CRUISE on Quantum Computing for Feature Selection in Recommender Systems

TL;DR

The paper tackles feature selection for recommender systems using quantum annealing by formulating the problem as a QUBO and enriching it with Counterfactual Analysis to emphasize features that causally improve ranking. It integrates MI/CMI with Counterfactual signals to guide feature selection (CAQUBO) and accommodates large feature sets through partitioning, evaluated on Item-KNN with nDCG as the performance metric. Results indicate meaningful gains over MI/CMI baselines, with performance sensitive to the balance parameter $\lambda$ and to the choice between quantum and classical annealing. The work demonstrates a practical pathway for quantum-assisted feature selection in recommendation contexts, while also highlighting opportunities to capture feature interactions more fully and to optimize partitioning schemes for larger-scale problems.

Abstract

Using Quantum Computers to solve problems in Recommender Systems that classical computers cannot address is a worthwhile research topic. In this paper, we use Quantum Annealers to address the feature selection problem in recommendation algorithms. This feature selection problem is a Quadratic Unconstrained Binary Optimization(QUBO) problem. By incorporating Counterfactual Analysis, we significantly improve the performance of the item-based KNN recommendation algorithm compared to using pure Mutual Information. Extensive experiments have demonstrated that the use of Counterfactual Analysis holds great promise for addressing such problems.