|
|
 |
| |
|
|
|
Statistical Characterization of Smooth Prime Constraints in RSA Variants for AI-Based Cryptanalysis |
|
|
|
PP: 955-961 |
|
|
doi:10.18576/jsap/140631
|
|
|
|
Author(s) |
|
|
|
Hamza Farhan Abu Owida,
Sulieman Mohammad,
Asokan Vasudevan,
Hanan Jadallah,
Mohammad Faleh Hunitie,
Yogeesh N.,
|
|
|
|
Abstract |
|
|
| The work we introduces a machine-learning-aided model of the optimization of elliptic curve parameters in prime
finite fields, which combines discrete optimization methods to explore the exponentially large search space of elliptic curves systematically. The suggested methodology imparts cryptographic characteristics, such as subgroup order, embedding degree, cofactor, and discriminant features, to a high-dimensional feature space, which is then exploited by the ensemble- based classifiers to generatively predict fidelity curve security. A discrete optimization layer is used to steer the selection of candidates, which increases the calculation speed and classification accuracy. Results in experiments with a synthetic data set of 10,000 elliptic curve images show that the framework results in 92% classification accuracy, and an F1-score of 0.91, which is better than the traditional rule-based and random selection methods. The comparative studies affirm that the hybrid model ensures that false positives/negatives are minimised in addition to having a high resistance to algebraic and statistical
attacks hence making it a scalable and adaptive paradigm of deploying secure elliptic curve in cryptography systems. |
|
|
|
|
|
|
|
