NSTL国家科技图书文献中心

1. From Explanation Correctness to Explanation Goodness: Only Provably Correct Explanations Can Save the World NSTL国家科技图书文献中心

Maike Schwammberger - 《Bridging the Gap Between AI and Reality》 - International Conference on Bridging the Gap between AI and Reality - 2025, - 307～317 - 共11页

摘要： stakeholder-tailored explanation* can be provided in a* | Explainability Engineering gets evermore | important in the era of self-learning and automated | systems. We motivate the necessity for interdisciplinary | research to engineer verifiably correct and good

关键词： Self-explainable software systems | Explanation correctness | Explanation goodness | EXplainable artificial intelligence | Trustworthy systems

2. Distribution-Aligned Sequential Counterfactual Explanation with Local Outlier Factor NSTL国家科技图书文献中心

Shoki Yamao | Ken Kobayashi ... - 《PRICAI 2024,Part I》 - Pacific Rim International Conference on Artificial Intelligence - 2025, - 243～256 - 共14页

摘要：Sequential counterfactual explanation* is one* | of the counter-factual explanation* methods* | counterfactual explanation* method that generates a realistic* | suggesting how to sequentially change the input feature | vector to obtain the desired prediction result from a

关键词： Counterfactual explanation | Local outlier factor | Explainable machine learning

3. Graph Neural Network Causal Explanation via Neural Causal Models NSTL国家科技图书文献中心

Arman Behnam | Binghui Wang - 《Computer Vision - ECCV 2024,Part LXI》 - European Conference on Computer Vision - 2025, - 410～427 - 共18页

摘要： GNN explainers in exact groundtruth explanation | Graph neural network (GNN) explainers identify | the important subgraph that ensures the prediction | for a given graph. Until now, almost all GNN | explainers are based on association, which is prone to

关键词： Graph neural network explanation | Neural causal model

4. On Spectral Properties of Gradient-Based Explanation Methods NSTL国家科技图书文献中心

Amir Mehrpanah | Erik Englesson ... - 《Computer Vision - ECCV 2024,Part LXXXVII》 - European Conference on Computer Vision - 2025, - 282～299 - 共18页

摘要： spectral perspectives to formally analyze explanation | choice of perturbation hyperparameters in explanation | Understanding the behavior of deep networks is | crucial to increase our confidence in their results | . Despite an extensive body of work for explaining their

关键词： Probabilistic machine learning | Gradient-based explanation methods | Probabilistic pixel attribution techniques | Explainability | Deep neural networks | Spectral analysis

5. Towards Reliable Drift Detection and Explanation in Text Data NSTL国家科技图书文献中心

Robert Feldhans | Barbara Hammer - 《Intelligent Data Engineering and Automated Learning - IDEAL 2024,Part I》 - International Conference on Intelligent Data Engineering and Automated Learning - 2025, - 301～312 - 共12页

摘要：When delivered to the market, machine learning | models face new data which are possibly subject to | novel characteristics - a phenomenon known as concept | drift. As this might lead to performance degradation | , it is necessary to detect such drift and, if

关键词： Drift explanation | Text data | Transformer | Visualization

6. Transparency vs Explanation of Machine Learning Algorithms: Perspectives from Recent Legal Proceedings NSTL国家科技图书文献中心

Uchenna Nnawuchi | Carlisle George ... - 《Progress in Artificial Intelligence,Part I》 - EPIA Conference on Artificial Intelligence - 2025, - 270～283 - 共14页

摘要： right to explanation* to address ML opacity, while* | and explanation* within the GDPR's regulatory* | , the lack of an explicit right to explanation* has led* | , explanation, and individual rights intersect in the realm of | The extensive use of Machine Learning (ML

关键词： Machine learning | Right to explanation | Transparency | Automated decision-making | GDPR

7. The Right to an Explanation Under the GDPR and the AI Act NSTL国家科技图书文献中心

Bjorn Aslak Juliusse... - 《MultiMedia Modeling,Part IV》 - International Conference on MultiMedia Modeling - 2025, - 184～197 - 共14页

摘要：, focusing on the right to explanation* for individual* | explanation in automated decision-making processes | The article provides a comprehensive overview | of European regulations, the GDPR and the AI Act | decisions inferred from high-risk AI systems and automated

关键词： The right to an explanation | EU law | XAI

8. Causal Explanation of Graph Neural Networks NSTL国家科技图书文献中心

Hichem Debbi - 《Intelligent Data Engineering and Automated Learning - IDEAL 2024,Part I》 - International Conference on Intelligent Data Engineering and Automated Learning - 2025, - 277～288 - 共12页

摘要： explanatory subgraphs is not sufficient explanation* tool* | this regard, we propose a causal explanation | with existing explanation* framework for GNNs, but* | Graph Neural Networks (GNNs) are currently | used in many real-world applications. With this

关键词： Graph neural networks (GNNs) | Causality | Explainability | Counterfactuals

9. Measuring Fairness in AI Explanations with LEADR: Local Explanation Amplification Disparity Ratio NSTL国家科技图书文献中心

Niloufar Shoeibi | Jonathan DeGange ... - 《Advances in Practical Applications of Agents,Multi-Agent Systems,and Digital Twins》 - International Conference on Practical Applications of Agents and Multi-Agent Systems - 2025, - 252～263 - 共12页

摘要：, the Local Explanation* Amplification Disparity Ratio* | We investigate the fairness of local | explanations in AI models by comparing the mean explanations | for privileged and unprivileged groups across | various datasets and model types. Specifically, we train

关键词： Fair XAI | Fair local AI explanation | Bias | Fairness | XAI | Bias amplification

10. Good Teachers Explain: Explanation-Enhanced Knowledge Distillation NSTL国家科技图书文献中心

Amin Parchami-Araghi | Moritz Bohle ... - 《Computer Vision - ECCV 2024,Part LXXIII》 - European Conference on Computer Vision - 2025, - 293～310 - 共18页

摘要： that our proposed 'explanation-enhanced' KD (e~2KD | Knowledge Distillation (KD) has proven | effective for compressing large teacher models into | smaller student models. While it is well known that | student models can achieve similar accuracies as the

摘要： stakeholder-tailored explanation* can be provided in a* | Explainability Engineering gets evermore | important in the era of self-learning and automated | systems. We motivate the necessity for interdisciplinary | research to engineer verifiably correct and good

关键词： Self-explainable software systems | Explanation correctness | Explanation goodness | EXplainable artificial intelligence | Trustworthy systems

摘要：Sequential counterfactual explanation* is one* | of the counter-factual explanation* methods* | counterfactual explanation* method that generates a realistic* | suggesting how to sequentially change the input feature | vector to obtain the desired prediction result from a

关键词： Counterfactual explanation | Local outlier factor | Explainable machine learning

摘要： GNN explainers in exact groundtruth explanation | Graph neural network (GNN) explainers identify | the important subgraph that ensures the prediction | for a given graph. Until now, almost all GNN | explainers are based on association, which is prone to

关键词： Graph neural network explanation | Neural causal model

摘要： spectral perspectives to formally analyze explanation | choice of perturbation hyperparameters in explanation | Understanding the behavior of deep networks is | crucial to increase our confidence in their results | . Despite an extensive body of work for explaining their

关键词： Probabilistic machine learning | Gradient-based explanation methods | Probabilistic pixel attribution techniques | Explainability | Deep neural networks | Spectral analysis

摘要：When delivered to the market, machine learning | models face new data which are possibly subject to | novel characteristics - a phenomenon known as concept | drift. As this might lead to performance degradation | , it is necessary to detect such drift and, if

关键词： Drift explanation | Text data | Transformer | Visualization

摘要： right to explanation* to address ML opacity, while* | and explanation* within the GDPR's regulatory* | , the lack of an explicit right to explanation* has led* | , explanation, and individual rights intersect in the realm of | The extensive use of Machine Learning (ML

关键词： Machine learning | Right to explanation | Transparency | Automated decision-making | GDPR

摘要：, focusing on the right to explanation* for individual* | explanation in automated decision-making processes | The article provides a comprehensive overview | of European regulations, the GDPR and the AI Act | decisions inferred from high-risk AI systems and automated

关键词： The right to an explanation | EU law | XAI

摘要： explanatory subgraphs is not sufficient explanation* tool* | this regard, we propose a causal explanation | with existing explanation* framework for GNNs, but* | Graph Neural Networks (GNNs) are currently | used in many real-world applications. With this

关键词： Graph neural networks (GNNs) | Causality | Explainability | Counterfactuals

摘要：, the Local Explanation* Amplification Disparity Ratio* | We investigate the fairness of local | explanations in AI models by comparing the mean explanations | for privileged and unprivileged groups across | various datasets and model types. Specifically, we train

关键词： Fair XAI | Fair local AI explanation | Bias | Fairness | XAI | Bias amplification

摘要： that our proposed 'explanation-enhanced' KD (e~2KD | Knowledge Distillation (KD) has proven | effective for compressing large teacher models into | smaller student models. While it is well known that | student models can achieve similar accuracies as the

关键词： Model compression | Faithful distillation | Interpretability

4008-161-200 800-990-8900

国家科技图书文献中心

摘要： stakeholder-tailored explanation can be provided in a | Explainability Engineering gets evermore | important in the era of self-learning and automated | systems. We motivate the necessity for interdisciplinary | research to engineer verifiably correct and good

关键词： Self-explainable software systems | Explanation correctness | Explanation goodness | EXplainable artificial intelligence | Trustworthy systems

摘要：Sequential counterfactual explanation is one | of the counter-factual explanation methods | counterfactual explanation method that generates a realistic | suggesting how to sequentially change the input feature | vector to obtain the desired prediction result from a

关键词： Counterfactual explanation | Local outlier factor | Explainable machine learning

摘要： GNN explainers in exact groundtruth explanation | Graph neural network (GNN) explainers identify | the important subgraph that ensures the prediction | for a given graph. Until now, almost all GNN | explainers are based on association, which is prone to

关键词： Graph neural network explanation | Neural causal model

摘要： spectral perspectives to formally analyze explanation | choice of perturbation hyperparameters in explanation | Understanding the behavior of deep networks is | crucial to increase our confidence in their results | . Despite an extensive body of work for explaining their

关键词： Probabilistic machine learning | Gradient-based explanation methods | Probabilistic pixel attribution techniques | Explainability | Deep neural networks | Spectral analysis

摘要：When delivered to the market, machine learning | models face new data which are possibly subject to | novel characteristics - a phenomenon known as concept | drift. As this might lead to performance degradation | , it is necessary to detect such drift and, if

关键词： Drift explanation | Text data | Transformer | Visualization

摘要： right to explanation to address ML opacity, while | and explanation within the GDPR's regulatory | , the lack of an explicit right to explanation has led | , explanation, and individual rights intersect in the realm of | The extensive use of Machine Learning (ML

关键词： Machine learning | Right to explanation | Transparency | Automated decision-making | GDPR

摘要：, focusing on the right to explanation for individual | explanation in automated decision-making processes | The article provides a comprehensive overview | of European regulations, the GDPR and the AI Act | decisions inferred from high-risk AI systems and automated

关键词： The right to an explanation | EU law | XAI

摘要： explanatory subgraphs is not sufficient explanation tool | this regard, we propose a causal explanation | with existing explanation framework for GNNs, but | Graph Neural Networks (GNNs) are currently | used in many real-world applications. With this

关键词： Graph neural networks (GNNs) | Causality | Explainability | Counterfactuals

摘要：, the Local Explanation Amplification Disparity Ratio | We investigate the fairness of local | explanations in AI models by comparing the mean explanations | for privileged and unprivileged groups across | various datasets and model types. Specifically, we train

关键词： Fair XAI | Fair local AI explanation | Bias | Fairness | XAI | Bias amplification

摘要： that our proposed 'explanation-enhanced' KD (e~2KD | Knowledge Distillation (KD) has proven | effective for compressing large teacher models into | smaller student models. While it is well known that | student models can achieve similar accuracies as the

关键词： Model compression | Faithful distillation | Interpretability

4008-161-200 800-990-8900

国家科技图书文献中心

摘要： stakeholder-tailored explanation* can be provided in a* | Explainability Engineering gets evermore | important in the era of self-learning and automated | systems. We motivate the necessity for interdisciplinary | research to engineer verifiably correct and good

摘要：Sequential counterfactual explanation* is one* | of the counter-factual explanation* methods* | counterfactual explanation* method that generates a realistic* | suggesting how to sequentially change the input feature | vector to obtain the desired prediction result from a

摘要： right to explanation* to address ML opacity, while* | and explanation* within the GDPR's regulatory* | , the lack of an explicit right to explanation* has led* | , explanation, and individual rights intersect in the realm of | The extensive use of Machine Learning (ML

摘要：, focusing on the right to explanation* for individual* | explanation in automated decision-making processes | The article provides a comprehensive overview | of European regulations, the GDPR and the AI Act | decisions inferred from high-risk AI systems and automated

摘要： explanatory subgraphs is not sufficient explanation* tool* | this regard, we propose a causal explanation | with existing explanation* framework for GNNs, but* | Graph Neural Networks (GNNs) are currently | used in many real-world applications. With this

摘要：, the Local Explanation* Amplification Disparity Ratio* | We investigate the fairness of local | explanations in AI models by comparing the mean explanations | for privileged and unprivileged groups across | various datasets and model types. Specifically, we train