Intelligent Systems, Advanced Learning Theory, Methodology, and Techniques: Mapping Black-Box Attack Metrics and Parameter Spaces in Machine Learning

Abstract

Investigating the Impact of Transformer Architectures on Traditional Security Paradigms1. Craft an evaluation framework for measuring the transferability of adversarial examples across different pre-trained model architectures.2. Demonstrate how transformers exhibit high transferability rates of adversarial examples against other model architectures.3. Show that the degree of transferability of adversarial examples is dependent on the finetuned dataset. The Space of Adversarial Strategies1. Develop a unified framework of attacks in adversarial machine learning.2. Evaluate the generalization of attacks introduced by our framework across threat models, datasets, and robust v. non-robust models.3. Determine the components that yield the most performant attack.

Open PDF

Document Details

Document Type
Technical Report
Publication Date
Sep 15, 2023
Accession Number
AD1230148

Entities

People

  • Patrick Drew McDaniel

Organizations

  • Pennsylvania State University

Tags

Fields of Study

  • Computer science

Readers

  • Distributed Systems and Data Platform Development
  • Irregular Warfare and Special Operations Cyberspace Operations against Adversarial Threats.
  • Neural Network Machine Learning.

Technology Areas

  • AI & ML
  • AI & ML - Machine Learning Algorithms
  • AI & ML - Neural Networks
  • Space