相关实验视频
Updated: May 12, 2025

06:15
Interactive Molecular Model Assembly with 3D Printing
Published on: August 13, 2020
9.8K
化学语言模型链接器:将文本和分子与模块化适配器混合在一起
Yifan Deng1,2, Spencer S Ericksen3, Anthony Gitter4,1,2
1Department of Computer Sciences, University of Wisconsin-Madison.
ArXiv
|May 5, 2025
概括
我们介绍ChemLML,一种轻量级的方法,用于从文本中生成分子. 这种方法利用了预训练的模型,降低了计算成本,与从头开始训练相比,提高了分子生成性能.
科学领域:
- 计算化学是一种计算化学.
- 人工智能的人工智能是人工智能.
- 药物发现 药物发现
背景情况:
- 大型语言模型 (LLM) 和多模式模型提供了从文本描述中生成新分子的潜力.
- 当前的多模式模型通常需要从头开始进行培训,消耗大量计算资源并限制可扩展性.
- 现有的方法缺乏有效地整合预训练过的文本和分子模型.
研究的目的:
- 提出一种基于轻量级适配器的新策略 (ChemLML) 来从文本描述中生成条件分子.
- 允许使用各种预训练的文本模型进行分子生成,而无需进行广泛的再训练.
- 调查分子表示 (SMILES与SELFIES) 对生成性能的影响.
主要方法:
- 开发了ChemLML,一种基于适配器的轻量级策略,将预训练过的文本和分子模型结合在一起.
- 在分子域的专门嵌入空间内运行,以保持域特异性.
- 训练最小的适配器参数来定制各种预训练的文本模型用于分子生成.
- 评估了SMILES和SELFIES分子表示对条件生成的影响.
主要成果:
- 通过利用预训练模型,ChemLML有效地从文本描述中生成分子.
- 与从头开始的培训相比,基于适配器的培训需要的参数要少得多.
- 分子表示的选择 (SMILES优先于SELFIES) 影响生成性能.
- 确定了PubChem数据集的评估问题,并提供了一个过的测试集.
结论:
- ChemLML提供了一种高效且可扩展的方法,用于从文本中生成条件分子.
- 基于适配器的策略显著降低了计算开销,并提高了模型的灵活性.
- 分子表示选择对于优化生成模型性能至关重要.
- 通过产生候选蛋白质抑制剂和膜透分子来证明实际应用.
相关概念视频
Molecular Models
37.5K
Physical models representing molecular architectures of chemical compounds play essential roles in understanding chemistry. The use of molecular models makes it easier to visualize the structures and shapes of atoms and molecules.
37.5K
Ligand Binding and Linkage
4.7K
Allosteric proteins have more than one ligand binding site; the binding of a ligand to any of these sites influences the binding of ligands to the other sites. When a protein is allosteric, its binding sites are called coupled or linked. In the case of enzymes, the site that binds to the substrate is known as the active site and the other site is known as the regulatory site. When a ligand binds to the regulatory site, this leads to conformational changes in the protein that can influence...
4.7K
Olefin Metathesis Polymerization: Acyclic Diene Metathesis (ADMET)
1.9K
Acyclic diene metathesis polymerization or ADMET polymerization involves cross-metathesis of terminal dienes, such as 1,8-nonadiene, to give linear unsaturated polymer and ethylene. As ADMET is a reversible process, the formed ethylene gas must be removed from the reaction mixture to complete the polymerization process.
Similar to cross-metathesis, ADMET also involves the formation of metallacyclobutane intermediate by [2+2] cycloaddition of one of the double bonds of a terminal diene with...
Similar to cross-metathesis, ADMET also involves the formation of metallacyclobutane intermediate by [2+2] cycloaddition of one of the double bonds of a terminal diene with...
1.9K
Polymer Classification: Architecture
2.6K
Polymers are classified as linear or branched on the basis of their chain architecture. The polymer chains in linear polymers have a long chain-like structure with minimal to no branching at all. Even if a polymer features large substituent groups on the monomer, which appear as branches to the skeleton, it is not considered a branched polymer. A branched polymer contains secondary polymer chains that arise from the main polymer chain. The branching occurs when the polymer growth shifts from...
2.6K
Cycloaddition Reactions: MO Requirements for Photochemical Activation
2.0K
Some cycloaddition reactions are activated by heat, while others are initiated by light. For example, a [2 + 2] cycloaddition between two ethylene molecules occurs only in the presence of light. It is photochemically allowed but thermally forbidden.
2.0K
Diels–Alder Reaction Forming Bridged Bicyclic Products: Stereochemistry
4.6K
Diels–Alder reactions between cyclic dienes locked in an s-cis configuration and dienophiles yield bridged bicyclic products.
4.6K

