Microsoft Research Podcast Abstracts: Heat Transfer and Deep Learning with Hongxia Hao and Bing Lv
May 8, 2025
Join Hongxia Hao, a senior researcher at Microsoft Research AI for Science, and Bing Lv, an associate professor of physics, as they delve into the frontier of heat transfer in inorganic crystals. They discuss how AI is revolutionizing material science by enhancing thermal conductivity beyond silicon. With insights from their groundbreaking research, they reveal a massive database of materials and how AI serves as a collaborative partner in discovering innovative thermal solutions. The future of material design looks promising, driven by generative AI and advanced methodologies.
AI Snips
Chapters
Transcript
Episode notes
Fundamental Limit of Heat Transfer
- The speed limit of heat transfer in solids is a fundamental physics question with practical impact on technology design.
- Thermal bottlenecks in dense, fast computing chips limit performance more than transistor density now.
AI Advances Materials Discovery
- Traditional materials discovery using quantum mechanics is limited by computational cost and scale.
- AI, especially deep learning, enables large-scale and efficient prediction of material properties to overcome this.
Methodology for Material Search
- Combine AI with quantum mechanics and computational brute force for efficient and accurate material property predictions.
- Use AI to screen vast chemical and structural spaces for stable, diverse crystal candidates.