Instruere

Deployer_Interface
Mining_Interface
Block_Addition

Inspiration

Instruere was inspired by the inefficiencies in traditional blockchain mining, the inaccessibility of AI training for smaller developers, and the underutilization of everyday devices like TPU-enabled smartphones. We saw an opportunity to combine AI training with blockchain, empowering users to contribute to machine learning while earning rewards.

What it does

Instruere decentralises AI model training by allowing participants to fine-tune AI models locally using their devices. Scripts are stored on IPFS, fetched by miners, and executed to produce machine-learning outputs. Successful contributions are verified through Hugging Face and rewarded with tokens via smart contracts.

How we built it

We built Instruere by combining Web3 technologies with federated machine learning. The platform uses IPFS for decentralized storage, Hugging Face for model validation, and blockchain smart contracts to manage rewards. The system supports Python script deployment through a one-click interface, leveraging both laptops and smartphones for training AI models.

Challenges we ran into

We faced challenges in integrating decentralised storage with smart contracts and ensuring compatibility across various devices. Additionally, managing computational load distribution based on device specifications and verifying model contributions through Hugging Face was technically demanding.

Accomplishments that we're proud of

We successfully created a one-click deployer interface that simplifies training script deployment, developed a reward mechanism that incentivises AI contributions and onboarded early users through educational sessions. The combination of blockchain and AI is a unique accomplishment we're excited about.

What we learned

We gained deeper insights into the complexities of combining AI and blockchain, particularly in areas of decentralized storage, federated learning, and multi-device compatibility. We also learned the importance of user education in driving adoption.

What's next for Instruere

Next, we plan to scale Instruere by expanding user participation, improving model validation processes, and optimizing the reward distribution system. We're also looking to integrate more diverse AI models and extend the platform to support additional devices for training, further pushing the boundaries of decentralized AI development.