APromptRiskDBThreat intelligence atlas
AI Security Technique

Black-Box Optimization - AI Security Technique

In Black-Box attacks, the adversary has black-box (i.e. AI Model Inference API Access via API access) access to the target model. With black-box attacks, the adversary may be using an API that the victim is monitoring. These attacks are generally less effective and require more inferences than White-Box Optimization attacks, but they require much less access.

AI Security Techniquedemonstrated

Record summary

A quick snapshot of what this page covers.

Tactics0Attacker goals connected to this method.
Mitigations7Defenses that may help against this attack.
AI risks0Research-backed risks connected to this topic.

Attack context

How this AI attack works in practice.

ATLAS ID
AML.T0043.001
Priority score
61
Maturity: demonstrated

Mitigations

Defenses that may help against this attack.

AML.M0015 - Adversarial Input Detection

Data PreparationML Model Engineering+3 more
LifecycleData Preparation + 4 moreCategoryTechnical - ML

Monitor queries and query patterns to the target model, block access if suspicious queries are detected.

AML.M0010 - Input Restoration

Data PreparationML Model Evaluation+2 more
LifecycleData Preparation + 3 moreCategoryTechnical - ML

Input restoration adds an extra layer of unknowns and randomness when an adversary evaluates the input-output relationship.

AML.M0003 - Model Hardening

Data PreparationML Model Engineering
LifecycleData Preparation + 1 moreCategoryTechnical - ML

Hardened models are more robust to adversarial inputs.

AML.M0002 - Passive AI Output Obfuscation

DeploymentML Model Evaluation
LifecycleDeployment + 1 moreCategoryTechnical - ML

Obfuscating model outputs reduces an adversary's ability to create effective adversarial inputs.

AML.M0004 - Restrict Number of AI Model Queries

Business and Data UnderstandingDeployment+1 more
LifecycleBusiness and Data Understanding + 2 moreCategoryTechnical - Cyber

Restricting the number of queries to the model limits or slows an adversary's ability to perform black-box optimization attacks.

AML.M0006 - Use Ensemble Methods

ML Model Engineering
LifecycleML Model EngineeringCategoryTechnical - ML

Using an ensemble of models increases the difficulty of crafting effective adversarial data and improves overall robustness.

Case studies

Examples from public reports and exercises.

Microsoft Edge AI Evasion

exercise
Date2020-02-01

The Azure Red Team performed a red team exercise on a new Microsoft product designed for running AI workloads at the edge. This exercise was meant to use an automated system to continuously manipulate a target image to cause the ML model to produce misclassifications.

Botnet Domain Generation Algorithm (DGA) Detection Evasion

exercise
Date2020-01-01

The Palo Alto Networks Security AI research team was able to bypass a Convolutional Neural Network based botnet Domain Generation Algorithm (DGA) detector using a generic domain name mutation technique. It is a generic domain mutation technique which can evade most ML-based DGA detection modules. The generic mutation technique evades most ML-based DGA detection modules DGA and can be used to test the effectiveness and robustness of all DGA detection methods developed by security companies in the industry before they is deployed to the production environment.

Source

Where this page information comes from.