Exploring Tunable Hyperparameters for Deep Neural Networks with Industrial ADME Data Sets
| Autor: | Christos A. Nicolaou, Steven Combs, Jie Shen, Prashant V. Desai, Suntara Cahya, Yadi Zhou, Jibo Wang |
|---|---|
| Rok vydání: | 2018 |
| Předmět: |
Quantitative structure–activity relationship
Computer science General Chemical Engineering Quantitative Structure-Activity Relationship Library and Information Sciences Machine learning computer.software_genre 01 natural sciences Field (computer science) Deep Learning Drug Discovery 0103 physical sciences Dropout (neural networks) Hyperparameter 010304 chemical physics Artificial neural network business.industry Deep learning General Chemistry 0104 chemical sciences Computer Science Applications 010404 medicinal & biomolecular chemistry Absorption Physicochemical Pharmaceutical Preparations Benchmark (computing) Artificial intelligence business computer Model building |
| Zdroj: | Journal of Chemical Information and Modeling. 59:1005-1016 |
| ISSN: | 1549-960X 1549-9596 |
| Popis: | Deep learning has drawn significant attention in different areas including drug discovery. It has been proposed that it could outperform other machine learning algorithms, especially with big data sets. In the field of pharmaceutical industry, machine learning models are built to understand quantitative structure-activity relationships (QSARs) and predict molecular activities, including absorption, distribution, metabolism, and excretion (ADME) properties, using only molecular structures. Previous reports have demonstrated the advantages of using deep neural networks (DNNs) for QSAR modeling. One of the challenges while building DNN models is identifying the hyperparameters that lead to better generalization of the models. In this study, we investigated several tunable hyperparameters of deep neural network models on 24 industrial ADME data sets. We analyzed the sensitivity and influence of five different hyperparameters including the learning rate, weight decay for L2 regularization, dropout rate, activation function, and the use of batch normalization. This paper focuses on strategies and practices for DNN model building. Further, the optimized model for each data set was built and compared with the benchmark models used in production. Based on our benchmarking results, we propose several practices for building DNN QSAR models. |
| Databáze: | OpenAIRE |
| Externí odkaz: |
|