Reward misalignment as an opportunity for materials discovery
Ryotaro Okabe, Shoki Kishida, Abhijatmedhi Chotrattanapituk, Dominic Bollettieri, Takuya Isogawa, Yongqiang Cheng, Mingda Li, Masaki Adachi
Abstract
Reward misalignment is usually regarded as a failure of artificial intelligence. In materials discovery with generative AI models, however, disagreement between predefined rewards and expert judgment may reveal materials that the rewards overlook. We investigate this possibility using electronic flat bands as a prime example. Experts mostly agree with a numerical bandwidth score, but their disagreements are systematic and cannot be explained by an estimated 5% rate of mistaken labels. We learn the experts' preferences from pairwise comparisons and use them to guide a crystal diffusion model. The learned preferences and numerical score agree on the flattest bands but diverge elsewhere. For example, the preferences favor KCo2As2 and CoGa, reported compounds with partially flat bands that the score would discard. The framework also extends to synthesizability and phonon properties. Our findings suggest that reward misalignment can expose opportunities for materials discovery beyond what fixed rewards are designed to find.
Create a lesson
Related papers
Emergent Fatigue Dynamics from Foundational Interatomic Potentials
Jingye Tan, Benjamin A. Jasperson, Eric Rothchild et al.
Constitutive Kolmogorov-Arnold Networks in Finite Element Model Updating: Closed-Form Hyperelastic Strain Energies from Heterogeneous Experiments
Kian P. Abdolazizi, Kevin Linka, Christian J. Cyron
Two distinct notions of speed in antiferromagnets
Paul Herrgen, Stephan Wust, Christin Schmitt et al.
Perfumer: A Large Language Model-based Agentic Framework Accelerating Structural Relaxation and Electronic Band Prediction for First-Principle Calculation
Zhiwei Li, Yuhao Zhou, Yikai Yang et al.
Unusual increase in electrical and thermal conductivity of severely deformed CrMnFeCoNi high-entropy alloy: Roles of disordering and nanostructuring in electron and phonon transport
Jacqueline Hidalgo-Jiménez, Payam Edalati, Md Amirul Islam et al.
Lattice dynamics study of (Bi1-xSbx)2Te3 compounds: Comparative compositional study of force constants and Grüneisen parameters
Baihong Sun, Chenxin Wei, Elissaios Stavrou