• Corpus ID: 245906143

Improving VAE based molecular representations for compound property prediction

  title={Improving VAE based molecular representations for compound property prediction},
  author={A. Tevosyan and L. Khondkaryan and H. Khachatrian and G. Tadevosyan and L. Apresyan and N. Babayan and H. Stopper and Z. Navoyan},
Collecting labeled data for many important tasks in chemoinformatics is time consuming and requires expensive experiments. In recent years, machine learning has been used to learn rich representations of molecules using large scale unlabeled molecular datasets and transfer the knowledge to solve the more challenging tasks with limited datasets. Variational autoencoders are one of the tools that have been proposed to perform the transfer for both chemical property prediction and molecular… 


Inductive Transfer Learning for Molecular Activity Prediction: Next-Gen QSAR Models with MolPMoFiT
The results showed the Molecular Prediction Model Fine-Tuning approach can achieve comparable or better prediction performances on all three datasets compared to state-of-the-art prediction techniques reported in the literature so far.
Convolutional Embedding of Attributed Molecular Graphs for Physical Property Prediction.
A convolutional neural network is employed for the embedding task of learning an expressive molecular representation by treating molecules as undirected graphs with attributed nodes and edges, and preserves molecule-level spatial information that significantly enhances model performance.
MoleculeNet: A Benchmark for Molecular Machine Learning
MoleculeNet benchmarks demonstrate that learnable representations are powerful tools for molecular machine learning and broadly offer the best performance, however, this result comes with caveats.
Using Rule-Based Labels for Weak Supervised Learning: A ChemNet for Transferable Chemical Property Prediction
This work develops an approach of using rule-based knowledge for training ChemNet, a transferable and generalizable deep neural network for chemical property prediction that learns in a weak-supervised manner from large unlabeled chemical databases.
Prediction of chemical compounds properties using a deep learning model
A new deep learning model, capable of conducting a preliminary screening of chemical compounds in-silico, is described, constructed using a variation autoencoder to generate chemical compound fingerprints, which have been used to create a regression model to predict their LogD property and a classificationmodel to predict binding in selected assays from the ChEMBL dataset.
CRNNTL: convolutional recurrent neural network and transfer learning for QSAR modelling
The convolutional recurrent neural network and transfer learning (CRNNTL) method inspired by the applications of polyphonic sound detection and electrocardiogram classification is proposed and efficient knowledge transfer is achieved to overcome data scarcity considering binding site similarity between different targets.
Transfer and Multi-task Learning in QSAR Modeling: Advances and Challenges
This review will present the main features of transfer and multi-task learning studies, as well as some applications and its potentiality in drug design projects.
Improved Prediction of Aqueous Solubility of Novel Compounds by Going Deeper With Deep Learning
Aqueous solubility is an important physicochemical property of compounds in anti-cancer drug discovery. Artificial intelligence solubility prediction tools have scored impressive performances by
A Survey of Multi‐task Learning Methods in Chemoinformatics
This work reviews the recent developments in multi‐learning approaches as well as cover the freely available tools and packages that can be used to perform joint data analyses for simultaneously predicting different ADMET and biological properties of molecules.
Synergy Effect between Convolutional Neural Networks and the Multiplicity of SMILES for Improvement of Molecular Prediction
The synergy effect between convolutional neural networks and the multiplicity of SMILES during training acts as a regulariser and therefore avoids overfitting and can be seen as ensemble learning when considered for testing.