Machine Learning for Detecting Steering in Qutrit-Pair States

• DOI: 10.48550/arxiv.2502.11365
• Publication Date: 2025-02-16
• Authors: Pu Wang; Zhongyan Li; Hui-Xian Meng
• Keywords: Qutrit

Only a few states in high-dimensional systems can be identified as (un)steerable using existing theoretical or experimental methods. We utilize semidefinite programming (SDP) to construct dataset for steerability detection qutrit-qutrit systems. For the full-information feature $F_1$, artificial neural networks achieve high classification accuracy and generalization, preform better than support vector machine. As engineering playing pivotal role, we introduce steering ellipsoid-like $F_2$, which significantly enhances performance of each our models. Given SDP method provides only sufficient condition detection, establish first rigorously constructed, accurately labeled based on foundations. This enables models exhibit outstanding generalization capabilities, independent choice features. applications, investigate boundaries isotropic partially entangled states, find new steerable states. work not advances application machine learning probing quantum but also deepens understanding itself.