Accurately predicting blood glucose (BG) levels of ICU patients is critical, as both hypoglycemia (BG < 70 mg/dL) and hyperglycemia (BG > 180 mg/dL) are associated with increased morbidity and mortality. We develop the Multi-source Irregular Time-Series Transformer (MITST), a novel machine learning-based model to forecast the next BG level, classifying it into hypoglycemia, hyperglycemia, or euglycemia (70-180 mg/dL). The irregularity and complexity of Electronic Health Record (EHR) data, spanning multiple heterogeneous clinical sources like lab results, medications, and vital signs, pose significant challenges for prediction tasks. MITST addresses these using hierarchical Transformer architectures, which include a feature-level, a timestamp-level, and a source-level Transformer. This design captures fine-grained temporal dynamics and allows learning-based data integration instead of traditional predefined aggregation. In a large-scale evaluation using the eICU database (200,859 ICU stays across 208 hospitals), MITST achieves an average improvement of 1.7% (p < 0.001) in AUROC and 1.8% (p < 0.001) in AUPRC over a state-of-the-art baseline. For hypoglycemia, MITST achieves an AUROC of 0.915 and an AUPRC of 0.247, both significantly higher than the baseline's AUROC of 0.862 and AUPRC of 0.208 (p < 0.001). The flexible architecture of MITST allows seamless integration of new data sources without retraining the entire model, enhancing its adaptability in clinical decision support. Although this study focuses on predicting BG levels, MITST can easily be extended to other critical event prediction tasks in ICU settings, offering a robust solution for analyzing complex, multi-source, irregular time-series data.

Generalizing to out-of-distribution data while being aware of model fairness is a significant and challenging problem in meta-learning. The goal of this problem is to find a set of fairness-aware invariant parameters of classifier that is trained using data drawn from a family of related training domains with distribution shift on non-sensitive features as well as different levels of dependence between model predictions and sensitive features so that the classifier can achieve good generalization performance on unknown but distinct test domains. To tackle this challenge, existing state-of-the-art methods either address the domain generalization problem but completely ignore learning with fairness or solely specify shifted domains with various fairness levels. This paper introduces an approach to fairness-aware meta-learning that significantly enhances domain generalization capabilities. Our framework, Fairness-Enhanced Meta-Learning for Domain Generalization (FEED), disentangles latent data representations into content, style, and sensitive vectors. This disentanglement facilitates the robust generalization of machine learning models across diverse domains while adhering to fairness constraints. Unlike traditional methods that focus primarily on domain invariance or sensitivity to shifts, our model integrates a fairness-aware invariance criterion directly into the meta-learning process. This integration ensures that the learned parameters uphold fairness consistently, even when domain characteristics vary widely. We validate our approach through extensive experiments across multiple benchmarks, demonstrating not only superior performance in maintaining high accuracy and fairness but also significant improvements over existing state-of-the-art methods in domain generalization tasks.

Image Quality of MRI brain scans is strongly influenced by within scanner head movements and the resulting image artifacts alter derived measures like brain volume and cortical thickness. Automated image quality assessment is key to controlling for confounding effects of poor image quality. In this study, we systematically test for the influence of image quality on univariate statistics and machine learning classification. We analyzed group effects of sex/gender on local brain volume and made predictions of sex/gender using logistic regression, while correcting for brain size. From three large publicly available datasets, two age and sex-balanced samples were derived to test the generalizability of the effect for pooled sample sizes of n=760 and n=1094. Results of the Bonferroni corrected t-tests over 3747 gray matter features showed a strong influence of low-quality data on the ability to find significant sex/gender differences for the smaller sample. Increasing sample size and more so image quality showed a stark increase in detecting significant effects in univariate group comparisons. For the classification of sex/gender using logistic regression, both increasing sample size and image quality had a marginal effect on the Area under the Receiver Operating Characteristic Curve for most datasets and subsamples. Our results suggest a more stringent quality control for univariate approaches than for multivariate classification with a leaning towards higher quality for classical group statistics and bigger sample sizes for machine learning applications in neuroimaging.

In today's data-intensive landscape, where high-dimensional datasets are increasingly common, reducing the number of input features is essential to prevent overfitting and improve model accuracy. Despite numerous efforts to tackle dimensionality reduction, most approaches apply a universal set of features across all classes, potentially missing the unique characteristics of individual classes. This paper presents the Explainable Class-Specific Naive Bayes (XNB) classifier, which introduces two critical innovations: 1) the use of Kernel Density Estimation to calculate posterior probabilities, allowing for a more accurate and flexible estimation process, and 2) the selection of class-specific feature subsets, ensuring that only the most relevant variables for each class are utilized. Extensive empirical analysis on high-dimensional genomic datasets shows that XNB matches the classification performance of traditional Naive Bayes while drastically improving model interpretability. By isolating the most relevant features for each class, XNB not only reduces the feature set to a minimal, distinct subset for each class but also provides deeper insights into how the model makes predictions. This approach offers significant advantages in fields where both precision and explainability are critical.

Real-world machine learning applications often face simultaneous covariate and semantic shifts, challenging traditional domain generalization and out-of-distribution (OOD) detection methods. We introduce Meta-learned Across Domain Out-of-distribution Detection (MADOD), a novel framework designed to address both shifts concurrently. MADOD leverages meta-learning and G-invariance to enhance model generalizability and OOD detection in unseen domains. Our key innovation lies in task construction: we randomly designate in-distribution classes as pseudo-OODs within each meta-learning task, simulating OOD scenarios using existing data. This approach, combined with energy-based regularization, enables the learning of robust, domain-invariant features while calibrating decision boundaries for effective OOD detection. Operating in a test domain-agnostic setting, MADOD eliminates the need for adaptation during inference, making it suitable for scenarios where test data is unavailable. Extensive experiments on real-world and synthetic datasets demonstrate MADOD's superior performance in semantic OOD detection across unseen domains, achieving an AUPR improvement of 8.48% to 20.81%, while maintaining competitive in-distribution classification accuracy, representing a significant advancement in handling both covariate and semantic shifts.

Recent advances in AI models have increased the integration of AI-based decision aids into the human decision making process. To fully unlock the potential of AI-assisted decision making, researchers have computationally modeled how humans incorporate AI recommendations into their final decisions, and utilized these models to improve human-AI team performance. Meanwhile, due to the ``black-box'' nature of AI models, providing AI explanations to human decision makers to help them rely on AI recommendations more appropriately has become a common practice. In this paper, we explore whether we can quantitatively model how humans integrate both AI recommendations and explanations into their decision process, and whether this quantitative understanding of human behavior from the learned model can be utilized to manipulate AI explanations, thereby nudging individuals towards making targeted decisions. Our extensive human experiments across various tasks demonstrate that human behavior can be easily influenced by these manipulated explanations towards targeted outcomes, regardless of the intent being adversarial or benign. Furthermore, individuals often fail to detect any anomalies in these explanations, despite their decisions being affected by them.

In this paper, we present XST-GCNN (eXplainable Spatio-Temporal Graph Convolutional Neural Network), a novel architecture for processing heterogeneous and irregular Multivariate Time Series (MTS) data. Our approach captures temporal and feature dependencies within a unified spatio-temporal pipeline by leveraging a GCNN that uses a spatio-temporal graph aimed at optimizing predictive accuracy and interoperability. For graph estimation, we introduce techniques, including one based on the (heterogeneous) Gower distance. Once estimated, we propose two methods for graph construction: one based on the Cartesian product, treating temporal instants homogeneously, and another spatio-temporal approach with distinct graphs per time step. We also propose two GCNN architectures: a standard GCNN with a normalized adjacency matrix and a higher-order polynomial GCNN. In addition to accuracy, we emphasize explainability by designing an inherently interpretable model and performing a thorough interpretability analysis, identifying key feature-time combinations that drive predictions. We evaluate XST-GCNN using real-world Electronic Health Record data from University Hospital of Fuenlabrada to predict Multidrug Resistance (MDR) in ICU patients, a critical healthcare challenge linked to high mortality and complex treatments. Our architecture outperforms traditional models, achieving a mean ROC-AUC score of 81.03 +- 2.43. Furthermore, the interpretability analysis provides actionable insights into clinical factors driving MDR predictions, enhancing model transparency. This work sets a benchmark for tackling complex inference tasks with heterogeneous MTS, offering a versatile, interpretable solution for real-world applications.

Backdoor attacks present a significant threat to the robustness of Federated Learning (FL) due to their stealth and effectiveness. They maintain both the main task of the FL system and the backdoor task simultaneously, causing malicious models to appear statistically similar to benign ones, which enables them to evade detection by existing defense methods. We find that malicious parameters in backdoored models are inactive on the main task, resulting in a significantly large empirical loss during the machine unlearning process on clean inputs. Inspired by this, we propose MASA, a method that utilizes individual unlearning on local models to identify malicious models in FL. To improve the performance of MASA in challenging non-independent and identically distributed (non-IID) settings, we design pre-unlearning model fusion that integrates local models with knowledge learned from other datasets to mitigate the divergence in their unlearning behaviors caused by the non-IID data distributions of clients. Additionally, we propose a new anomaly detection metric with minimal hyperparameters to filter out malicious models efficiently. Extensive experiments on IID and non-IID datasets across six different attacks validate the effectiveness of MASA. To the best of our knowledge, this is the first work to leverage machine unlearning to identify malicious models in FL. Code is available at \url{https://github.com/JiiahaoXU/MASA}.

Credit card fraud detection is a critical concern for financial institutions, intensified by the rise of contactless payment technologies. While deep learning models offer high accuracy, their lack of explainability poses significant challenges in financial settings. This paper explores the integration of fuzzy logic into Deep Symbolic Regression (DSR) to enhance both performance and explainability in fraud detection. We investigate the effectiveness of different fuzzy logic implications, specifically {\L}ukasiewicz, G\"odel, and Product, in handling the complexity and uncertainty of fraud detection datasets. Our analysis suggest that the {\L}ukasiewicz implication achieves the highest F1-score and overall accuracy, while the Product implication offers a favorable balance between performance and explainability. Despite having a performance lower than state-of-the-art (SOTA) models due to information loss in data transformation, our approach provides novelty and insights into into integrating fuzzy logic into DSR for fraud detection, providing a comprehensive comparison between different implications and methods.

We aim to evaluate Large Language Models (LLMs) for embodied decision making. While a significant body of work has been leveraging LLMs for decision making in embodied environments, we still lack a systematic understanding of their performance because they are usually applied in different domains, for different purposes, and built based on different inputs and outputs. Furthermore, existing evaluations tend to rely solely on a final success rate, making it difficult to pinpoint what ability is missing in LLMs and where the problem lies, which in turn blocks embodied agents from leveraging LLMs effectively and selectively. To address these limitations, we propose a generalized interface (Embodied Agent Interface) that supports the formalization of various types of tasks and input-output specifications of LLM-based modules. Specifically, it allows us to unify 1) a broad set of embodied decision-making tasks involving both state and temporally extended goals, 2) four commonly-used LLM-based modules for decision making: goal interpretation, subgoal decomposition, action sequencing, and transition modeling, and 3) a collection of fine-grained metrics which break down evaluation into various types of errors, such as hallucination errors, affordance errors, various types of planning errors, etc. Overall, our benchmark offers a comprehensive assessment of LLMs' performance for different subtasks, pinpointing the strengths and weaknesses in LLM-powered embodied AI systems, and providing insights for effective and selective use of LLMs in embodied decision making.