The rapid advancements in artificial intelligence (AI) are catalyzing transformative changes in atomic modeling, simulation, and design. AI-driven potential energy models have demonstrated the capability to conduct large-scale, long-duration simulations with the accuracy of ab initio electronic structure methods. However, the model generation process remains a bottleneck for large-scale applications. We propose a shift towards a model-centric ecosystem, wherein a large atomic model (LAM), pre-trained across multiple disciplines, can be efficiently fine-tuned and distilled for various downstream tasks, thereby establishing a new framework for molecular modeling. In this study, we introduce the DPA-2 architecture as a prototype for LAMs. Pre-trained on a diverse array of chemical and materials systems using a multi-task approach, DPA-2 demonstrates superior generalization capabilities across multiple downstream tasks compared to the traditional single-task pre-training and fine-tuning methodologies. Our approach sets the stage for the development and broad application of LAMs in molecular and materials simulation research.

Cadmium selenide (CdSe) is an inorganic semiconductor with unique optical and electronic properties that make it useful in various applications, including solar cells, light-emitting diodes, and biofluorescent tagging. In order to synthesize high-quality crystals and subsequently integrate them into devices, it is crucial to understand the atomic scale crystallization mechanism of CdSe. Unfortunately, such studies are still absent in the literature. To overcome this limitation, we employed an enhanced sampling-accelerated active learning approach to construct a deep neural potential with ab initio accuracy for studying the crystallization of CdSe. Our brute-force molecular dynamics simulations revealed that a spherical-like nucleus formed spontaneously and stochastically, resulting in a stacking disordered structure where the competition between hexagonal wurtzite and cubic zinc blende polymorphs is temperature-dependent. We found that pure hexagonal crystal can only be obtained approximately above 1430 K, which is 35 K below its melting temperature. Furthermore, we observed that the solidification dynamics of Cd and Se atoms were distinct due to their different diffusion coefficients. The solidification process was initiated by lower mobile Se atoms forming tetrahedral frameworks, followed by Cd atoms occupying these tetrahedral centers and settling down until the third-shell neighbor of Se atoms sited on their lattice positions. Therefore, the medium-range ordering of Se atoms governs the crystallization process of CdSe. Our findings indicate that understanding the complex dynamical process is the key to comprehending the crystallization mechanism of compounds like CdSe, and can shed lights in the synthesis of high-quality crystals.(c) 2023 Published by Elsevier Ltd on behalf of The editorial office of Journal of Materials Science & Technology.

The rapid development of artificial intelligence (AI) is driving significant changes in the field of atomic modeling, simulation, and design. AI-based potential energy models have been successfully used to perform large-scale and long-time simulations with the accuracy of ab initio electronic structure methods. However, the model generation process still hinders applications at scale. We envision that the next stage would be a model-centric ecosystem, in which a large atomic model (LAM), pre-trained with as many atomic datasets as possible and can be efficiently fine-tuned and distilled to downstream tasks, would serve the new infrastructure of the field of molecular modeling. We show that DPA-2 can accurately represent a diverse range of chemical systems and materials, enabling high-quality simulations and predictions with significantly reduced efforts compared to traditional methods. Our approach paves the way for a universal large atomic model that can be widely applied in molecular and material simulation research, opening new opportunities for scientific discoveries and industrial applications.

Data: The complete collection of datasets employed in this research is encapsulated within the archive file data-v1.3.tgz. This encompasses both the upstream datasets for pre-training and downstream datasets for fine-tuning, all in DeePMD format. We recommend creating a new directory and employing the command 'tar -xzvf data-v1.3.tgz' to extract the data files. Inside each dataset contained in subdirectories (e.g., Domains, Metals, H2O, and Others), one will find: A README file A 'train' directory (included if utilized in upstream pre-training) train.json -- A list of file paths for training systems test.json -- A list of file paths for testing systems A 'downstream' directory (included if utilized in downstream fine-tuning) train.json -- A list of file paths for training systems test.json -- A list of file paths for testing systems *Main data files comprising various structures *Additional processing scripts The root directory contains train.json and downstream.json files that amalgamate the respective upstream and downstream splits mentioned above. The datasets used in this study are described in Section S1 of the Supplementary Materials and are readily accessible on AIS Square, which provides extensive details. Code: The 'code' directory, extractable from the archive Code_model_script.tgz, includes the DeePMD-kit's source code, which is based on PyTorch (2.0) Version. Installation and usage instructions can be found within the README file located in deepmd-pytorch-devel.zip. Model: Within the 'model' directory, also found in the extracted Code_model_script.tgz, resides the multi-task pre-trained DPA-2 model utilized in this research. Accompanying the model is its configuration file, input.json, which details the simultaneous pre-training of this model across 18 upstream datasets with shared descriptor parameters for 1 million steps. Scripts: The 'scripts' directory, part of the uncompressed Code_model_script.tgz, comprises all the scripts used for training, fine-tuning (learning curve analysis), and distillation in this work: 1. Upstream_single_task_training: Contains individual training scripts for DPA-2, Gemnet-OC, Equiformer-V2, Nequip, and Allegro, corresponding to the 18 upstream datasets. 2. Downstream_lcurve_workflow: Includes code and input files to evaluate the learning curves, including tests for DPA-2 fine-tuning transferability across 15 downstream datasets, as depicted in Figure 3 of the manuscript. 3. Distillation_workflow: Provides input files for distilling the fine-tuned DPA-2 models in datasets such as H2O-PBE0TS-MD, SSE-PBE-D, and FerroEle-D, as illustrated in Figure 4 of the manuscript. It is important to note that the scripts in 'Upstream_single_task_training' require the installation of deepmd-pytorch and other related models from their respective repositories (Gemnet-OC and Equiformer-V2: here [commit hash: 9bc9373], Nequip: here [commit hash: dceaf49, tag: v0.5.6], Allegro: here [commit hash: 22f673c]). The scripts in 'Downstream_lcurve_workflow' and 'Distillation_workflow' leverage Dflow—a Python framework for constructing scientific computing workflow—and dpgen2, the 2nd generation of the Deep Potential GENerator, both of which are repositories in the Deep Modeling Community.

Cadmium selenide (CdSe) is an inorganic semiconductor with unique optical and electronic properties that made it useful in various applications, including solar cells, light-emitting diodes, and biofluorescent tagging. In order to synthesize high-quality crystals and subsequently integrate them into devices, it is crucial to understand the atomic scale crystallization mechanism of CdSe. Unfortunately, such studies are still absent in the literature.To overcome this limitation, we employed an enhanced sampling-accelerated active learning approach to construct a deep neural potential with ab initio accuracy for studying the crystallization of CdSe.Our brute-force molecular dynamics simulations revealed that a spherical-like nucleus formed spontaneously and stochastically, resulting in a stacking disordered structure where the competition between hexagonal wurtzite and cubic zinc blende polymorphs is temperature-dependent. We found that pure hexagonal crystal can only be obtained approximately above 1430 K, which is 35 K below its melting temperature. We observed that the solidification dynamics of Cd and Se atoms were distinct due to their different diffusion coefficients. The solidification process was initiated by lower mobile Se atoms forming tetrahedral frameworks, followed by Cd atoms occupying these tetrahedral centers and settling down until the third-shell neighbor of Se atoms sited on their lattice positions. Therefore, the medium-range ordering of Se atoms governs the crystallization process of CdSe. Our findings indicate that understanding the complex dynamical process is the key to comprehending the crystallization mechanism of compounds like CdSe, and can shed lights in the synthesis of high-quality crystals.