Chemical Property Prediction Under Experimental Biases
share
The ability to predict the chemical properties of compounds is crucial in discovering novel materials and drugs with specific desired characteristics. Recent significant advances in machine learning technologies have enabled automatic predictive modeling from past experimental data reported in the literature.However, these datasets are often biased due to various reasons, such as experimental plans and publication decisions, and the prediction models trained using such biased datasets often suffer from over-fitting to the biased distributions and perform poorly on subsequent uses.The present study focuses on mitigating bias in the experimental datasets. To this purpose, we adopt two techniques from causal inference and domain adaptation combined with graph neural networks capable of handling molecular structures.The experimental results in four possible bias scenarios show that the inverse propensity scoring-based method makes solid improvements, while the domain-invariant representation learning approach fails.
READ FULL TEXT
