Federated Learning Implemented on iOS: AI Assistant Trains While Being Used, User Data Privacy is Controllable and Can Also Earn on-chain Incentives

[Coin World] According to official news, a certain company has become the first in the world to implement Federated Learning (FedML) technology in the iOS App scenario, realizing a new paradigm of "training while using" for AI Agents on users' local devices. The private data generated through interaction with the AI virtual assistant will be directly used for personalized model fine-tuning, and users will receive on-chain incentives, with data remaining on the device, privacy controllable, and earnings obtainable. Currently, the platform's daily trading volume has exceeded 600,000 transactions, with a cumulative paid user base of over 300,000. Users can experience Prompt-to-earn and multi-Agent task collaboration, while developers can publish Agents and earn points and Token revenue based on usage effectiveness. The company is building a decentralized intelligent agent network of "user co-construction + model co-training" to provide a new computational paradigm and economic model for AI collaboration.