www.app.spectral.finance
Open in
urlscan Pro
2600:9000:266e:7200:17:5dda:3f00:93a1
Public Scan
Submitted URL: https://app.spectral.finance/tit%C2%A1f5%7D
Effective URL: https://www.app.spectral.finance/tit%C2%A1f5%7D/
Submission: On April 09 via api from US — Scanned from DE
Effective URL: https://www.app.spectral.finance/tit%C2%A1f5%7D/
Submission: On April 09 via api from US — Scanned from DE
Form analysis
0 forms found in the DOMText Content
Mainnet Alpha - v0.1.0 (Starburst) Sign In Build ML Models on SpectralBuild ML Models on SpectralJoin our first challenge to earn significant rewards and become an integral part of our Machine Intelligence Network. Join Challenge + First ChallengeLive First Challenge is LiveBuild a model to improve on-chain capital efficiency. Predict the probability of accounts entering liquidation on the world's largest DeFi lending protocols. Leverage custom datasets created from multiple years of transaction events. Join Challenge +View Challenge Leaderboard Spectral Web3 Credit Scoring Active challenge You are challenged to create a ML model that accurately forecasts borrower liquidation on the Aave and Compound lending protocols. $100,000 in rewards $50,000 additional rewards for top early submissions 85% ongoing revenue sharing for model consumption Data and tooling provided by * * * * * * * * * * * * * * * * * * * * * * * * * * Our Modeling Process: a Step-by-Step GuideFrom setting up your model to earning ongoing payouts. What you do 01 02 03 Configuration Create an account, install our CLI, and fetch the necessary training dataset. Model Training Using PyTorch neural network or logistic regression, train your model based on the provided dataset. Submission Submit your model's predicted liquidation probabilities, target labels, and feature values used. What Spectral does 04 05 06 Evaluation We will evaluate models' predictions against a weighted average of seven validation metrics. Verification We will use zk-proofs to ensure model integrity, while keeping model details confidential. Consumption We will aggregate top inferences for third-party use and award Modelers a share of consumption revenue. 01 02 03 04 05 06 Configuration Create an account, install our CLI, and fetch the necessary training dataset. Model Training Using PyTorch neural network or logistic regression, train your model based on the provided dataset. Submission Submit your model's predicted liquidation probabilities, target labels, and feature values used. Evaluation We will evaluate models' predictions against a weighted average of seven validation metrics. Verification We will use zk-proofs to ensure model integrity, while keeping model details confidential. Consumption We will aggregate top inferences for third-party use and award Modelers a share of consumption revenue. Get StartedLearn more on how to participate in our challenges. Modeler Handbook Getting Started Model training Submission Platform Overview Platform Overview Platform FAQs Spectral Tooling Spectral CLI Spectral SDK Challenge Timeline Why you should model on SpectralEarn enduring rewards and safeguard your IP as a Spectral Modeler Join Challenge + Win major bounties with tamper-proof validationGet substantial rewards from web3 companies when your models meet performance benchmarks. Benefit from provable, transparent, merit-based on-chain evaluations that ensure a level playing field. Shield your model IP for fair compensation with zkMLPreserve your intellectual property and model integrity with zero-knowledge ML verification. Secure sustainable, equitable compensation for your expertise and contributions. Earn ongoing, usage-based incomeUnlock long-term revenue streams from continuous consumption of your models. Made possible by challenges designed for perpetual participation. Upcoming Challenges Coming 2024 Trading & Price PredictionForecast short-term ETH/USDC returns in Uniswap's bustling V3 pool. Uniswap TrendsAccurately predict future transaction trends and liquidity patterns n the Uniswap platform. Sports ForecastingElevate sports analytics by predicting individual player stats and team performances in the NBA. Do you have an idea for a challenge?Connect with us to host it on Spectral. And hurry! We’re offering to fully sponsor challenge bounties for a limited number of early Creator partners. Contact Us Join the Spectral community on your modeling journey Join our SlackJoin our Discord Unlock Your Full Potential in the Web3 Space Get direct support from the Spectral DS team Attend office hours to get answers to your questions Connect with other participants and experts in challenge-related discussions Questions? Feedback?Find us on Slack and Discord. SlackDiscord BlogDocsContactPrivacy Policy Copyright © 2024 Spectral