Blockchain Federated Learning for Sustainable Retail: Reducing Waste through Collaborative Demand Forecasting

TL;DR

This work tackles privacy barriers to cross-retailer demand forecasting in grocery retail, which contributes to food waste. It proposes a Blockchain-enabled Federated Learning framework with Secure Aggregation, Global Differential Privacy, and IPFS for verifiable, privacy-preserving training, formalized by the FL objective $\min f(w) = \sum_{k=1}^N \frac{n_k}{n} F_k(w)$. Empirical results on the Walmart dataset show Federated Learning achieving performance close to centralized training and significantly better than isolated models, while reducing over-provisioning error and waste. The study also assesses blockchain platform costs, finding Layer-2 solutions offer cost benefits but that scalability and security trade-offs remain, pointing to incentives and standardization as key areas for practical deployment.

Abstract

Effective demand forecasting is crucial for reducing food waste. However, data privacy concerns often hinder collaboration among retailers, limiting the potential for improved predictive accuracy. In this study, we explore the application of Federated Learning (FL) in Sustainable Supply Chain Management (SSCM), with a focus on the grocery retail sector dealing with perishable goods. We develop a baseline predictive model for demand forecasting and waste assessment in an isolated retailer scenario. Subsequently, we introduce a Blockchain-based FL model, trained collaboratively across multiple retailers without direct data sharing. Our preliminary results show that FL models have performance almost equivalent to the ideal setting in which parties share data with each other, and are notably superior to models built by individual parties without sharing data, cutting waste and boosting efficiency.

Figures (4)

• Figure 1: Different categorizations of FL: A) Horizontal FL, B) Vertical FL, C) Federated Transfer Learning
• Figure 2: Weekly sales of multiple stores in the Walmart data
• Figure 3: Overview of the final architecture representing the correlation between the various entities and the lifecycle of the Federated Learning system, from local training to validation with Blockchain.
• Figure 4: Comparison of MSE for different learning approaches.