A novel approach to explain the black-box nature of machine learning in compressive strength predictions of concrete using Shapley additive explanations (SHAP)

Ekanayake, I.U.; Meddage, D. P. P.; Rathnayake, Upaka

doi:10.1016/j.cscm.2022.e01059

Cited by 105 publications

(53 citation statements)

References 77 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Although ML has provided important advances in diagnosing autism, considerable challenges must be addressed. Many methods for classification lack interpretability, which is disadvantageous, especially for the understanding of medical data [28,29]. Also, according to Table I, [25,27], small data sets are quite common [30][31][32][33], which might cause unreliable results.…”

Section: Introductionmentioning

confidence: 99%

Diagnosis of autism spectrum disorder based on functional brain networks and machine learning

Alves

Toutain

Aguiar

et al. 2022

Preprint

View full text Add to dashboard Cite

Autism spectrum disorder is a multifactorial neurodevelopmental disorder with high genetic heterogeneity. Studies of brain networks in autism can provide new insights into the dynamics of information processing in individuals who suffer from such a condition. This paper proposes a method for automatic diagnosis of autism based on fMRI time series and machine learning algorithms. We verify that the left ventral posterior cingulate cortex region reduces the functional connectivity of the brain area in patients with autism spectrum disorder. Also, the brain networks of patients with autism spectrum disorder show more segregation, lower distribution of information, and less connectivity. Our methodology accurately differentiates control and autistic subjects providing an area under the curve close to higher than 95%.

show abstract

Section: Introductionmentioning

confidence: 99%

Diagnosis of autism spectrum disorder based on functional brain networks and machine learning

Alves

Toutain

Aguiar

et al. 2022

Preprint

View full text Add to dashboard Cite

show abstract

“…Many researchers have implemented this technique to interpret the black-box nature of the ML-based models. Ekanayake et al have implemented a decision tree, adaptive boost (AdaBoost), and extreme gradient boost (XGBoost) in the SHAP framework . Zaki et al utilized SHAP to study the optical properties of glass by implying the ML model .…”

Section: Introductionmentioning

confidence: 99%

“…Ekanayake et al have implemented a decision tree, adaptive boost (AdaBoost), and extreme gradient boost (XGBoost) in the SHAP framework. 38 Zaki et al utilized SHAP to study the optical properties of glass by implying the ML model. 39 Onsree et al investigated the effect of features on the accuracy of the ML model using SHAP.…”

Section: Introductionmentioning

confidence: 99%

Multiobjective Bayesian Optimization Framework for the Synthesis of Methanol from Syngas Using Interpretable Gaussian Process Models

et al. 2022

View full text Add to dashboard Cite

Methanol production has gained considerable interest on the laboratory and industrial scale as it is a renewable fuel and an excellent hydrogen energy storehouse. The formation of synthesis gas (CO/H 2 ) and the conversion of synthesis gas to methanol are the two basic catalytic processes used in methanol production. Machine learning (ML) approaches have recently emerged as powerful tools in reaction informatics. Inspired by these, we employ Gaussian process regression (GPR) to the model conversion of carbon monoxide (CO) and selectivity of the methanol product using data sets obtained from experimental investigations to capture uncertainty in prediction values. The results indicate that the proposed GPR model can accurately predict CO conversion and methanol selectivity as compared to other ML models. Further, the factors that influence the predictions are identified from the best GPR model employing "Shapley Additive exPlanations" (SHAP). After interpretation, the essential input features are found to be the inlet mole fraction of CO (Y(CO, in)) and the net inlet flow rate (Fin(nL/min)) for our best prediction GPR models, irrespective of our data sets. These interpretable models are employed for Bayesian optimization in a weighted multiobjective framework to obtain the optimal operating points, namely, maximization of both selectivity and conversion.

show abstract

“…As brain data are characterized by high complexity and highly correlated brain regions, ML algorithms have been widely used as a important tool capable of detecting acute and permanent abnormalities in the brain [61][62][63]. On the other hand, ML shows a lack of interpretability and a black box nature that is an especially disadvantageous general limitation when it comes to understanding medical data [64,65]. In the last years new techniques have emerged to help in the interpretation of machine learning results.…”

Section: Introductionmentioning

confidence: 99%

Application of machine learning and complex network measures to an EEG dataset from DMT experiments

Alves

Toutain

Porto

et al. 2022

Preprint

View full text Add to dashboard Cite

There is a growing interest in the medical use of psychedelic substances as preliminary studies using them for psychiatric disorders have shown positive results. In particularly, one of these substances is N,N-dimethyltryptamine (DMT) an agonist serotonergic psychedelic that can induce profound alterations in state of consciousness. In this work, we propose a computational method based on machine learning as an exploratory tool to reveal DMT-induced changes in brain activity using EEG data and provide new insights into the mechanisms of action of this psychedelic substance. To answer these questions, we propose a two-class classification based on (A) the connectivity matrix or (B) complex network measures derived from it as input to a support vector machine We found that both approaches were able to automatically detect changes in the brain activity, with case (B) showing the highest AUC (89%), indicating that complex network measurements best capture the brain changes that occur due to DMT use. In a second step, we ranked the features that contributed most to this result. For case (A) we found that differences in the high alpha, low beta, and delta frequency band were most important to distinguish between the state before and after DMT inhalation, which is consistent with results described in the literature. Further, the connection between the temporal (TP8) and central cortex (C3) and between the precentral gyrus (FC5) and the lateral occipital cortex (T8) contributed most to the classification result. The connection between regions TP8 and C3 has been found in the literature associated with finger movements that might have occurred during DMT consumption. However, the connection between cortical regions FC5 and P8 has not been found in the literature and is presumably related to emotional, visual, sensory, perceptual, and mystical experiences of the volunteers during DMT consumption. For case (B) closeness centrality was the most important complex network measure. Moreover, we found larger communities and a longer average path length with the use of DMT and the opposite in its absence indicating that the balance between functional segregation and integration was disrupted. This findings supports the idea that cortical brain activity becomes more entropic under psychedelics. Overall, a robust computational workflow has been developed here with an interpretability of how DMT (or other psychedelics) modify brain networks and insights into their mechanism of action. Finally, the same methodology applied here may be useful in interpreting EEG time series from patients who consumed other psychedelic drugs and can help obtain a detailed understanding of functional changes in the neural network of the brain as a result of drug administration.

show abstract

A novel approach to explain the black-box nature of machine learning in compressive strength predictions of concrete using Shapley additive explanations (SHAP)

Cited by 105 publications

References 77 publications

Diagnosis of autism spectrum disorder based on functional brain networks and machine learning

Diagnosis of autism spectrum disorder based on functional brain networks and machine learning

Multiobjective Bayesian Optimization Framework for the Synthesis of Methanol from Syngas Using Interpretable Gaussian Process Models

Application of machine learning and complex network measures to an EEG dataset from DMT experiments

Contact Info

Product

Resources

About