Using SKALE-ML

What is SKALE-ML

SKALE-ML is a unique tool that allows SKALE developers to combine trained machine learning models with smart contracts. It provides a method to forward propagate data and control smart contract logic based on model outcomes.

  • Use TensorFlow/Keras ML models to control dApp logic.

  • Decentralize and store trained models using SKALE Filestorage.

  • Leverage SKALE’s zero-cost gas to perform forward propagation in EVM.

Example ideas of what you can build with SKALE-ML

  • Autonomously filter file uploads using a decentralized ResNet neural network.

  • Identify malicious transactions and pause a bridge using a KNN classifier.

  • Generate NFT images using a decentralized GAN model.

  • Re-balance a DAO treasury using a portfolio optimizer.

  • Build an IMA chatbot using an NLP model.

SKALE-ML @ EthBoston
AnonHero @ EthBoston

How it works

  1. Developers build and train a model using TensorFlow/Keras.

  2. Developers upload a trained h5 model and file converter to a SKALE Chain’s filestorage.

  3. SKALE-ML contracts and precompiles run the predict() method on users' inputs.

  4. SKALE-ML forward propagates input data through the h5 model and outputs a file with results.

  5. Endpoints call getResults() to receive the prediction output.

Want to know more?

Reach out to the SKALE core team: https://discord.gg/skale

Edit on Github