PulseAugur
EN
LIVE 05:06:53

New defense Random Logit Scaling protects AI models from adversarial attacks

By PulseAugur Editorial · [2 sources] ·

Researchers have introduced Random Logit Scaling (RLS), a new defense mechanism designed to protect deep neural networks against black-box score-based adversarial example attacks. RLS functions as a post-processing step that can be applied to existing machine learning models with minimal integration effort. By outputting randomly scaled logits, RLS aims to confuse attackers while preserving the model's accuracy and minimizing distortions in confidence scores. The proposed defense has demonstrated a significant reduction in the success rates of current state-of-the-art attacks. AI

IMPACT Enhances the security and reliability of AI models against sophisticated adversarial attacks.

RANK_REASON The cluster contains an academic paper detailing a new defense mechanism for machine learning models.

Read on arXiv cs.LG →

AI-generated summary · Google Gemini · from 2 sources. How we write summaries →

New defense Random Logit Scaling protects AI models from adversarial attacks

COVERAGE [2]

  1. arXiv cs.AI TIER_1 English(EN) · Hamid Dashtbani, Mehdi Dousti Gandomani, AmirMahdi Sadeghzadeh ·

    Random Logit Scaling: Defending Deep Neural Networks Against Black-Box Score-Based Adversarial Example Attacks

    arXiv:2607.14921v1 Announce Type: cross Abstract: Machine learning models are increasingly adapted in various domains. However, adversarial examples pose a significant threat to the reliable deployment of these models. In recent years, some powerful adversarial example attacks ha…

  2. arXiv cs.LG TIER_1 English(EN) · AmirMahdi Sadeghzadeh ·

    Random Logit Scaling: Defending Deep Neural Networks Against Black-Box Score-Based Adversarial Example Attacks

    Machine learning models are increasingly adapted in various domains. However, adversarial examples pose a significant threat to the reliable deployment of these models. In recent years, some powerful adversarial example attacks have been proposed for the fast and query-efficient …

RELATED ENTITIES

RELATED TOPICS