48485.2024.10447968"},{"key":"e_1_2_14_23_2","doi-asserted-by":"publisher","DOI":"10.1109\/TASLP.2024.3407571"},{"key":"e_1_2_14_24_2","doi-asserted-by":"crossref","unstructured":"YerlikayaF. A.andBahtiyar\u015e. A Textual Clean-Label Backdoor Attack Strategy Against Spam Detection 2021 14th International Conference on Security of Information and Networks (SIN) 2021 Edinburgh United Kingdom IEEE 1\u20138.","DOI":"10.1109\/SIN54109.2021.9699173"},{"key":"e_1_2_14_25_2","doi-asserted-by":"publisher","DOI":"10.1007\/s11633-022-1377-5"},{"key":"e_1_2_14_26_2","doi-asserted-by":"crossref","unstructured":"LiZ. ZengY. XiaP. LiuL. FuZ. andLiB. Large Language Models are Good Attackers: Efficient and Stealthy Textual Backdoor Attacks 2024.","DOI":"10.2139\/ssrn.5365101"},{"key":"e_1_2_14_27_2","unstructured":"ZhangR. LiH. andWenR. et al.Instruction Backdoor Attacks Against Customized LLMs Proceedings of the 33rd USENIX Conference on Security Symposium 2024 USENIX Association 1849\u20131866."},{"key":"e_1_2_14_28_2","first-page":"100938","volume-title":"Advances in Neural Information Processing Systems. 37","author":"Yang W.","year":"2024"},{"key":"e_1_2_14_29_2","doi-asserted-by":"crossref","unstructured":"ZhaoS. JiaM. andGuoZ. et al.A Survey of Recent Backdoor Attacks and Defenses in Large Language Models 2025.","DOI":"10.36227\/techrxiv.172832726.62863760\/v1"},{"key":"e_1_2_14_30_2","doi-asserted-by":"publisher","DOI":"10.1109\/TDSC.2020.3021407"},{"key":"e_1_2_14_31_2","doi-asserted-by":"publisher","DOI":"10.1016\/j.cviu.2024.104208"},{"key":"e_1_2_14_32_2","doi-asserted-by":"crossref","unstructured":"ZhangM. JiS. CaiH. DongH. ZhangP. andLiY. Audio Steganography Based Backdoor Attack for Speech Recognition Software 2024 IEEE 48th Annual Computers Software and Applications Conference (COMPSAC) 2024 Osaka Japan IEEE 1208\u20131217.","DOI":"10.1109\/COMPSAC61105.2024.00161"},{"key":"e_1_2_14_33_2","doi-asserted-by":"publisher","DOI":"10.1109\/TDSC.2021.3055844"},{"key":"e_1_2_14_34_2","doi-asserted-by":"publisher","DOI":"10.18653\/v1\/2021.emnlp-main.659"},{"key":"e_1_2_14_35_2","doi-asserted-by":"crossref","unstructured":"PeiH. JiaJ. GuoW. LiB. andSongD. TextGuard: Provable Defense against Backdoor Attacks on Text Classification 2023.","DOI":"10.14722\/ndss.2024.24090"},{"key":"e_1_2_14_36_2","doi-asserted-by":"crossref","unstructured":"ZhaiS. ShenQ. andChenX. et al.NCL: Textual Backdoor Defense Using Noise-Augmented Contrastive Learning ICASSP 2023 - 2023 IEEE International Conference on Acoustics Speech and Signal Processing (ICASSP) 2023 Rhodes Island Greece IEEE 1\u20135 https:\/\/doi.org\/10.1109\/ICASSP49357.2023.10095007.","DOI":"10.1109\/ICASSP49357.2023.10095007"},{"key":"e_1_2_14_37_2","doi-asserted-by":"publisher","DOI":"10.1016\/j.neucom.2021.04.105"},{"key":"e_1_2_14_38_2","first-page":"5009","volume-title":"Advances in Neural Information Processing Systems. 35","author":"Cui G.","year":"2022"},{"key":"e_1_2_14_39_2","doi-asserted-by":"publisher","DOI":"10.18653\/v1\/D13-1170"},{"key":"e_1_2_14_40_2","doi-asserted-by":"publisher","DOI":"10.1609\/icwsm.v11i1.14955"},{"key":"e_1_2_14_41_2","first-page":"4171","volume-title":"Proceedings of the 2019 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies, Volume 1 (Long and Short Papers)","author":"Devlin J.","year":"2019"},{"key":"e_1_2_14_42_2","article-title":"Identifying Vulnerabilities in the Machine Learning Model Supply Chain","author":"Gu T.","year":"2017","journal-title":"ArXiv"},{"key":"e_1_2_14_43_2","doi-asserted-by":"publisher","DOI":"10.1145\/3460120.3485370"},{"key":"e_1_2_14_44_2","doi-asserted-by":"crossref","unstructured":"QiJ. YuY. WangL. andLiuJ. K\u2009\u2217-Means: An Effective and Efficient K-Means Clustering Algorithm 2016 IEEE International Conferences on Big Data and Cloud Computing (BDCloud) Social Computing and Networking (SocialCom) Sustainable Computing and Communications (SustainCom) (BDCloud-SocialCom-SustainCom) 2016 Atlanta GA USA IEEE 242\u2013249 https:\/\/doi.org\/10.1109\/BDCloud-SocialCom-SustainCom.2016.46 2-s2.0-85001074602.","DOI":"10.1109\/BDCloud-SocialCom-SustainCom.2016.46"},{"key":"e_1_2_14_45_2","doi-asserted-by":"crossref","unstructured":"SinghH. V. GirdharA. andDahiyaS. A Literature Survey Based on DBSCAN Algorithms 2022 6th International Conference on Intelligent Computing and Control Systems (ICICCS) 2022 Madurai India IEEE 751\u2013758 https:\/\/doi.org\/10.1109\/ICICCS53718.2022.9788440.","DOI":"10.1109\/ICICCS53718.2022.9788440"},{"key":"e_1_2_14_46_2","doi-asserted-by":"publisher","DOI":"10.1109\/TNNLS.2024.3403155"}],"container-title":["IET Information Security"],"original-title":[],"language":"en","link":[{"URL":"https:\/\/ietresearch.onlinelibrary.wiley.com\/doi\/pdf\/10.1049\/ise2\/8880733","content-type":"application\/pdf","content-version":"vor","intended-application":"text-mining"},{"URL":"https:\/\/ietresearch.onlinelibrary.wiley.com\/doi\/full-xml\/10.1049\/ise2\/8880733","content-type":"application\/xml","content-version":"vor","intended-application":"text-mining"},{"URL":"https:\/\/ietresearch.onlinelibrary.wiley.com\/doi\/pdf\/10.1049\/ise2\/8880733","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2026,3,8]],"date-time":"2026-03-08T22:34:24Z","timestamp":1773009264000},"score":1,"resource":{"primary":{"URL":"https:\/\/ietresearch.onlinelibrary.wiley.com\/doi\/10.1049\/ise2\/8880733"}},"subtitle":[],"editor":[{"given":"Naghmeh","family":"Moradpoor","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"editor"}]}],"short-title":[],"issued":{"date-parts":[[2025,1]]},"references-count":46,"journal-issue":{"issue":"1","published-print":{"date-parts":[[2025,1]]}},"alternative-id":["10.1049\/ise2\/8880733"],"URL":"https:\/\/doi.org\/10.1049\/ise2\/8880733","archive":["Portico"],"relation":{},"ISSN":["1751-8709","1751-8717"],"issn-type":[{"value":"1751-8709","type":"print"},{"value":"1751-8717","type":"electronic"}],"subject":[],"published":{"date-parts":[[2025,1]]},"assertion":[{"value":"2025-01-16","order":0,"name":"received","label":"Received","group":{"name":"publication_history","label":"Publication History"}},{"value":"2025-10-08","order":2,"name":"accepted","label":"Accepted","group":{"name":"publication_history","label":"Publication History"}},{"value":"2025-11-10","order":3,"name":"published","label":"Published","group":{"name":"publication_history","label":"Publication History"}}],"article-number":"8880733"}}