相关参考文献
注意:仅列出部分参考文献,下载原文获取全部文献信息。
Article
Chemistry, Multidisciplinary
Seokhyun Moon et al.
Summary: Recently, deep neural network (DNN)-based drug-target interaction (DTI) models have gained attention for their high accuracy and affordable computational costs. However, these models still face challenges in generalization for in silico drug discovery. To enhance generalization in the DTI model, two key strategies are proposed: predicting atom-atom pairwise interactions using physics-informed equations parameterized with neural networks, and including a broader range of binding poses and ligands in the training data. The proposed model, PIGNet, outperforms previous methods in docking and screening powers, and also provides insights for further ligand optimization through visualization of predicted affinities.
Article
Chemistry, Physical
Zijie Li et al.
Summary: Molecular Dynamics (MD) simulation is an important tool for studying the dynamics and structure of matter. Researchers have developed a model called GAMD that uses deep learning to predict the forces in a system, bypassing the evaluation of potential energy. Their experimental results show that GAMD performs competitively in large-scale simulations.
JOURNAL OF CHEMICAL PHYSICS
(2022)
Article
Engineering, Multidisciplinary
Sina Amini Niaki et al.
Summary: The study introduces a Physics-Informed Neural Network (PINN) to simulate the thermal-chemical evolution of a composite material curing in an autoclave. By optimizing deep neural network (DNN) parameters using a physics-based loss function, the research solves coupled differential equations, designs a PINN with two disconnected subnetworks, and develops a sequential training algorithm. The approach incorporates explicit discontinuities at the composite-tool interface and enforces known physical behavior in the loss function to enhance solution accuracy.
COMPUTER METHODS IN APPLIED MECHANICS AND ENGINEERING
(2021)
Article
Multidisciplinary Sciences
Qi Wang et al.
Summary: The inverse design of glassy materials for specific properties is challenging due to their disordered, non-prototypical structure. Wang and Zhang propose a data-driven property-oriented approach for the design of glassy materials with desired functionalities.
NATURE COMMUNICATIONS
(2021)
Article
Multidisciplinary Sciences
Seungwoong Ha et al.
Summary: The study introduces AgentNet, a model-free data-driven framework utilizing deep neural networks to reveal and analyze the hidden interactions in complex systems. AgentNet successfully captured a wide variety of simulated complex systems and demonstrated potential applications with real bird flock data.
SCIENTIFIC REPORTS
(2021)
Article
Chemistry, Physical
Cheol Woo Park et al.
Summary: GNNFF framework accurately predicts atomic forces using machine learning, achieving high performance and computational speed across various material systems, and accurately predicting the forces of large MD systems after training on smaller systems. The Li diffusion coefficient obtained through MD simulation using this framework shows good accuracy compared to AIMD.
NPJ COMPUTATIONAL MATERIALS
(2021)
Review
Computer Science, Artificial Intelligence
Kenneth Atz et al.
Summary: Geometric deep learning, based on neural network architectures that process symmetry information, shows promise in molecular modeling applications. This review provides a detailed overview of molecular GDL, its applications, challenges, and future opportunities, highlighting the importance of geometric representations in molecular deep learning. Kenneth Atz and colleagues review the current progress and challenges of geometric deep learning in molecular sciences, emphasizing the significance of spatial structure information in molecules.
NATURE MACHINE INTELLIGENCE
(2021)
Article
Computer Science, Artificial Intelligence
Oscar Mendez-Lucio et al.
Summary: DeepDock, developed by Mendez-Lucio and colleagues, utilizes geometric deep learning to predict binding conformations of ligands to protein targets, providing guidance for molecular optimization. The method learns a statistical potential based on distance likelihood, tailored for each ligand-target pair, and can reproduce experimental binding conformations effectively. This approach showcases the use of artificial intelligence to enhance structure-based drug design.
NATURE MACHINE INTELLIGENCE
(2021)
Review
Physics, Applied
George Em Karniadakis et al.
Summary: Physics-informed learning seamlessly integrates data and mathematical models through neural networks or kernel-based regression networks for accurate inference of realistic and high-dimensional multiphysics problems. Challenges remain in incorporating noisy data seamlessly, complex mesh generation, and addressing high-dimensional problems.
NATURE REVIEWS PHYSICS
(2021)
Article
Computer Science, Artificial Intelligence
Shigeyuki Matsumoto et al.
Summary: Cryo-electron microscopy (cryo-EM) is a powerful tool for determining protein structures at atomic-level resolution. The DEFMap deep learning approach can extract dynamics hidden in cryo-EM density maps, providing valuable insights into protein function and structural changes. This multidisciplinary approach combines experimental data, molecular dynamics simulations, and deep learning techniques to study protein dynamics.
NATURE MACHINE INTELLIGENCE
(2021)
Article
Physics, Multidisciplinary
V Bapst et al.
Article
Chemistry, Physical
Oliver T. Unke et al.
JOURNAL OF CHEMICAL THEORY AND COMPUTATION
(2019)
Article
Chemistry, Multidisciplinary
Jiang Wang et al.
ACS CENTRAL SCIENCE
(2019)
Review
Physics, Multidisciplinary
Giuseppe Carleo et al.
REVIEWS OF MODERN PHYSICS
(2019)
Article
Chemistry, Physical
K. T. Schuett et al.
JOURNAL OF CHEMICAL PHYSICS
(2018)
Review
Multidisciplinary Sciences
Alexander Radovic et al.
Review
Materials Science, Multidisciplinary
W. Gregory Sawyer et al.
ANNUAL REVIEW OF MATERIALS RESEARCH, VOL 44
(2014)
Review
Chemistry, Physical
Edward P. Furlani