0xDeCA10B

Why we included this project

Microsoft's 0xDeCA10B tackles a hard problem: how to let many parties contribute data to a shared machine learning model without any single owner controlling it. The whole lifecycle runs on a blockchain, so adding data goes through an incentive mechanism that may require a stake, a data handler that stores records on-chain, and a training step that updates the model. Predictions are then free for anyone to query. If you are exploring decentralized or collaborative ML, or curious about how smart contracts can govern model updates and reward contributors, this is a concrete reference implementation with a working demo and simulation, not just a white paper. It is a research prototype rather than a production service, so treat it as a design template for incentive-based collaborative training, not something to deploy as-is.