Coarse ethics: how to ethically assess explainable artificial intelligence

Izumo, Takashi; Weng, Yueh-Hsuan

doi:10.1007/s43681-021-00091-y

Cited by 5 publications

(2 citation statements)

References 40 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…XAI should not reduce people's responsibilities or obligations to AI. XAI should not cause people to develop emotional attachment or dependence on AI [51,52]. Another limitation of this study is that it is only a computational theoretical study.…”

Section: Discussionmentioning

confidence: 99%

PROCAL: A SHap-based approach to explain Random Forest predictions of COVID-19- Procalcitonin levels

Naldan,

Kırboğa,

Işık

et al. 2024

Preprint

View full text Add to dashboard Cite

In COVID-19, the inflammatory cytokine storm is a critical factor that increases the severity of the disease. Procalcitonin (PCT) is a costly, time-consuming and important biomarker involved in the cytokine storm that exacerbates the severity of COVID-19. This study aims to develop an algorithm that can predict the PCT value in an explainable and interpretable way using explainable artificial intelligence (XAI) methods. The dataset consists of 1068 COVID-19 patients registered at Erzurum Regional Research Center in Turkey between March 2020 and March 2021 (ethical decision number: 2023/3–17). The Permutation Feature Significance (PFI) method was used to identify essential features and build the model. Among the seven-machine learning (ML) models, RandomForestClassifier performed best. RandomForestClassifier's performance metrics training accuracy: 0.89, test accuracy: 0.88, precision: 0.91, recall: 0.88, F-1 score: 0.88, Brier score: 0.11, AUC (area under the curve): 0.935, confidence intervals: 0.877, 0.883. The importance of the features in the model's predictions was analysed with the Shapley additive annotation (SHap) method integrated into the model. The results showed that LDH U/L, CRP mg/L and lymphocytes are important in predicting PCT. This study showed that PCT plays a vital role in assessing the condition of COVID-19 patients, and XAI methods can be helpful in this assessment.

show abstract

Section: Discussionmentioning

confidence: 99%

PROCAL: A SHap-based approach to explain Random Forest predictions of COVID-19- Procalcitonin levels

Naldan,

Kırboğa,

Işık

et al. 2024

Preprint

View full text Add to dashboard Cite

show abstract

“…Furthermore, utilizing eight distinct classifiers allowed us to test a variety of simple (but sophisticated enough to suit a relationship between input and output well) to complex models [73]. In addition, to effectively cover the ethics of XAI, we considered sufficiently high coverage and order preservation [74]. Using multiple sizes of feature sets and varied combinations of them allowed us to cover the most available features while also studying each feature set separately.…”

Section: Discussionmentioning

confidence: 99%

Prediction of Cognitive Decline in Parkinson’s Disease Using Clinical and DAT SPECT Imaging Features, and Hybrid Machine Learning Systems

et al. 2023

View full text Add to dashboard Cite

Background: We aimed to predict Montreal Cognitive Assessment (MoCA) scores in Parkinson’s disease patients at year 4 using handcrafted radiomics (RF), deep (DF), and clinical (CF) features at year 0 (baseline) applied to hybrid machine learning systems (HMLSs). Methods: 297 patients were selected from the Parkinson’s Progressive Marker Initiative (PPMI) database. The standardized SERA radiomics software and a 3D encoder were employed to extract RFs and DFs from single-photon emission computed tomography (DAT-SPECT) images, respectively. The patients with MoCA scores over 26 were indicated as normal; otherwise, scores under 26 were indicated as abnormal. Moreover, we applied different combinations of feature sets to HMLSs, including the Analysis of Variance (ANOVA) feature selection, which was linked with eight classifiers, including Multi-Layer Perceptron (MLP), K-Neighbors Classifier (KNN), Extra Trees Classifier (ETC), and others. We employed 80% of the patients to select the best model in a 5-fold cross-validation process, and the remaining 20% were employed for hold-out testing. Results: For the sole usage of RFs and DFs, ANOVA and MLP resulted in averaged accuracies of 59 ± 3% and 65 ± 4% for 5-fold cross-validation, respectively, with hold-out testing accuracies of 59 ± 1% and 56 ± 2%, respectively. For sole CFs, a higher performance of 77 ± 8% for 5-fold cross-validation and a hold-out testing performance of 82 + 2% were obtained from ANOVA and ETC. RF+DF obtained a performance of 64 ± 7%, with a hold-out testing performance of 59 ± 2% through ANOVA and XGBC. Usage of CF+RF, CF+DF, and RF+DF+CF enabled the highest averaged accuracies of 78 ± 7%, 78 ± 9%, and 76 ± 8% for 5-fold cross-validation, and hold-out testing accuracies of 81 ± 2%, 82 ± 2%, and 83 ± 4%, respectively. Conclusion: We demonstrated that CFs vitally contribute to predictive performance, and combining them with appropriate imaging features and HMLSs can result in the best prediction performance.

show abstract

Introduction to Coarse Ethics: Tradeoff Between the Accuracy and Interpretability of Explainable Artificial Intelligence

Izumo

2024

Studies in Neuroscience, Psychology and Behavioral Economics

View full text Add to dashboard Cite

Coarse ethics: how to ethically assess explainable artificial intelligence

Cited by 5 publications

References 40 publications

PROCAL: A SHap-based approach to explain Random Forest predictions of COVID-19- Procalcitonin levels

PROCAL: A SHap-based approach to explain Random Forest predictions of COVID-19- Procalcitonin levels

Prediction of Cognitive Decline in Parkinson’s Disease Using Clinical and DAT SPECT Imaging Features, and Hybrid Machine Learning Systems

Introduction to Coarse Ethics: Tradeoff Between the Accuracy and Interpretability of Explainable Artificial Intelligence

Contact Info

Product

Resources

About