Introduction

With the rapid advancement of artificial intelligence (AI), the need for data privacy protection, training efficiency, and decentralization has become increasingly prominent. Traditional AI model training methods often rely on centralized data collection and processing, which poses risks of data breaches and misuse and limits the diversity and innovation of AI models. At the same time, the maturity of blockchain technology, with its decentralized and tamper-proof characteristics, offers new opportunities for the AI field. Against this backdrop, FLock.io has emerged, dedicated to building a decentralized AI training and model-sharing platform to address the challenges faced by current AI development.

What is Flock.io?

FLock.io is a decentralized AI training and model-sharing platform based on blockchain technology. By integrating Federated Learning and blockchain, it allows users to collaboratively participate in AI model training and optimization while preserving data privacy. FLock.io has introduced the concept of “Federated Learning Blocks (FLocks),” leveraging blockchain as a coordination platform between data holders for machine learning, ensuring that data remains stored locally and private. This innovative approach aims to break down data silos, foster knowledge sharing, and accelerate the iteration and innovation of AI models.

Source: https://www.flock.io/whitepaper

Project Background

Team Members

The Flock.io team consists of several experts from the Web2 domain:

• Jiahao Sun | Founder and Chief Executive Officer (CEO)
• Vincent Z. Wang | Chief Financial Officer and General Counsel (CFO & General Counsel)
• In addition, the team includes multiple technical experts and researchers with extensive experience in AI, blockchain, and data security. Together, they strongly support FLock’s technological development and product innovation.

Funding Status

Flock.io has raised a total of $9 million, which includes:

• Seed Round: $6 million, led by Lightspeed Faction and Tagus Capital, with DCG, OKX Ventures, Inception Capital, and Volt Capital participation.
• Strategic Round: $3 million, led by DCG (Digital Currency Group), the parent company of Grayscale, with follow-on investment from Lightspeed Faction. Strategic partners such as Animoca Brands, Fenbushi Capital, GnosisVC, GSR Ventures, OKCoinJapan, Bas1s Ventures, and A41 also participated.

Technical Architecture

The FLock system consists of four key components: the blockchain layer, the AI layer, the AI marketplace, and various participants, each playing a crucial role in the system’s functionality and security.

Blockchain Layer: The Foundation for Stakeholder Participation and Reward Distribution

The blockchain layer employs smart contracts to enable participants to securely lock their stakes, fostering a trusted and transparent environment. This mechanism incentivizes participation by distributing rewards based on contributions, encouraging greater engagement and a more active community. The inherent security features of the blockchain layer prevent fraudulent activities, ensuring the integrity of stakes and reward distribution. It is a critical component that supports model security and enhances resilience against malicious attacks. By leveraging smart contracts, the system automates an efficient and fair reward process, reducing human error and ensuring timely and equitable reward distribution.

AI Layer: Infrastructure for Decentralized Training, Knowledge Extraction, and Monetization From Data

The AI layer provides the infrastructure for decentralized model training, allowing participants to contribute computing power and data while being rewarded via blockchain-based incentives. This layer supports traditional machine learning (ML) model training paradigms, enabling users to optimize models directly on their own devices using private or public datasets. By utilizing Federated Learning (FL) methods, the AI layer enables thousands of participants to collaboratively train global models while ensuring that no local data is transferred at any stage of the training process, thus preserving data sovereignty.

AI Marketplace: A Platform for Optimized Model Deployment and Trading

Once trained and optimized, AI models can be deployed and traded within the marketplace. This platform offers a comprehensive environment for deploying ML models, making them accessible within blockchain network virtual machines (VMs). By integrating with these networks, the marketplace facilitates the seamless execution and inference of complex ML models, providing real-time, scalable, and secure solutions.

Source: https://www.flock.io/whitepaper

Participants: Various Roles in the FLock Ecosystem

• Task Creators: Individuals or organizations that define training tasks and submit them to the platform. They provide datasets and set training objectives.
• Training Nodes: AI developers responsible for training and optimizing AI models. They participate in tasks by staking tokens and earn rewards or penalties based on model performance.
• Validators: Participants who evaluate the models submitted by training nodes, ensuring their quality and performance. They also stake tokens and receive rewards based on their evaluation results.
• Delegators: Users who delegate their tokens to training nodes or validators to support their efforts and share in the rewards.
• Model Hosts: Responsible for deploying and managing trained models in the AI marketplace, providing inference services to users and earning revenue based on service quality.

Tokenomics

Flock’s tokenomics are designed to establish a fair and sustainable ecosystem by incentivizing participation and contributions through its token reward mechanisms. The total supply of $FLOCK tokens is 1 billion, with a linear unlocking model. Currently, 128 million $FLOCK tokens have been unlocked. The token allocation is as follows:

• Community: 47%
• Team & Advisors: 17%
• Investors: 16.3%
• Ecosystem: 19.7%

Source: https://docs.flock.io/flock-tokenomics/token-allocations

Token Use Cases

• Staking Participation: Users can stake tokens to participate in roles such as training nodes and validators, earning corresponding rewards.
• Fee Payments: Tokens are required to pay for operations such as model training, validation, and deployment on the platform.
• Governance Voting: Token holders can participate in governance decisions, voting on technical upgrades, protocol changes, and other platform-related matters.
• Model Purchases: Tokens can be used to buy and utilize AI models within the AI marketplace.

Risk Warning
According to Gate.io data as of January 13, 2024, the current price of $FLOCK is $0.2846, representing a maximum decline of approximately 69% from its all-time high of $0.899.

Source: https://www.gate.io/zh/price/flock.io-flock

FLOCK is currently available for spot trading on Gate.io. Trade Now: https://www.gate.io/zh/trade/FLOCK_USDT

Ecosystem Status

According to Flock Explore, the ecosystem has currently created 707 AI models, with 52 training nodes, 119 validator nodes, and 614 delegators jointly executing and maintaining AI model training.

As for participants, those operating as delegators can achieve an annual yield of over 1,148%.

Source: https://train.flock.io/explore

Conclusion

As an innovative decentralized AI training and model-sharing platform, FLock.io demonstrates tremendous potential in the AI and blockchain fields with its unique technical architecture and incentive mechanisms. It effectively protects data privacy and accelerates AI model iteration and innovation, providing high-quality AI solutions for developers and enterprises. However, the project still faces various challenges regarding technology, market adoption, and regulatory compliance. In the future, with continuous technological advancements and market maturation, FLock is expected to play an increasingly significant role in the AI space, driving the development and application of decentralized AI solutions.