Multitask deep learning with dynamic task balancing for quantum mechanical properties prediction

Yang, Ziduo; Zhong, Weihe; Lv, Qiujie; Chen, Calvin Yu-Chian

doi:10.1039/d1cp05172e

Cited by 6 publications

(4 citation statements)

References 35 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…GNNs cannot be fully trusted without understanding and verifying their working mechanisms, , which limits their application in drug discovery scenarios. In this section, we conduct two visual explanation experiments to rationalize GIGN.…”

Section: Visual Explanations For Gignmentioning

confidence: 99%

Geometric Interaction Graph Neural Network for Predicting Protein–Ligand Binding Affinities from 3D Structures (GIGN)

Yang

Zhong

et al. 2023

J. Phys. Chem. Lett.

Self Cite

View full text Add to dashboard Cite

Predicting protein–ligand binding affinities (PLAs) is a core problem in drug discovery. Recent advances have shown great potential in applying machine learning (ML) for PLA prediction. However, most of them omit the 3D structures of complexes and physical interactions between proteins and ligands, which are considered essential to understanding the binding mechanism. This paper proposes a geometric interaction graph neural network (GIGN) that incorporates 3D structures and physical interactions for predicting protein–ligand binding affinities. Specifically, we design a heterogeneous interaction layer that unifies covalent and noncovalent interactions into the message passing phase to learn node representations more effectively. The heterogeneous interaction layer also follows fundamental biological laws, including invariance to translations and rotations of the complexes, thus avoiding expensive data augmentation strategies. GIGN achieves state-of-the-art performance on three external test sets. Moreover, by visualizing learned representations of protein–ligand complexes, we show that the predictions of GIGN are biologically meaningful.

show abstract

Section: Visual Explanations For Gignmentioning

confidence: 99%

Geometric Interaction Graph Neural Network for Predicting Protein–Ligand Binding Affinities from 3D Structures (GIGN)

Yang

Zhong

et al. 2023

J. Phys. Chem. Lett.

Self Cite

View full text Add to dashboard Cite

show abstract

“…In the main task, the loss function is based on the extraction of affective dimensions from the FA, and tasks of the same category are learned at the same speed. Hence, the tasks of each category have the same degree of importance and receive the same a ention, balancing the loss and avoiding seesaw phenomena [50][51][52]. MAE is used for the affective regression task and categorical cross-entropy loss for the classification task.…”

Section: Joint Loss Functionmentioning

confidence: 99%

Multitask Learning-Based Affective Prediction for Videos of Films and TV Scenes

Su,

Lin,

Zhang

et al. 2024

Applied Sciences

View full text Add to dashboard Cite

Film and TV video scenes contain rich art and design elements such as light and shadow, color, composition, and complex affects. To recognize the fine-grained affects of the art carrier, this paper proposes a multitask affective value prediction model based on an attention mechanism. After comparing the characteristics of different models, a multitask prediction framework based on the improved progressive layered extraction (PLE) architecture (multi-headed attention and factor correlation-based PLE), incorporating a multi-headed self-attention mechanism and correlation analysis of affective factors, is constructed. Both the dynamic and static features of a video are chosen as fusion input, while the regression of fine-grained affects and classification of whether a character exists in a video are designed as different training tasks. Considering the correlation between different affects, we propose a loss function based on association constraints, which effectively solves the problem of training balance within tasks. Experimental results on a self-built video dataset show that the algorithm can give full play to the complementary advantages of different features and improve the accuracy of prediction, which is more suitable for fine-grained affect mining of film and TV scenes.

show abstract

“…GNNs cannot be fully trusted without understanding and verifying their inner working mechanisms, which limits their application in drug discovery scenarios. 48,49 In this study, we conducted two visual explanation-related experiments to rationalize the SA-DDI. First, to investigate how the atom hidden vectors evolved during the learning process, we obtained the similarity coefficient between atom pairs by measuring the Pearson correlation coefficient for those hidden vectors.…”

Section: Visual Explanations For Sa-ddimentioning

confidence: 99%

Learning size-adaptive molecular substructures for explainable drug–drug interaction prediction by substructure-aware graph neural network

Yang

Zhong

et al. 2022

Chem. Sci.

Self Cite

View full text Add to dashboard Cite

show abstract

Multitask deep learning with dynamic task balancing for quantum mechanical properties prediction

Abstract: Predicting quantum mechanical properties (QMPs) is greatly important for the innovation of material and chemistry science. Multitask deep learning models had been widely used in QMPs prediction. However, existing multitask...

Cited by 6 publications

References 35 publications

Geometric Interaction Graph Neural Network for Predicting Protein–Ligand Binding Affinities from 3D Structures (GIGN)

Geometric Interaction Graph Neural Network for Predicting Protein–Ligand Binding Affinities from 3D Structures (GIGN)

Multitask Learning-Based Affective Prediction for Videos of Films and TV Scenes

Learning size-adaptive molecular substructures for explainable drug–drug interaction prediction by substructure-aware graph neural network

Contact Info

Product

Resources

About