ML2SC: Deploying Machine Learning Models as Smart Contracts on the Blockchain

TL;DR

This paper introduces ML2SC, a tool that automatically translates PyTorch ML models into Solidity smart contracts, enabling on-chain deployment and inference with transparent, tamper-proof execution on blockchain.

Contribution

ML2SC is the first automated translator from PyTorch models to Solidity, facilitating secure on-chain ML inference with accurate cost modeling and maintained classification accuracy.

Findings

Gas costs increase linearly with model size and complexity.

On-chain inference accuracy matches off-chain PyTorch results.

ML2SC successfully deploys and updates models on blockchain.

Abstract

With the growing concern of AI safety, there is a need to trust the computations done by machine learning (ML) models. Blockchain technology, known for recording data and running computations transparently and in a tamper-proof manner, can offer this trust. One significant challenge in deploying ML Classifiers on-chain is that while ML models are typically written in Python using an ML library such as Pytorch, smart contracts deployed on EVM-compatible blockchains are written in Solidity. We introduce Machine Learning to Smart Contract (ML2SC), a PyTorch to Solidity translator that can automatically translate multi-layer perceptron (MLP) models written in Pytorch to Solidity smart contract versions. ML2SC uses a fixed-point math library to approximate floating-point computation. After deploying the generated smart contract, we can train our models off-chain using PyTorch and then further transfer the acquired weights and biases to the smart contract using a function call. Finally, the model inference can also be done with a function call providing the input. We mathematically model the gas costs associated with deploying, updating model parameters, and running inference on these models on-chain, showing that the gas costs increase linearly in various parameters associated with an MLP. We present empirical results matching our modeling. We also evaluate the classification accuracy showing that the outputs obtained by our transparent on-chain implementation are identical to the original off-chain implementation with Pytorch.