Modeling Industrial ADMET Data with Multitask Networks

By Steven M. Kearnes, Brian B. Goldman, Vijay S. Pande. 2016

58 citations recorded by the source corpus.

Deep learning methods such as multitask neural networks have recently been applied to ligand-based virtual screening and other drug discovery applications. Using a set of industrial ADMET datasets, we compare neural networks to standard baseline models and analyze multitask learning effects with both random cross-validation and a more relevant temporal validation scheme. We confirm that multitask learning can provide modest benefits over single-task models and show that smaller datasets tend to benefit more than larger datasets from multitask learning. Additionally, we find that adding massive amounts of side information is not guaranteed to improve performance relative to simpler multitask learning. Our results emphasize that multitask effects are highly dataset-dependent, suggesting the use of dataset-specific models to maximize overall performance.

Subfields: Biomedical Engineering, Computational Theory and Mathematics, Materials Chemistry, Model Architectures, Computational Chemistry & Drug Discovery

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Biomedical Engineering

1. Abhinav Vishnu
2. Garrett B. Goh
3. Nathan Oken Hodas
4. Charles Siegel
5. Jure Leskovec
6. Nathan Baker
7. Percy Liang
8. Vijay S. Pande

In Computational Theory and Mathematics

1. Charles Siegel
2. Abhinav Vishnu
3. Garrett B. Goh
4. Nathan Oken Hodas
5. Jure Leskovec
6. Marinka Zitnik
7. Nathan Baker
8. Bowen Liu

In Materials Chemistry

1. Charles Siegel
2. Nathan Oken Hodas
3. Abhinav Vishnu
4. Garrett B. Goh
5. Jure Leskovec
6. Nathan Baker
7. Marinka Zitnik
8. Bowen Liu

In Model Architectures

1. Zhiyong Lu
2. Charles Siegel
3. Abhinav Vishnu
4. Percy Liang
5. Joseph Gomes
6. Jure Leskovec
7. Anne E Carpenter
8. Yanjun Qi

In Computational Chemistry & Drug Discovery

1. Abhinav Vishnu
2. Garrett B. Goh
3. Nathan Oken Hodas
4. Charles Siegel
5. Jure Leskovec
6. Bowen Liu
7. Joseph Gomes
8. Marinka Zitnik

Citations in the corpus

• Opportunities and obstacles for deep learning in biology and medicine (2017, background): “Other studies have shown the limitations of transfer learning when the prediction tasks are insufficiently related [296,311].”
• Pre-training Graph Neural Networks (2019, background): “This phenomenon is referred to as negative transfer [35, 43], which poses a significant problem when deploying pre-trained models to real world applications, and has been previously observed for multi-task learning of molecular property prediction tasks [39, 40, 54, 22].”
• Machine Learning in Drug Discovery: A Review (2021, unknown): “with ADME/T experimental data, and outcome demonstrated better performance in multi-task model (Kearnes et al. 2016).”
• Chemception: A Deep Neural Network with Minimal Chemistry Knowledge Matches the Performance of Expert-developed QSAR/QSPR Models (2017, methodology): “These human-expert features are then used by machine learning algorithms for QSAR and QSPR modeling, which in recent years have also included deep neural networks.(9-23) However, we observed that even in the most recent chemistry literature, deep learning in chemistry still falls under this traditio”