Surrogate Modeling via Factorization Machine and Ising Model with Enhanced Higher-Order Interaction Learning

2025-07-02

AI Summary

This paper addresses the limited capability of existing methods in modeling high-order feature interactions for drug combination effect prediction. We propose a novel surrogate model that integrates Factorization Machines (FM) with the Ising model. Its core innovation lies in introducing slack variables to unify traditionally decoupled FM parameter learning and Ising energy minimization into a single, joint iterative optimization step—thereby explicitly capturing second- and higher-order feature interactions. Furthermore, we incorporate a quantum-annealing-inspired optimization strategy to enhance computational efficiency. Extensive experiments across multiple drug combination benchmarks demonstrate that our method significantly outperforms state-of-the-art baselines—including DeepSynergy and NIMFA—with AUC improvements of 3.2–5.8%. These results validate its superior capacity to model complex synergistic and antagonistic effects, as well as the potential benefits of quantum-inspired optimization.

Technology Category

• Machine Learning: Quantum Machine Learning
• Search and Optimization: Learning to Search
• Constraint Satisfaction and Optimization: Satisfiability Modulo Theories

Application Category

• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs
• User Modeling, Personalization and Recommendation: Explainable and interpretable methods for personalization
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking

Abstract

Recently, a surrogate model was proposed that employs a factorization machine to approximate the underlying input-output mapping of the original system, with quantum annealing used to optimize the resulting surrogate function. Inspired by this approach, we propose an enhanced surrogate model that incorporates additional slack variables into both the factorization machine and its associated Ising representation thereby unifying what was by design a two-step process into a single, integrated step. During the training phase, the slack variables are iteratively updated, enabling the model to account for higher-order feature interactions. We apply the proposed method to the task of predicting drug combination effects. Experimental results indicate that the introduction of slack variables leads to a notable improvement of performance. Our algorithm offers a promising approach for building efficient surrogate models that exploit potential quantum advantages.

Problem

• Enhances surrogate modeling with higher-order interaction learning
• Unifies two-step process into single integrated step
• Improves drug combination effect prediction performance

Innovation

• Enhanced surrogate model with slack variables
• Unified FM and Ising model in one step
• Improved higher-order interaction learning