SIGNAL//DESK
AI securitysrc: OWASP LLM Top 10

Adversarial example

An input crafted with small, often imperceptible perturbations that causes a model to misclassify or misbehave, exploiting the brittleness of learned decision b

An input crafted with small, often imperceptible perturbations that causes a model to misclassify or misbehave, exploiting the brittleness of learned decision boundaries.

An input crafted with small, often imperceptible perturbations that causes a model to misclassify or misbehave, exploiting the brittleness of learned decision boundaries.

evolution

  1. 2004 · history
    Discovery of adversarial spam filtering

    Dalvi et al. demonstrated that machine learning-based spam filters could be bypassed by attackers modifying email content.

  2. 2013 · history
    Formalization of adversarial examples

    Szegedy et al. identified that deep neural networks are highly susceptible to small, non-random perturbations that cause misclassification.

  3. 2014 · history
    Fast Gradient Sign Method (FGSM)

    Goodfellow et al. introduced a computationally efficient method to generate adversarial examples, highlighting the linear nature of neural networks.

  4. 2016 · history
    Universal adversarial perturbations

    Moosavi-Dezfooli et al. proved that a single, image-agnostic perturbation can cause misclassification across a wide range of inputs.

  5. 2017 · history
    Physical world attacks

    Researchers demonstrated that adversarial examples could be printed or applied to physical objects to fool autonomous systems in real-world conditions.


← all terms