Adaptive Data Perturbation for Authentication Risk Assessment

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Existing information processing techniques fail to effectively assess and mitigate risks in authentication processing due to adversarial samples, which can lead to erroneous assessments.

Innovation Solution

An information processing apparatus and method that determines a perturbing target and applies a perturbation based on surrounding element values, assessing the risk in authentication processing using a perturbed adversarial sample.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Reliability

If a perturbation is applied to adversarial samples to assess authentication risk, then the ability to detect authentication vulnerabilities is improved, but the perturbation may be perceived as abnormal changes affecting the authenticity assessment

Engineering Contradiction:

Improveauthentication risk assessment accuracyVSAvoidperceptibility of perturbation

Core Design Contradiction:

ReliabilityVSObject-affected harmful factors

Solution Approach 1:

The patent applies different perturbation magnitudes to different elements (pixels) of the input data based on their importance. Elements surrounding the target element have their perturbation magnitudes considered, creating localized quality variations that make the perturbation less perceptible while maintaining effective risk assessment capability

Inventive Principle:

Principle #3Local quality

Solution Approach 2:

The patent dynamically adjusts the perturbation magnitude parameter based on the values of surrounding elements. By changing the perturbation parameter adaptively rather than using a fixed approach, the system achieves imperceptible changes that still effectively test authentication vulnerability

Inventive Principle:

Principle #35Parameter changes

2Reliability

If a large perturbation is applied to ensure effective risk assessment, then the detection capability is improved, but the perturbation becomes obviously different from the original data

Engineering Contradiction:

Improverisk assessment effectivenessVSAvoidauthenticity of data

Core Design Contradiction:

ReliabilityVSMeasurement precision

Solution Approach 1:

The patent implements local quality by considering the values of elements surrounding the target element when determining perturbation magnitude. This creates a localized adaptation where the perturbation is tailored to the specific context of each element, making it less noticeable while maintaining assessment effectiveness

Inventive Principle:

Principle #3Local quality

Solution Approach 2:

The patent applies partial action by using smaller, more subtle perturbations rather than large, obvious changes. The perturbation magnitude is carefully controlled to be just sufficient for effective risk assessment without creating obviously different data, avoiding excessive action that would compromise authenticity perception

Inventive Principle:

Principle #16Partial or excessive action

Data Source

PatentUS20250316077A1Information processing apparatus, information processing method, and non-transitory recording medium

Publication Date: 2025.10.09 NEC CORP

AI summary

An information processing apparatus includes: a perturbing position determination unit that determines an element serving as a perturbing target in first information; a perturbing unit that applies a perturbation with magnitude based on values of one or more elements surrounding the element serving as the perturbing target, to the element serving as the perturbing target; and an assessment unit that assesses a risk in authentication processing based on a result of the authentication processing of matching the first information to which the perturbation is applied, with second information that is different from the first information.