Secure Viterbi Decoding Using Encrypted Cost Values
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing decoding algorithms, such as the Viterbi algorithm, do not address the concern of limiting information disclosure during the decoding process in communication systems, and they do not enable secure computation of the most likely sequence of states without revealing model parameters or observed outputs.
Innovation Solution
A method is provided where an encrypted model is used to determine cost values for a sequence of observed outputs, allowing one party to compute the most likely sequence without revealing the observed sequence or model parameters, utilizing secure function evaluation and homomorphic public key encryption to perform the Viterbi algorithm with reasonable computational complexity.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If traditional decoding algorithms (e.g., Viterbi algorithm) are used to determine the most likely sequence of states, then decoding accuracy is improved, but information disclosure increases because model parameters and observed outputs are exposed
Solution Approach 1:
The patent introduces encrypted representations of model parameters and cost values as intermediaries between the decoding algorithm and the actual data. The Viterbi algorithm operates on encrypted cost values without exposing the underlying model parameters or observed outputs, thereby maintaining decoding accuracy while preventing information disclosure through cryptographic protection layers
Solution Approach 2:
The patent transforms the traditional Viterbi algorithm by changing the parameter representation from plaintext to encrypted form. Cost values and model parameters are encoded using cryptographic functions, allowing the algorithm to perform computations on transformed parameters while preserving the mathematical relationships needed for accurate decoding without exposing the original parameter values
2Reliability
If encrypted model parameters are used to perform secure decoding, then information security is improved, but computational complexity increases
Solution Approach 1:
The patent extracts the computationally intensive cryptographic operations from the main decoding loop and performs them in separate preprocessing and postprocessing stages. The core Viterbi dynamic programming operations work with encrypted values using simpler homomorphic operations, separating the security-critical encryption/decryption steps from the iterative decoding computations to reduce overall computational complexity
Solution Approach 2:
The patent performs preliminary encryption of model parameters and preliminary computation of encrypted cost values before executing the main Viterbi decoding algorithm. By preparing the encrypted data structures in advance, the actual decoding process only requires efficient homomorphic operations on pre-processed values, reducing the real-time computational burden while maintaining security
3Loss of information
If the Viterbi algorithm is modified to perform secure function evaluation, then privacy protection is improved, but algorithm complexity increases
Solution Approach 1:
The patent segments the Viterbi algorithm into distinct cryptographic modules: encrypted model parameter storage, encrypted cost value computation, secure minimum finding, and decrypted result extraction. Each segment handles a specific aspect of the secure decoding process, allowing the complex secure function evaluation to be implemented through modular, manageable components that can be optimized independently
Data Source
AI summary
Techniques are disclosed for securely classifying or decoding data. By way of example, a method of determining a most likely sequence for a given data set comprises a computer system associated with a first party performing the following steps. An encrypted model is obtained from a second party. The encrypted model is utilized to determine cost values associated with a particular sequence of observed outputs associated with the given data set. The cost values are sent to the second party. At least one index of a minimum cost value determined by the second party from the cost values sent thereto is obtained from the second party. A minimum cost sequence resulting from the at least one index is determined as the most likely sequence.