All You Need to Know About OORT

Intermediate7/16/2024, 4:24:57 AM
OORT 是一個成立於 2021 年五月的去中心化 AI 的雲平台,提供的一系列服務包括數據收集和標註、數據儲存和模型訓練,旨在透過去中心化網絡幫助 AI 開發者,為其搭建更公平高效的 AI 應用奠定基礎,涵蓋數據收集和標註、模型訓練和微調等關鍵階段,建構一個以信任爲核心的生態系統,確保人類與 AI 能夠共同繁榮。

OORT Vision: Facilitating Trustworthy AI for Humanity

Since 2023, AI technologies like ChatGPT, Google Gemini, and OpenAI Sora have profoundly impacted us. They are, however, poorly understood as to how they collect and use data or even the biases that may be present.

Imagine an AI trained mainly on data from one specific group. This could make the AI unintentionally favor, create biases, or disadvantage certain groups when making decisions. To avoid this, it is essential to use varied data from different groups. Therefore, AI can be fair and inclusive.

As these technologies become more integral parts of our lives, trust matters most. As humans spend time together and understand each other’s strengths with respect to capabilities than we do with technology, so does our trust build over time. For us to trust AI better we should study its operations, ensure that it uses unbiased data, and also have a variety of perspectives involved in its design. This approach emphasizes transparency and diverse input, which is crucial in making AI that everybody can rely upon and gain from.

As we navigate these advancements, our goal at OORT is simple: to build an ecosystem where trust is paramount, ensuring humans and AI can thrive together.

OORT Mission:Cloud for Decentralized AI

Founded in May 2021, OORT is a cloud platform for decentralized AI. Our network utilizes global resources ranging from data centers to edge devices, with a proprietary blockchain-based verification layer. OORT offers a suite of services, including data collection and labeling, data storage, and model training. We aim to help AI builders through decentralized networks and lay the foundation for a better future where AI development happens more equitably and efficiently.

In other words, OORT proposes a new way of developing AI that’s open and transparent, involving key stages such as data collection and labeling, model training, and fine-tuning. By involving the community in these stages, we aim to make AI easier to understand and promote collaboration from the start. Our goal is to democratize AI development, making it more inclusive and building trust through everyone’s active participation.

How Centralized AI (CeAI) Works


Figure 1. Centralized Al. All stages are fully controlled by centralized organizations and viewed as a Blackbox for the public.

First, let’s explore how centralized AI operates (refer to Figure 1). Companies like OpenAI and Google run their AI systems in a centralized manner, using their own data centers and conducting undisclosed processes. Users don’t know how each stage of AI development happens or how data is used to train and enhance these models.

How Decentralized AI (DeAI) Works

In contrast, OORT utilizes blockchain technology to redefine each stage, as illustrated in Figure 2.


Figure 2. Decentralized Al. Each stage is contributed by community, transparent to the public.

Data Collection: Anyone, anywhere in the world, can share data using their own laptops or smartphones, and they’ll get rewarded for sharing.

‍Data Preprocessing: People can use their smartphones and PCs to help process data such as data labeling for AI projects. After that process ends, the data will be updated to OORT decentralized cloud, while an identifier (hash) for the processed dataset is stored on a blockchain.

Data Model Design: Members of the community work together to create AI models that anyone can use and build upon, thereby making innovation easier and more accessible to all.

Model Training and Fine-Tuning: With a fraction of the cost required to access services such as AWS or Google Cloud, AI engineers can now connect to under-utilized global computing networks. More importantly, the training process such as the time stamp of training and the dataset used for training will be all recorded on our blockchain, which enhances transparency and accountability.

Model Deployment: In contrast to relying on big servers like those owned by AWS, AI app & software developers can now enable users to deploy models directly on their devices, such as smartphones and PCs. This allows users to use AI applications offline, ensuring that personal data remains secure and is not transmitted to third parties. By leveraging the capabilities of their own devices, users can enjoy the flexibility and convenience of AI tools anytime and anywhere.

OORT Products

At OORT, we enable developers to build AI in a transparent, decentralized, and trustworthy way. Unlike other systems developed in isolation, OORT products involves the entire community. This means that instead of being mere spectators, everyone can actively participate as we strive to improve the state of the art in AI. We are determined to democratize artificial intelligence so that its evolution reflects a broad range of perspectives. As of June 2024, OORT has served over 10 thousand enterprise and individual customers such as DELL, Lenovo Image and Binance Smart Chain.

OORT Datahub

It assists AI projects by collecting and labeling data provided by the global OORT community. Blockchain technology has been used to maintain an unambiguous, permanent record of every move in this process.

OORT Storage

Decentralized storage is achieved by combining global servers like Seagate, Tencent Cloud, and our own edge device Deimos. It works like Amazon AWS S3, but it’s 60% cheaper.

OORT Compute

The integration of geo-distributed GPUs makes them cost-effective for AI model training and inference. Coming in 2024.

OORT’s Decentralized Physical Infrastructure (DePin)

OORT’s decentralized infrastructure builds the foundation of its products. OORT’s physical decentralized infrastructure (DePin) consists of two main components: geo-distributed physical nodes that provide storage and computing capabilities, and OORT mainnet (Olympus Protocol) that connects these nodes and ensures these nodes operate securely and reliably.


Figure 3. OORT Architecture

Three-Types Physical Nodes

OORT’s DePin system consists of three types of physical nodes:

Super nodes: These are servers from OORT partners like Seagate and Tencent Cloud. They handle tasks such as training AI models and storing large amounts of data.

Archiving nodes: These nodes from Filecoin and Arweave store data securely. They’re used mainly for backing up AI datasets.

Edge nodes: These are devices owned by users, such as OORT’s “Deimos” edge devices, laptops, and PCs. They handle tasks like storing data locally, labeling it, and running AI models for quick decision-making.

Olympus Protocol

The Olympus Protocol is a proprietary blockchain technology of OORT that can be deployed by any party interested. The detailed technical description and performance are given in the Yellow Paper. Connecting all DePin devices globally across OORT, Olympus uses its specific proof of honesty (PoH) consensus algorithm that ensures each node works transparently and utilizes network resources optimally in a decentralized manner.

Learn more about the proof of honesty consensus algorithm.

OORT Tokenomics

OORT has its own native utility token within its system. It is used to pay for services, can be staked for rewards, and enables community members to make decisions.


Figure 4. OORT Tokenomics

Payment to OORT Products: You can use OORT products by paying fiat currencies via credit cards. However, the use of OORT tokens will always attract discounts on these purchases.

Staking: There are two ways in which you can stake OORT tokens:

  1. One can put them in staking pools for rewards such as DeFi, OORT Earn and DEX liquidity pools.
  2. One can delegate them through bonding pools (e.g., OORT Boost) to node providers who use them as collateral for their commitment to the network. One will receive a share of mining rewards from the node provider

Governance: In OORT, community members propose and make major decisions such as fee setting and rule establishment among others. For instance, if you want to propose something, you need to deposit some tokens that are refunded once your proposal gets approved; otherwise, they go to the node providers actively supporting the network.

OORT’s Competitive Edge in the Decentralized AI Space

In decentralized AI space, there are different projects focusing on various aspects such as managing AI data, training AI models and developing AI algorithms. Why does OORT stand out?

  • OORT provides a one-stop-shop democratic platform where any kind of AI project can find everything needed from start to end with focus on AI transparent and cost-effective.
  • OORT’s three products collaborate seamlessly to achieve optimal outcomes and the best enterprise-grade experience in AI development.

Investors and Team

As of Jan 2024, OORT has raised $10M funds from institutional investors such as Taisu Ventures, (aka, Emurgo - Cardano Venture Arm), Red Beard Ventures (funded by A16Z and Animoca Brands), Angelist, Trinity Venture Capital, Linkvc and Waterdrip Capital, and individual investors from JP Morgan, Columbia University, Cornell University, Ausum Ventures and ISKER Group. OORT also received grants from Microsoft & Google,.

Based in United States, OORT core members come from world-renowned organizations, such as Columbia University, Qualcomm, AT&T, J.P. Morgan, etc.

Disclaimer:

  1. This article is reprinted from [oortfoundation]. All copyrights belong to the original author [oortfoundation]. If there are objections to this reprint, please contact the Gate Learn team, and they will handle it promptly.
  2. Liability Disclaimer: The views and opinions expressed in this article are solely those of the author and do not constitute any investment advice.
  3. Translations of the article into other languages are done by the Gate Learn team. Unless mentioned, copying, distributing, or plagiarizing the translated articles is prohibited.

All You Need to Know About OORT

Intermediate7/16/2024, 4:24:57 AM
OORT 是一個成立於 2021 年五月的去中心化 AI 的雲平台,提供的一系列服務包括數據收集和標註、數據儲存和模型訓練,旨在透過去中心化網絡幫助 AI 開發者,為其搭建更公平高效的 AI 應用奠定基礎,涵蓋數據收集和標註、模型訓練和微調等關鍵階段,建構一個以信任爲核心的生態系統,確保人類與 AI 能夠共同繁榮。

OORT Vision: Facilitating Trustworthy AI for Humanity

Since 2023, AI technologies like ChatGPT, Google Gemini, and OpenAI Sora have profoundly impacted us. They are, however, poorly understood as to how they collect and use data or even the biases that may be present.

Imagine an AI trained mainly on data from one specific group. This could make the AI unintentionally favor, create biases, or disadvantage certain groups when making decisions. To avoid this, it is essential to use varied data from different groups. Therefore, AI can be fair and inclusive.

As these technologies become more integral parts of our lives, trust matters most. As humans spend time together and understand each other’s strengths with respect to capabilities than we do with technology, so does our trust build over time. For us to trust AI better we should study its operations, ensure that it uses unbiased data, and also have a variety of perspectives involved in its design. This approach emphasizes transparency and diverse input, which is crucial in making AI that everybody can rely upon and gain from.

As we navigate these advancements, our goal at OORT is simple: to build an ecosystem where trust is paramount, ensuring humans and AI can thrive together.

OORT Mission:Cloud for Decentralized AI

Founded in May 2021, OORT is a cloud platform for decentralized AI. Our network utilizes global resources ranging from data centers to edge devices, with a proprietary blockchain-based verification layer. OORT offers a suite of services, including data collection and labeling, data storage, and model training. We aim to help AI builders through decentralized networks and lay the foundation for a better future where AI development happens more equitably and efficiently.

In other words, OORT proposes a new way of developing AI that’s open and transparent, involving key stages such as data collection and labeling, model training, and fine-tuning. By involving the community in these stages, we aim to make AI easier to understand and promote collaboration from the start. Our goal is to democratize AI development, making it more inclusive and building trust through everyone’s active participation.

How Centralized AI (CeAI) Works


Figure 1. Centralized Al. All stages are fully controlled by centralized organizations and viewed as a Blackbox for the public.

First, let’s explore how centralized AI operates (refer to Figure 1). Companies like OpenAI and Google run their AI systems in a centralized manner, using their own data centers and conducting undisclosed processes. Users don’t know how each stage of AI development happens or how data is used to train and enhance these models.

How Decentralized AI (DeAI) Works

In contrast, OORT utilizes blockchain technology to redefine each stage, as illustrated in Figure 2.


Figure 2. Decentralized Al. Each stage is contributed by community, transparent to the public.

Data Collection: Anyone, anywhere in the world, can share data using their own laptops or smartphones, and they’ll get rewarded for sharing.

‍Data Preprocessing: People can use their smartphones and PCs to help process data such as data labeling for AI projects. After that process ends, the data will be updated to OORT decentralized cloud, while an identifier (hash) for the processed dataset is stored on a blockchain.

Data Model Design: Members of the community work together to create AI models that anyone can use and build upon, thereby making innovation easier and more accessible to all.

Model Training and Fine-Tuning: With a fraction of the cost required to access services such as AWS or Google Cloud, AI engineers can now connect to under-utilized global computing networks. More importantly, the training process such as the time stamp of training and the dataset used for training will be all recorded on our blockchain, which enhances transparency and accountability.

Model Deployment: In contrast to relying on big servers like those owned by AWS, AI app & software developers can now enable users to deploy models directly on their devices, such as smartphones and PCs. This allows users to use AI applications offline, ensuring that personal data remains secure and is not transmitted to third parties. By leveraging the capabilities of their own devices, users can enjoy the flexibility and convenience of AI tools anytime and anywhere.

OORT Products

At OORT, we enable developers to build AI in a transparent, decentralized, and trustworthy way. Unlike other systems developed in isolation, OORT products involves the entire community. This means that instead of being mere spectators, everyone can actively participate as we strive to improve the state of the art in AI. We are determined to democratize artificial intelligence so that its evolution reflects a broad range of perspectives. As of June 2024, OORT has served over 10 thousand enterprise and individual customers such as DELL, Lenovo Image and Binance Smart Chain.

OORT Datahub

It assists AI projects by collecting and labeling data provided by the global OORT community. Blockchain technology has been used to maintain an unambiguous, permanent record of every move in this process.

OORT Storage

Decentralized storage is achieved by combining global servers like Seagate, Tencent Cloud, and our own edge device Deimos. It works like Amazon AWS S3, but it’s 60% cheaper.

OORT Compute

The integration of geo-distributed GPUs makes them cost-effective for AI model training and inference. Coming in 2024.

OORT’s Decentralized Physical Infrastructure (DePin)

OORT’s decentralized infrastructure builds the foundation of its products. OORT’s physical decentralized infrastructure (DePin) consists of two main components: geo-distributed physical nodes that provide storage and computing capabilities, and OORT mainnet (Olympus Protocol) that connects these nodes and ensures these nodes operate securely and reliably.


Figure 3. OORT Architecture

Three-Types Physical Nodes

OORT’s DePin system consists of three types of physical nodes:

Super nodes: These are servers from OORT partners like Seagate and Tencent Cloud. They handle tasks such as training AI models and storing large amounts of data.

Archiving nodes: These nodes from Filecoin and Arweave store data securely. They’re used mainly for backing up AI datasets.

Edge nodes: These are devices owned by users, such as OORT’s “Deimos” edge devices, laptops, and PCs. They handle tasks like storing data locally, labeling it, and running AI models for quick decision-making.

Olympus Protocol

The Olympus Protocol is a proprietary blockchain technology of OORT that can be deployed by any party interested. The detailed technical description and performance are given in the Yellow Paper. Connecting all DePin devices globally across OORT, Olympus uses its specific proof of honesty (PoH) consensus algorithm that ensures each node works transparently and utilizes network resources optimally in a decentralized manner.

Learn more about the proof of honesty consensus algorithm.

OORT Tokenomics

OORT has its own native utility token within its system. It is used to pay for services, can be staked for rewards, and enables community members to make decisions.


Figure 4. OORT Tokenomics

Payment to OORT Products: You can use OORT products by paying fiat currencies via credit cards. However, the use of OORT tokens will always attract discounts on these purchases.

Staking: There are two ways in which you can stake OORT tokens:

  1. One can put them in staking pools for rewards such as DeFi, OORT Earn and DEX liquidity pools.
  2. One can delegate them through bonding pools (e.g., OORT Boost) to node providers who use them as collateral for their commitment to the network. One will receive a share of mining rewards from the node provider

Governance: In OORT, community members propose and make major decisions such as fee setting and rule establishment among others. For instance, if you want to propose something, you need to deposit some tokens that are refunded once your proposal gets approved; otherwise, they go to the node providers actively supporting the network.

OORT’s Competitive Edge in the Decentralized AI Space

In decentralized AI space, there are different projects focusing on various aspects such as managing AI data, training AI models and developing AI algorithms. Why does OORT stand out?

  • OORT provides a one-stop-shop democratic platform where any kind of AI project can find everything needed from start to end with focus on AI transparent and cost-effective.
  • OORT’s three products collaborate seamlessly to achieve optimal outcomes and the best enterprise-grade experience in AI development.

Investors and Team

As of Jan 2024, OORT has raised $10M funds from institutional investors such as Taisu Ventures, (aka, Emurgo - Cardano Venture Arm), Red Beard Ventures (funded by A16Z and Animoca Brands), Angelist, Trinity Venture Capital, Linkvc and Waterdrip Capital, and individual investors from JP Morgan, Columbia University, Cornell University, Ausum Ventures and ISKER Group. OORT also received grants from Microsoft & Google,.

Based in United States, OORT core members come from world-renowned organizations, such as Columbia University, Qualcomm, AT&T, J.P. Morgan, etc.

Disclaimer:

  1. This article is reprinted from [oortfoundation]. All copyrights belong to the original author [oortfoundation]. If there are objections to this reprint, please contact the Gate Learn team, and they will handle it promptly.
  2. Liability Disclaimer: The views and opinions expressed in this article are solely those of the author and do not constitute any investment advice.
  3. Translations of the article into other languages are done by the Gate Learn team. Unless mentioned, copying, distributing, or plagiarizing the translated articles is prohibited.
Start Now
Sign up and get a
$100
Voucher!