Gallil Maimon, Lior Rokach: A universal adversarial policy for text classifiers. Neural Networks 153: 282-291 (2022)