Newest papers

PYTHIA, a transferable machine learning framework for simultaneous defect and adsorption site optimization

Ioannis Kouroudis, Zhaosu Gu, Poonam, Leon Muller, Alessio Gagliardi · 2026 · arXiv (Cornell University)

DOI: 10.48550/arxiv.2610.31539 · cited by 0 works (OpenAlex) · open-access PDF

Spanish

El aprendizaje de marcos de máquina ha permeado la optimización del material, proporcionando un alto valor. Sin embargo, su dependencia de un gran número de muestras limita su aplicabilidad a familias de materiales relativamente exploradas. En este artículo ampliamos la aplicabilidad del aprendizaje de máquina a materiales nuevos y subestudados a través del marco PYTHIA (Probabilistic Yet Transferable Hierarchical Interaction Architecture). Pythia utiliza dos de los más potentes conceptos de aprendizaje de máquina, modelos fundamentales y aprendizaje de transferencia. Empleamos un modelo de base complejo entrenado en un gran conjunto de datos de moléculas químicamente similares, sobre un objetivo que está físicamente cerca de nuestra predicción. Luego, congelamos los pesos y realizamos el alineamiento subespacial con el uso de una capa de transformador y añadimos la cu

Machine translation (Meta M2M100); the English original follows.

Abstract

Machine learning has permeated material optimization, providing high value. Nevertheless, its re- liance on a high number of samples limits its applicability to relatively explored material families. In this paper we extend the applicability of machine learning to novel and understudied materials through the PYTHIA (Probabilistic Yet Transferable Hierarchical Interaction Architecture) framework. Pythia utilizes two of the most powerful machine learning concepts, foundational models and transfer learning. We employ a complex foundational model trained on a large dataset of chemically similar molecules, on a target that is physically close to our prediction target. Then, we freeze the weights and perform subspace alignment with the use of a transformer layer and add uncertainty quantification with gaussian processes. These two layers can be trained on a limited dataset yet allow us to leverage the gained knowl- edge and the studied materials and complex architectures. The resulting model is used as a surrogate for a Bayesian Optimization scheme. To showcase the robustness of the framework, we applied it to the investigation of adsorption and defect optimization of two distinct systems. Not only were we able to correctly and efficiently identify the optimal adsorption sites of even complex a

Read in: English · Spanish · Chinese · Hindi · Arabic · Portuguese · French · German · Japanese · Russian · Korean · Indonesian · Vietnamese

Topics: Bayesian optimization, Leverage (statistics), Machine learning, Artificial intelligence, Computer science, Robustness (evolution)

Sources: OpenAlex (CC0), arXiv, Crossref. Agent Exchange