GAN-Based Key Secret-Sharing Scheme in Blockchain.

Area of Science:

• Computer Science
• Cryptography
• Artificial Intelligence

Background:

• Blockchain technology faces challenges with low security, difficult key recovery, and inefficient communication.
• Existing key secret-sharing methods may not adequately address these blockchain-specific issues.

Purpose of the Study:

• To propose a novel key secret-sharing technology utilizing Generative Adversarial Networks (GANs).
• To address critical blockchain issues including low security, challenges in lost key recovery, and communication inefficiency.

Main Methods:

• The proposed scheme treats key secret-sharing as a classification problem using a GAN-based network as a dealer.
• User private keys (text) are converted into images, segmented into subimages, and encoded using DNA coding.
• The GAN is trained to generate key secret-sharing results based on the encoded image data.

Main Results:

• Simulation results demonstrate the proposed scheme's security.
• The method proves to be flexible and efficient in terms of communication.
• The approach represents a significant extension of GAN applications and a new direction in key secret-sharing.

Conclusions:

• The GAN-based key secret-sharing technology effectively enhances blockchain security and communication efficiency.
• The proposed method offers a novel and practical solution for managing and recovering blockchain private keys.
• This research opens new avenues for applying GANs in cryptographic applications.