Ma, T., Zhao, H., Tang, L., Xue, M., Liu, J. (2025). Efficient black-box attack with surrogate models and multiple universal adversarial perturbations. Scientific Reports. https://doi.org/10.1038/s41598-025-87529-z