ML Encoder Parameter Encryption for Plausible Decoy Decryption
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing data encryption methods fail to provide robust protection against malicious actors who can decrypt sensitive information using brute force or other methods, and there is a need for techniques that obscure sensitive data while maintaining a plausible appearance of decrypted data.
Innovation Solution
Utilizing machine learning models to modify encoder and decoder parameters with unique keys, allowing data to be encapsulated in hidden compartments, where different keys reveal different plausible data sets, thereby protecting the actual sensitive data.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If traditional data encryption methods are used, then data protection is provided, but the encryption can be compromised by brute force or other methods
Solution Approach 1:
The patent changes the parameters of machine learning models (weights and biases) using encryption keys to protect data. Instead of traditional encryption, the system modifies model parameters to encode data, making decryption computationally intensive for attackers while maintaining legitimate access for authorized users.
Solution Approach 2:
The patent replaces traditional mechanical encryption systems with machine learning-based encryption. The encoder and decoder models use neural network parameters to perform encryption and decryption operations, substituting conventional cryptographic mechanisms with AI-driven processes.
2Reliability
If machine learning models are used to encrypt data, then robust protection against brute force is achieved, but the complexity of the encryption system increases
Solution Approach 1:
The patent makes the machine learning models serve multiple functions: they perform both the encryption and decryption operations. The same encoder model that encrypts data can be used to decrypt it with the correct key, eliminating the need for separate encryption and decryption systems and reducing overall complexity.
Solution Approach 2:
The patent uses copies of the same machine learning model for both encoding and decoding operations. Instead of requiring different complex systems, the same model architecture is replicated and configured with different parameters for encryption and decryption functions.
3Reliability
If data is encrypted using machine learning models, then sensitive information remains protected, but the resource intensity for determining data authenticity increases
Solution Approach 1:
The patent makes the encryption system dynamic by allowing the same data to be encrypted with different keys to produce different plausible data sets. This dynamic encryption approach prevents attackers from determining data authenticity through static analysis, as the encrypted data changes with each key while maintaining the appearance of valid data.
Data Source
AI summary
A first password is received by a password encoder which uses the first password to generate a first key. This first key is used to modify weights and biases of an encoder to result in a modified encoder. Further, weights and biases of a decoder operating in tandem with the encoder based can be modified based on a second key to result in a modified decoder. First data is received which encapsulates second data in a hidden compartment. The first data is encoded by the modified encoder to result to generate an embedding. The modified decoder decodes the embedding to result in a representation of the second data which, in turn, can be provided to a consuming application or process.The first data can be input into the encoder and the decoder prior to those components being modified to result in a representation of the first data.