Practical Application and Evaluation of Atomic Swaps for Blockchain-based Recommender Systems

Abstract

Recommender Systems are very popular tools within the online community, suggesting to their users a big variety of items like products, videos, music and locations to visit. Moreover, users actively populate these systems sending and reading opinions under the form of reviews, and potentially obtaining a reward for their activities. However, such systems typically rely on a central authority that acts as a trusted party having total control over the system. Decentralized Recommender Systems have been proposed to solve such issue distributing the control and responsibility on the hands of their users, but leading to risks in case of disputes or misbehaviour. Based on a general architecture of a Decentralized Recommender System, in this paper we identify the potential unfair exchanges that may rise during the activity between two users, and we propose a solution based on the concept of atomic swaps inherited from the blockchain technology. Finally, we provide an attack model to show that the proposed solution creates fair processes.