Inactive

Machine Learning (ML) techniques are susceptible to adversarial deception at training time and when deployed. Similarly, humans are susceptible to being deceived by adversarially falsified media (imag... Machine Learning (ML) techniques are susceptible to adversarial deception at training time and when deployed. Similarly, humans are susceptible to being deceived by adversarially falsified media (images, video, audio, text) or other information. The consequences may be significant in both cases, and deception plays an increasingly central role in information-based attacks. The Reverse Engineering of Deceptions (RED) effort will develop techniques that automatically reverse engineer the toolchains behind attacks such as multimedia falsification, adversarial machine learning (ML) attacks, or other information deception attacks. The tools by which those attacks are accomplished and the adversaries behind such attacks are often unclear; recovering the tools and processes used to create an attack provides information that may aid in identifying an adversary. RED will seek to develop techniques that support the automated identification of attack toolchains and the development and maintenance of scalable databases of attack toolchains.