Linear ML model outperforms Transformer for quantum chaotic dynamics

A new research paper explores the effectiveness of machine learning models in simulating complex quantum chaotic dynamics. The study compares a nonlinear Transformer model with a simple linear model called DLinear, using a PXP chain of Rydberg-atom arrays as a benchmark. While the Transformer performed well in ergodic regimes, DLinear demonstrated superior accuracy across various dynamical regimes, including quantum many-body scarring, suggesting that simple linear mappings can effectively forecast complex quantum evolution. AI

IMPACT Suggests that complex quantum dynamics can be forecast with simple linear models, potentially simplifying computational challenges in quantum physics.

RANK_REASON Research paper published on arXiv detailing machine learning models for quantum dynamics. [lever_c_demoted from research: ic=1 ai=1.0]

Read on arXiv cs.LG →

• arXiv
• DLinear
• Francesco Perciavalle
• PXP chain
• Rydberg-atom arrays
• Transformer

• paper
• other

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