Calibrated uncertainty for molecular property prediction using ensembles of message passing neural networks
Research output: Contribution to journal › Journal article › Research › peer-review
Standard
Calibrated uncertainty for molecular property prediction using ensembles of message passing neural networks. / Busk, Jonas; Jørgensen, Peter Bjørn; Bhowmik, Arghya; Schmidt, Mikkel N.; Winther, Ole; Vegge, Tejs.
In: Machine Learning: Science and Technology, Vol. 3, No. 1, 015012, 2022.Research output: Contribution to journal › Journal article › Research › peer-review
Harvard
APA
Vancouver
Author
Bibtex
}
RIS
TY - JOUR
T1 - Calibrated uncertainty for molecular property prediction using ensembles of message passing neural networks
AU - Busk, Jonas
AU - Jørgensen, Peter Bjørn
AU - Bhowmik, Arghya
AU - Schmidt, Mikkel N.
AU - Winther, Ole
AU - Vegge, Tejs
PY - 2022
Y1 - 2022
N2 - Data-driven methods based on machine learning have the potential to accelerate computational analysis of atomic structures. In this context, reliable uncertainty estimates are important for assessing confidence in predictions and enabling decision making. However, machine learning models can produce badly calibrated uncertainty estimates and it is therefore crucial to detect and handle uncertainty carefully. In this work we extend a message passing neural network designed specifically for predicting properties of molecules and materials with a calibrated probabilistic predictive distribution. The method presented in this paper differs from previous work by considering both aleatoric and epistemic uncertainty in a unified framework, and by recalibrating the predictive distribution on unseen data. Through computer experiments, we show that our approach results in accurate models for predicting molecular formation energies with well calibrated uncertainty in and out of the training data distribution on two public molecular benchmark datasets, QM9 and PC9. The proposed method provides a general framework for training and evaluating neural network ensemble models that are able to produce accurate predictions of properties of molecules with well calibrated uncertainty estimates.
AB - Data-driven methods based on machine learning have the potential to accelerate computational analysis of atomic structures. In this context, reliable uncertainty estimates are important for assessing confidence in predictions and enabling decision making. However, machine learning models can produce badly calibrated uncertainty estimates and it is therefore crucial to detect and handle uncertainty carefully. In this work we extend a message passing neural network designed specifically for predicting properties of molecules and materials with a calibrated probabilistic predictive distribution. The method presented in this paper differs from previous work by considering both aleatoric and epistemic uncertainty in a unified framework, and by recalibrating the predictive distribution on unseen data. Through computer experiments, we show that our approach results in accurate models for predicting molecular formation energies with well calibrated uncertainty in and out of the training data distribution on two public molecular benchmark datasets, QM9 and PC9. The proposed method provides a general framework for training and evaluating neural network ensemble models that are able to produce accurate predictions of properties of molecules with well calibrated uncertainty estimates.
KW - molecular property prediction
KW - machine learning potential
KW - uncertainty quantification
KW - uncertainty calibration
KW - message passing neural network
KW - graph neural network
KW - ensemble model
KW - DESIGN
U2 - 10.1088/2632-2153/ac3eb3
DO - 10.1088/2632-2153/ac3eb3
M3 - Journal article
VL - 3
JO - Machine Learning: Science and Technology
JF - Machine Learning: Science and Technology
SN - 2632-2153
IS - 1
M1 - 015012
ER -
ID: 288267480