Asymptotically false-positive-maximizing attack on non-binary Tardos codes

TL;DR

This paper analyzes a specific attack on non-binary Tardos fingerprinting codes, revealing how it maximizes false accusations and providing insights into code length requirements and transition phenomena.

Contribution

It generalizes a method to include various collusion strategies and analytically derives properties of an attack that maximizes false positives.

Findings

Identifies an attack that asymptotically maximizes false accusation probabilities.

Provides numerical results on code length thresholds for the attack.

Explains abrupt transitions in false-positive probabilities.

Abstract

We use a method recently introduced by Simone and Skoric to study accusation probabilities for non-binary Tardos fingerprinting codes. We generalize the pre-computation steps in this approach to include a broad class of collusion attack strategies. We analytically derive properties of a special attack that asymptotically maximizes false accusation probabilities. We present numerical results on sufficient code lengths for this attack, and explain the abrupt transitions that occur in these results.