TY - CONF A1 - George, Janine T1 - High-throughput and automated bonding Analysis N2 - Talk as a part of the LOBSTER CECAM SCHOOL. This talk introduced all participants to automation tools around the software LOBSTER. T2 - CECAM LOBSTER School CY - Aalto, Finland DA - 12.03.2024 KW - Automation KW - Workflows KW - Bonding Analysis KW - Materials Design KW - Chemically Complex Materials PY - 2024 AN - OPUS4-59672 LA - eng AD - Bundesanstalt fuer Materialforschung und -pruefung (BAM), Berlin, Germany ER - TY - CONF A1 - George, Janine T1 - High-throughput Approaches for Materials Understanding and Design N2 - Bonds and local atomic environments are crucial descriptors of material properties. They have been used to create design rules and heuristics and as features in machine learning of materials properties.[1] Implementations and algorithms (e.g., ChemEnv and LobsterEnv) for identifying local atomic environments based on geometrical characteristics and quantum-chemical bonding analysis are nowadays available.[2,3] Fully automatic workflows and analysis tools have been developed to use quantum-chemical bonding analysis on a large scale.[3,4] The lecture will demonstrate how our tools, that assess local atomic environments and perform automatic bonding analysis, help to develop new machine learning models and a new intuitive understanding of materials.[5,6] Furthermore, the general trend toward automation in computational materials science and some of our recent contributions will be discussed.[7–10] T2 - International Materials Science and Engineering Congress - MSE 2024 CY - Darmstadt, Germany DA - 24.09.2024 KW - Automation KW - High-throughput KW - Chemically Complex Materials KW - Thermal Properties KW - Phonons KW - Bonding Analysis PY - 2024 AN - OPUS4-61118 LA - eng AD - Bundesanstalt fuer Materialforschung und -pruefung (BAM), Berlin, Germany ER - TY - CONF A1 - George, Janine T1 - High-throughput approaches for materials understanding and design N2 - Bonds and local atomic environments are crucial descriptors of material properties. They have been used to create design rules and heuristics and as features in machine learning of materials properties.[1] Implementations and algorithms (e.g., ChemEnv and LobsterEnv) for identifying local atomic environments based on geometrical characteristics and quantum-chemical bonding analysis are nowadays available.[2,3] Fully automatic workflows and analysis tools have been developed to use quantum-chemical bonding analysis on a large scale.[3,4] The lecture will demonstrate how our tools, that assess local atomic environments and perform automatic bonding analysis, help to develop new machine learning models and a new intuitive understanding of materials.[5,6] Furthermore, the general trend toward automation in computational materials science and some of our recent contributions will be discussed.[7–11] T2 - TCO 2024 - Transparent Conductive Oxides, Fundamentals and Applications CY - Leipzig, Germany DA - 23.09.2024 KW - Automation KW - Materials discovery KW - Machine Learned Interatomic Potentials KW - Workflows KW - Chemically Complex Materials KW - Bonding Analysis PY - 2024 AN - OPUS4-61153 LA - eng AD - Bundesanstalt fuer Materialforschung und -pruefung (BAM), Berlin, Germany ER - TY - CONF A1 - George, Janine T1 - New opportunities through an efficient combination of machine learning and high-throughput computing N2 - The underlying data is crucial for machine learning (ML) tasks.[1] Ab initio data is often used as a target (occasionally as features[2]). High-throughput calculations and automation make it possible to generate such data as efficiently as possible and with uniform standards.[3,4] The high-throughput data of the Materials Project has recently been used to train new foundation interatomic potentials[5]. Several high-throughput frameworks have been developed in recent years. This presentation will introduce the atomate2/jobflow[3,4] framework that is used and developed by researchers around the Materials Project. In the future, it will be used to generate the data for the Materials Project. Atomate2 allows both the use of DFT and ML potentials within one framework. Building upon this framework, we have developed a package with the possibility to train and benchmark ML potentials automatically. Currently, the package can be used to train ML interatomic potentials sufficient to predict harmonic phononic properties of materials. Typical foundation models in this area now achieve acceptable results but are still far from routinely replacing DFT.[5,6] Building on the promising results from the ref where we explored data generation strategies for accurate phononic properties[7], we will present new fully automated workflows for training and benchmarking ML interatomic potentials with force predictions that are accurate enough to compute harmonic phonons in very good agreement with DFT for different crystal structures of the same composition. T2 - GAP/(M)ACE Developers & Users Meeting 2024 CY - Berlin, Germany DA - 17.09.2024 KW - Automation KW - Workflows KW - Machine Learned Interatomic Potentials KW - Thermal Properties KW - Bonding Analysis KW - Chemically Complex Materials PY - 2024 AN - OPUS4-61099 LA - eng AD - Bundesanstalt fuer Materialforschung und -pruefung (BAM), Berlin, Germany ER - TY - CONF A1 - George, Janine T1 - Data-Driven Approaches for Materials Understanding and Design N2 - Data-Driven Approaches for Materials Understanding and Design Analyzing the crystal structure (including coordination environments, bond distances, and angles) and chemical bonding are standardly used in chemistry to understand material properties. In this talk, I will show how we can use this information in machine learning of material properties. For this purpose, I will present our software tools ChemEnv[1] for analyzing coordination environments and for automation of orbital-based bonding analysis with LOBSTER[2] (LobsterPy[3] and density-functional theory workflows in atomate2[4]). Enabled by these software tools, we have built interpretable machine-learned models for magnetic and vibrational properties that allow us to test and further develop intuitive rules.[5,6] T2 - Institutskolloqium Institut für Anorganische Chemie und Kristallographie CY - Leipzig, Germany DA - 23.10.2024 KW - Automation KW - Bonding Analysis KW - Materials Design KW - Chemically Complex Materials KW - Phonons KW - Thermal Properties KW - Synthesizability KW - Machine Learning PY - 2024 AN - OPUS4-61445 LA - eng AD - Bundesanstalt fuer Materialforschung und -pruefung (BAM), Berlin, Germany ER - TY - CONF A1 - George, Janine T1 - High-throughput and data-driven approaches for materials understanding and design N2 - Bonds and local atomic environments are crucial descriptors of material properties. They have been used to create design rules and heuristics and as features in machine learning of materials properties.[1] Implementations and algorithms (e.g., ChemEnv and LobsterEnv) for identifying local atomic environments based on geometrical characteristics and quantum-chemical bonding analysis are nowadays available.[2,3] Fully automatic workflows and analysis tools have been developed to use quantum-chemical bonding analysis on a large scale.[3,4] The lecture will demonstrate how our tools, which assess local atomic environments and perform automatic bonding analysis, help to develop new machine-learning models and a new intuitive understanding of materials.[5,6] Beyond this, the general trend toward automation in computational materials science and some of our recent contributions will be discussed.[7–11] The focus will be especially on the interplay of DFT and machine-learned interatomic potentials will present new fully automated workflows for training and benchmarking machine-learned interatomic potentials with force predictions accurate enough to compute harmonic phonons in very good agreement with DFT. Semi-automated fine-tuning of existing ML models for phonon properties will also be discussed. T2 - XXIV Chilean Physics Symposium CY - Temucu, Chile DA - 20.11.2024 KW - Automation KW - Machine Learning KW - Chemically Complex Materials KW - Machine-Learned Interatomic Potentials KW - Finetuning PY - 2024 AN - OPUS4-61749 LA - eng AD - Bundesanstalt fuer Materialforschung und -pruefung (BAM), Berlin, Germany ER - TY - CONF A1 - George, Janine T1 - Data-Driven Materials Design N2 - Implementations and algorithms, such as ChemEnv and LobsterEnv, for identifying local atomic environments based on geometrical characteristics and quantum-chemical bonding analysis are now available. Fully automatic workflows and analysis tools have been developed to use quantumchemical bonding analysis on a large scale. The first part of the lecture will demonstrate how our tools, which assess local atomic environments and perform automatic bonding analysis, help develop new machine-learning models and a new intuitive understanding of material properties. Many new universal machine-learned interatomic potentials, such as MACE-MP-0, have been developed. The second part of the lecture will showcase how these potentials, in combination with DFT, could significantly accelerate our research. The focus will be on the interplay of DFT and machine-learned interatomic potentials, presenting new fully automated workflows for training, finetuning, and implemented benchmarking these potentials, in our software autoplex (https://github.com/autoatml/autoplex). Additionally, I will show how to train new interatomic potentials from scratch by exploring potential energy surfaces extensively, offering a method to enhance current universal machine-learned potentials. Beyond this, the general trend toward automation in computational materials science and some of our recent contributions will be discussed. T2 - AI4AM2025 CY - Donostia, Spain DA - 08.04.2025 KW - Automation KW - Machine Learning KW - Chemically Complex Materials KW - Thermal Conductivity KW - Bonding Analysis KW - Workflows KW - Machine Learned Interatomic Potentials PY - 2025 AN - OPUS4-62950 LA - eng AD - Bundesanstalt fuer Materialforschung und -pruefung (BAM), Berlin, Germany ER -