Non-Rival Data as Rival Products: An Encapsulation-Forging Approach for Data Synthesis

TL;DR

This paper tackles the data-sharing dilemma by transforming non-rival data into rival data with asymmetric utility using Encapsulation-Forging (EnFo). EnFo first encapsulates predictive knowledge into a designated key model and then forges a small synthetic dataset that overfits this model, yielding data that delivers value primarily to the intended algorithm ($| ilde{\mathcal{D}}|=M$, $M\ll N$). The approach achieves remarkable sample efficiency, robust privacy protection (low Monte Carlo attack risk), and resistance to unauthorized augmentation, while enabling practical extensions for data augmentation and cross-model adaptability. Collectively, EnFo offers a principled, technically enforceable mechanism for controlled data sharing in co-opetition contexts and has broad implications for privacy, efficiency, and strategic data collaboration in the digital economy.

Abstract

The non-rival nature of data creates a dilemma for firms: sharing data unlocks value but risks eroding competitive advantage. Existing data synthesis methods often exacerbate this problem by creating data with symmetric utility, allowing any party to extract its value. This paper introduces the Encapsulation-Forging (EnFo) framework, a novel approach to generate rival synthetic data with asymmetric utility. EnFo operates in two stages: it first encapsulates predictive knowledge from the original data into a designated ``key'' model, and then forges a synthetic dataset by optimizing the data to intentionally overfit this key model. This process transforms non-rival data into a rival product, ensuring its value is accessible only to the intended model, thereby preventing unauthorized use and preserving the data owner's competitive edge. Our framework demonstrates remarkable sample efficiency, matching the original data's performance with a fraction of its size, while providing robust privacy protection and resistance to misuse. EnFo offers a practical solution for firms to collaborate strategically without compromising their core analytical advantage.

Figures (10)

• Figure 1: EnFo for data synthesis
• Figure 2: Statistical Divergence Results on Monte Carlo Datasets (Higher Is Better)
• Figure 3: Accuracy-Privacy Results on Monte Carlo Datasets
• Figure 4: Accuracy-Volume Results on Monte Carlo Datasets
• Figure 5: Data Velocity Scalability Results on Monte Carlo Datasets