Forward the Original Title ‘解读 Allora 白皮书：自我改进的去中心化 AI 网络’

Meme is currently rampant in the market, and the AI track has entered a short rest period.

However, with Nvidia’s performance soaring and more AI industry events coming in the second half of the year, encrypted AI projects are still worthy of attention.

There is a new trend on the rise —-The combination of zkML (zero-knowledge machine learning) and AI agents. The former verifies the correctness of AI calculation results while ensuring privacy and security; the latter realizes automated task execution and decision-making through smart contracts and decentralized networks.

Some old encryption projects will take advantage of this new trend to adjust their business directions in an attempt to gain more value in the new cycle.

Allora Network is one of them.

Yesterday, AlloraOfficially announced its latest technical white paper, positioning itself as a “self-improving decentralized AI network” also means that the project business is moving closer to narrative hot spots.

At the same time, the project also announced its points incentive plan in May, which is of great interest to both hair-lovers and Alpha hunters.

As the AI track is already crowded, what makes Allora unique? Considering that its technical white paper is relatively complex, we have interpreted and analyzed it, and presented the key value points and project introduction to you in a more popular way.

The old problem of AI resource monopoly

Judging from the Allora white paper, the project is mainly aimed at old problems in the current AI field: computing power, algorithms and data are concentrated in the hands of a few giants, and resource monopoly is not conducive to the optimal state of machine learning (ML).

Allora believes that the key to building optimal machine intelligence is to maximize the number of connections in the network, allowing different data sets and algorithms to be freely combined in the network to obtain the most relevant insights.

Therefore, we need a form of swarm intelligence that can connect large data sets and inference algorithms.

In short, in existing encrypted AI projects, the cooperation between different models is not good enough, and there are also problems with the incentive methods. The models are either isolated or not closely connected and effective enough, resulting in unsatisfactory final reasoning results.

Vitalik also mentioned before, “A higher-level mechanism is needed to judge the performance of different AIs so that AI can participate as players.”

Allora’s goal is to enable nodes in the decentralized AI network to collaborate better through a better incentive structure; at the same time, introduce more intelligent ways to identify contextual details to improve the effectiveness of machine learning models, thereby achieving more efficient intelligence Reasoning and judgment.

Allora: Introducing context awareness and differentiated incentives to improve model performance

Specifically, how does Allora achieve a “better decentralized AI network”?

The key highlight is thatContext-aware and differentiated incentive structures.These innovations enable the network to deliver optimal inference results in any environment while providing fair rewards for each participant’s unique contributions.

But these two words sound a bit mysterious. We might as well take a look at the participants of the Allora network first.

Participants in the Allora network include workers, evaluators and consumers, each role has its specific responsibilities and roles:

1. Workers: Provide AI inference results and predict the loss value of other workers’ inference results.
2. Reputers: Evaluate the quality of inference results and predicted loss values ​​provided by workers.
3. Consumers: They request and pay to infer results from the network.

a network interact through a coordinator (Topic Coordinator):

• consumerRequest inference results from the network and pay a fee to get them.
• workerProvides inference results and a loss value for predicting the inference results of other workers. The coordinator synthesizes this information to generate more accurate inference results.
• evaluatorBased on the inference results and predicted loss values ​​provided by workers, evaluations are conducted using real data to ensure the fairness of the evaluation and are rewarded based on their consensus with other evaluators.

Through the design of these three roles, an efficient decentralized machine intelligence network is achieved, achieving the goal of optimizing resource utilization and improving inference accuracy. It is essentially a system that achieves self-improvement and fair rewards through role division and incentive mechanisms. design.

After understanding these three types of roles, it will be easier to look at Allora’s context awareness and differentiated incentive design.

Infer the synthesis mechanism

Allora’s inference synthesis mechanism is the key to its realization of decentralized machine intelligence. It is achieved through the following steps:

1. Inference Task: Each worker generates inference results using its own dataset and model.
2. Forecasting Task: Each worker predicts the loss value of the inference results of other workers. These predicted loss values represent the worker’s expected performance under current conditions.
3. Context-Aware Inference: The network uses the prediction loss value provided by the worker to generate a context-aware prediction inference result through a weighted average. These weighted averages take into account historical and context-dependent accuracy.
4. Network Inference: The final network inference is generated by combining the worker’s inference results with context-aware predicted inference results.

The key to this mechanism is that it not only assesses the historical accuracy of the model like other crypto projects, but also takes into account the current context, thereby achieving the best combination of inferences and improving the intelligence of the overall network.

Differentiated reward mechanism

At the same time, Allora introduces a differentiated reward mechanism to ensure that each participant’s contribution is fairly recognized:

1. worker rewards: Assigned based on their contribution to inference and prediction tasks, incentivizing them to provide high-quality data and predictions.
2. Reviewer rewards: Allocate rewards based on its closeness to the consensus and shares held to ensure the accuracy and fairness of the evaluation.
3. Overall reward distribution: The reward mechanism not only encourages participants’ positive contributions, but also avoids excessive concentration of a single participant through decentralized design.

Some solutions currently in use on Allora:

• AI Price Prediction:Provides precise, real-time asset price information critical to advanced financial primitives.
• Vault powered by artificial intelligence: Enable developers to implement advanced DeFi strategies and increase earning potential.
• Artificial Intelligence Risk Modeling: Allows protocols to build more secure systems to deal with external risks.
• AnyML: Provides easy integration of any machine learning model so that anyone (not just machine learning engineers) can build more powerful products using decentralized AI.

Token economy

The Allora network uses its native token ALLO to facilitate the exchange of value between network participants. Specific uses of ALLO tokens include:

1. Buy inference results: Users can use ALLO tokens to purchase inference results generated by the network. Allora adopts a “what are you willing to pay” (PWYW) model, allowing users to independently decide the ALLO fee to pay for inference.
2. Pay the participation fee: ALLO tokens can be used to pay for creating topics or participating in the network (as a worker, evaluator or network validator). Participation fees are variable.
3. pledge: Evaluators and network validators can use ALLO tokens for staking, and other token holders can also delegate their tokens to evaluators or network validators. The staking evaluators, verifiers and their delegators will receive ALLO rewards.
4. Incentive payment: The network uses ALLO tokens to pay rewards to participants. For workers, these rewards are proportional to their unique contribution to the accuracy of the network. For raters and network validators, these rewards are proportional to their stake and consensus.

Token value

The token economics in the Allora network are designed to ensure the intrinsic value and stability of tokens:

1. fee income: All fees collected by the network will be added to the network treasury to pay for reward issuance. This means that in practice, Network Depot will decay more slowly than a simple exponential decay, maintaining a high APY
2. Token recycling: Fees collected from network usage first pay out rewards before new tokens are minted. This means that depending on market dynamics, the circulating supply of ALLO can increase (corresponding to inflation) or decrease (corresponding to deflation)
3. Smooth issuance mechanism: By applying an exponential moving average, token issuance is smoothed, thus avoiding a sharp drop in APY when the main token is unlocked, ensuring that token holders continue to stake their tokens.
   

However, the white paper did not mention the release date and details of the token. For more information, you need to pay attention to its social media trends.

The resources behind Allora

The above content does not actually mention the zkML technology mentioned at the beginning of the article. It seems that Allora has nothing to do with this technology.

But behind Allora, the old project Upshot is a core contributor to Allora development.

Upshot enhances Allora’s capabilities by deploying its flagship price prediction model, which provides AI-driven price information for more than 400 million assets, on the network. The most accurate forecasts from the model have historically shown confidence levels of 95-99%.

Additionally, the model’s output can be accessed via zkPredictor (The largest on-chain zkML application to date) is provided to enable applications to consume the output in a cryptographically verifiable manner.

At the same time, Upshot also received US$22 million in financing in 2022 led by Polychain, Framework, CoinFund and Blockchain Capital. The direction at that time was to use technology to do real-time NFT asset evaluation. Now with the rise of AI, the track has also changed. , but the technology accumulated previously has also been applied to the new Allora.

Roadmap and testnet incentives

Judging from previous information on Allora’s official blog, the project’s launch is divided into three stages:

• Testnet Phase 1: Mid-February 2024
• Testnet Phase 2: Mid-March 2024
• Mainnet: Early Q2 2024

At this point in time, it seems that the project progress has been delayed, but it is still in the stage before the main network is launched.

In order to build momentum and allow more people to use it, Allora also launched the first phase of its testnet incentive plan on May 17. You can also earn points by participating in on-chain and off-chain activities to gain more airdrop expectations in the future.

Specific activities that can earn points include:

On-chain activities

1. Create topics: Identify and define specific issues or areas of interest within the network, engaging other actors to develop and deliver solutions.
2. Introduce machine learning models: Add machine learning models to the network for others to use.
3. Use Allora Powered Apps: Participate in apps and services that leverage Allora’s machine intelligence capabilities

Off-chain activities

1. Community involvement: Follow Allora on Twitter and join the Discord and Telegram groups.
2. Participate in the community: Participate in select community events and activities to support the Allora network.

Currently, activities that are easy for ordinary users to participate in can be found on the Galxe event page. Interested players canClick here to participate

Generally speaking, Allora is an encryption project with certain technological innovation, background resources and capability reuse. It can follow the trend in the transformation of AI hot spots and maximize its use of its capabilities to expand new business directions. At least it can ensure that it attracts new attention. Never be left behind in a war.

As for how high the upper limit is, firstly, it depends on waiting for the AI wind to blow again, and secondly, it depends on more operational methods of the project in the future.

Disclaimer：

1. This article is reprinted from [Techflow]. Forward the Original Title ‘解读 Allora 白皮书：自我改进的去中心化 AI 网络’.All copyrights belong to the original author [TechFlow]. 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.