Paid with Models: Optimal Contract Design for Collaborative Machine Learning
This addresses the challenge of incentivizing fair collaboration in machine learning for participants, though it appears incremental as it builds on existing contract theory applied to a new context.
The paper tackles the problem of rent-seeking behaviors in collaborative machine learning by designing optimal contracts that reward participants with models of varying accuracy based on contributions, and it proposes a transformation to simplify the non-convex optimization into a solvable convex problem, with analysis and numerical experiments exploring benefits and welfare implications.
Collaborative machine learning (CML) provides a promising paradigm for democratizing advanced technologies by enabling cost-sharing among participants. However, the potential for rent-seeking behaviors among parties can undermine such collaborations. Contract theory presents a viable solution by rewarding participants with models of varying accuracy based on their contributions. However, unlike monetary compensation, using models as rewards introduces unique challenges, particularly due to the stochastic nature of these rewards when contribution costs are privately held information. This paper formalizes the optimal contracting problem within CML and proposes a transformation that simplifies the non-convex optimization problem into one that can be solved through convex optimization algorithms. We conduct a detailed analysis of the properties that an optimal contract must satisfy when models serve as the rewards, and we explore the potential benefits and welfare implications of these contract-driven CML schemes through numerical experiments.