Linguistic steganalysis via multi-task with crossing generative-natural domain

Huiqing You, Lingyun Xiang, Chunfang Yang, Xiaobo Shen

Abstract: Highlights•A cross generative-natural domain multi-task model is proposed for steganalysis.•Transfer learning captures features distinguishing generative text from natural ones.•Multi-task learning reduces sample bias, boosts cross-domain steganalysis performance.