Exploration of Security and Privacy Challenges through Adversarial Weight Perturbation in Deep Learning Models