TopP-S: Persistent homology-based multi-task deep neural networks for simultaneous predictions of partition coefficient and aqueous solubility.
Kedi WuZhixiong ZhaoRenxiao WangGuo-Wei WeiPublished in: J. Comput. Chem. (2018)
Keyphrases
- multi task
- persistent homology
- neural network
- multi task learning
- learning tasks
- multi class
- topological features
- multiple tasks
- learning problems
- morse theory
- gaussian processes
- transfer learning
- computational geometry
- feature selection
- sparse learning
- competitive learning
- information gain
- text classification
- support vector machine