New framework auto-tunes SVMs on quantum annealers

This paper introduces a novel framework for optimizing Support Vector Machines (SVMs) that utilize Quadratic Unconstrained Binary Optimization (QUBO) models on quantum-inspired annealers. The framework employs Optuna for auto-tuning, addressing parameters related to numerical representation, kernel geometry, and constraint penalties. Tested across Fixstars Amplify Annealing Engine, Toshiba SQBM+, and Fujitsu Digital Annealer, the method demonstrated mean accuracy gains of 0.8 to 2.1 percentage points over traditional grid search on classification tasks with varying levels of label noise. AI

IMPACT This research could improve the efficiency and accuracy of quantum-inspired machine learning algorithms.

RANK_REASON The item is an academic paper detailing a new methodology for machine learning optimization. [lever_c_demoted from research: ic=1 ai=1.0]

Read on arXiv cs.LG →

• Fixstars Amplify Annealing Engine
• Fujitsu Digital Annealer
• Gaussian process
• Optuna
• QUBO
• support vector machine
• Takahiro Katagiri
• Toshiba SQBM+
• TPE

• paper
• infra

AI-generated summary · Google Gemini · from 1 sources. How we write summaries →