{"about":{"site":"https://codewithpapers.app","non_affiliation":"Code with Papers and Syntology are not affiliated with, endorsed by, or sponsored by Papers with Code, Meta, or the pwc-archive mirror.","licence":"CC BY-SA 4.0","licence_url":"https://creativecommons.org/licenses/by-sa/4.0/legalcode","attribution":"https://codewithpapers.app/attribution","modified":"archive material modified by Syntology; see the attribution page"},"url":"/paper/interpretable-adversarial-perturbation-in","title":"Interpretable Adversarial Perturbation in Input Embedding Space for Text","arxiv_id":"1805.02917","date":"2018-05-08","proceeding":null,"authors":["Motoki Sato","Jun Suzuki","Hiroyuki Shindo","Yuji Matsumoto"],"abstract":"Following great success in the image processing field, the idea of\nadversarial training has been applied to tasks in the natural language\nprocessing (NLP) field. One promising approach directly applies adversarial\ntraining developed in the image processing field to the input word embedding\nspace instead of the discrete input space of texts. However, this approach\nabandons such interpretability as generating adversarial texts to significantly\nimprove the performance of NLP tasks. This paper restores interpretability to\nsuch methods by restricting the directions of perturbations toward the existing\nwords in the input embedding space. As a result, we can straightforwardly\nreconstruct each input with perturbations to an actual text by considering the\nperturbations to be the replacement of words in the sentence while maintaining\nor even improving the task performance.","url_abs":"http://arxiv.org/abs/1805.02917v1","url_pdf":"http://arxiv.org/pdf/1805.02917v1.pdf","source":{"archive":"pwc-archive (Hugging Face), CC BY-SA 4.0","snapshot":"2025-07-28","licence_url":"https://creativecommons.org/licenses/by-sa/4.0/legalcode","row_kind":"abstracts"},"code_links":[{"paper_slug":"interpretable-adversarial-perturbation-in","repo_url":"https://github.com/aonotas/interpretable-adv","is_official":1,"mentioned_in_paper":1,"mentioned_in_github":1,"framework":"none","reach":{"status":"ok"}},{"paper_slug":"interpretable-adversarial-perturbation-in","repo_url":"https://github.com/aonotas/interpretable_adv","is_official":1,"mentioned_in_paper":1,"mentioned_in_github":1,"framework":"none","reach":{"status":"ok"}}],"tasks":[{"task_slug":"sentence","task_name":"Sentence"}],"methods":[{"method_slug":"interpretability","method_name":"Interpretability"}],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"atlas_url":"https://app.syntology.ai/?focus=1805.02917","mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}