Split-n-Chain: Privacy-Preserving Multi-Node Split Learning with Blockchain-Based Auditability
Mukesh Sahani, Binanda Sengupta
TL;DR
This work tackles privacy in distributed deep learning by moving beyond standard federated learning to a privacy-preserving multi-node split-learning framework called Split-n-Chain, which distributes neural-network layers across data-source nodes, intermediate nodes, and a terminal node while recording computations on a blockchain. It introduces a bootstrapping phase that estimates intermediate-node compute via PoW to determine layer distribution using power fractions $CPF_j$, and then trains the split model with forward/backward passes, where each layer’s outputs and gradients are committed as transactions to a blockchain governed by Algorand consensus. An incentive mechanism rewards intermediate nodes proportional to the number of parameters they train, provided the training loss converges below a threshold $ au$, thereby discouraging malicious behavior. Experimental evaluation on MNIST in both a local setup and a LAN demonstrates that the training loss trajectories remain close to monolithic training, while the overhead from bootstrapping and inter-node communication scales with the number of intermediate nodes, indicating practical viability for privacy-preserving multi-node deep learning.
Abstract
Deep learning, when integrated with a large amount of training data, has the potential to outperform machine learning in terms of high accuracy. Recently, privacy-preserving deep learning has drawn significant attention of the research community. Different privacy notions in deep learning include privacy of data provided by data-owners and privacy of parameters and/or hyperparameters of the underlying neural network. Federated learning is a popular privacy-preserving execution environment where data-owners participate in learning the parameters collectively without leaking their respective data to other participants. However, federated learning suffers from certain security/privacy issues. In this paper, we propose Split-n-Chain, a variant of split learning where the layers of the network are split among several distributed nodes. Split-n-Chain achieves several privacy properties: data-owners need not share their training data with other nodes, and no nodes have access to the parameters and hyperparameters of the neural network (except that of the respective layers they hold). Moreover, Split-n-Chain uses blockchain to audit the computation done by different nodes. Our experimental results show that: Split-n-Chain is efficient, in terms of time required to execute different phases, and the training loss trend is similar to that for the same neural network when implemented in a monolithic fashion.