☆ 4.6 Article

Explainable artificial intelligence: A taxonomy and guidelines for its application to drug discovery

WILEY INTERDISCIPLINARY REVIEWS-COMPUTATIONAL MOLECULAR SCIENCE (2023)

相关参考文献

注意：仅列出部分参考文献，下载原文获取全部文献信息。

Article Computer Science, Artificial Intelligence

Explainability in Graph Neural Networks: A Taxonomic Survey

Hao Yuan et al.

Summary: Deep learning methods have achieved impressive performance on various AI tasks, but their interpretability is limited. Recent progress has been made in explaining deep models on images, texts, and graph data, especially graph neural networks (GNNs). However, there is a lack of unified GNN explainability methods and standard evaluation benchmarks. This survey provides a unified and taxonomic view of current GNN explainability methods and establishes a testbed with datasets, algorithms, and evaluation metrics. Comprehensive experiments are conducted to compare and analyze the performance of different techniques.

IEEE TRANSACTIONS ON PATTERN ANALYSIS AND MACHINE INTELLIGENCE (2023)

添加到收藏夹

Article Engineering, Chemical

An accurate and interpretable deep learning model for environmental properties prediction using hybrid molecular representations

Jun Zhang et al.

Summary: This study developed an accurate and interpretable deep neural network (AI-DNN) model for predicting lipophilicity. A hybrid method of molecular representation, combining directed message passing neural networks and fixed molecule-level features, was employed to capture the local and global features of molecules. The proposed model demonstrated promising predictive accuracy and discriminative power in structural and stereoisomers. The use of Monte Carlo Tree Search allowed for interpretation of the model, which is important in fields with a high demand for interpretable deep networks, such as green solvent design and drug discovery.

AICHE JOURNAL (2022)

添加到收藏夹

Article Biochemical Research Methods

Explainable multimodal machine learning model for classifying pregnancy drug safety

Guy Shtar et al.

Summary: This study proposes an explainable machine learning model for classifying pregnancy drug safety based on multimodal data. By processing over 100,000 textual responses, a labeled drug dataset was created and structured textual information was incorporated into the model. The findings suggest the safety of Varenicline and Mebeverine during pregnancy, while Meloxicam poses a higher risk.

BIOINFORMATICS (2022)

添加到收藏夹

Article Biochemical Research Methods

DeepDDS: deep graph neural network with attention mechanism to predict synergistic drug combinations

Jinxian Wang et al.

Summary: In this paper, we proposed a deep learning model based on graph neural network and attention mechanism to identify drug combinations that can effectively inhibit the viability of specific cancer cells. The model, called DeepDDS, achieved better performance than other methods in predicting drug synergy. Additionally, we explored the interpretability of the model and found important chemical substructures of drugs. DeepDDS is considered an effective tool for prioritizing synergistic drug combinations for further experimental validation.

BRIEFINGS IN BIOINFORMATICS (2022)

添加到收藏夹

Article Biochemical Research Methods

Graph Convolutional Networks for Drug Response Prediction

Tuan Nguyen et al.

Summary: This study proposes a novel method called GraphDRP based on graph convolutional networks for drug response prediction and finds that graph representation can improve prediction performance.

IEEE-ACM TRANSACTIONS ON COMPUTATIONAL BIOLOGY AND BIOINFORMATICS (2022)

添加到收藏夹

Article Chemistry, Medicinal

Benchmarking Molecular Feature Attribution Methods with Activity Cliffs

Jose Jimenez-Luna et al.

Summary: Feature attribution techniques are popular in explainable AI and can help understand the important parts of inputs for specific predictions. In molecular design, these techniques involve coloring molecular graphs, aiding chemists in decision-making. However, quantitative evaluation of such coloring methods has mainly focused on substructure identification tasks.

JOURNAL OF CHEMICAL INFORMATION AND MODELING (2022)

添加到收藏夹

Article Chemistry, Medicinal

PlayMolecule Glimpse: Understanding Protein-Ligand Property Predictions with Interpretable Neural Networks

Alejandro Varela-Rial et al.

Summary: In this study, the authors present a new application to visualize the contribution of each input atom to the prediction made by a convolutional neural network, aiding in the interpretability of such predictions. The results suggest that the model is able to learn meaningful chemistry signals from the data, but it has also exposed the inaccuracies of the current model, serving as a guideline for further optimization.

JOURNAL OF CHEMICAL INFORMATION AND MODELING (2022)

添加到收藏夹

Article Chemistry, Medicinal

MolGPT: Molecular Generation Using a Transformer-Decoder Model

Viraj Bagal et al.

Summary: The application of deep learning techniques in inverse molecular design for drug design has gained significant attention. In this study, we train a transformer-decoder model, MolGPT, on the next token prediction task using masked self-attention, and demonstrate that it performs on par with other modern machine learning frameworks for generating valid, unique, and novel druglike molecules. We also show that the model can be trained conditionally to control multiple properties of the generated molecules.

JOURNAL OF CHEMICAL INFORMATION AND MODELING (2022)

添加到收藏夹

Article Chemistry, Medicinal

Interpretation of Structure-Activity Relationships in Real-World Drug Design Data Sets Using Explainable Artificial Intelligence

Tobias Harren et al.

Summary: In silico models based on Deep Neural Networks (DNNs) have shown promise in predicting activities and properties of new molecules. However, their black-box nature hinders understanding of which structural features are crucial for activity. Explainable Artificial Intelligence (XAI) methods have recently gained popularity to address this gap. In this study, we apply and compare multiple XAI methods to lead optimization data sets with known structure-activity relationships (SARs) and available X-ray crystal structures. Our results show that combining DNN models with powerful interpretation methods, particularly SHAP-based methods, yields easily understandable and comprehensive interpretations. A novel visualization scheme using atom-based heatmaps provides valuable insights into the underlying SAR.

JOURNAL OF CHEMICAL INFORMATION AND MODELING (2022)

添加到收藏夹

Article Biochemistry & Molecular Biology

ABCpred: a webserver for the discovery of acetyl- and butyryl-cholinesterase inhibitors

Aijaz Ahmad Malik et al.

Summary: Alzheimer's disease is a common form of dementia associated with cognitive decline and language impairment. Inhibition of cholinesterase enzymes, particularly acetylcholinesterase and butyrylcholinesterase, shows promise as a therapeutic approach for AD. A large-scale classification structure-activity relationship model was developed to predict cholinesterase inhibitory activities and identify important substructures influencing their activities.

MOLECULAR DIVERSITY (2022)

添加到收藏夹

Article Chemistry, Multidisciplinary

MGraphDTA: deep multiscale graph neural network for explainable drug-target binding affinity prediction

Ziduo Yang et al.

Summary: In this study, a deep multiscale graph neural network called MGraphDTA was proposed for predicting drug-target affinity (DTA). The method combines dense connection and super-deep GNN to capture the local and global structure of compounds simultaneously. A novel visual explanation method, Grad-AAM, was also developed to analyze the model from a chemical perspective.

CHEMICAL SCIENCE (2022)

添加到收藏夹

Article Chemistry, Multidisciplinary

Model agnostic generation of counterfactual explanations for molecules

Geemi P. Wellawatte et al.

Summary: A major challenge in deep learning in chemistry is the lack of interpretability, which hinders the deployment of AI models. Counterfactuals provide a rationale for model predictions and insights into chemical structures. This work demonstrates a universal model-agnostic approach that can explain predictions of any black-box model.

CHEMICAL SCIENCE (2022)

添加到收藏夹

Review Chemistry, Multidisciplinary

Application advances of deep learning methods for de novo drug design and molecular dynamics simulation

Qifeng Bai et al.

Summary: De novo drug design involves building novel ligands by assembling atoms or fragments, while molecular dynamics simulation studies the interaction mechanism between ligands and receptors based on molecular force field. Deep learning methods and interpretable machine learning can improve the efficiency and accuracy of drug design and simulations. Further application of deep learning and IML can promote technical development in drug discovery.

WILEY INTERDISCIPLINARY REVIEWS-COMPUTATIONAL MOLECULAR SCIENCE (2022)

添加到收藏夹

Review Chemistry, Multidisciplinary

A Systematic Review of Explainable Artificial Intelligence in Terms of Different Application Domains and Tasks

Mir Riyanul Islam et al.

Summary: This paper presents a systematic literature review of recent developments in explainable artificial intelligence (XAI) methods and evaluation metrics for different application domains and tasks. The study finds that XAI methods are mostly used in safety-critical domains, deep learning and ensemble models are popular, visual explanations are preferred, and more attention is needed to generate explanations for general users in sensitive domains.

APPLIED SCIENCES-BASEL (2022)

添加到收藏夹

Review Computer Science, Artificial Intelligence

Machine Learning in Drug Discovery: A Review

Suresh Dara et al.

Summary: This review provides an overview of the application of machine learning in drug discovery, emphasizing its role in accelerating the research process, reducing clinical trial risks and expenditures. It also discusses the challenges and issues faced in this field.

ARTIFICIAL INTELLIGENCE REVIEW (2022)

添加到收藏夹

Article Biochemical Research Methods

Ensemble modeling with machine learning and deep learning to provide interpretable generalized rules for classifying CNS drugs with high prediction power

Tzu-Hui Yu et al.

Summary: The trade-off between predictability and interpretability of machine learning and deep learning models is a growing concern in central nervous system-related quantitative structure-activity relationship analysis. We developed a protocol combining ML and DL to generate simple rules that are easy to interpret with high prediction power.

BRIEFINGS IN BIOINFORMATICS (2022)

添加到收藏夹

Article Computer Science, Software Engineering

A survey of visual analytics for Explainable Artificial Intelligence methods

Gulsum Alicioglu et al.

Summary: Deep learning models have achieved impressive performance in various domains, but their black-box nature calls for explanations. Explainable Artificial Intelligence (XAI) provides tools and techniques to reveal the behavior and decision-making mechanisms of these models. Visualization techniques help present model and prediction explanations in an understandable and interpretable way. This survey reviews the current trends and challenges of interpreting deep learning models using XAI methods and proposes future research directions in visual analytics.

COMPUTERS & GRAPHICS-UK (2022)

添加到收藏夹

Article Biochemical Research Methods

Discovery of moiety preference by Shapley value in protein kinase family using random forest models

Yu-Wei Huang et al.

Summary: This study successfully constructed accurate prediction models for kinase family inhibitors using various features and machine learning methods. It also discovered common and specific inhibitor features between different kinase families, providing new insights for inhibitor design.

BMC BIOINFORMATICS (2022)

添加到收藏夹

Review Biology

Generative machine learning for de novo drug discovery: A systematic review

Dominic D. Martinelli

Summary: Recent research suggests that machine learning algorithms can generate novel drug-like molecules, revolutionizing drug discovery. This systematic literature review examines machine learning models, challenges, and encoding methods in molecular design. The review identified six prominent algorithms and eight central challenges. Statistical analysis and visualization show the evolution of machine learning approaches in drug design over the past five years. Future opportunities and reservations are also discussed.

COMPUTERS IN BIOLOGY AND MEDICINE (2022)

添加到收藏夹

Article Computer Science, Artificial Intelligence

Counterfactuals and causability in explainable artificial intelligence: Theory, algorithms, and applications

Yu-Liang Chou et al.

Summary: Deep learning models have high performance but are hard to interpret due to complexity, sparking interest in model-agnostic methods for transparency. Counterfactual algorithms have potential for causability, but current ones lack grounding in causal theory. This paper highlights challenges and new directions for promoting causability in explainable AI through model-agnostic approaches.

INFORMATION FUSION (2022)

添加到收藏夹

Article Chemistry, Medicinal

Structure-Aware Multimodal Deep Learning for Drug-Protein Interaction Prediction

Penglei Wang et al.

Summary: In this study, we propose a structure-aware multimodal deep DPI prediction model (STAMP-DPI), which accurately predicts drug-protein interactions by training on a carefully curated industry-scale benchmark dataset. The model combines the feature representations of molecules and proteins, effectively capturing the interaction features between them using graph neural networks and pretrained embeddings. Experimental results demonstrate that STAMP-DPI outperforms existing methods on multiple datasets and has interpretability.

JOURNAL OF CHEMICAL INFORMATION AND MODELING (2022)

添加到收藏夹

Article Chemistry, Multidisciplinary

DeSIDE-DDI: interpretable prediction of drug-drug interactions using drug-induced gene expressions

Eunyoung Kim et al.

Summary: This article introduces a deep learning-based framework for DDI prediction with drug-induced gene expression signatures to improve prediction accuracy and provide interpretability for the model.

JOURNAL OF CHEMINFORMATICS (2022)

添加到收藏夹

Article Chemistry, Multidisciplinary

Collaborative Approach between Explainable Artificial Intelligence and Simplified Chemical Interactions to Explore Active Ligands for Cyclin-Dependent Kinase 2

Tomomi Shimazaki et al.

Summary: This study proposes a collaborative approach using explainable artificial intelligence and simplified chemical interaction scores to efficiently search for active ligands bound to the target receptor in drug discovery virtual screening. By employing machine learning and simplified scoring functions, the research demonstrates improved classification ability of machine learning models and highlights important residues of the target receptor.

ACS OMEGA (2022)

添加到收藏夹

Article Biochemical Research Methods

AttentionSiteDTI: an interpretable graph-based model for drug-target interaction prediction using NLP sentence-level relation classification

Mehdi Yazdani-Jahromi et al.

Summary: In this study, an interpretable graph-based deep learning prediction model called AttentionSiteDTI is introduced to address the problem of drug-target interaction prediction. The model utilizes protein binding sites and a self-attention mechanism to identify the most contributive binding sites. Experimental results show that AttentionSiteDTI outperforms current state-of-the-art models on three benchmark datasets and exhibits high generalizability on new proteins. The agreement between computationally predicted and experimentally observed drug-target interactions demonstrates the potential of the proposed method as an effective pre-screening tool in drug repurposing applications.

BRIEFINGS IN BIOINFORMATICS (2022)

添加到收藏夹

Article Neurosciences

Designing for Confidence: The Impact of Visualizing Artificial Intelligence Decisions

Alexander John Karran et al.

Summary: Explainable artificial intelligence aims to improve the Human-AI trust relationship by translating, simplifying, and visualizing AI decisions. Two studies show that morphological clarity and information presentation order have an impact on trust. The results indicate that low morphological clarity positively impacts trust in AI and the order of information presentation influences user trust.

FRONTIERS IN NEUROSCIENCE (2022)

添加到收藏夹

Article Pharmacology & Pharmacy

Do AutoML-Based QSAR Models Fulfill OECD Principles for Regulatory Assessment? A 5-HT1A Receptor Case

Natalia Czub et al.

Summary: The study presents the best-in-class predictive model for the serotonin 1A receptor affinity and validates it according to OECD guidelines. This can help simplify the drug discovery and development process, improving efficiency.

PHARMACEUTICS (2022)

添加到收藏夹

Article Computer Science, Artificial Intelligence

Quantitative evaluation of explainable graph neural networks for molecular property prediction

Jiahua Rao et al.

Summary: In this study, XAI methods were quantitatively assessed on GNN models using five molecular property benchmarks. The results showed that XAI methods can provide reliable and informative answers for medicinal chemists in identifying key substructures, leading to improved molecular property predictions.

PATTERNS (2022)

添加到收藏夹

Article Computer Science, Artificial Intelligence

Higher-Order Explanations of Graph Neural Networks via Relevant Walks

Thomas Schnake et al.

Summary: This paper introduces a new method for explaining graph neural networks, which can extract relevant input graph traversals that contribute to the prediction using higher-order expansions and nested attribution scheme. It has practical applications in areas such as sentiment analysis of text data, structure-property relationships in quantum chemistry, and image classification.

IEEE TRANSACTIONS ON PATTERN ANALYSIS AND MACHINE INTELLIGENCE (2022)

添加到收藏夹

Article Computer Science, Artificial Intelligence

Knowledge-guided deep learning models of drug toxicity improve interpretation

Yun Hao et al.

Summary: DTox is an interpretation framework for identifying toxicity pathways, predicting compound response to toxicity assays, improving interpretability, and discovering different mechanisms of toxicity.

PATTERNS (2022)

添加到收藏夹

Article Computer Science, Information Systems

Toward Practical Usage of the Attention Mechanism as a Tool for Interpretability

Martin Tutek et al.

Summary: Natural language processing (NLP) has been greatly influenced by the neural revolution in artificial intelligence. Attention mechanisms provide transparency to previously black-box recurrent neural network (RNN) models, but recent research questions their faithfulness. This study presents a regularization technique to improve the faithfulness of attention-based explanations, showing consistent improvements across various datasets and models.

IEEE ACCESS (2022)

添加到收藏夹

Article Chemistry, Medicinal

Explainable Deep Relational Networks for Predicting Compound- Protein Affinities and Contacts

Mostafa Karimi et al.

Summary: This study utilizes deep learning to improve the interpretability of compound-protein affinity prediction by defining intermolecular contacts. By embedding protein sequences and compound graphs with joint attentions, as well as introducing three methodological advances, it achieves high accuracy in affinity prediction. Compared to other methods, our models show better performance in both affinity prediction and contact prediction.

JOURNAL OF CHEMICAL INFORMATION AND MODELING (2021)

添加到收藏夹

Article Computer Science, Artificial Intelligence

Notions of explainability and evaluation approaches for explainable artificial intelligence

Giulia Vilone et al.

Summary: XAI has seen significant growth due to the widespread application of machine learning, particularly deep learning. Studies have identified various notions and requirements for explanations to be easily understandable and actionable, proposing different approaches for evaluation. However, there is no general consensus among scholars on how to define explanations and assess their validity and reliability.

INFORMATION FUSION (2021)

添加到收藏夹

Article Chemistry, Medicinal

Explainable Machine Learning for Property Predictions in Compound Optimization

Raquel Rodriguez-Perez et al.

Summary: The text discusses the application of machine learning in medicinal chemistry, stating that ML is often used in applications with large datasets while in compound optimization, understanding the decision-making process of ML models is crucial. It notes that only a few ML methods are interpretable, and suggests that explanatory approaches can be applied to gain insights into complex ML model decisions.

JOURNAL OF MEDICINAL CHEMISTRY (2021)

添加到收藏夹

Article Computer Science, Software Engineering

Explainable Matrix - Visualization for Global and Local Interpretability of Random Forest Classification Ensembles

Mario Popolin Neto et al.

Summary: Classification models are crucial tools in various domains, but visualizing large and complex models like Random Forest remains a challenge. The new ExMatrix method utilizes a matrix-like visual metaphor to enhance interpretability and analysis of massive rule-based models.

IEEE TRANSACTIONS ON VISUALIZATION AND COMPUTER GRAPHICS (2021)

添加到收藏夹

Article Chemistry, Medicinal

Coloring Molecules with Explainable Artificial Intelligence for Preclinical Relevance Assessment

Jose Jimenez-Luna et al.

Summary: This study improves modeling transparency of graph neural network models in drug discovery by applying explainable artificial intelligence methods. It provides important insights into molecular features and structural elements, helps identify key features of drug properties, and offers a deeper understanding of drug-target interactions.

JOURNAL OF CHEMICAL INFORMATION AND MODELING (2021)

添加到收藏夹

Article Chemistry, Medicinal

Transferable Multilevel Attention Neural Network for Accurate Prediction of Quantum Chemistry Properties via Multitask Learning

Ziteng Liu et al.

Summary: DeepMoleNet is a multilevel attention neural network that efficiently predicts 12 properties, including dipole moment, HOMO, and Gibbs free energy, demonstrating good transferability for high-throughput screening of chemical species in both equilibrium and nonequilibrium molecular spaces.

JOURNAL OF CHEMICAL INFORMATION AND MODELING (2021)

添加到收藏夹

Review Biochemistry & Molecular Biology

A review on compound-protein interaction prediction methods: Data, format, representation and model

Sangsoo Lim et al.

Summary: The review comprehensively covers topics related to computational prediction of compound protein interactions, highlighting the accumulation and curation of CPI data, the power of computational methods to analyze complex data, and the success in improved prediction quality through sophisticated techniques and high-quality information.

COMPUTATIONAL AND STRUCTURAL BIOTECHNOLOGY JOURNAL (2021)

添加到收藏夹

Article Multidisciplinary Sciences

Explainable machine learning predictions of dual-target compounds reveal characteristic structural features

Christian Feldmann et al.

Summary: Compounds with defined multi-target activity are important in drug discovery, and explainable machine learning can uncover structural motifs characteristic of dual-target compounds, improving predictive performance and interpretability.

SCIENTIFIC REPORTS (2021)

添加到收藏夹

Article Computer Science, Artificial Intelligence

New Insights Into Drug Repurposing for COVID-19 Using Deep Learning

Chun Yen Lee et al.

Summary: The article discusses drug repurposing strategies for COVID-19 and proposes improvements to current deep learning approaches for identifying and repurposing drugs. The enhanced DL approach aims to extract features from big data effectively, providing a transparent model with high prediction accuracy for repurposing drugs not only for COVID-19 but also for other complex diseases.

IEEE TRANSACTIONS ON NEURAL NETWORKS AND LEARNING SYSTEMS (2021)

添加到收藏夹

Article Biochemical Research Methods

SumGNN: multi-typed drug interaction prediction via efficient knowledge graph summarization

Yue Yu et al.

Summary: The paper introduces a new method called SumGNN, which utilizes a knowledge summarization graph neural network to improve multi-typed DDI pharmacological effect prediction using large biomedical knowledge graphs. Results show significant performance gains, particularly in low data relation types. Additionally, SumGNN provides interpretable predictions through generated reasoning paths for each prediction.

BIOINFORMATICS (2021)

添加到收藏夹

Article Biochemical Research Methods

An effective self-supervised framework for learning expressive molecular global representations to drug discovery

Pengyong Li et al.

Summary: This study introduces a novel molecular pre-training graph-based deep learning framework, called MPG, which learns molecular representations from large-scale unlabeled molecules. MolGNet, the powerful GNN for modeling molecular graph, can capture valuable chemical insights to produce interpretable representations. The pre-trained MolGNet can be fine-tuned and easily applied to a wide range of drug discovery tasks through simple structural adjustments.

BRIEFINGS IN BIOINFORMATICS (2021)

添加到收藏夹

Article Pharmacology & Pharmacy

Feature importance of machine learning prediction models shows structurally active part and important physicochemical features in drug design

Katsunori Sasahara et al.

Summary: The study aimed to design new chemical entities with desirable pharmacokinetic profiles by obtaining indicators of physicochemical parameters and structurally active sites through machine learning models. Prediction models for metabolic stability and substrate recognition, as well as models showing structurally active sites, were developed efficiently to optimize lead compounds.

DRUG METABOLISM AND PHARMACOKINETICS (2021)

添加到收藏夹

Article Chemistry, Medicinal

MSA-Regularized Protein Sequence Transformer toward Predicting Genome-Wide Chemical-Protein Interactions: Application to GPCRome Deorphanization

Tian Cai et al.

Summary: This study develops a new deep learning framework, DISAE, to predict chemical binding to evolutionarily divergent unannotated proteins, improving generalizability and outperforming existing methods. By incorporating evolutionary information into self-supervised learning of protein sequences, DISAE shows promising results in exploring the chemical landscape of entire sequenced genomes.

JOURNAL OF CHEMICAL INFORMATION AND MODELING (2021)

添加到收藏夹

Review Biochemistry & Molecular Biology

Trends in Deep Learning for Property-driven Drug Design

Jannis Born et al.

Summary: The combination of high-throughput screening and deep learning has accelerated drug development in the pharmaceutical industry. Deep generative models have emerged as a powerful tool to explore chemical space and expedite drug discovery. Future challenges include building multimodal conditional generative models, integrating diverse knowledge sources for mapping biochemical properties to target structures, and incorporating deep learning systems into experimental workflows.

CURRENT MEDICINAL CHEMISTRY (2021)

添加到收藏夹

Article Computer Science, Artificial Intelligence

Molecular generative Graph Neural Networks for Drug Discovery

Pietro Bongini et al.

Summary: Drug discovery is a costly and time-consuming field of research, but computational methods, particularly machine learning, can help reduce these costs. This paper presents a molecular graph generator based on graph neural network modules, which shows promising applications in generating new molecules with potential uses.

NEUROCOMPUTING (2021)

添加到收藏夹

Article Multidisciplinary Sciences

Feature importance correlation from machine learning indicates functional relationships between proteins and similar compound binding characteristics

Raquel Rodriguez-Perez et al.

Summary: This study introduces a new approach that uses compound activity predictions to uncover relationships between target proteins, showing feature importance correlation analysis and functional relationships between proteins. The approach does not rely on specific representations, algorithms, or metrics and has potential practical applications in drug discovery.

SCIENTIFIC REPORTS (2021)

添加到收藏夹

Article Chemistry, Multidisciplinary

Evidential Deep Learning for Guided Molecular Property Prediction and Discovery

Ava P. Soleimany et al.

Summary: This paper introduces a new approach to uncertainty quantification for neural network-based molecular structure-property prediction using evidential deep learning, which enables calibrated predictions, sample-efficient training, and improved experimental validation rates in the chemical and physical sciences.

ACS CENTRAL SCIENCE (2021)

添加到收藏夹

Proceedings Paper Computer Science, Artificial Intelligence

Bayesian Deep Basis Fitting for Depth Completion with Uncertainty

Chao Qu et al.

Summary: This study investigates uncertainty estimation for image-guided depth completion using Bayesian evidence framework. The Bayesian Deep Basis Fitting (BDBF) approach can predict high-quality uncertainty estimates and enable depth completion with few or no sparse measurements. Controlled experiments show that BDBF produces better uncertainty estimates with accurate depth prediction.

2021 IEEE/CVF INTERNATIONAL CONFERENCE ON COMPUTER VISION (ICCV 2021) (2021)

添加到收藏夹

Article Genetics & Heredity

Predicting anticancer hyperfoods with graph convolutional networks

Guadalupe Gonzalez et al.

Summary: This study used unsupervised and supervised algorithms to predict cancer-beating molecules within food. By representing each drug as a graph and using a graph neural network model, the model outperformed the baseline approach in predicting anticancer therapeutics. The model demonstrated the ability to capture biological pathway knowledge and showed interpretability.

HUMAN GENOMICS (2021)

添加到收藏夹

Article Chemistry, Medicinal

Pushing the Boundaries of Molecular Representation for Drug Discovery with the Graph Attention Mechanism

Zhaoping Xiong et al.

JOURNAL OF MEDICINAL CHEMISTRY (2020)

添加到收藏夹

Article Chemistry, Medicinal

Interpretation of Compound Activity Predictions from Complex Machine Learning Models Using Local Approximations and Shapley Values

Raquel Rodriguez-Perez et al.

JOURNAL OF MEDICINAL CHEMISTRY (2020)

添加到收藏夹

Article Medicine, General & Internal

Estimated Research and Development Investment Needed to Bring a New Medicine to Market, 2009-2018

Olivier J. Wouters et al.

JAMA-JOURNAL OF THE AMERICAN MEDICAL ASSOCIATION (2020)

添加到收藏夹

Article Biology

hERG-Att: Self-attention-based deep neural network for predicting hERG blockers

Hyunho Kim et al.

COMPUTATIONAL BIOLOGY AND CHEMISTRY (2020)

添加到收藏夹

Article Computer Science, Artificial Intelligence

Semi-supervised regression trees with application to QSAR modelling

Jurica Levatic et al.

EXPERT SYSTEMS WITH APPLICATIONS (2020)

添加到收藏夹

Article Biochemistry & Molecular Biology

Revealing cytotoxic substructures in molecules using deep learning

Henry E. Webel et al.

JOURNAL OF COMPUTER-AIDED MOLECULAR DESIGN (2020)

添加到收藏夹

Article Biochemistry & Molecular Biology

MONN: A Multi-objective Neural Network for Predicting Compound-Protein Interactions and Affinities

Shuya Li et al.

CELL SYSTEMS (2020)

添加到收藏夹

Article Biochemical Research Methods

GCN-BMP: Investigating graph representation learning for DDI prediction task

Xin Chen et al.

METHODS (2020)

添加到收藏夹

Article Oncology

Predicting Drug Response and Synergy Using a Deep Learning Model of Human Cancer Cells

Brent M. Kuenzi et al.

CANCER CELL (2020)

添加到收藏夹

Article Computer Science, Interdisciplinary Applications

Multi-view self-attention for interpretable drug-target interaction prediction

Brighter Agyemang et al.

JOURNAL OF BIOMEDICAL INFORMATICS (2020)

添加到收藏夹

Article Chemistry, Medicinal

SMARTS.plus - A Toolbox for Chemical Pattern Design

Christiane Ehrt et al.

MOLECULAR INFORMATICS (2020)

添加到收藏夹

Review Chemistry, Multidisciplinary

Interpretable machine learning as a tool for scientific discovery in chemistry

Richard Dybowski

NEW JOURNAL OF CHEMISTRY (2020)

添加到收藏夹

Review Computer Science, Artificial Intelligence

Drug discovery with explainable artificial intelligence

Jose Jimenez-Luna et al.

NATURE MACHINE INTELLIGENCE (2020)

添加到收藏夹

Article Computer Science, Theory & Methods

A Survey of Methods for Explaining Black Box Models

Riccardo Guidotti et al.

ACM COMPUTING SURVEYS (2019)

添加到收藏夹

Article Chemistry, Medicinal

Optimal Piecewise Linear Regression Algorithm for QSAR Modelling

Jonathan Cardoso-Silva et al.

MOLECULAR INFORMATICS (2019)

添加到收藏夹

Article Computer Science, Artificial Intelligence

Explanation in artificial intelligence: Insights from the social sciences

Tim Miller

ARTIFICIAL INTELLIGENCE (2019)

添加到收藏夹

Article Biochemical Research Methods

DeepAffinity: interpretable deep learning of compound-protein affinity through unified recurrent and convolutional neural networks

Mostafa Karimi et al.

BIOINFORMATICS (2019)

添加到收藏夹

Article Chemistry, Medicinal

Interpretation of QSAR Models by Coloring Atoms According to Changes in Predicted Activity: How Robust Is It?

Robert P. Sheridan

JOURNAL OF CHEMICAL INFORMATION AND MODELING (2019)

添加到收藏夹

Editorial Material Chemistry, Medicinal

Machine Learning in Drug Discovery

Guenter Klambauer et al.

JOURNAL OF CHEMICAL INFORMATION AND MODELING (2019)

添加到收藏夹

Article Chemistry, Medicinal

Target-Specific Prediction of Ligand Affinity with Structure-Based Interaction Fingerprints

Florian Leidner et al.

JOURNAL OF CHEMICAL INFORMATION AND MODELING (2019)

添加到收藏夹

Article Medicine, Research & Experimental

Toward Explainable Anticancer Compound Sensitivity Prediction via Multimodal Attention-Based Convolutional Encoders

Matteo Manica et al.

MOLECULAR PHARMACEUTICS (2019)

添加到收藏夹

Article Computer Science, Artificial Intelligence

Neural-based approaches to overcome feature selection and applicability domain in drug-related property prediction

Maria Virginia Sabando et al.

APPLIED SOFT COMPUTING (2019)

添加到收藏夹

Article Biochemical Research Methods

A Random Forest Model to Predict the Activity of a Large Set of Soluble Epoxide Hydrolase Inhibitors Solely Based on a Set of Simple Fragmental Descriptors

Jamal Shamsara

COMBINATORIAL CHEMISTRY & HIGH THROUGHPUT SCREENING (2019)

添加到收藏夹

Article Chemistry, Medicinal

Deep Confidence: A Computationally Efficient Framework for Calculating Reliable Prediction Errors for Deep Neural Networks

Isidro Cortes-Ciriano et al.

JOURNAL OF CHEMICAL INFORMATION AND MODELING (2019)

添加到收藏夹

Proceedings Paper Computer Science, Artificial Intelligence

Explainability Methods for Graph Convolutional Neural Networks

Phillip E. Pope et al.

2019 IEEE/CVF CONFERENCE ON COMPUTER VISION AND PATTERN RECOGNITION (CVPR 2019) (2019)

添加到收藏夹

Article Computer Science, Artificial Intelligence

Stop explaining black box machine learning models for high stakes decisions and use interpretable models instead

Cynthia Rudin

NATURE MACHINE INTELLIGENCE (2019)

添加到收藏夹

Article Computer Science, Artificial Intelligence

Descriptor Free QSAR Modeling Using Deep Learning With Long Short-Term Memory Neural Networks

Suman K. Chakravarti et al.

FRONTIERS IN ARTIFICIAL INTELLIGENCE (2019)

添加到收藏夹

Article Biochemical Research Methods

The revival of the Gini importance?

Stefano Nembrini et al.

BIOINFORMATICS (2018)

添加到收藏夹

Review Pharmacology & Pharmacy

Computational modeling of human oral bioavailability: what will be next?

Miguel Angel Cabrera-Perez et al.

EXPERT OPINION ON DRUG DISCOVERY (2018)

添加到收藏夹

Article Computer Science, Software Engineering

Visual Analytics for Explainable Deep Learning

Jaegul Choo et al.

IEEE COMPUTER GRAPHICS AND APPLICATIONS (2018)

添加到收藏夹

Article Chemistry, Medicinal

Multi-Descriptor Read Across (MuDRA): A Simple and Transparent Approach for Developing Accurate Quantitative Structure-Activity Relationship Models

Vinicius M. Alves et al.

JOURNAL OF CHEMICAL INFORMATION AND MODELING (2018)

添加到收藏夹

Article Chemistry, Medicinal

KDEEP: Protein-Ligand Absolute Binding Affinity Prediction via 3D-Convolutional Neural Networks

Jose Jimenez et al.

JOURNAL OF CHEMICAL INFORMATION AND MODELING (2018)

添加到收藏夹

Article Chemistry, Multidisciplinary

Interpretation of ANN-based QSAR models for prediction of antioxidant activity of flavonoids

Petar Zuvela et al.

JOURNAL OF COMPUTATIONAL CHEMISTRY (2018)

添加到收藏夹

Article Biochemical Research Methods

Predicting adverse drug reactions through interpretable deep learning framework

Sanjoy Dey et al.

BMC BIOINFORMATICS (2018)

添加到收藏夹

Article Chemistry, Medicinal

Comparison of the Predictive Performance and Interpretability of Random Forest and Linear Models on Benchmark Data Sets

Richard L. Marchese Robinson et al.

JOURNAL OF CHEMICAL INFORMATION AND MODELING (2017)

添加到收藏夹

Article Chemistry, Medicinal

Interpretation of Quantitative Structure Activity Relationship Models: Past, Present, and Future

Pavel Polishchuk

JOURNAL OF CHEMICAL INFORMATION AND MODELING (2017)

添加到收藏夹

Article Computer Science, Artificial Intelligence

Explaining nonlinear classification decisions with deep Taylor decomposition

Gregoire Montavon et al.

PATTERN RECOGNITION (2017)

添加到收藏夹

Article Biochemical Research Methods

Decision Trees for Continuous Data and Conditional Mutual Information as a Criterion for Splitting Instances

Georgios Drakakis et al.

COMBINATORIAL CHEMISTRY & HIGH THROUGHPUT SCREENING (2016)

添加到收藏夹

Article Social Sciences, Interdisciplinary

How the machine 'thinks': Understanding opacity in machine learning algorithms

Jenna Burrell

BIG DATA & SOCIETY (2016)

添加到收藏夹

Article Chemistry, Medicinal

A Machine Learning Approach to Explain Drug Selectivity to Soluble and Membrane Protein Targets

Eva Freyhult et al.

MOLECULAR INFORMATICS (2015)

添加到收藏夹

Article Computer Science, Information Systems

Determining molecular predictors of adverse drug reactions with causality analysis based on structure learning

Mei Liu et al.

JOURNAL OF THE AMERICAN MEDICAL INFORMATICS ASSOCIATION (2014)

添加到收藏夹

Article Chemistry, Medicinal

Integrative Chemical-Biological Read-Across Approach for Chemical Hazard Classification

Yen Low et al.

CHEMICAL RESEARCH IN TOXICOLOGY (2013)

添加到收藏夹

Article Chemistry, Medicinal

Beyond the Scope of Free-Wilson Analysis: Building Interpretable QSAR Models with Machine Learning Algorithms

Hongming Chen et al.

JOURNAL OF CHEMICAL INFORMATION AND MODELING (2013)

添加到收藏夹

Article Chemistry, Multidisciplinary

The rise of fragment-based drug discovery

Christopher W. Murray et al.

NATURE CHEMISTRY (2009)

添加到收藏夹

Article Chemistry, Multidisciplinary

Reoptimization of MDL keys for use in drug discovery

JL Durant et al.

JOURNAL OF CHEMICAL INFORMATION AND COMPUTER SCIENCES (2002)

添加到收藏夹

Article Ecology

Illuminating the black box: a randomization approach for understanding variable contributions in artificial neural networks

JD Olden et al.

ECOLOGICAL MODELLING (2002)

添加到收藏夹

© Peeref 2019-2024. All rights reserved.