Abstract
Decentralized Intelligence Network (DIN) is a theoretical framework designedto address challenges in AI development, particularly focusing on datafragmentation and siloing issues. It facilitates effective AI training withinsovereign data networks by overcoming barriers to accessing diverse datasources, leveraging: 1) personal data stores to ensure data sovereignty, wheredata remains securely within Participants' control; 2) a scalable federatedlearning protocol implemented on a public blockchain for decentralized AItraining, where only model parameter updates are shared, keeping data withinthe personal data stores; and 3) a scalable, trustless cryptographic rewardsmechanism on a public blockchain to incentivize participation and ensure fairreward distribution through a decentralized auditing protocol. This approachguarantees that no entity can prevent or control access to training data orinfluence financial benefits, as coordination and reward distribution aremanaged on the public blockchain with an immutable record. The frameworksupports effective AI training by allowing Participants to maintain controlover their data, benefit financially, and contribute to a decentralized,scalable ecosystem that leverages collective AI to develop beneficialalgorithms.