Inspiration
Our inspiration for integrating AI inference with blockchain technology comes from the emerging need to enhance decentralized applications (DApps) with real-time, data-driven insights. By utilizing Chainlink, we aim to create a reliable bridge between private, secure AI computations and blockchain ecosystems, enabling DApps to leverage predictive analytics and other AI functionalities.
What it does
2PM.Network leverages AI to process and infer valuable insights from encrypted data, then uses Chainlink to securely and verifiably bring these AI inference results onto blockchain networks. This integration enables DApps, such as those used for market predictions, to utilize real-time AI-driven data without compromising on privacy or security.
How we built it
We built 2PM.Network using a combination of privacy-preserving computation technologies and blockchain integration. AI models are trained on encrypted data using techniques such as federated learning and homomorphic encryption to ensure privacy. We then employ Chainlink oracles to provide a secure and trustless mechanism for transferring AI inference results to various blockchain platforms. This is facilitated by developing external adapters in Chainlink that convert our AI results into blockchain-readable formats.
Challenges we ran into
One of the main challenges was ensuring the integrity and privacy of the data throughout the process, from training AI models on sensitive data to securely transmitting inference results via Chainlink.
Accomplishments that we're proud of
We are particularly proud of our ability to seamlessly integrate advanced AI capabilities with blockchain technology, creating a new paradigm for privacy-preserving, AI-driven DApps. Our successful deployment of external adapters in Chainlink that accurately relay AI inference to blockchain networks stands as a testament to our innovative approach.
What we learned
Throughout the development process, we gained substantial knowledge about the functionality of external adapters and the broader capabilities of Chainlink. We learned how to effectively use Chainlink to bridge the gap between off-chain AI computations and on-chain applications, ensuring both the reliability and security of the data transferred.
What's next for 2PM.Network
Looking forward, we plan to expand our use of external data sources for both training and inference tasks, further enriching the AI models and the accuracy of their predictions. This will involve more complex integrations with Chainlink to handle a variety of data types and formats. Additionally, we aim to enhance our platform's scalability and introduce more features to support a wider range of DApps, thereby strengthening our position at the intersection of AI and blockchain technology.
Log in or sign up for Devpost to join the conversation.