ROBENS: A Robust Ensemble System for Password Strength Classification

Weak passwords create vulnerabilities that attackers exploit through password-guessing techniques. This weakness arises from predictable patterns that users often employ when creating passwords, such as common dictionary words, familiar phrases, and sequential characters. In this study, we present ROBENS (ROBust ENSemble system for password strength), a machine learning–based framework for reliable password strength assessment under data challenges. Specifically, we investigate the effects of limited, imbalanced, and noisy datasets on model performance, and we propose strategies for improving generalization across different password corpora. Experimental results demonstrate that ROBENS, through ensemble-based methods, achieves improved robustness compared with traditional rule-based metrics. Our findings highlight the importance of data quality, feature engineering, and model design in constructing dependable password strength evaluators that enhance authentication security.