Reproduce “Extended Factorization Machine Annealing for Rapid Discovery of Transparent Conducting Materials” on OpenResearch
Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.
Universal app for Apple silicon and Intel. Requires macOS 11 or later.
About this paper
Daisuke Makino, Tatsuya Goto, Yoshinori Suga
The research from Jij Inc. and Toyota Motor Corporation introduces an Extended Factorization Machine Annealing (FMA) framework to accelerate the discovery of transparent conducting materials (TCMs), specifically within the (AlₓGaᵧIn𝓏)₂O₃ system. This framework achieves faster and more accurate material searches compared to conventional optimization methods, also demonstrating capabilities for multi-objective material design.
Key insights
- Binarizing continuous variables and incorporating a sophisticated QUBO design enables efficient application of annealing algorithms to complex, high-dimensional materials design problems.
- Balancing exploitation of promising regions (via Hopfield penalty) and robust exploration of the broader search space (via adaptive random flips) is critical for rapid discovery and avoiding local minima.
- An optimal discretization bit length exists, where an 8-bit representation provided the best balance between parameter precision and search space manageability for efficient convergence.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.