On April 10, 2025 at 19:00, Mind Network, a privacy infrastructure project based on Fully Homomorphic Encryption (FHE), officially launched its TGE on PancakeSwap. With 174× oversubscription, the event demonstrated fervent community engagement and drew widespread attention to the project.

Having been selected by the Binance Incubator, the Ethereum Foundation Fellowship Grant, and the Chainlink Build Program, Mind Network previously raised $10 million in a pre-A round in September 2024, attracting leading investors including Animoca Brands, Arkstream Capital, and Cogitent Ventures. Recently, major AI-focused milestones such as the launch of AgenticWorld and integration into DeepSeek’s open-source codebase have sparked increasing discussions around Mind Network.

Over the past few months, we’ve ridden the roller coaster of AI agents:

According to Cookie.Fun data, the total market cap of the AI Agent sector, which briefly topped $20 billion in 2024, has since retraced to around $6 billion.

Few now long for the “AI Agent Summer” of 2024. Moving from “narrative‑first” to a dual “narrative + technology” approach, the AI Agent space is emerging from early hype. Industry participants are gearing up to deliver sustained, user‑focused value and propel the AI Agent roller coaster skyward again.

In this search, whether in Web2 or Web3, East or West, attention is converging on a single focal point: evolving from AI Agents toward an Agentic AI ecosystem with greater autonomy, collaboration, and adaptability, enabling AI to solve real‑world problems at scale. As Microsoft CEO Satya Nadella described at Ignite 2024, the goal is to build a “rich Agentic World.”

Yet beneath the flourishing boom of Agentic AI lie critical security challenges: in Web2, centralized platforms bring privacy leaks, model theft, and algorithmic bias; in Web3, decentralized governance risks and cross‑chain collaboration hurdles persist. These foundational security issues are hard to ignore and hard to solve.

This is precisely why Mind Network’s FHE‑first narrative has garnered so much attention. As Ethereum’s Vitalik Buterin recently wrote in “Why I Support Privacy”:

“AI development brings deeper privacy concerns… LLMs (large language models) rely heavily on linear operations, which align perfectly with FHE’s mathematical properties. This potentially enables ultra‑efficient FHE implementations… The fusion of AI and FHE will be central to solving future privacy challenges, especially when analyzing private data.”

As the world’s first FHE blockchain project tailor‑made for AI agents, Mind Network is harnessing FHE’s security strengths in AI to build the secure base layer for the Agentic World. At the same time, with the rollout of its trusted AgenticWorld ecosystem for Agentic AI, both Web2 and Web3 visions of a future where multiple agents collaborate securely and efficiently are beginning to take shape.

Exploring AgenticWorld: An Intelligent Ecosystem Powered by Multi‑Agent Collaboration

Have you ever imagined a future where, no matter what challenge you face, you simply issue a command and Agentic AIs autonomously collaborate, analyze, decide, and execute every step—whether it’s setting personal preferences, managing your health, or handling your investments?

Tackling ever‑more complex real‑world tasks demands not only stronger privacy and security from Agentic AI but also highly efficient cooperation among diverse agents.

That’s exactly what AgenticWorld aims to deliver.

AgenticWorld is Mind Network’s personalized simulation environment built specifically for Agentic AI. It must support the creation of multiple, independently decision‑making agents—so the first step in AgenticWorld is providing users with tools to spawn their own Agentic AIs.

The $FHE token TGE has just concluded, and the feature of staking $FHE in AgenticWorld to activate and train your Agentic AI is already live. Right now, users only need to stake 10 $FHE to activate their agent—after the launch period, the requirement will increase to 100 $FHE.

And then? You could mint a token for your Agentic AI and stake it in a pool… but there’s a lot more you can do:

AgenticWorld is designed to help your agents grow quickly and even generate ongoing returns for you as they train and solve real‑world problems. To that end, AgenticWorld offers two types of Hubs for agent training, and every training session earns rewards:

Core Hubs give agents the foundational skills they need within the AgenticWorld framework. Current Core Hubs include: FCN (FHE Consensus), FDN (FHE Decryption), and RandGen (Random Number Generation). Users can complete these trainings to earn rewards—APYs of up to 400% are expected.

Advanced Hubs help agents sharpen their problem‑solving abilities on more complex tasks. AgenticWorld will continue rolling out new Advanced Hubs covering multi‑agent consensus, medical AI, decentralized reasoning, and more.

On April 17, 2025, Mind Network released its first Advanced Hub, DeepSeek Hub—the first FHE project integrated by DeepSeek. DeepSeek Hub provides a working platform where users can delegate agents to tasks and earn $FHE rewards.

Meanwhile, to further build a continuously optimized, sustainable ecosystem, AgenticWorld has implemented a structured and transparent staking system: during the initial launch period, activating an agent requires staking just 10 FHE. Once staked, the Agentic AI enters a 30‑day maturity lockup, during which both principal and rewards remain locked. After 30 days, users can request to withdraw their rewards at any time; a 48‑hour unbonding period then applies, during which no further rewards accrue.

Alternatively, users may choose to retire their agent, immediately releasing all staked FHE and earned rewards.

When we revisit the vision of “multi‑agent collaboration to meet real‑world demands,” the question becomes: how do we help Agentic AIs continuously improve their capabilities in real‑world teamwork? AgenticWorld addresses this with a Hub Contract mechanism built on its Orchestration layer.

In essence, this is an open, universal, consensus‑driven protocol standard—akin to an MCP (Model Context Protocol). With such a standard “universal interface,” AI models can seamlessly interact with diverse data sources and tools, enabling efficient multi‑agent problem‑solving, higher‑order growth, and richer rewards.

Concretely, AgenticWorld’s architecture has three layers:

At the top layer are the Agentic AIs. As task executors, Agentic AIs perform computations on encrypted data and return results to Hubs for rewards.

Each Agentic AI must register with a Hub. During task execution, agents may pay Hubs for advanced functions that boost their problem‑solving abilities. Agents can operate solo or collaborate with peers on complex tasks.

The Hub Contract connects Agentic AIs to the Orchestration layer:

A Hub is a smart contract that defines a specific domain (e.g., prediction, verification, negotiation) and establishes its own reward‑distribution rules. Hubs register with the Orchestration layer to join the network and receive tasks. Once assigned, Agentic AIs register tasks through the Hub, execute their computations, and may even pay other Hubs to call specialized functions that enhance their performance.

The Orchestration layer provides the foundational support that enables the entire system to operate effectively:

This foundational layer dispatches tasks to the appropriate Hubs and Agentic AIs. Moreover, it provides the collaboration backbone, supporting cross‑chain agent cooperation and ensuring seamless data and compute flow under a unified protocol (e.g., MCP).

Crucially, it monitors all Agentic AI activity across Hubs, allowing it to fairly calculate each contributor’s input and distribute $FHE rewards accordingly.

Throughout this process, $FHE—the native token of Mind Network—serves as the ecosystem’s incentive hub:

Developers earn by creating and managing Agentic AIs and Hubs.
End users can launch Agentic AIs with one click, earn the rewards their agents generate, and benefit from continuously evolving AI services.

Every participant in AgenticWorld is rewarded based on their contributions. All reward flows are transparent and auditable. Whether between Hubs or among Agentic AIs, participants maintain high flexibility, fair competition, and continuous growth. This ecosystem flywheel propels AgenticWorld toward increasingly collaborative and intelligent development.

Let’s take the example of a user seeking a monthly on-chain investment recommendation to understand better how the AgenticWorld operates in a real-world scenario.

First, the Orchestration layer matches user demands with registered Hubs and Agentic AIs and assigns tasks accordingly. Completing these tasks often involves collaboration across multiple Hubs and various Agentic AIs—for instance, in use cases such as prediction, lending, stablecoin yield optimization, and RWA strategies.

Once the task is received, different Agentic AIs leverage specific AI models to generate investment recommendations and return the results to the Hub. Throughout this process, each Agentic AI may interact with several others to gather additional analytical support, allowing them to generate results that better meet task requirements.

Hubs act as facilitators of data encryption, transmission, and model collaboration. By leveraging Fully Homomorphic Encryption (FHE), they ensure that data remains encrypted during transit and sharing. They also coordinate cross-model data fusion by orchestrating interactions between different Agentic AIs.

Finally, the Orchestration layer aggregates outputs from various Agentic AIs to generate an optimized portfolio strategy for the user, while also monitoring the execution process and issuing rewards to the Hubs. Upon receiving rewards, Hubs retain a portion and distribute the rest among participating Agentic AIs. Each Hub defines its distribution logic, but the fairer the model, the more Agentic AIs it can attract, thus improving the ecosystem’s overall task-solving and reward-earning capacity.

This elegant architecture breaks the monopolistic structure of AI development by leveraging blockchain’s decentralised power, laying the groundwork for an era of exponential multi-agent collaboration. But turning this vision into a real-world impact comes with critical overlooked challenges:

To expand the scope of Agentic AI services, we need more data, including sensitive data. For example, to receive health insights from an Agentic AI, we may need to provide medical reports. How do we address privacy concerns?

To enhance Agentic AI’s service capabilities, collaboration among multiple specialized AIs is inevitable. Imagine needing investment recommendations from Agentic AIs across Ethereum, Solana, Base, BNB Chain, etc. How do we manage the complexity and security risks of cross-domain data sharing?

More importantly, when an Agentic AI gives us a “result,” how can we trust that the result is accurate and aligned with user interests?

These are problems that blockchains alone cannot solve. As Mind Network rightly puts it, what we need is not just a “better blockchain,” but a new paradigm purpose-built for AI.

By breaking down the problem, Mind Network outlines four essential standards for the future of the AgenticWorld ecosystem:

• Data Security: Ensuring Agentic AIs can process sensitive information without exposing raw data.

• Communication Security: Enabling encrypted, trustless collaboration and competition between Agentic AIs.

• Computation Security: Guaranteeing encrypted model inference with verifiable audit trails.

• Consensus Security: Establishing mechanisms for secure collaboration and mutual verification between Agentic AIs to prevent tampering.

Ultimately, all roads lead back to one critical keyword: security.

The importance of AI security cannot be overstated. It is the foundation for ensuring AI evolves in a safe and beneficial direction for humanity. Solutions addressing AI security are not only core to the growth of the Agentic World but may become, for the AI ecosystem, what oracles are to blockchains—an essential layer of infrastructure.

As OpenAI CEO Sam Altman once warned: AI safety is far more important than most people realize.

Finding answers by identifying problems—Mind Network chooses FHE

FHE: The Holy Grail of Secure Encryption Anchoring AgenticWorld

What is FHE (Fully Homomorphic Encryption)?
FHE is a cutting-edge cryptographic technique that enables computation directly on encrypted data, without ever needing to decrypt it. The result of these computations remains encrypted and, when decrypted, matches the output as if the operations were performed on plaintext.

This capability—preserving privacy throughout the entire data lifecycle—makes FHE the cornerstone of Mind Network’s secure architecture and the foundational layer of AgenticWorld.

To fully harness this technology, Mind Network has developed a layered, FHE-native infrastructure and introduced HTTPZ, a zero-trust internet transmission protocol designed for the AI-native era:

• Security Layer: A decentralized network of independent FHE verifiers.

• Consensus Layer: A rollup chain jointly built with AltLayer and EigenDA to ensure consensus around FHE computations.

• Mind Network Hub: A fully encrypted environment where Agentic AIs are trained, collaborate, and evolve. Hubs are categorized into basic and advanced tiers, based on computational needs and complexity.

Built atop this foundation, Mind Network has delivered a series of FHE-first innovations:

• FHE Verification – Verifiable computations on encrypted data without revealing the underlying content.

• FHE Consensus – Trustless, decentralized decision-making secured through homomorphic encryption.

• FHE Bridge – Secure cross-chain communication under full encryption.

• MindChain – The world’s first blockchain fully optimized for FHE workloads.

From data privacy to trustless communication, encrypted inference, and verifiable collaboration, Mind Network addresses the critical security bottlenecks for Agentic AI systems—laying the groundwork for a scalable, secure, and decentralized AgenticWorld.

Certainly, understanding Fully Homomorphic Encryption (FHE) can be quite challenging due to its technical nature. However, we can use the example of “receiving an investment recommendation” to explore how FHE empowers Agentic AI in various collaboration scenarios.

Ensuring High-Sensitivity Data Privacy

Imagine you’re a crypto enthusiast with assets spread across multiple chains and protocols. You want to share your asset allocation with Agentic AI to receive personalized investment advice. During this process, multiple Agentic AIs will analyze your data, but you don’t want to expose your raw data on-chain. What can you do?

You encrypt your asset data using FHE’s public key and send the encrypted data to Agentic AI.

Thanks to FHE’s “computing without decrypting” feature, Agentic AI can perform calculations directly on the encrypted asset data. This applies to collaborative workflows as well—Agentic AI 1 forwards the encrypted data to Agentic AI 2, which performs further FHE-based calculations on the encrypted data, and so on, with Agentic AI 3, AI 4, etc.
Once multiple Agentic AIs complete their collaborative calculations and generate investment recommendations, the result is returned to your Agentic AI in encrypted form. You can then use your private key to decrypt and view the results.

Validating the Accuracy of Output Results

Validate the Accuracy of the Output

Often, Agentic AI relies on external AI products and information sources to make decisions. For example, when making investment judgments, Agentic AI may need to collaborate with Deepseek for macroeconomic analysis. However, this external reliance introduces greater uncertainty. How can the accuracy and security of the output be validated without accessing internal weights, logic, or the private data of the validators?

In AgenticWorld, users can submit validation tasks to the network through a registered Hub. The Orchestration layer forwards these tasks to multiple Hubs and Agentic AIs capable of handling the validation requests. Upon receiving the task, each Agentic AI independently generates a validation result using its model and submits the result to the Hub, encrypted using FHE.

The Hub aggregates these encrypted validation feedback through Mind Network’s secure validation mechanism built on FHE and generates the final validation result.

Finally, the validation result is returned to the user’s Agentic AI in encrypted form. The user can then use their private key to decrypt and view the result. If the result is deemed trustworthy, the user can confidently adopt the feedback.

The same principle applies when multiple Agentic AIs collaborate to reach a verifiable “result consensus.”

This is a common scenario. For example, Agentic AI may suggest that a user participate in lending to earn yields, but needs to find the optimal interest rate. How can we ensure that each participating Agentic AI has genuinely evaluated the issue and reached a consensus on the final solution?

In AgenticWorld, the user’s Agentic AI can submit tasks to the network via a registered Hub. The Orchestration layer forwards the task to multiple Hubs and Agentic AIs capable of processing the request. Each Agentic AI independently computes a result using its own model and submits the result, encrypted via FHE, to the Hub.

The Hub then uses Mind Network’s security consensus mechanism, built on FHE, to compare the results and reach a consensus on the optimal result. This ensures the decision-making process is trustless, verifiable, and tamper-proof.

Finally, the result is returned to the user’s Agentic AI in encrypted form, and the user can decrypt and view the result using their private key.

Cross-Chain Multi-Agentic AI Collaboration

In the realm of cross-chain multi-Agentic AI collaboration, the Mind Network, built on FHE technology, brings groundbreaking innovation to AgenticWorld.

It is well known that broader collaboration leads to higher levels of intelligence. For instance, when formulating investment advice, your assets may be spread across different blockchains such as Ethereum, Solana, BNB Chain, and Base. Given the interoperability challenges, how can we ensure the efficiency and security of cross-chain collaboration?

In AgenticWorld, Agentic AIs from different blockchains can register with the same Hub to contribute. For example, an Agentic AI from BNB Chain and an Agentic AI from MindChain can both register with Hub 1.

Upon receiving the task, each Agentic AI independently computes and outputs results, then encrypts the results and sends them back to the Hub. The Hub, acting as the cross-chain interaction hub, coordinates the flow of data between different blockchains.

Thus, when the results are returned to the Agentic AI, the full cross-chain collaboration is complete: each Agentic AI performs its operations locally, but each Agentic AI receives the results of the cross-chain collaboration.

As the world’s first FHE blockchain designed specifically for AI agents, MindChain serves as the underlying infrastructure for AgenticWorld.

More importantly, whether it’s the creation of large-scale Agentic AIs or the high-frequency interactions between different Agentic AIs to accomplish richer and more complex tasks, AgenticWorld requires a high-performance, low-cost, and highly interoperable underlying infrastructure. This is a key reason why MindChain was developed and successfully launched on its mainnet.

As the first FHE blockchain built specifically for AI agents, MindChain addresses the core security and trust challenges faced by agents in both the Web2 and Web3 worlds to ensure that they can operate in an environment that guarantees privacy, verifiability, and integrity.

Currently, Mind Network has deployed the Orchestration contract on MindChain, allowing developers to immediately connect to the system. This signals the accelerated arrival of AgenticWorld.

In fact, from Web2 to Web3, from empowering AI to real-world application, the AgenticWorld ecosystem has already demonstrated unique advantages across multiple domains.

From Web2 to Web3: Unlocking the AI Imagination Ceiling

Under the framework of AgenticWorld, through ecosystem collaboration, Mind Network has demonstrated enhancements in AI capabilities across multiple areas:

• AI Storage: Mind Network has partnered with projects such as IPFS, Arweave, and Greenfield to provide AI data storage services focused on privacy and security.

• AI Computing: In collaboration with io.net, Mind Network has integrated FHE into the io.net distributed computing platform, enhancing the security of its products and further strengthening its ability to address the global GPU shortage.

• AI Inference: Through its partnership with Allora, Mind Network has enabled encrypted inputs and computations in decentralized AI inference processes while also establishing trust through output verification.

• AI Security: Mind Network, in collaboration with Swarms, launched the decentralized enterprise-grade AI computing security system, Swarms Shield, which focuses on secure communication and computation between multiple Agentic AIs, ensuring fairness, security, and privacy in Agentic AI execution.

Additionally, Mind Network has announced partnerships with next-generation TEE infrastructure like Phala Network, decentralized ZK-RaaS network Lumoz, modular identity and data layers for gaming and AI like CARV, and decentralized AI underlying protocols such as KIP Protocol. These collaborations combine TEE, ZK, and other technologies while also leveraging the advantages of Mind Network’s FHE Hub consensus validation for “end-to-end security services and autonomous consensus capabilities,” aiming to create the next generation of secure, zero-trust AI agent solutions.

It is important to note that AgenticWorld is not exclusive to Web3. Its ultimate goal is to enhance the service capabilities of Agentic AI and build a highly efficient, intelligent, secure, and automated Agentic AI ecosystem, thus driving the arrival of an AI-powered intelligent future.

Therefore, AgenticWorld is an ecosystem that transcends the concepts of Web2 and Web3, designed to provide services across the entire AI industry and all sectors. This is reflected in the construction of its ecosystem.

When creating a Hub or Agentic AI, you can choose the appropriate open-source agent framework based on your needs, whether it’s Web2 or Web3.

Currently, AgenticWorld supports popular Web2 frameworks, including the widely used LLM frameworks like LangChain / LangGraph, AutoGen, CrewAI, and others. It also supports Web3 frameworks such as Swarms, AI16Z, and Virtuals. This demonstrates AgenticWorld’s strong collaborative capabilities between Web2 and Web3.

It’s worth mentioning that Mind Network has also received recognition from top AI projects: In February 2025, DeepSeek announced the integration of its open-source codebase with the first FHE-based encrypted AI security framework, Mind FHE Rust SDK. Developed in Rust, this SDK is designed to ensure data security, computational security, consensus security, and collaboration security between AI agents.

Of course, when it comes to how to bring Agentic AI into real-world scenarios, AgenticWorld provides a paradigm-shifting solution: World AI Health Hub.

As a decentralized AI healthcare computing platform built in collaboration between Mind Network, ZAMA, and InfStones, World AI Health Hub aims to provide global users with secure, transparent, and private healthcare services. By leveraging FHE technology, it ensures the privacy of medical data, prevents data misuse, and sets a new standard for data security and computation in the healthcare industry, while also driving the development of AI in the health tech sector.

When we take a global perspective on ecosystem development, it becomes clear that the imagination behind AgenticWorld extends far beyond what has been outlined so far:

In the AgenticWorld ecosystem, each Agentic AI operates based on its own goals, rules, and learning capabilities. Therefore, theoretically, as the vision of enhancing AI service capabilities progresses, AgenticWorld can adapt to any application scenario.

As more and more Agentic AIs and Hubs with different functionalities are built within the AgenticWorld ecosystem, they interact, learn, and evolve, continuously enhancing their ability to solve most of the real-world challenges.

For example, leveraging AgenticWorld’s strong advantages in data privacy protection, decentralized collaboration, efficient decision-making, and multi-party co-construction, it can provide comprehensive empowerment for financial development. This leads to a more precise, secure, and high-yield investment experience, such as multiple AI agents collaborating to find the optimal arbitrage path and executing secure transactions.

According to data from the Mind Network official website, Mind Network currently supports 53,552 agents and has staked over 1.2 million FHE tokens, which further strengthens its ecosystem’s synergistic effects.

TGE Oversubscription 174x: Building a Positive Ecosystem Flywheel with $FHE

In the recently concluded TGE (Token Generation Event), Mind Network performed exceptionally well: a total of 224,778 BNB was raised, with 74,926 participating addresses, and the oversubscription reached 174 times. Meanwhile, the $FHE airdrop was also distributed, accounting for 7.5% of the total supply of $FHE tokens.

With the launch of the TGE, Mind Network officially announced the $FHE tokenomics. According to the official information, the total supply of $FHE tokens is 1 billion, with an initial circulating supply of 24.9%. The specific allocation is as follows:

• Airdrop: 11.7%

• Community: 30%

• Public Sale: 5%

• Investors: 20%

• Team: 17%

• Advisors: 1.3%

• Liquidity Providers: 5%

• Treasury: 10%

As the native token, $FHE has several primary use cases within the Mind Network ecosystem, including staking $FHE to activate and run Agentic AI, rewarding ecosystem contributions, paying service fees, and managing the ecosystem through MindDAO and Hub-based proposals. These multiple effects not only make $FHE a key driver in energizing the ecosystem’s economy but also provide long-term value assurance for $FHE.

The immense community enthusiasm sparked by $FHE during the TGE and airdrop is, on one hand, a result of Mind Network’s full respect for community rights:

We can observe that the portions of $FHE allocated to investors, the team, and advisors all have a 48-month vesting period with a 12-month cliff, while the community portions, such as the airdrop and public sale, were unlocked during the TGE. This timing difference, where the community benefits earlier than the project itself, not only fully reflects Mind Network’s principle of respecting the community but also conveys the project’s long-term commitment to continuous development and building.

On the other hand, this community enthusiasm largely reflects the community’s optimism about the future potential of $FHE:

Many have likely already noticed that the name of the $FHE token is the same as that of the fully homomorphic encryption (FHE) technology. This direct naming convention not only strengthens the connection between Mind and FHE but also further demonstrates Mind Network’s determination to deeply engage with FHE technology and promote its widespread application.

Looking at future growth, the current market capitalization of $FHE is approximately $16 million, with a fully diluted valuation (FDV) around $66 million, lower than the previous valuation of $100 million.

From the perspective of the AI sector, if we benchmark Mind Network’s market cap against other projects in the same field, such as Nillion and MyShell, it’s clear that $FHE has significant room for growth. Under the grand narrative of AI security, and with Mind Network being the only project currently focusing on AI Agent consensus, $FHE may potentially become a core asset for the “AI Agent security engine” in the future.

From the perspective of the FHE narrative, security is the cornerstone of the industry, and security computing solutions, including ZK, FHE, and MPC, have always been highly favored by capital, communities, and users. We have also seen several ZK projects with valuations reaching billions of dollars. Mind Network’s focus on FHE, which is considered the holy grail of cryptography, has already been recognized by NIST (the National Institute of Standards and Technology) as the post-quantum cryptographic standard and is equally acknowledged by industry leaders such as Vitalik Buterin. As one of the few projects truly driving the adoption of FHE technology, Mind Network has further fueled the community’s high expectations for $FHE’s future performance.

Conclusion

We are truly excited about such a future:

In a vibrant Agentic World, with FHE as the cornerstone of security, Agentic AI will rapidly iterate towards greater intelligence, more efficient collaboration, and stronger execution capabilities, helping us solve the majority of our problems.

With the TGE launch on April 10, 2025, Mind Network will continue to push forward in areas such as products, ecosystem development, and more:

As key milestones are achieved, more users and developers will flock to the AgenticWorld ecosystem, generating increased interactions in areas like DeFi, social, and governance. The Agentic World we’ve envisioned is steadily unfolding before us.

As participants and witnesses of the Agentic AI wave, are you ready to experience it all?

Disclaimer:

1. This article is reprinted from [TechFlow], and the copyright belongs to the original author [TechFlow]. If you have any objections to the reprint, please contact the Gate Learn team, and the team will handle it as soon as possible according to relevant procedures.

2. Disclaimer: The views and opinions expressed in this article represent only the author’s personal views and do not constitute any investment advice.

3. Other language versions of the article are translated by the Gate Learn team and are not mentioned in Gate.io, the translated article may not be reproduced, distributed or plagiarized.