AINeutralarXiv – CS AI · 18h ago6/10
🧠
Agentic multi-fidelity learning of quasiparticle and excitonic properties
Researchers introduce an agent-guided multi-fidelity machine learning framework that corrects numerical instabilities in GW-Bethe-Salpeter calculations for simulating electronic and optical properties of strained MoS2-WS2 bilayers. The approach uses confidence-weighted structural agents and Gaussian process corrections to improve accuracy of quasiparticle gaps and exciton binding energies while preserving physical strain dependence.