Coin.AI: A Proof-of-Useful-Work Scheme for Blockchain-Based Distributed Deep Learning

Area of Science:

• Computer Science
• Artificial Intelligence
• Blockchain Technology

Background:

• Cryptocurrencies like Bitcoin utilize proof-of-work (PoW) for block generation, often involving energy-intensive cryptographic problem-solving.
• The primary drawback of traditional PoW is its substantial energy consumption with no tangible real-world utility beyond network security.

Purpose of the Study:

• To introduce a theoretical proof-of-useful-work (PoUW) scheme for a cryptocurrency named Coin.AI.
• To propose an alternative to energy-intensive PoW by integrating deep learning model training into the mining process.

Main Methods:

• The Coin.AI system mandates miners to train deep learning models; block generation is contingent upon achieving a performance threshold.
• A distributed verification mechanism allows nodes to efficiently validate trained models, triggering block creation.
• A proof-of-storage scheme is introduced to incentivize users for hosting deep learning models.

Main Results:

• The proposed PoUW system offers a theoretical framework for energy-efficient cryptocurrency mining.
• The system enables efficient model verification by network nodes.
• It establishes a mechanism for rewarding storage providers within the network.

Conclusions:

• Coin.AI presents a viable alternative to traditional PoW, aligning cryptocurrency mining with productive computational tasks.
• The proposed system has the potential to democratize access to artificial intelligence by leveraging distributed computing resources.