Quantitative evaluation of molecular generation performance of graph-based GANs

被引:0
|
作者
Zhang, Jinli [1 ]
Wang, Zhenbo [1 ]
Jiang, Zongli [1 ]
Wu, Man [2 ]
Li, Chen [3 ]
Yamanishi, Yoshihiro [3 ]
机构
[1] Beijing Univ Technol, Fac Informat Technol, Beijing, Peoples R China
[2] Keio Univ, Dept Informat & Comp Sci, Yokohama, Japan
[3] Nagoya Univ, Grad Sch Informat, Nagoya, Japan
关键词
Quantitative evaluation; Molecular generation; Generative adversarial network; Reinforcement learning; MODEL;
D O I
10.1007/s11219-024-09671-7
中图分类号
TP31 [计算机软件];
学科分类号
081202 ; 0835 ;
摘要
Deep generative models have been widely used in molecular generation tasks because they can save time and cost in drug development compared with traditional methods. Previous studies based on generative adversarial network (GAN) models typically employ reinforcement learning (RL) to constrain chemical properties, resulting in efficient and novel molecules. However, such models have poor performance in generating molecules due to instability in training. Therefore, quantitative evaluation of existing molecular generation models, especially GAN models, is necessary. This study aims to evaluate the performance of discrete GAN models using RL in molecular generation tasks and explore the impact of different factors on model performance. Through evaluation experiments on QM9 and ZINC datasets, the results show that noise sampling distributions, training epochs, and training data volumes can affect the performance of molecular generation. Finally, we provide strategies for stable training and improved performance for GAN models.
引用
收藏
页码:791 / 819
页数:29
相关论文
共 50 条
  • [1] Exploring Graph Traversal Algorithms in Graph-Based Molecular Generation
    Mercado, Rocio
    Bjerrum, Esben J.
    Engkvist, Ola
    JOURNAL OF CHEMICAL INFORMATION AND MODELING, 2022, 62 (09) : 2093 - 2100
  • [2] Towards Performance Evaluation of Graph-Based Representation
    Jouili, Salim
    Tabbone, Salvatore
    GRAPH-BASED REPRESENTATIONS IN PATTERN RECOGNITION, 2011, 6658 : 72 - 81
  • [3] Performance evaluation of decision tree graph-based induction
    Geamsakul, Warodom
    Matsuda, Takashi
    Yoshida, Tetsuya
    Motoda, Hiroshi
    Washio, Takashi
    2003, Springer Verlag (2843):
  • [4] Performance evaluation of decision tree graph-based induction
    Geamsakul, W
    Matsuda, T
    Yoshida, T
    Motoda, H
    Washio, T
    DISCOVERY SCIENCE, PROCEEDINGS, 2003, 2843 : 128 - 140
  • [5] THE GENERATION OF MOLECULAR-STRUCTURES FROM A GRAPH-BASED QSAR EQUATION
    KIER, LB
    HALL, LH
    QUANTITATIVE STRUCTURE-ACTIVITY RELATIONSHIPS, 1993, 12 (04): : 383 - 388
  • [6] Graph-Based Mutations for Music Generation
    Kanani, Maziar
    O'Leary, Sean
    McDermott, James
    PROCEEDINGS OF THE 2023 GENETIC AND EVOLUTIONARY COMPUTATION CONFERENCE COMPANION, GECCO 2023 COMPANION, 2023, : 1916 - 1919
  • [7] A Graph-Based Generation of Virtual Grids
    Grabska, Ewa
    Palacz, Wojciech
    Strug, Barbara
    Slusarczyk, Grazyna
    PARALLEL PROCESSING AND APPLIED MATHEMATICS, PT I, 2012, 7203 : 451 - 460
  • [8] Graph-based generation of referring expressions
    Krahmer, E
    van Erk, S
    Verleg, A
    COMPUTATIONAL LINGUISTICS, 2003, 29 (01) : 53 - 72
  • [9] Graph-based molecular Pareto optimisation
    Verhellen, Jonas
    CHEMICAL SCIENCE, 2022, 13 (25) : 7526 - 7535
  • [10] Graph-based Molecular Representation Learning
    Guo, Zhichun
    Guo, Kehan
    Nan, Bozhao
    Tian, Yijun
    Iyer, Roshni G.
    Ma, Yihong
    Wiest, Olaf
    Zhang, Xiangliang
    Wang, Wei
    Zhang, Chuxu
    Chawla, Nitesh V.
    PROCEEDINGS OF THE THIRTY-SECOND INTERNATIONAL JOINT CONFERENCE ON ARTIFICIAL INTELLIGENCE, IJCAI 2023, 2023, : 6638 - 6646