Neural Code Obfuscation for Reverse-Engineering Resistance

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing code obfuscation techniques, including neural networks, are vulnerable to reverse engineering due to their determinate nature and opacity, making them ineffective in whitebox security scenarios.

Innovation Solution

Utilizing a neural network to approximate an obfuscation function, creating an obfuscated code portion with additional outputs that do not correspond to the original function inputs, thereby increasing complexity and security.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Reliability

If code obfuscation techniques are applied to protect intellectual property, then security against reverse engineering is improved, but code complexity and processing overhead increase

Engineering Contradiction:
ImprovesecurityVSAvoidcode complexity
Core Design Contradiction:
ReliabilityVSDevice complexity

Solution Approach 1:

The patent replaces traditional mechanical/code-based obfuscation methods with a neural network system. The neural network automatically performs function identification, transformation code generation, and validation, substituting manual or rule-based obfuscation processes with an intelligent system that reduces complexity while maintaining security.

Inventive Principle:
Principle #28Mechanics substitution (Replace mechanical system)

Solution Approach 2:

The obfuscation system performs self-validation through the neural network, which automatically verifies that transformed code maintains equivalent functionality. The system self-corrects and self-optimizes by learning from validation results, reducing the need for external verification and manual intervention.

Inventive Principle:
Principle #25Self-service

2Reliability

If traditional obfuscation methods are used, then some level of protection is achieved, but they fail to prevent determined reverse engineers and add significant overhead

Engineering Contradiction:
Improveprotection effectivenessVSAvoidprocessing efficiency
Core Design Contradiction:
ReliabilityVSProductivity

Solution Approach 1:

The patent replaces traditional obfuscation methods with a neural network-based system that learns optimal obfuscation strategies. This intelligent system processes code more efficiently than rule-based methods while providing adaptive protection that evolves against new reverse engineering techniques.

Inventive Principle:
Principle #28Mechanics substitution (Replace mechanical system)

Solution Approach 2:

The neural network dynamically adjusts obfuscation parameters based on the specific code being protected and threat patterns. This allows the system to optimize the balance between protection strength and processing overhead, applying stronger obfuscation only where necessary rather than uniformly across all code.

Inventive Principle:
Principle #35Parameter changes

3Reliability

If complex obfuscation transformations are applied, then security against analysis is improved, but the transformed code becomes harder to debug and maintain

Engineering Contradiction:
Improvesecurity against analysisVSAvoiddebuggability
Core Design Contradiction:
ReliabilityVSEase of operation

Solution Approach 1:

The patent implements a feedback mechanism where the neural network validates transformed code against the original functionality. This feedback loop ensures that obfuscation transformations preserve semantic meaning and control flow, making the code more maintainable while still providing security against analysis.

Inventive Principle:
Principle #23Feedback

Solution Approach 2:

The system adjusts obfuscation parameters to balance security and debuggability. By controlling the degree and type of transformation applied, the system maintains enough structure for debugging while applying sufficient obfuscation to protect against analysis.

Inventive Principle:
Principle #35Parameter changes

Data Source

PatentEP4127981B1Systems, methods, and storage media for creating secured transformed code from input code using a neural network to obscure a function
Publication Date: 2026.04.29 IRDETO BV
  • EP4127981B1 patent drawingFigure 1
  • EP4127981B1 patent drawingFigure 2
  • EP4127981B1 patent drawingFigure 3

AI summary

Systems, methods, and storage media for creating secured transformed code from input code, the input code having at least one code function that includes at least one function value are disclosed. Exemplary implementations may: receive input code; apply an obfuscation algorithm to at least a portion of a selected code function of the input code to thereby create an obfuscated code portion having at least one obfuscated value that is different from the at least one function value; and store the obfuscated code portion on non-transient computer media to create obfuscated code having substantially the same function as the input code.