<p><strong>Driving rewards to open source dependencies with a market of AIs as the engine and humans as the steering wheel, </strong>deep funding requires 3 components:</p><p><br></p><ul><li><strong>A data layer</strong> mapping all contributions to an ecosystem</li><li><strong>An open contest</strong> to crowdsource the relative value between contributions</li><li><strong>A jury</strong> to express preferences on a randomly selected set of contributions </li></ul><p><br></p><p>The model(s) best aligning with jury spot checks distribute funds across the dependency graph, thus <strong>scaling high quality human judgement</strong></p><p><br></p><p><img src="https://giveth.mypinata.cloud/ipfs/QmfMgtuPsdDRUASe8uNDeUn7sFb3W9f21xnQU8Qdf8VC3s"></p><p><br></p><p>In most funding rounds, selection procedures require a committee or donors to have an opinion on all eligible projects.</p><p><br></p><p>In deep funding, the jury votes carefully on a few projects. These selections serve as test data to choose a winning model that decides allocations to all projects, using a data science competition approach (like the ones on Kaggle or Pond) where anyone can submit models giving weights to projects in a round. </p><p><br></p><h1>Pilot Implementations</h1><p><br></p><p>Deep Funding kicked off with a <a href="https://huggingface.co/spaces/DeepFunding/PredictiveFundingChallengeforOpenSourceDependencies" rel="noopener noreferrer" target="_blank">competition</a> using funding data from 2019-2024 from Open Collective, Gitcoin, Optimism and Octant. Participants needed to build open source AI or metrics based models to predict funding received by projects in these past rounds.</p><p><br></p><p>You can see our proof of work <a href="https://research.allo.capital/t/submission-of-entries-to-the-deep-funding-mini-contest/22/17" rel="noopener noreferrer" target="_blank">here</a> - 15 high quality submissions from participants that are being judged for $20k in prizes!</p><p><br></p><h1>What's Next</h1><p><br></p><p>The next contest asks participants to submit machine learning or metrics based model that gives 15,000 weights indicating the contribution of ~5000 open source repos to the ethereum universe</p><p><br></p><p>The model that best aligns with juror rankings for a subset of these dependencies gets to allocate $170,000 based on their weights</p><p><br></p><p><img src="https://giveth.mypinata.cloud/ipfs/QmZJZttu3Z44HdfuzeFmDcGo6oj52pb2oMRXa4Ej5dRxac"></p><p>Simultaneously, we are spreading the meme of predictive funding challenges where models guess the amount received by a project before the round even starts. Proof of Work: <a href="https://gov.gitcoin.co/t/gcp-xxx-predictive-funding-competition-for-gg23/19902/6" rel="noopener noreferrer" target="_blank">https://gov.gitcoin.co/t/gcp-xxx-predictive-funding-competition-for-gg23/19902/6</a></p><p><br></p><p>Your support will help the 6 partners involved in Deep Funding;</p><p><br></p><p>VoiceDeck as StoryTeller Open Source Observer as Wizard Pairwise as Referee</p><p>Drips as Sprinkler Pollen Labs as Accelerator Allo as Weaver</p><p><br></p><p><br></p>