ETH London 2023 was more than just an event for FLock.io; it was an opportunity to connect, engage, and celebrate with our vibrant community right in our backyard. With our core team members based in London and as a proud sponsor , we were thrilled to be at the heart of the action. We had a packed schedule during the ETH London week, and here’s a recap of the highlights.

🌃 Decentralised Machine Learning (DML) Night

The DML Night, hosted by FLock.io, was a gathering of minds passionate about the evolving world of decentralized machine learning. The symposium was tailored for both builders and researchers, providing a platform to delve deep into the intricacies and potential of DML.

Jiahao from FLock.io kickstarted the evening with his opening remarks. Zhipeng Wang, a PhD candidate at Imperial College London and a FLock Scholar, delved into the advancements in federated learning with his talk on Zero-Knowledge Proof-based Gradient Aggregation. Alexander Dante Camuto, co-founder of EZKL, demystified the realm of Zero-Knowledge Machine Learning.

The highlight of the evening was the panel discussion on ZK & DML. Zeo from HomeDAO expertly moderated the session, with Jiahao, Dante, and Harry sharing the panel. The discourse was rich, touching upon pressing topics like the challenges in AIxWeb3 hiring, the unique propositions of their respective companies, and their collective vision for the future of decentralised ML. The panel highlighted recent progress: Gensyn is building its testnet, and EZKL has released its codebase, enabling new interesting applications of zero-knowledge proofs. The panelists also explored potential synergies, especially concerning FLock’s training framework, Gensyn’s decentralized computation, and EZKL’s zk libraries.

A notable point of discussion was the continued significance of London as a hub of innovation, especially in the realms of AI and web3. The panelists also shared their perspectives on the current trends in AIxWeb3, dissecting the nuances of Data layers, compute layers, training layers, and inference layers.

💻FLock AIxWeb3 Tool Bounty Hackathon

In collaboration with TRIAS and TusimaNetwork, the hackathon showcased the innovative spirit of the blockchain community. Our bounty was centered on the innovative applications of our fine-tuned model. We invited the community to embark on one of the following tracks: Integrating FLock with TON for Next-Gen Telegram Bots, Unlocking the Power of Intent with FLock, and the Freestyle Innovation Track. The response was overwhelming with 10 submissions, out of which 8 projects were identified as completed. Engaging with the hackers during the judging panel was enlightening, and the motivation behind their projects was truly inspiring. Here are our standout winners:

• Quantum Oracle — We love the idea of integration of quantum computing with blockchain, leveraging AI for code generation!
• Chat Champion — A novel approach to visualizing user engagement within communities using sentiment analysis, redefining the capabilities of a TG bot.
• Wsydom — Envisioned as the notebook for blockchain, this dev tool simplifies coding. Despite the model’s data limitations, they innovatively improved code generation.
• Lester — An on-chain “expense system” that optimizes transaction processes, identifying the quickest and most cost-effective payment routes in an era of multiple bridges and chains.

Congratulations to the winners and looking forward to building together!

ETH London 2023 was a celebration of community and innovation. FLock.io, with its roots in London, is proud to be part of this dynamic ecosystem. As we move forward, our community remains our guiding star, leading us towards a decentralized, AI-driven future. We’re excited about the journey ahead and look forward to seeing you in Istanbul next.

About FLock.io

FLock.io is a decentralized and permissionless platform for co-owned AI models and Dapps. By harnessing the synergies of Federated Learning and blockchain technologies, we address the growing demands of models and potential data breach threats, guaranteeing secure model training without revealing underlying source data, and fair rewards to data contribution and community collaboration.

🌐Website｜👾Discord｜🐦Twitter｜📺Youtube｜📖Medium｜👥Telegram Chat Group｜📢Telegram Channel