相关实验视频

Updated: May 10, 2026

High-throughput Synthesis of Carbohydrates and Functionalization of Polyanhydride Nanoparticles

High-throughput Synthesis of Carbohydrates and Functionalization of Polyanhydride Nanoparticles

Published on: July 6, 2012

研究计算机辅助合成规划的效率

Peter B R Hartog^1,2, Annie M Westerlund¹, Igor V Tetko²

¹Molecular AI, Discovery Sciences, R&D, AstraZeneca, Pepparedsleden 1, 431 83 Mölndal, Sweden.

Journal of chemical information and modeling

|January 31, 2025

概括

为化学合成规划优化机器学习模型是关键. 虽然更快的单步预测很重要,但这些预测的多样性和信心更显著地影响了整体的多步搜索效率.

科学领域:

计算化学是一种计算化学.
人工智能在药物发现中的作用
机器学习用于合成规划.

背景情况:

机器学习 (ML) 模型对于有效的化学合成路径生成至关重要.
当前的回复合成模型往往太慢,无法在实践中应用.
减少ML模型的推断时间和碳足迹是一个重大挑战.

研究的目的:

为了减少回归合成的Chemformer模型的推断时间.
研究替代变压器架构,知识蒸 (KD) 和超参数优化.
评估这些优化对单步和多步搜索效率的影响.

主要方法:

评估了与知识蒸 (KD) 密切相关的变压器架构.
调查的基于特征和基于响应的KD.
基于推断时间和模型准确性的优化超参数.
评估了单步预测速度和多步搜索效率.

主要成果:

在使用KD时,替代变压器架构的性能不佳.
减少模型尺寸和提高单步速度是重要的,但不够.
多步骤搜索效率更多地受到单步预测的多样性和信心的影响.
仅仅知识蒸并没有显著提高多步效率.

结论:

进一步的研究应该将KD与其他技术相结合,以改善合成规划.
反合成中的多步搜索效率是复杂的,并且受到超出单步模型速度的因素的影响.
在基于蒙特卡洛的回复合成中平衡勘探和开采是至关重要的.

更多相关视频

Curation of Computational Chemical Libraries Demonstrated with Alpha-Amino Acids

Curation of Computational Chemical Libraries Demonstrated with Alpha-Amino Acids

Published on: April 13, 2022

Author Spotlight: Accelerating Discovery in Microporous Material Chemistry

Author Spotlight: Accelerating Discovery in Microporous Material Chemistry

Published on: October 6, 2023

相关概念视频

Structure-Activity Relationships and Drug Design

Structure-Activity Relationships and Drug Design

Drug design is a dynamic field that involves discovering and developing new medications based on specific biological targets. This process heavily relies on structure-activity relationships (SAR) and quantitative structure-activity relationships (QSAR) to guide the design and optimization of efficient drugs.
SAR studies the intricate relationship between a drug's chemical structure and biological activity. It focuses on understanding how modifications to a drug's structure can influence its...

Methods of Medium Optimization

Methods of Medium Optimization

Optimizing growth media enhances microbial proliferation and maximizes product yield. Statistical experimental design methodologies provide structured and reproducible approaches, offering progressively higher levels of robustness and efficiency.The One-Factor-at-a-Time (OFAT) MethodThe One-Factor-at-a-Time (OFAT) method involves adjusting a single variable while keeping all others constant. However, it cannot detect interactions between variables, often leading to suboptimal outcomes when...