Info
- Utilizzare la checkbox di selezione a fianco di ciascun documento per attivare le funzionalità di stampa, invio email, download nei formati disponibili del (i) record.
Explainable Artificial Intelligence : Second World Conference, XAI 2024, Valletta, Malta, July 17-19, 2024, Proceedings, Part I
| Explainable Artificial Intelligence : Second World Conference, XAI 2024, Valletta, Malta, July 17-19, 2024, Proceedings, Part I |
| Autore | Longo Luca |
| Edizione | [1st ed.] |
| Pubbl/distr/stampa | Cham : , : Springer International Publishing AG, , 2024 |
| Descrizione fisica | 1 online resource (508 pages) |
| Altri autori (Persone) |
LapuschkinSebastian
SeifertChristin |
| Collana | Communications in Computer and Information Science Series |
| ISBN | 3-031-63787-9 |
| Formato | Materiale a stampa  |
| Livello bibliografico | Monografia |
| Lingua di pubblicazione | eng |
| Nota di contenuto |
Intro -- Preface -- Organization -- Contents - Part I -- Intrinsically Interpretable XAI and Concept-Based Global Explainability -- Seeking Interpretability and Explainability in Binary Activated Neural Networks -- 1 Introduction -- 2 Related Works and Positioning -- 3 Notation -- 4 Dissecting Binary Activated Neural Networks -- 5 The BGN (Binary Greedy Network) Algorithm -- 5.1 Learning Shallow Networks -- 5.2 Properties of the BGN Algorithm -- 5.3 Improvements and Deeper Networks -- 6 SHAP Values for BANNs: Inputs, Neurons and Weights -- 7 Numerical Experiments -- 7.1 Predictive Accuracy Experiments -- 7.2 Pruning Experiments -- 7.3 Interpretability and Explainability of 1-BANNs -- 8 Conclusion -- References -- Prototype-Based Interpretable Breast Cancer Prediction Models: Analysis and Challenges -- 1 Introduction -- 2 Related Work -- 3 Preliminaries -- 4 Prototype Evaluation Framework -- 5 Experimental Setup -- 5.1 Datasets -- 5.2 Training Details -- 5.3 Prototype Evaluation Framework Setup -- 5.4 Visualization of Prototypes -- 6 Results and Discussion -- 6.1 Performance Comparison of Black-Box Vs Prototype-Based Models -- 6.2 Local and Global Visualization of Prototypes -- 6.3 Automatic Quantitative Evaluation of Prototypes -- 7 Conclusion and Future Work -- References -- Evaluating the Explainability of Attributes and Prototypes for a Medical Classification Model -- 1 Introduction -- 2 Related Work -- 3 Methods and Materials -- 3.1 Dataset -- 3.2 xAI-Model -- 4 Experiments -- 4.1 Aim of Study -- 4.2 Survey Setup -- 4.3 Participants -- 5 Results and Discussion -- 5.1 Radiologists' Attitude Towards AI. -- 5.2 RQ1: How Does the Explanation Affect the User's Performance? -- 5.3 RQ2: How Does the Explanation Affect the Trust in the Model? -- 5.4 RQ3: Are Attribute-Based Explanation (scores, Prototypes) Helpful? -- 5.5 Limitations -- 6 Conclusion.
References -- Revisiting FunnyBirds Evaluation Framework for Prototypical Parts Networks -- 1 Introduction -- 2 Related Works -- 3 Methods -- 3.1 FunnyBirds -- 3.2 Summed Similarity Maps (SSM) for More Precise Interface Functions -- 4 Experimental Setup -- 5 Results -- 5.1 Metrics Scores for Attribution Maps Based on Bounding Boxes or Similarity Maps -- 5.2 Various Backbones of ProtoPNet -- 6 Conclusions -- References -- CoProNN: Concept-Based Prototypical Nearest Neighbors for Explaining Vision Models -- 1 Introduction -- 2 Related Work -- 3 Methods -- 3.1 CoProNN -- 3.2 Prototype Images via Stable Diffusion -- 3.3 Nearest Neighbors as Explanations -- 3.4 Coarse and Fine Grained Classification Tasks -- 3.5 Evaluation Without Humans-in-the-Loop -- 3.6 Quantitative User Study -- 3.7 Qualitative User Study -- 4 Results -- 4.1 Explanations via Task-Specific Concept-Based Prototypes -- 4.2 Explanations via Task-Unspecific Concepts -- 4.3 Quantitative User Study -- 4.4 Results Qualitative User Study -- 5 Discussion -- 5.1 Improved Task Specificity of CoProNN Concepts -- 5.2 Interpretation User Study Results -- 5.3 Applying CoProNN to Your Own Use Case -- 5.4 Limitations and Extensions -- 6 Conclusion -- References -- Unveiling the Anatomy of Adversarial Attacks: Concept-Based XAI Dissection of CNNs -- 1 Introduction -- 2 Related Work -- 3 Background -- 3.1 Adversarial Attacks -- 3.2 Concept Discovery with Matrix Factorization -- 3.3 Concept Comparison -- 4 Experimental Setup -- 4.1 Models -- 4.2 Data -- 4.3 Layer Selection -- 5 Experimental Results -- 5.1 Adversarial Attacks Impact on Latent Space Representations -- 5.2 Concept Discovery in Adversarial Samples -- 5.3 Concept Analysis of Adversarial Perturbation -- 6 Conclusion and Outlook -- References -- AutoCL: AutoML for Concept Learning -- 1 Introduction -- 2 Background -- 3 Related Work -- 4 AutoCL. 4.1 Feature Selection -- 4.2 Hyperparameter Optimization -- 5 Evaluation -- 5.1 Experimental Setup -- 5.2 Feature Selection Results -- 5.3 Hyperparameter Optimization Results -- 5.4 AutoCL Results -- 6 Discussion -- 7 Conclusion -- References -- Locally Testing Model Detections for Semantic Global Concepts -- 1 Introduction -- 2 Related Work -- 2.1 Local Input Attribution -- 2.2 Global Concept Encodings -- 2.3 Combining Local and Global Approaches -- 3 Local Concept-Based Attributions -- 3.1 Global-to-Local Concept Attribution -- 3.2 Applicability in Object Detection -- 4 Quantification Measures -- 4.1 Concept Localization -- 4.2 Faithfulness Testing -- 5 Results -- 5.1 Experimental Setting -- 5.2 Local Concept Attribution -- 5.3 Evaluating Concept Usage -- 5.4 Localization Quantification -- 5.5 Faithfulness Evaluation -- 6 Testing for Erroneous Feature Correlation -- 7 Conclusion -- A Selection of Concepts -- B Implementation Details and Color Coding -- C Comparison to CRP -- D Criteria for the Qualitative Evaluation -- E Limitations -- F Additional Visualizations -- References -- Knowledge Graphs for Empirical Concept Retrieval -- 1 Introduction -- 2 Related Work -- 2.1 Concept-Based Explainability Methods -- 2.2 What Is a Concept? -- 2.3 Knowledge Graphs -- 3 Methods -- 3.1 Concepts from Knowledge Graphs -- 3.2 Retrieval of a Concept Database -- 3.3 Concept Activation Vectors and Regions -- 3.4 Machine Learning Models -- 3.5 Accuracy and Robustness -- 3.6 Alignment of Concepts and Sub-concepts -- 4 Results -- 4.1 Knowledge Graphs Can Assist the Definition of Data-Driven Concepts -- 4.2 Robustness of Knowledge Graph Derived Concepts -- 4.3 Alignment of Human and Machine Representations -- 5 Conclusion -- References -- Global Concept Explanations for Graphs by Contrastive Learning -- 1 Introduction -- 2 Related Work -- 3 Background. 3.1 MEGAN: Multi-Explanation Graph Attention Network -- 3.2 Definition of Graph Concept Explanations -- 4 Methods -- 4.1 Extended Network Architecture -- 4.2 Contrastive Explanation Learning -- 4.3 Concept Clustering -- 4.4 Prototype Optimization -- 4.5 Hypothesis Generation -- 5 Computational Experiments -- 5.1 Synthetic Datasets -- 5.2 Real-World Datasets -- 6 Limitations -- 7 Conclusion -- References -- Generative Explainable AI and Verifiability -- Augmenting XAI with LLMs: A Case Study in Banking Marketing Recommendation*-9pt -- 1 Introduction -- 1.1 Research Questions and Contribution -- 2 State of the Art -- 2.1 NLG for Explainability -- 3 Existing Environments -- 3.1 Human Roles and Responsibilities -- 3.2 Pipeline for Generating Commercial Recommendation -- 3.3 Pipeline for Generating Rule-Based Natural Language Explanations -- 4 Proposed Methods -- 4.1 Pipeline for Generating Automated Natural Language Explanations -- 4.2 Evaluation Methods -- 4.3 Statistical Analysis -- 5 Results -- 6 Discussion and Conclusion -- References -- Generative Inpainting for Shapley-Value-Based Anomaly Explanation -- 1 Introduction -- 2 Related Work -- 3 Methodology -- 3.1 Shapley Value Explanations and Replacement Values -- 3.2 Perturbation with Generative Inpainting -- 3.3 Tabular Diffusion with TabDDPM -- 3.4 Generative Inpainting for Diffusion Models -- 4 Experiments -- 4.1 Data, Anomaly Detectors, and Metrics -- 4.2 Generative Models and Inpainting -- 4.3 Results -- 5 Conclusion -- References -- Challenges and Opportunities in Text Generation Explainability -- 1 Introduction -- 2 Background and Related Work -- 2.1 Text Generation -- 2.2 Explainability Methods -- 2.3 Attribution-Based Methods -- 3 Dataset Creation -- 3.1 Human-Centered Explanations -- 3.2 Perturbation-Based Datasets -- 3.3 Tracing the Effect of Perturbations -- 4 Explanation Design. 4.1 Challenges Originating from the Language Model -- 4.2 Challenges Originating from the Text Data -- 5 Explanation Evaluation -- 5.1 Static Evaluation with Accuracy -- 5.2 Static Evaluation with Faithfulness -- 5.3 Contrastive Evaluation with Coherency -- 5.4 Characterization of Explainability Methods -- 6 Conclusion -- References -- NoNE Found: Explaining the Output of Sequence-to-Sequence Models When No Named Entity Is Recognized -- 1 Introduction -- 2 Related Work -- 2.1 Disaster Risk Management -- 2.2 Seq2seq NER Approach -- 2.3 NER Explanations -- 2.4 Seq2seq Explanations -- 3 Methods -- 4 Experimental Setup -- 4.1 Datasets -- 4.2 Experiments -- 5 Results -- 5.1 Model Results -- 5.2 Validation Results -- 5.3 Insights from NoNE Explanations -- 6 Conclusion and Future Work -- References -- Notion, Metrics, Evaluation and Benchmarking for XAI -- Benchmarking Trust: A Metric for Trustworthy Machine Learning -- 1 Introduction -- 2 Aspects of Trustworthy Machine Learning -- 2.1 Fairness -- 2.2 Robustness -- 2.3 Integrity -- 2.4 Explainability -- 2.5 Safety -- 3 Measures of Quantification and Operationalization -- 3.1 Quantifying the Notion of Trust -- 3.2 Experimental Design -- 4 Experimental Results -- 5 Conclusion and Outlook -- References -- Beyond the Veil of Similarity: Quantifying Semantic Continuity in Explainable AI -- 1 Introduction -- 2 Explainable AI -- 3 Related Work -- 4 Semantic Continuity -- 4.1 Proof-of-Concept Experiment -- 4.2 From Perfect Predictor to Imperfect Predictor -- 4.3 Synthesis of the Human Facial Dataset -- 5 Experimental Setup -- 5.1 Shape Dataset -- 5.2 Synthesis Facial Dataset -- 5.3 Software -- 6 Results -- 6.1 Proof-of-Concept Results: Shape Dataset -- 6.2 Synthesis Facial Dataset -- 7 Conclusions and Outlook -- References -- Conditional Calibrated Explanations: Finding a Path Between Bias and Uncertainty. 1 Introduction. |
| Record Nr. | UNINA-9910872185403321 |
Longo Luca
|
||
| Cham : , : Springer International Publishing AG, , 2024 | ||
| Materiale a stampa | ||
| Lo trovi qui: Univ. Federico II | ||
| ||
Explainable Artificial Intelligence : Second World Conference, XAI 2024, Valletta, Malta, July 17-19, 2024, Proceedings, Part II
| Explainable Artificial Intelligence : Second World Conference, XAI 2024, Valletta, Malta, July 17-19, 2024, Proceedings, Part II |
| Autore | Longo Luca |
| Edizione | [1st ed.] |
| Pubbl/distr/stampa | Cham : , : Springer International Publishing AG, , 2024 |
| Descrizione fisica | 1 online resource (0 pages) |
| Altri autori (Persone) |
LapuschkinSebastian
SeifertChristin |
| Collana | Communications in Computer and Information Science Series |
| ISBN | 3-031-63797-6 |
| Formato | Materiale a stampa |
| Livello bibliografico | Monografia |
| Lingua di pubblicazione | eng |
| Nota di contenuto |
Intro -- Preface -- Organization -- Contents - Part II -- XAI for Graphs and Computer Vision -- Model-Agnostic Knowledge Graph Embedding Explanations for Recommender Systems -- 1 Introduction -- 1.1 Problem Setting and Objective -- 1.2 Overview of Main Findings and Contributions -- 2 Related Work -- 3 Methodology -- 3.1 Baselines -- 3.2 Metrics -- 3.3 Multi-Attribute Utility Theory (MAUT) -- 4 Embedding Proposal Algorithm -- 5 Results -- 6 Future Work, Limitations, and Open Directions -- 7 Conclusion -- References -- Graph-Based Interface for Explanations by Examples in Recommender Systems: A User Study -- 1 Introduction -- 2 Related Work -- 3 Generating Graph-Based Explanations-by-Examples -- 3.1 Obtaining the Explanation Links -- 3.2 Visualising Explanations Through Interactive Graphs -- 4 User Study -- 4.1 Methodology -- 4.2 User Study Results -- 5 Conclusions -- References -- Explainable AI for Mixed Data Clustering -- 1 Introduction -- 2 Related Work -- 3 Entropy-Based Cluster Explanations for Mixed Data -- 4 Evaluation -- 5 Discussion -- 6 Conclusion -- References -- Explaining Graph Classifiers by Unsupervised Node Relevance Attribution -- 1 Introduction -- 2 Preliminaries -- 2.1 Deep Graph Networks -- 2.2 XAI Methods for Graphs -- 2.3 Benchmarking XAI Methods for Graphs -- 2.4 Assessing XAI Methods for Graphs in Real-World Contexts -- 3 Methods -- 3.1 Unsupervised Attribution Strategies for the Realistic Setting -- 3.2 A Measure of Quality for the Relevance Assignment -- 4 Experiments -- 4.1 Objective -- 4.2 Experimental Details -- 5 Results -- 6 Conclusion -- References -- Explaining Clustering of Ecological Momentary Assessment Data Through Temporal and Feature Attention -- 1 Introduction -- 2 Related Work -- 2.1 Clusters' Descriptive Representation -- 2.2 Explanations on TS Clustering -- 3 Review on Challenges of Explaining MTS Data.
3.1 Clustering Explanations -- 4 Framework for Clustering Explanations -- 4.1 EMA Data -- 4.2 EMA Clustering -- 4.3 Proposed Framework for Clustering Explanations -- 5 Analysis and Results -- 5.1 Performance Evaluation -- 5.2 Cluster-Level Explanations Through Temporal-Attention -- 5.3 Cluster-Level Explanations Through Feature-Attention -- 5.4 Individual-Level Explanations -- 6 Discussion -- 6.1 The Role of the Multi-aspect Attention -- 6.2 The Role of the Multi-level Analysis -- 6.3 The Impact of Utilizing a Pre-defined Clustering Result -- 7 Conclusion -- References -- Graph Edits for Counterfactual Explanations: A Comparative Study -- 1 Introduction -- 2 Related Work -- 3 Method -- 3.1 The Importance of Graph Machine Learning -- 4 Experiments -- 4.1 Experimental Setup -- 4.2 Quantitative Results -- 4.3 Qualitative Results -- 5 Conclusion -- References -- Model Guidance via Explanations Turns Image Classifiers into Segmentation Models -- 1 Introduction -- 1.1 Relation to Previous Works -- 1.2 Limitations -- 2 Unrolled Heatmap Architectures -- 2.1 LRP Basics -- 2.2 Unrolled LRP Architectures for Convolutional Classifiers -- 2.3 Losses and Training -- 2.4 Relation to Previous Formal Analyses and Standard Architectures -- 3 Unrolled Heatmap Architectures for Segmentation: Results -- 4 Conclusion -- References -- Understanding the Dependence of Perception Model Competency on Regions in an Image -- 1 Importance of Understanding Model Competency -- 2 Background and Related Work -- 2.1 Uncertainty Quantification -- 2.2 Out-of-Distribution (OOD) Detection -- 2.3 Explainable Image Classification -- 2.4 Explainable Competency Estimation -- 3 Approach for Understanding Model Competency -- 3.1 Estimating Model Competency -- 3.2 Identifying Regions Contributing to Low Competency -- 4 Method Evaluation and Analysis -- 4.1 Metrics for Comparison. 4.2 Dataset 1: Lunar Environment -- 4.3 Dataset 2: Speed Limit Signs -- 4.4 Dataset 3: Outdoor Park -- 4.5 Analysis of Results -- 5 Conclusions -- 6 Limitations and Future Work -- A Data Sources and Algorithm Parameters -- B Comparison to Class Activation Maps -- B.1 Dataset 1: Lunar Environment -- B.2 Dataset 2: Speed Limit Signs -- B.3 Dataset 3: Outdoor Park -- References -- A Guided Tour of Post-hoc XAI Techniques in Image Segmentation -- 1 Introduction -- 2 Categorization of XAI Techniques for Image Segmentation -- 3 Review of XAI Techniques for Image Segmentation -- 3.1 Local XAI -- 3.2 Evaluation of Local XAI Methods -- 3.3 A Comparative Analysis of Local XAI Methods -- 3.4 Global XAI -- 4 Tools for Practitioners -- 5 Discussion -- 6 Conclusion -- A Reviewed XAI Algorithms -- References -- Explainable Emotion Decoding for Human and Computer Vision -- 1 Introduction -- 2 Related Works -- 2.1 Explainable Computer Vision -- 2.2 Brain Decoding: Machine Learning on fMRI Data -- 2.3 Emotion Decoding for Human and Computer Vision -- 3 Experimental Setup -- 3.1 Frames, fMRI and Labels -- 3.2 Machine Learning on Movie Frames -- 3.3 Machine Learning on fMRI Data -- 3.4 XAI for Emotion Decoding -- 3.5 CNN-Humans Attentional Match: A Comparative Analysis -- 4 Experimental Results -- 4.1 Machine Learning on Movie Frames -- 4.2 Machine Learning on fMRI Data -- 4.3 Explainability for fMRI-Based Models -- 4.4 Comparative Analysis -- 5 Conclusions -- References -- Explainable Concept Mappings of MRI: Revealing the Mechanisms Underlying Deep Learning-Based Brain Disease Classification -- 1 Introduction -- 2 Methodology -- 2.1 Dataset -- 2.2 Preprocessing -- 2.3 Standard Classification Network -- 2.4 Relevance-Guided Classification Network -- 2.5 Training -- 2.6 Model Selection -- 2.7 Bootstrapping Analysis -- 2.8 Concepts Identification. 2.9 Relevance-Weighted Concept Map Representation -- 2.10 R2* and Relevance Region of Interest Analysis -- 3 Results -- 4 Discussion -- 5 Conclusion -- References -- Logic, Reasoning, and Rule-Based Explainable AI -- Template Decision Diagrams for Meta Control and Explainability -- 1 Introduction -- 2 Related Work -- 3 Foundations -- 4 Template Decision Diagrams -- 4.1 Hierarchical Decision Diagrams by Templates -- 4.2 Standard Template Boxes -- 5 Templates for Self-adaptive Systems and Meta Control -- 5.1 An Overview of the Case Study -- 5.2 Modeling the Case Study with Template DDs -- 6 Improving Control Strategy Explanations -- 6.1 Explainability Metrics for Template DDs -- 6.2 Decision Diagram Refactoring -- 6.3 Implementation and Evaluation -- 7 Conclusion -- References -- A Logic of Weighted Reasons for Explainable Inference in AI -- 1 Introduction -- 2 Weighted Default Justification Logic -- 2.1 Justification Logic Preliminaries -- 2.2 Weighted Default Justification Logic -- 2.3 Example -- 3 Strong Weighted Default Justification Logic -- 3.1 Preliminaries -- 3.2 Strong Weighted Default Justification Logic -- 3.3 Example -- 3.4 Justification Default Graphs -- 4 WDJL and WDJL+ as Explainable Neuro-Symbolic Architectures -- 5 Related Work -- 5.1 Numeric Inference Graphs -- 5.2 Numeric Argumentation Frameworks -- 6 Conclusions and Future Work -- References -- On Explaining and Reasoning About Optical Fiber Link Problems -- 1 Introduction -- 2 Literature Review -- 3 Dataset Overview -- 4 Explanations Pipeline Architecture -- 4.1 Data Aggregation -- 4.2 Data Cleansing and Normalisation -- 4.3 Data Transformation -- 4.4 ML Training -- 4.5 AI Explainability -- 5 Experimental Results -- 5.1 Model Performance -- 5.2 Model Explainability -- 6 Conclusion -- A Appendix A -- References. Construction of Artificial Most Representative Trees by Minimizing Tree-Based Distance Measures -- 1 Introduction -- 2 Methods -- 2.1 Random Forests -- 2.2 Variable Importance Measures (VIMs) -- 2.3 Selection of Most Representative Trees (MRTs) -- 2.4 Construction of Artificial Most Representative Trees (ARTs) -- 2.5 Simulation Design -- 2.6 Benchmarking Experiment -- 3 Results -- 3.1 Prediction Performance -- 3.2 Included Variables -- 3.3 Computation Time -- 3.4 Influence of Tuning Parameters -- 3.5 Tree Depth -- 3.6 Benchmarking -- 4 Discussion -- Appendix -- References -- Decision Predicate Graphs: Enhancing Interpretability in Tree Ensembles -- 1 Introduction -- 2 Literature Review -- 3 Decision Predicate Graphs -- 3.1 Definition -- 3.2 From Ensemble to a DPG -- 3.3 DPG Interpretability -- 4 Empirical Results and Discussion -- 4.1 DPG: Iris Insights -- 4.2 Comparing to the Graph-Based Solutions -- 4.3 Potential Improvements -- 5 Conclusion -- References -- Model-Agnostic and Statistical Methods for eXplainable AI -- Observation-Specific Explanations Through Scattered Data Approximation -- 1 Introduction -- 2 Methodology -- 2.1 Observation-Specific Explanations -- 2.2 Surrogate Models Using Scattered Data Approximation -- 2.3 Estimation of the Observation-Specific Explanations -- 3 Application -- 3.1 Simulated Studies -- 3.2 Real-World Application -- 4 Discussion -- References -- CNN-Based Explanation Ensembling for Dataset, Representation and Explanations Evaluation -- 1 Introduction -- 2 Related Works -- 3 The Concept of CNN-Based Ensembled Explanations -- 3.1 Experimental Setup for Training -- 3.2 Ablation Studies -- 3.3 Method Evaluation -- 4 Metrics for Representation, Dataset and Explanation Evaluation -- 4.1 Representation Evaluation -- 4.2 Dataset Evaluation -- 4.3 Explanations Evaluation -- 5 Conclusions and Future Works -- References. Local List-Wise Explanations of LambdaMART. |
| Record Nr. | UNINA-9910872194103321 |
|
Longo Luca
|
||
| Cham : , : Springer International Publishing AG, , 2024 | ||
| Materiale a stampa | ||
| Lo trovi qui: Univ. Federico II | ||
| ||
Explainable Artificial Intelligence : Second World Conference, XAI 2024, Valletta, Malta, July 17-19, 2024, Proceedings, Part III
| Explainable Artificial Intelligence : Second World Conference, XAI 2024, Valletta, Malta, July 17-19, 2024, Proceedings, Part III |
| Autore | Longo Luca |
| Edizione | [1st ed.] |
| Pubbl/distr/stampa | Cham : , : Springer International Publishing AG, , 2024 |
| Descrizione fisica | 1 online resource (471 pages) |
| Altri autori (Persone) |
LapuschkinSebastian
SeifertChristin |
| Collana | Communications in Computer and Information Science Series |
| ISBN | 3-031-63800-X |
| Formato | Materiale a stampa |
| Livello bibliografico | Monografia |
| Lingua di pubblicazione | eng |
| Nota di contenuto |
Intro -- Preface -- Organization -- Contents - Part III -- Counterfactual Explanations and Causality for eXplainable AI -- Sub-SpaCE: Subsequence-Based Sparse Counterfactual Explanations for Time Series Classification Problems -- 1 Introduction -- 2 Related Work -- 3 Proposed Method -- 3.1 Problem Formulation -- 3.2 Sub-SpaCE: Subsequence-Based Sparse Counterfactual Explanations -- 4 Numerical Experiments -- 4.1 Set Up -- 4.2 Results -- 4.3 Ablation Study -- 5 Conclusions and Future Work -- References -- Human-in-the-Loop Personalized Counterfactual Recourse -- 1 Introduction -- 2 Related Work -- 3 Problem Statement -- 4 Framework -- 4.1 Personalized Counterfactual Generation -- 4.2 Preference Modeling -- 4.3 Preference Estimation -- 4.4 HIP-CORE Framework -- 4.5 Complexity of User Feedback -- 4.6 Limitations -- 5 Experiments -- 5.1 Experimental Setting -- 5.2 Overall Performance -- 5.3 Model-Agnostic Validation -- 5.4 Study on the Number of Iterations -- 5.5 Study on the Number of Decimal Places -- 5.6 Discussion and Ethical Implications -- 6 Conclusions -- A Appendix -- References -- COIN: Counterfactual Inpainting for Weakly Supervised Semantic Segmentation for Medical Images -- 1 Introduction -- 2 Related Works -- 2.1 Weakly Supervised Semantic Segmentation -- 2.2 Counterfactual Explanations -- 3 Counterfactual Approach for WSSS -- 3.1 Method Formulation -- 3.2 Image Generation Architecture -- 3.3 Loss Function for Training GAN -- 4 Experiments -- 4.1 Datasets -- 4.2 Evaluation -- 4.3 Implementation Details -- 4.4 Comparison with Modified Singla et al.* Method -- 5 Results -- 5.1 Ablation Experiments -- 6 Discussion -- 7 Conclusion -- A.1 Loss Function for Dual-Conditioning in Singla et al.* -- A.2 Synthetic Anomaly Generation -- A.3 Original vs Perturbation-Based Generator.
A.4 Influence of Skip Connections on the Generated Images Quality -- A.5 Counterfactual Explanation vs Counterfactual Inpainting Segmentation Accuracy -- References -- Enhancing Counterfactual Explanation Search with Diffusion Distance and Directional Coherence -- 1 Introduction -- 2 Related Work -- 3 Incorporating Novel Biases in Counterfactual Search -- 3.1 Using Diffusion Distance to Search for More Feasible Transitions -- 3.2 Directional Coherence -- 3.3 Bringing Feasibility and Directional Coherence into Counterfactual Objective Function -- 3.4 Evaluation Metrics -- 4 Experiments -- 4.1 Synthetic Datasets -- 4.2 Datasets with Continuous Features -- 4.3 Classification Datasets with Mix-Type Features -- 4.4 Benchmarking with Other Frameworks -- 5 Results -- 5.1 Diffusion Distance and Directional Coherence on Synthetic and Diabetes Datasets -- 5.2 Comparison of CoDiCE with Other Counterfactual Methods on Various Datasets -- 5.3 Ablation Experiments -- 6 Discussion -- 7 Conclusion -- A Appendix -- References -- CountARFactuals - Generating Plausible Model-Agnostic Counterfactual Explanations with Adversarial Random Forests -- 1 Introduction -- 2 Related Work -- 3 Background -- 3.1 Multi-objective Counterfactual Explanations -- 3.2 Generative Modeling and Adversarial Random Forests -- 4 Methods -- 4.1 Algorithm [alg::mocspsarf]1: Integrating ARF into MOC -- 4.2 Algorithm [alg::onlyspsarf]2: ARF Is All You Need -- 5 Experiments -- 5.1 Data-Generating Process -- 5.2 Competing Methods -- 5.3 Evaluation Criteria -- 5.4 Results -- 6 Real Data Example -- 7 Discussion -- A Algorithm [alg::mocspsarf]1: Integrating ARF into MOC -- B Algorithm [alg::onlyspsarf]2: ARF Is All You Need -- C Synthetic Data -- C.1 Illustrative Datasets -- C.2 Randomly Generated DGPs -- D Additional Empirical Results -- References. Causality-Aware Local Interpretable Model-Agnostic Explanations -- 1 Introduction -- 2 Related Works -- 3 Background -- 4 Causality-Aware LIME -- 5 Experiments -- 5.1 Datasets and Classifiers -- 5.2 Comparison with Related Works -- 5.3 Evaluation Measures -- 5.4 Results -- 6 Conclusion -- References -- Evaluating the Faithfulness of Causality in Saliency-Based Explanations of Deep Learning Models for Temporal Colour Constancy -- 1 Introduction -- 2 Related Work -- 3 Proposed Neural Architectures -- 4 Original Methodology of the Tests -- 4.1 Adaptation of the Original Tests -- 5 Experimental Setup -- 6 Method -- 7 Results -- 7.1 Preliminary Accuracy Investigation -- 7.2 Test WP1 -- 7.3 Test WP2 -- 7.4 Discussion -- 8 Conclusions and Future Work -- A Extended Results of the Experiments -- References -- CAGE: Causality-Aware Shapley Value for Global Explanations -- 1 Introduction -- 2 Preliminaries and Notation -- 2.1 Causal Models and Interventions -- 2.2 Shapley Additive Global Importance -- 3 Causality-Aware Global Explanations -- 3.1 Global Causal Shapley Values -- 3.2 Properties of Global Causal Feature Importance -- 3.3 Computing Causal Shapley Values -- 4 Experiments -- 4.1 Experiments on Synthetic Data -- 4.2 Explanations on Alzheimer Data -- 5 Related Work -- 6 Discussion -- 7 Conclusion -- A Data - Generating Causal Models -- A.1 Direct-Cause structure -- A.2 Markovian Structure -- A.3 Mixed structure -- References -- Fairness, Trust, Privacy, Security, Accountability and Actionability in eXplainable AI -- Exploring the Reliability of SHAP Values in Reinforcement Learning -- 1 Introduction -- 1.1 Shapley Values -- 1.2 Shapley Values for ML - SHAP -- 1.3 Contributions -- 2 Related Work -- 3 Benchmark Environments -- 4 Experiment 1: Dependency of KernelSHAP on Background Data -- 4.1 KernelSHAP and Background Data -- 4.2 Experimental Setup. 4.3 Robustness of KernelSHAP -- 5 Experiment 2: Empirical Evaluation of SHAP-Based Feature Importance -- 5.1 Generalized Feature Importance -- 5.2 Experimental Setup -- 5.3 Performance Drop Vs. Feature Importance -- 6 Interpretation of SHAP Time Dependency in RL -- 7 Conclusion and Outlook -- References -- Categorical Foundation of Explainable AI: A Unifying Theory -- 1 Introduction -- 2 Explainable AI Theory: Requirements -- 2.1 Category Theory: A Framework for (X)AI Processes -- 2.2 Institution Theory: A Framework for Explanations -- 3 Categorical Framework of Explainable AI -- 3.1 Abstract Learning Processes -- 3.2 Concrete Learning and Explaining Processes -- 4 Impact on XAI and Key Findings -- 4.1 Finding #1: Our Framework Models Existing Learning Schemes and Architectures -- 4.2 Finding #2: Our Framework Enables a Formal Definition of ``explanation'' -- 4.3 Finding #3: Our Framework Provides a Theoretical Foundation for XAI Taxonomies -- 4.4 Finding #4: Our Framework Emphasizes Commonly Overlooked Aspects of Explanations -- 5 Discussion -- A Elements of Category Theory -- A.1 Monoidal Categories -- A.2 Cartesian and Symmetric Monoidal Categories -- A.3 Feedback Monoidal Categories -- A.4 Free Categories -- A.5 Institutions -- References -- Investigating Calibrated Classification Scores Through the Lens of Interpretability -- 1 Introduction -- 2 Formal Setup -- 3 Desiderata for Calibration -- 3.1 Interplay of Strict Properties -- 4 Relaxed Desiderata for Calibration -- 4.1 Analysis of Cell Merging -- 4.2 Analysis of Average Label Assignment -- 5 Experimental Evaluation of Decision Tree Based Models -- 6 Concluding Discussion -- A Exploring the Probabilistic Count (PC) -- B Critiquing the Expected Calibration Error -- C Empirically Motivating the Probability Deviation Error -- References. XentricAI: A Gesture Sensing Calibration Approach Through Explainable and User-Centric AI -- 1 Introduction -- 2 Background and Related Work -- 2.1 User-Centric XAI Techniques -- 2.2 Gesture Sensing Model Calibration Using Experience Replay -- 3 XAI for User-Centric and Customized Gesture Sensing -- 3.1 Gesture Sensing Algorithm and Feature Design -- 3.2 Model Calibration Using Experience Replay -- 3.3 Anomalous Gesture Detection and Characterization -- 4 Experiments -- 4.1 Implementation Settings -- 4.2 Experimental Results -- 5 Conclusion -- Appendix -- References -- Toward Understanding the Disagreement Problem in Neural Network Feature Attribution -- 1 Introduction -- 2 Background and Related Work -- 3 Understanding the Explanation's Distribution -- 4 Do Feature Attribution Methods Attribute? -- 4.1 Impact of Data Preprocessing -- 4.2 Faithfulness of Effects -- 4.3 Beyond Feature Attribution Toward Importance -- 5 Discussion -- 6 Conclusion -- A Appendix -- A.1 COMPAS Dataset -- A.2 Simulation Details -- A.3 Model Performance -- References -- ConformaSight: Conformal Prediction-Based Global and Model-Agnostic Explainability Framework -- 1 Introduction -- 2 Preliminaries -- 2.1 Explainability -- 2.2 Uncertainty Estimation and Quantification -- 2.3 Conformal Prediction -- 3 Related Work -- 4 Methodology -- 4.1 ConformaSight Structure and Mechanism -- 4.2 ConformaSight in Practice: A Sample Scenario -- 4.3 Computational Complexity of the ConformaSight -- 5 Experiments and Evaluations -- 5.1 Experimental Settings -- 6 Results and Discussion -- 7 Conclusion, Limitations and Future Work -- References -- Differential Privacy for Anomaly Detection: Analyzing the Trade-Off Between Privacy and Explainability -- 1 Introduction -- 2 Related Work -- 2.1 Privacy-Preserving Anomaly Detection -- 2.2 Explainable Anomaly Detection. 2.3 Impact of Privacy on Explainability. |
| Record Nr. | UNINA-9910872195903321 |
|
Longo Luca
|
||
| Cham : , : Springer International Publishing AG, , 2024 | ||
| Materiale a stampa | ||
| Lo trovi qui: Univ. Federico II | ||
| ||
Explainable Artificial Intelligence : Second World Conference, XAI 2024, Valletta, Malta, July 17-19, 2024, Proceedings, Part IV
| Explainable Artificial Intelligence : Second World Conference, XAI 2024, Valletta, Malta, July 17-19, 2024, Proceedings, Part IV |
| Autore | Longo Luca |
| Edizione | [1st ed.] |
| Pubbl/distr/stampa | Cham : , : Springer International Publishing AG, , 2024 |
| Descrizione fisica | 1 online resource (480 pages) |
| Altri autori (Persone) |
LapuschkinSebastian
SeifertChristin |
| Collana | Communications in Computer and Information Science Series |
| ISBN | 3-031-63803-4 |
| Formato | Materiale a stampa |
| Livello bibliografico | Monografia |
| Lingua di pubblicazione | eng |
| Nota di contenuto |
Intro -- Preface -- Organization -- Contents - Part IV -- Explainable AI in Healthcare and Computational Neuroscience -- SRFAMap: A Method for Mapping Integrated Gradients of a CNN Trained with Statistical Radiomic Features to Medical Image Saliency Maps -- 1 Introduction -- 2 Related Work -- 3 Design and Methodology -- 3.1 The Approach -- 3.2 The Experiment -- 3.3 Evaluation of Saliency Maps -- 4 Results and Discussion -- 4.1 Discussion of Results -- 5 Conclusions and Future Work -- References -- Transparently Predicting Therapy Compliance of Young Adults Following Ischemic Stroke -- 1 Introduction -- 2 Related Studies -- 3 Materials and Methods -- 3.1 Participants and Clinical Settings -- 3.2 Cognitive Assessment -- 3.3 Computer-Based Rehabilitation Therapy -- 3.4 Modelling -- 3.5 Explanation Methods -- 4 Results -- 4.1 Experimental Data -- 4.2 Therapy Compliance Prediction -- 4.3 Explanation Reports -- 5 Discussion -- 6 Conclusions -- References -- Precision Medicine for Student Health: Insights from Tsetlin Machines into Chronic Pain and Psychological Distress -- 1 Introduction -- 2 Tsetlin Machines -- 3 Related Work -- 3.1 Pain and Psychological Distress -- 3.2 Explainable AI -- 4 Materials and Methods -- 4.1 The SHoT2018 Study -- 4.2 Models and Analyses -- 5 Results and Discussion -- 5.1 Performance -- 5.2 Interpretability Analysis -- 6 Conclusions and Future Work -- A Literal Frequency in the Tsetlin Machine -- References -- Evaluating Local Explainable AI Techniques for the Classification of Chest X-Ray Images -- 1 Introduction -- 2 Previous Work -- 2.1 Explainable AI for X-Ray Imaging -- 2.2 Evaluation Metrics for XAI -- 3 Explainable AI Techniques -- 4 Analyzed Dataset -- 5 Proposed Metrics -- 6 Results and Evaluation -- 7 Conclusions -- References -- Feature Importance to Explain Multimodal Prediction Models. a Clinical Use Case.
1 Introduction -- 2 Related Work -- 2.1 Short-Term Complication Prediction -- 2.2 Multimodal Prediction Models -- 2.3 Explainability -- 3 Materials and Methods -- 3.1 Dataset -- 3.2 Machine Learning Models -- 3.3 Training Procedure and Evaluation -- 3.4 Explanation -- 4 Results -- 4.1 Model Performance -- 4.2 Explainability -- 5 Discussion -- 6 Conclusion -- A Hyperparameters -- B Detailed Feature Overview -- References -- Identifying EEG Biomarkers of Depression with Novel Explainable Deep Learning Architectures -- 1 Introduction -- 2 Methods -- 2.1 Description of Dataset -- 2.2 Description of Model Development -- 2.3 Description of Explainability Analyses Applied to All Models -- 2.4 Description of Approach for Characterization of M2 and M3 Filters -- 2.5 Description of Novel Activation Explainability Analyses for M2 and M3 -- 2.6 Key Aspects of Approach -- 3 Results and Discussion -- 3.1 M1-M3: Model Performance Analysis -- 3.2 M1-M3: Post Hoc Explainability Analysis -- 3.3 M2-M3: Characterization of Extracted Features -- 3.4 M2-M3: Post Hoc Spatial Activation Analysis -- 3.5 M2-M3: Post Hoc of Activation Correlation Analysis -- 3.6 Summary of MDD-Related Findings -- 3.7 Limitations and Future Work -- 4 Conclusion -- References -- Increasing Explainability in Time Series Classification by Functional Decomposition -- 1 Introduction -- 2 Background and Related Work -- 3 Method -- 4 Case Study -- 4.1 Sensor Model -- 4.2 Simulator -- 5 Application -- 5.1 Instantiation of the Generic Methodology -- 5.2 Influence of Data Representation and Decompositions -- 5.3 Influence of the Chunking -- 5.4 Datasets -- 6 Realization and Evaluation -- 6.1 Training and Testing of the Chunk Classifier -- 6.2 Training and Testing of the Velocity Estimator -- 6.3 Robustness Analysis of the Chunk Classifier -- 6.4 Testing of the Complete System -- 7 Explanations. 7.1 Dataset-Based Explanations -- 7.2 Visual Explanations -- 8 Conclusion and Future Work -- References -- Towards Evaluation of Explainable Artificial Intelligence in Streaming Data -- 1 Introduction -- 2 Related Work -- 2.1 Consistency, Fidelity and Stability of Explanations -- 3 Methodology -- 4 A Case Study with the Iris Dataset -- 5 Results Analysis -- 5.1 XAI Metric: Agreement (Consistency) Between Explainers -- 5.2 XAI Metric: Lipschitz and Average Stability -- 5.3 Comparison of Stability Metrics -- 5.4 Detailed Stability Comparison for Anomalies A1 and A2 -- 5.5 Quantification of Differences in Stability Between Ground Truth and Black-Box Explainers -- 6 Conclusion -- 7 Future Work -- References -- Quantitative Evaluation of xAI Methods for Multivariate Time Series - A Case Study for a CNN-Based MI Detection Model -- 1 Introduction -- 2 Background and State of the Art -- 2.1 Multivariate Time Series -- 2.2 MI Detection Use Case -- 2.3 Related Work -- 3 Methodology -- 3.1 Explanations for Time Series Data -- 3.2 Truthfulness Analysis -- 3.3 Stability Analysis -- 3.4 Consistency Analysis -- 4 Experimental Results -- 4.1 Results of the Truthfulness Analysis -- 4.2 Results of the Stability Analysis -- 4.3 Results of the Consistency Analysis -- 5 Discussion -- 6 Conclusion -- References -- Explainable AI for Improved Human-Computer Interaction and Software Engineering for Explainability -- Influenciæ: A Library for Tracing the Influence Back to the Data-Points -- 1 Introduction -- 2 Attributing Model Behavior Through Data Influence -- 2.1 Notation -- 2.2 Influence Functions -- 2.3 Kernel-Based Influence -- 2.4 Tracing Influence Throughout the Training Process -- 3 API -- 4 Conclusion -- References -- Explainability Engineering Challenges: Connecting Explainability Levels to Run-Time Explainability -- 1 Introduction -- 2 Explainability Terminology. 3 MAB-EX Framework for Self-Explainable Systems -- 4 Explainability Requirements in Software-Intensive Systems -- 5 Integration of Explainability Levels into the MAB-EX Framework -- 6 The Role of eXplainable DRL in Explainability Engineering -- 7 Conclusion -- References -- On the Explainability of Financial Robo-Advice Systems -- 1 Introduction -- 2 Related Work -- 3 Background -- 3.1 Financial Robo-Advice Systems -- 3.2 XAI and the Law -- 3.3 EU Regulations Relevant to Financial Robo-Advice Systems: Scopes and Notions -- 4 Proposed Methodology -- 5 Robo-Advice Systems -- 6 Legal Compliance Questions for Robo-Advice Systems -- 7 Case Studies -- 7.1 Requested Financial Information -- 7.2 Personas -- 7.3 Results: Robo-Generated Financial Advice -- 8 Threats to Validity -- 9 Discussion -- 10 Conclusion and Future Work -- References -- Can I Trust My Anomaly Detection System? A Case Study Based on Explainable AI -- 1 Introduction -- 2 Literature Review -- 3 Preliminaries -- 3.1 VAE-GAN Models -- 3.2 Semi-supervised Anomaly Detection Using Variational Models -- 3.3 Explaining Anomaly Maps Using Model-Agnostic XAI Methods -- 3.4 Comparing Explained Anomalies Against a Ground Truth -- 4 Experimental Evaluation -- 5 Conclusions -- References -- Explanations Considered Harmful: The Impact of Misleading Explanations on Accuracy in Hybrid Human-AI Decision Making -- 1 Introduction -- 2 Background and Related Work -- 3 How Explanations Can Be Misleading -- 4 Methods -- 5 Results -- 5.1 Impact on Accuracy -- 5.2 Impact on Confidence -- 6 Discussion -- 7 Conclusion -- References -- Human Emotions in AI Explanations -- 1 Introduction -- 2 Related Literature -- 3 Method -- 4 Results -- 5 Robustness Check -- 6 Discussion -- 7 Conclusion -- References -- Study on the Helpfulness of Explainable Artificial Intelligence -- 1 Introduction -- 2 Measuring Explainability. 2.1 Approaches for Measuring Explainability -- 2.2 User Studies on the Performance of XAI -- 3 An Objective Methodology for Evaluating XAI -- 3.1 Objective Human-Centered XAI Evaluation -- 3.2 Image Classification and XAI Methods -- 3.3 Survey Design -- 4 Survey Results -- 4.1 Questionnaire Responses -- 4.2 Qualitative Feedback -- 5 Discussion -- 6 Conclusion -- Appendix A Additional Visualizations -- Appendix B Demographic Overview of Participants -- References -- Applications of Explainable Artificial Intelligence -- Pricing Risk: An XAI Analysis of Irish Car Insurance Premiums -- 1 Introduction -- 2 Background and Related Work -- 3 Data and Methods -- 4 Results -- 5 Discussion and Conclusion -- References -- Exploring the Role of Explainable AI in the Development and Qualification of Aircraft Quality Assurance Processes: A Case Study -- 1 Introduction -- 1.1 Background -- 1.2 Related Work -- 2 Description of Use Case -- 3 XAI Methods Applied to Use Case -- 4 Insights in the xAI Results -- 4.1 Experiment Results -- 4.2 xAI on New Model -- 5 Discussing xAI w.r.t. Development and Qualifiability -- 6 Conclusion -- References -- Explainable Artificial Intelligence Applied to Predictive Maintenance: Comparison of Post-Hoc Explainability Techniques -- 1 Introduction -- 2 Proposed Methodology -- 3 Post-Hoc Explainability Techniques -- 3.1 Impurity-Based Feature Importance -- 3.2 Permutation Feature Importance -- 3.3 Partial Dependence Plot (PDP) -- 3.4 Accumulated Local Effects (ALE) -- 3.5 Shapley Additive Explanations (SHAP) -- 3.6 Local Interpretable Model-Agnostic Explanations (LIME) -- 3.7 Anchor -- 3.8 Individual Conditional Expectation (ICE) -- 3.9 Discussion on Implementation and Usability of the Techniques -- 4 Conclusions -- References. A Comparative Analysis of SHAP, LIME, ANCHORS, and DICE for Interpreting a Dense Neural Network in Credit Card Fraud Detection. |
| Record Nr. | UNINA-9910872189603321 |
|
Longo Luca
|
||
| Cham : , : Springer International Publishing AG, , 2024 | ||
| Materiale a stampa | ||
| Lo trovi qui: Univ. Federico II | ||
| ||