Performance-Driven QUBO for Recommender Systems on Quantum Annealers

2024-10-20

arXiv.org

AI Summary

To address the feature selection challenge in recommender systems, this paper proposes CAQUBO—a novel method that tightly integrates counterfactual analysis with Quadratic Unconstrained Binary Optimization (QUBO) modeling. CAQUBO quantifies the causal effects of individual and combined features on recommendation performance, constructing an interpretable QUBO coefficient matrix that guides D-Wave quantum annealers to efficiently identify the optimal feature subset. By establishing an explicit causal mapping from features to recommendation accuracy, CAQUBO overcomes the interpretability and performance limitations of conventional black-box, quantum-inspired heuristics. Extensive experiments across multiple benchmark datasets demonstrate that CAQUBO achieves an average 7.2% improvement in recommendation accuracy over existing quantum annealing–based feature selection methods, empirically validating the efficacy of the “quantum–causal synergy” paradigm.

Technology Category

• Machine Learning: Quantum Machine Learning
• Reasoning under Uncertainty: Causality
• Search and Optimization: Metareasoning and Metaheuristics

Application Category

• User Modeling, Personalization and Recommendation: Attacks and countermeasures in recommendation systems
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Graph Algorithms and Modeling for the Web: Algorithms and analysis for incomplete, noisy, or partially observed Web-related graphs

Abstract

We propose Counterfactual Analysis Quadratic Unconstrained Binary Optimization (CAQUBO) to solve QUBO problems for feature selection in recommender systems. CAQUBO leverages counterfactual analysis to measure the impact of individual features and feature combinations on model performance and employs the measurements to construct the coefficient matrix for a quantum annealer to select the optimal feature combinations for recommender systems, thereby improving their final recommendation performance. By establishing explicit connections between features and the recommendation performance, the proposed approach demonstrates superior performance compared to the state-of-the-art quantum annealing methods. Extensive experiments indicate that integrating quantum computing with counterfactual analysis holds great promise for addressing these challenges.

Problem

• Optimize feature selection in recommender systems
• Measure feature impact via counterfactual analysis
• Enhance quantum annealing performance for recommendations

Innovation

• CAQUBO uses counterfactual analysis for feature impact
• Quantum annealers optimize feature selection via QUBO
• Explicit feature-performance links enhance recommendation accuracy