X2-GNN:一个物理信息传递神经网络与自然泛化能力大和复杂的分子的神经网络
Zhanfeng Wang1, Wenhao Zhang1, Minghong Jiang1
1Collaborative Innovation Center of Chemistry for Energy Materials, Shanghai Key Laboratory of Molecular Catalysis and Innovative Materials, MOE Key Laboratory of Computational Physical Sciences, Department of Chemistry, Fudan University, Shanghai 200438, China.
The journal of physical chemistry letters
|December 13, 2024
概括
一个新的图形神经网络,X2-GNN,通过整合物理见解来改善分子性质预测. 这种模型有效地对更大的分子进行了概括,显示了计算化学和材料科学的前景.
科学领域:
- 计算化学是一种计算化学.
- 机器学习是机器学习.
- 材料科学是一种材料科学.
背景情况:
- 神经网络擅长分子性质预测,但难以将其推广到更大的分子.
- 增加的分子大小导致更大的结构多样性和复杂的相互作用,挑战现有的模型.
研究的目的:
- 开发一个名为X2-GNN的E(3) 恒定图形神经网络 (GNN),用于增强分子性质预测.
- 通过结合物理见解来提高神经网络从小到大分子的概括能力.
主要方法:
- 引入了X2-GNN,一个传递GNN的E(3) 不变消息.
- 综合原子轨道重叠积分和核心哈密尔顿数,提供物理洞察力.
- 采用注意力机制来提高学习效率.
主要成果:
- 当在较小的数据集 (QM9) 上训练时,X2-GNN证明了对更大的分子 (数十个重原子) 的有效泛化.
- 在分子性质预测中实现了可信的每原子误差率.
- 在潜在能量表面建模和在次秒内预测债券解离能量的高精度.
结论:
- X2-GNN 展示了分子性质预测的可扩展性和广泛适用性.
- 将数据驱动方法与电子结构理论知识相结合,对于推进计算化学至关重要.
相关概念视频
Synaptic Signaling
74.8K
Neurons communicate at synapses, or junctions, to excite or inhibit the activity of other neurons or target cells, such as muscles. Synapses may be chemical or electrical.
74.8K
Neuronal Communication
785
Neurons, the fundamental units of the brain and nervous system, communicate through complex electrochemical signals that underpin all cognitive and bodily functions. This communication is primarily facilitated by a process involving the generation and propagation of an action potential along the axon of the neuron. When the internal electrical charge of a neuron surpasses a certain threshold, an action potential is triggered. This rapid change in voltage travels swiftly along the axon to the...
785
Types of Signaling Molecules
10.1K
In multicellular organisms, many molecules transmit signals between cells to pass information. These signals vary in complexity and include small peptides, nucleotides, steroids, fatty acid derivatives, and dissolved gases such as nitric oxide. Some signaling molecules diffuse through the plasma membrane to act locally between neighboring cells or travel long distances. Others remain attached to the cell surface, transmitting information to other cells only when they make contact. In some...
10.1K
Multi-pass Transmembrane Proteins and β-barrels
5.2K
In multi-pass transmembrane proteins, the polypeptide chain crosses the membrane more than once. The transmembrane polypeptide chain either forms an α-helix or β-strand structure. α-Helix containing multi-pass transmembrane proteins are ubiquitous, whereas β-strand containing ones are mainly found in gram-negative bacteria, mitochondria, and chloroplasts.
α-Helix containing multi-pass transmembrane proteins
Multi-pass transmembrane proteins such as...
α-Helix containing multi-pass transmembrane proteins
Multi-pass transmembrane proteins such as...
5.2K
The Synapse
123.3K
Neurons communicate with one another by passing on their electrical signals to other neurons. A synapse is the location where two neurons meet to exchange signals. At the synapse, the neuron that sends the signal is called the presynaptic cell, while the neuron that receives the message is called the postsynaptic cell. Note that most neurons can be both presynaptic and postsynaptic, as they both transmit and receive information.
123.3K
Two-Dimensional (2D) NMR: Overview
619
The 1D NMR spectrum of large and complex molecules like natural products has complicated splitting patterns and overlapping signals, which can be easily interpreted using 2-dimensional (2D) NMR. Unlike 1D NMR, 2D NMR has two frequency axes that provide the coupling information between the nucleus A and nucleus B in a molecule. The process from which 2D spectra are obtained has four steps.
The first step is the preparation period, during which nucleus A is excited with a radiofrequency pulse....
The first step is the preparation period, during which nucleus A is excited with a radiofrequency pulse....
619


