在机器学习中使用文献衍生数据进行产量预测的挑战:关于Pd催化碳化反应的案例研究
Dong-Zhi Li1, Xue-Qing Gong1,2
1Centre for Computational Chemistry, School of Chemistry and Molecular Engineering, East China University of Science and Technology, 130 Meilong Road, Shanghai 200237, China.
The journal of physical chemistry. A
|November 20, 2024
概括
机器学习模型使用实验数据准确预测反应产量,但与文献数据作斗争. 在各种文献中训练有素的模型在准确性和概括性方面扎,这表明当前方法的局限性.
科学领域:
- 计算化学是一种计算化学.
- 化学信息学 化学信息学
- 机器学习在化学中的应用.
背景情况:
- 机器学习 (ML) 模型在训练高通量实验或计算数据时,在预测反应产量方面表现出很高的准确性.
- 然而,当使用从科学文献中提取的数据时,ML模型的性能显著下降,这表明预测能力存在差距.
研究的目的:
- 调查ML模型的性能和局限性,用于预测Palladium催化碳化反应的产量,使用文献数据.
- 确定在异质文献数据上训练的ML模型的精度和通用性降低的因素.
主要方法:
- 从科学文献中编制了2512个Pd催化碳化反应的数据集.
- 在这个数据集上训练和评估机器学习模型,重点关注不同数据子集和来源的性能.
- 使用统计指标,包括R平方 (R2),来评估模型的准确性和通用性.
主要成果:
- 表现最好的ML模型在文献衍生数据集上只获得了0.51的R2.
- 模型的有效性在很大程度上仅限于来自相同文献来源的狭窄数据子集内的预测.
- 当模型应用于更广泛,异质的数据集时,性能显著下降,突出了数据相似性和小样本大小的问题.
结论:
- 目前的ML技术在准确预测化学反应产量方面存在固有的局限性,来自各种文献数据.
- 依赖特定来源的数据相似性和小样本大小导致模型对波动敏感,影响稳定性,准确性和通用性.
- 需要更先进的方法来有效地处理科学文献中发现的化学数据的复杂性和异质性,以实现可靠的产量预测.
相关概念视频
01:26Limitations of Friedel–Crafts Reactions
5.2K
Several restrictions limit the use of Friedel–Crafts reactions. First, the halogen in the alkyl halide must be attached to an sp3-hybridized carbon for the Friedel–Crafts reactions to occur. Vinyl or aryl halides do not react since the carbocations formed are unstable under the reaction conditions. Second, Friedel–Crafts alkylation is susceptible to carbocation rearrangement, and the major products obtained have a rearranged carbon skeleton. In contrast, the acylium ion is...
5.2K
01:33Carboxylic Acids to Methylesters: Alkylation using Diazomethane
2.1K
Carboxylic acids react with diazomethane in an ether solvent via alkylation at the carboxylate oxygen atom to give methyl esters of the corresponding acid with excellent yields.
2.1K
02:22Reaction Yield
50.9K
The theoretical yield of a reaction is the amount of product estimated to form based on the stoichiometry of the balanced chemical equation. The theoretical yield assumes the complete conversion of the limiting reactant into the desired product. The amount of product that is obtained by performing the reaction is called the actual yield, and it may be less than or (very rarely) equal to the theoretical yield.
50.9K
02:17Reduction of Alkenes: Asymmetric Catalytic Hydrogenation
3.2K
Catalytic hydrogenation of alkenes is a transition-metal catalyzed reduction of the double bond using molecular hydrogen to give alkanes. The mode of hydrogen addition follows syn stereochemistry.
The metal catalyst used can be either heterogeneous or homogeneous. When hydrogenation of an alkene generates a chiral center, a pair of enantiomeric products is expected to form. However, an enantiomeric excess of one of the products can be facilitated using an enantioselective reaction or an...
The metal catalyst used can be either heterogeneous or homogeneous. When hydrogenation of an alkene generates a chiral center, a pair of enantiomeric products is expected to form. However, an enantiomeric excess of one of the products can be facilitated using an enantioselective reaction or an...
3.2K
01:18β-Dicarbonyl Compounds via Crossed Claisen Condensations
3.0K
Crossed Claisen condensations are base-promoted reactions between two different ester molecules producing β-dicarbonyl compounds. The reaction involving esters, with both containing α hydrogen, results in a mixture of four different products that are difficult to isolate. This reduces the synthetic utility of the reaction.
3.0K
01:17Ziegler–Natta Chain-Growth Polymerization: Overview
3.2K
Ziegler–Natta polymerization is another form of addition or chain‐growth polymerization used for synthesizing linear polymers over branched polymers. The catalyst used for polymerization is the Ziegler–Natta catalyst, named after Karl Ziegler and Giulio Natta, who developed it in 1953. This catalyst is an organometallic complex of titanium tetrachloride and triethyl aluminum, with the active form of the catalyst being an alkyl titanium compound. Using the Ziegler–Natta...
3.2K