Machine learning models predict optimal thermoelectric materials using thermal conductivity ratio

Researchers have developed a data-driven framework using machine learning to accelerate the discovery of thermoelectric materials. The framework identifies materials with a specific lattice-to-total thermal conductivity ratio of approximately 0.5, which is linked to the phonon-glass electron-crystal design concept. This approach screens thousands of compounds to find those with ultralow thermal conductivity and guides optimization strategies for improved thermoelectric performance. AI

Summary written by gemini-2.5-flash-lite from 1 source. How we write summaries →

IMPACT Provides a new ML-driven screening and optimization framework for materials science, potentially accelerating discovery in energy harvesting technologies.

RANK_REASON Academic paper detailing a new data-driven framework for materials discovery using machine learning.

Read on arXiv cs.LG →

• Yifan Sun
• arXiv
• machine learning

• paper
• infra