# Introduction

Monte Carlo lets you run AI models with a cloud API or smart contract on DePIN with reliability, without having to manage your own infrastructure. 

You can run open-source models that other people have published, or package and publish your own models.

Leveraging the reputation system and on-chain validation pattern, we coordinate computing power to offer more affordable and reliable AI services for AI applications and smart contracts.

We aim to establish an open network that bridges the gap between computing power and developers, **making AI accessible to everyone**.

### Official Links

Homepage: <https://montecarlo.io>

Docs: <https://docs.montecarlo.io>

<br>


# Modular AI

Monte Carlo has proposed a novel modular architecture to provide reliable and affordable AI services for applications and smart contracts.

The token on the chain is key to the modular architecture, and it will trigger the actions of offchain worker such as coordinator, validator and orchestrator.

<figure><img src="https://155903546-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2F7MVCn8VzlTz60MGwqYP4%2Fuploads%2FcxZhkkL7huxKbTnjztH1%2FXnip2024-04-04_15-27-31.jpg?alt=media&amp;token=d2c8206e-ace0-4f4a-8bf3-e65ed539b0da" alt=""><figcaption></figcaption></figure>

There are mainly three types of offchain workers:

* **Coordinator**: Distribute AI instructions to nodes that can complete them, such as inferencing.
* **Validator**: Verify the results of task execution to avoid malicious nodes based on the optimistic algorithm.
* **Orchestrator**: Orchestrate complex AI tasks via instruction flow to meet the demand for task combinations.


# Coordinator and Validator

The task will be published to the Monte Carlo blockchain in the form of the Task Scheduling Token (TST) and then scheduled by the off-chain coordinator for the miner to run.

The Task Scheduling Token Standard is a token standard customized for task scheduling. The task scheduling information will be stored on the chain, and the parameters and the output results will be stored off-chain.

<figure><img src="https://155903546-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2F7MVCn8VzlTz60MGwqYP4%2Fuploads%2FVM9Zokj8WgilqDEXMVqA%2FXnip2024-04-04_16-28-00.jpg?alt=media&amp;token=a85be4f4-b450-45d4-a721-d22d11607021" alt=""><figcaption></figcaption></figure>

### Reputation-based scheduling

There is a reputation mechanism for miners, which serves as an important reference standard for scheduling.

When registering with Miners, they need to submit their benchmark data as the basis for scheduling. Subsequent scheduling and reward allocation will be centered around the benchmark for reputation rating.

New devices will default to a full reputation, ensuring fair scheduling. Each time a task takes longer than the benchmark, it will reduce reputation. The reputation will gradually recover through subsequent qualified tasks.

For those who need highly reliable results and don't have high requirements of timing, we can prioritize the scheduling of miners with good reputations. For those with high time requirements, they can increase the gas price to match higher computing miners.

### Consistency-based Validation

We can obtain the same output with the same input by fixing the random seed, thereby verifying that the output result is computed by the model. This has certain requirements for the model, it needs to adopt standardized data preprocessing technology, model architecture and training process. The existing mainstream AI models such as LLaMA and Stable Diffusion can all ensure consistency.

Optimistically, we assume that the committed result is correct. There exists a validation period during which the validators can verify the results. If the validation is passed, the provided result will be valid and accepted.

We will cooperate with DePHY, continuously collect the metrics data of the device, and detect outliers for the validator to find the fake miner more accurately. In the future, we will also consider incorporating opML for on-chain result verification.

### Weighted-based Rewarding

A computing task can be allocated to multiple miners to ensure the reliability of task scheduling.&#x20;

The system will reward based on the miner's benchmark and actual execution time.&#x20;

If the results are verified to be incorrect, there will be no reward. The staked tokens will also be slashed as a penalty.&#x20;

If the time consumed is longer than others, or the time spent is more than the benchmark's commitment, it will lead to a reduction in the weight of the reward.&#x20;

At the end of the validation period, rewards are automatically given to the miners.

<br>

<br>

<br>


# DePHY

DePHY provides an open-source hardware solution, supplemented with SDKs and tools, which significantly cuts down both the manufacturing and blockchain network communication expenses.

Through a deep integration with DePHY, Monte Carlo can operate node synchronization at the 500ms level. In addition, DePHY employs Soulbound DID for devices with secure, tamper-proof hardware, and uses ZK technology in Oracles, guaranteeing traceable, confidential, and verifiable network messages.

### Official Links

Homepage: <https://dephy.io/>


# Press Kit

Welcome to the Monte Carlo Press Kit section. Here, you will find a comprehensive collection of resources and materials designed to assist you in keeping up-to-date with Monte Carlo's latest developments. We appreciate your interest in our platform and encourage you to use the information provided below to communicate our brand accurately and effectively.

{% file src="/files/Au5PFEESEaxyzl4rQdre" %}

We appreciate your interest in Monte Carlo and look forward to collaborating with you. If you require further assistance or have specific design-related requests, please do not hesitate to contact our team.