Generating high-fidelity synthetic tabular data remains a critical challenge for enhancing data availability in privacy-sensitive and lowresource domains.Recent approaches leverage LLMs by representing table rows as sequences, yet suffer from two fundamental limitations:(1) they model feature dependencies densely, introducing spurious correlations; and (2) they assume static relationships between features, ignoring how these dependencies vary with feature values.To overcome these limitations, we introduce SAGE (Sparse Adaptive Guidance), a novel LLM-based generation framework that enforces sparse and dynamic dependency guidance.SAGE discretizes features into value-aware pseudo-features and constructs a mutual information-based sparse dependency graph.This graph adaptively guides generation through explicit context selection or implicit logit correction, enabling LLMs to focus on truly relevant information during synthesis.Our extensive experiments across six datasets and multiple tasks reveal that SAGE not only improves data fidelity and downstream utility, boosting F1 scores by 10% compared to previous LLM-based methods, but also reduces policy violations by one point.These results highlight the importance of adaptive structure in tabular data generation and provide new insights into context-sensitive control of LLMs. 1 * Equal contribution. 1 Our code is publicly available at https://github.com/ ShuoYangtum/SAGE.

Accurate and continuous monitoring of physiological stress is crucial, especially for patients with neurodegenerative diseases. Traditional monitoring methods, such as Electrocardiogram (ECG), are often invasive and limited in duration, while data from lightweight wearable devices, though more practical for seamless monitoring, typically suffers from significant quality degradation compared to clinical-grade measurements. The challenge lies in developing a robust, long-term, and patient-friendly stress monitoring system that overcomes the limitations of conventional approaches and the accuracy compromises of current wearables. Such a system must also provide actionable, interpretable insights for clinicians and adapt to individual patient variability. This manuscript introduces a methodology for seamless stress level monitoring by leveraging UniTS, a foundational model for time series. Our approach redefines stress detection as an anomaly detection problem, establishing a personalized baseline for each patient’s physiological behavior. Furthermore, to enhance clinical utility and trust, the system integrates a Large Language Model (LLM) to generate human-readable explanations for detected anomalies. The proposed UniTS-based methodology demonstrates superior performance, outperforming 12 top-performing methods on three benchmark datasets. Crucially, it achieves performance comparable to that obtained from more invasive, clinical-grade devices (like ECG) even when utilizing data from lightweight wearable devices, thereby enabling truly seamless monitoring. Furthermore, the system has been successfully tested in a real-world environment, in the context of a project to monitor elderly patients with cognitive disorders in their homes. This work presents an advancement in physiological stress monitoring by offering a personalized, explainable, and continuously adaptive system. We extend and fine-tune UniTS to support contextual anomaly detection and LLM-driven explainability, addressing critical gaps in current healthcare monitoring, fostering enhanced clinician control, improved system predictability, and facilitating long-term, real-world applicability for patients with neurodegenerative conditions. • Introduces a novel, personalized approach to stress detection by reformulating it as an anomaly detection task using a foundational time-series model (UniTS). • Demonstrates that lightweight wearable devices can achieve accuracy comparable to clinical-grade sensors like ECG, enabling seamless, long-term patient monitoring. • Integrates explainability through Large Language Models (LLMs), providing human-readable, clinician-oriented insights into detected anomalies. • Outperforms 12 state-of-the-art models across multiple benchmark datasets. • Validated in a real-world pilot study involving elderly patients with neurodegenerative disorders, the system shows high precision and robustness, supporting its deployment in home-care environments for vulnerable populations.

Despite the widespread use of LLMs due to their superior performance in various tasks, their high computational costs often lead potential users to opt for the pretraining-finetuning pipeline. However, biases prevalent in manually constructed datasets can introduce spurious correlations between tokens and labels, creating so-called shortcuts and hindering the generalizability of fine-tuned models. Existing debiasing methods often rely on prior knowledge of specific dataset biases, which is challenging to acquire a priori. We propose RAZOR (Rewriting And Zero-bias Optimization Refinement), a novel, unsupervised, and data-focused debiasing approach based on text rewriting for shortcut mitigation. RAZOR leverages LLMs to iteratively rewrite potentially biased text segments by replacing them with heuristically selected alternatives in a shortcut space defined by token statistics and positional information. This process aims to align surface-level text features more closely with diverse label distributions, thereby promoting the learning of genuine linguistic patterns. Compared with unsupervised SoTA models, RAZOR improves by 3.5% on the FEVER and 6.5% on MNLI and SNLI datasets according to the F1 score. Additionally, RAZOR effectively mitigates specific known biases, reducing bias-related terms by x2 without requiring prior bias information, a result that is on par with SoTA models that leverage prior information. Our work prioritizes data manipulation over architectural modifications, emphasizing the pivotal role of data quality in enhancing model performance and fairness. This research contributes to developing more robust evaluation benchmarks for debiasing methods by incorporating metrics for bias reduction and overall model efficacy.

We introduce AI on the Pulse, a real-world-ready anomaly detection system that continuously monitors patients using a fusion of wearable sensors, ambient intelligence, and advanced AI models. Powered by UniTS, a state-of-the-art (SoTA) universal time-series model, our framework autonomously learns each patient's unique physiological and behavioral patterns, detecting subtle deviations that signal potential health risks. Unlike classification methods that require impractical, continuous labeling in real-world scenarios, our approach uses anomaly detection to provide real-time, personalized alerts for reactive home-care interventions. Our approach outperforms 12 SoTA anomaly detection methods, demonstrating robustness across both high-fidelity medical devices (ECG) and consumer wearables, with a ~22% improvement in F1 score. However, the true impact of AI on the Pulse lies in @HOME, where it has been successfully deployed for continuous, real-world patient monitoring. By operating with non-invasive, lightweight devices like smartwatches, our system proves that high-quality health monitoring is possible without clinical-grade equipment. Beyond detection, we enhance interpretability by integrating LLMs, translating anomaly scores into clinically meaningful insights for healthcare professionals.

Financial markets are complex systems characterized by high statistical noise, nonlinearity, volatility, and constant evolution. Thus, modeling them is extremely hard. Here, we address the task of generating realistic and responsive Limit Order Book (LOB) market simulations, which are fundamental for calibrating and testing trading strategies, performing market impact experiments, and generating synthetic market data. We propose a novel TRAnsformer-based Denoising Diffusion Probabilistic Engine for LOB Simulations (TRADES). TRADES generates realistic order flows as time series conditioned on the state of the market, leveraging a transformer-based architecture that captures the temporal and spatial characteristics of high-frequency market data. There is a notable absence of quantitative metrics for evaluating generative market simulation models in the literature. To tackle this problem, we adapt the predictive score, a metric measured as an MAE, to market data by training a stock price predictive model on synthetic data and testing it on real data. We compare TRADES with previous works on two stocks, reporting a ×3.27 and ×3.48 improvement over SoTA according to the predictive score, demonstrating that we generate useful synthetic market data for financial downstream tasks. Furthermore, we assess TRADES’s market simulation realism and responsiveness, showing that it effectively learns the conditional data distribution and successfully reacts to an experimental agent, giving sprout to possible calibrations and evaluations of trading strategies and market impact experiments. To perform the experiments, we developed DeepMarket, the first open-source Python framework for LOB market simulation with deep learning. In our repository, we include a synthetic LOB dataset composed of TRADES’s generated simulations.

Tabular data is critical across diverse domains, yet high-quality datasets remain scarce due to privacy concerns and the cost of collection.Contemporary approaches adopt large language models (LLMs) for tabular augmentation, but exhibit two major limitations: (1) dense dependency modeling among tabular features that can introduce bias, and (2) high computational overhead in sampling.To address these issues, we propose SPADA (for SPArse Dependencydriven Augmentation), a lightweight generative framework that explicitly captures sparse dependencies via an LLM-induced graph.We treat each feature as a node and synthesize values by traversing the graph, conditioning each feature solely on its parent nodes.We explore two synthesis strategies: a non-parametric method using Gaussian kernel density estimation, and a conditional normalizing flow model that learns invertible mappings for conditional density estimation.Experiments on four datasets show that SPADA reduces constraint violations by 4% compared to diffusion-based methods and accelerates generation by nearly 9,500× over LLM-based baselines.1

Counterfactual explainability seeks to uncover model decisions by identifying minimal changes to the input that alter the predicted outcome. This task becomes particularly challenging for graph data due to preserving structural integrity and semantic meaning. Unlike prior approaches that rely on forward perturbation mechanisms, we introduce Graph Inverse Style Transfer (GIST), the first framework to re-imagine graph counterfactual generation as a backtracking process, leveraging spectral style transfer. By aligning the global structure with the original input spectrum and preserving local content faithfulness, GIST produces valid counterfactuals as interpolations between the input style and counterfactual content. Tested on 8 binary and multi-class graph classification benchmarks, GIST achieves a remarkable +7.6% improvement in the validity of produced counterfactuals and significant gains (+45.5%) in faithfully explaining the true class distribution. Additionally, GIST’s backtracking mechanism effectively mitigates overshooting the underlying predictor’s decision boundary, minimizing the spectral differences between the input and the counterfactuals. These results challenge traditional forward perturbation methods, offering a novel perspective that advances graph explainability.

Shortcut learning undermines model generalization to out-of-distribution data. While the literature attributes shortcuts to biases in superficial features, we show that imbalances in the semantic distribution of sample embeddings induce spurious semantic correlations, compromising model robustness. To address this issue, we propose SCISSOR (Semantic Cluster Intervention for Suppressing ShORtcut), a Siamese network-based debiasing approach that remaps the semantic space by discouraging latent clusters exploited as shortcuts. Unlike prior data-debiasing approaches, SCISSOR eliminates the need for data augmentation and rewriting. We evaluate SCISSOR on 6 models across 4 benchmarks: Chest-XRay and Not-MNIST in computer vision, and GYAFC and Yelp in NLP tasks. Compared to several baselines, SCISSOR reports +5.3 absolute points in F1 score on GYAFC, +7.3 on Yelp, +7.7 on Chest-XRay, and +1 on Not-MNIST. SCISSOR is also highly advantageous for lightweight models with $\tilde$9.5% improvement on F1 for ViT on computer vision datasets and $\tilde$11.9% for BERT on NLP. Our study redefines the landscape of model generalization by addressing overlooked semantic biases, establishing SCISSOR as a foundational framework for mitigating shortcut learning and fostering more robust, bias-resistant AI systems.

Visualization Recommendation Systems (VRSs) are a novel and challenging field of study aiming to help generate insightful visualizations from data and support non-expert users in information discovery. Among the many contributions proposed in this area, some systems embrace the ambitious objective of imitating human analysts to identify relevant relationships in data and make appropriate design choices to represent these relationships with insightful charts. We denote these systems as "agnostic" VRSs since they do not rely on human-provided constraints and rules but try to learn the task autonomously. Despite the high application potential of agnostic VRSs, their progress is hindered by several obstacles, including the absence of standardized datasets to train recommendation algorithms, the difficulty of learning design rules, and defining quantitative criteria for evaluating the perceptual effectiveness of generated plots. This article summarizes the literature on agnostic VRSs and outlines promising future research directions.

Counterfactual Explanation (CE) techniques have garnered attention as a means to provide insights to the users engaging with AI systems. While extensively researched in domains such as medical imaging and autonomous vehicles, Graph Counterfactual Explanation (GCE) methods have been comparatively under-explored. GCEs generate a new graph similar to the original one, with a different outcome grounded on the underlying predictive model. Among these GCE techniques, those rooted in generative mechanisms have received relatively limited investigation despite demonstrating impressive accomplishments in other domains, such as artistic styles and natural language modelling. The preference for generative explainers stems from their capacity to generate counterfactual instances during inference, leveraging autonomously acquired perturbations of the input graph. Motivated by the rationales above, our study introduces RSGG-CE, a novel Robust Stochastic Graph Generator for Counterfactual Explanations able to produce counterfactual examples from the learned latent space considering a partially ordered generation sequence. Furthermore, we undertake quantitative and qualitative analyses to compare RSGG-CE's performance against SoA generative explainers, highlighting its increased ability to engendering plausible counterfactual candidates.

Automated systems must adapt to evolving environments, yet many struggle with drift phenomena affecting healthcare, finance, and cybersecurity domains.The DELTA workshop addresses this by distinguishing between data and concept drift, aiming to create a practical, human-centric framework for managing drift.The workshop seeks innovative drift detection, prediction, and analysis solutions by uniting researchers and practitioners.DELTA fosters collaboration to advance the understanding and management of drift in dynamic data landscapes by featuring keynotes, paper presentations, interactive sessions, and discussions.

We present GRACIE (Graph Recalibration and Adaptive Counterfactual Inspection and Explanation), a novel approach for generative classification and counterfactual explanations of dynamically changing graph data. We study graph classification problems through the lens of generative classifiers. We propose a dynamic, self-supervised latent variable model that updates by identifying plausible counterfactuals for input graphs and recalibrating decision boundaries through contrastive optimization. Unlike prior work, we do not rely on linear separability between the learned graph representations to find plausible counterfactuals. Moreover, GRACIE eliminates the need for stochastic sampling in latent spaces and graph-matching heuristics. Our work distills the implicit link between generative classification and loss functions in the latent space, a key insight to understanding recent successes with this architecture. We further observe the inherent trade-off between validity and pulling explainee instances towards the central region of the latent space, empirically demonstrating our theoretical findings. In extensive experiments on synthetic and real-world graph data, we attain considerable improvements, reaching ~99% validity when sampling sets of counterfactuals even in the challenging setting of dynamic data landscapes.

Real-time monitoring of human behaviours, especially in e-Health applications, has been an active area of research in the past decades. On top of IoT-based sensing environments, anomaly detection algorithms have been proposed for the early detection of abnormalities. Gradual change procedures, commonly referred to as drift anomalies, have received much less attention in the literature because they represent a much more challenging scenario than sudden temporary changes (point anomalies). In this article, we propose, for the first time, a fully unsupervised real-time drift detection algorithm named DynAmo, which can identify drift periods as they are happening. DynAmo comprises a dynamic clustering component to capture the overall trends of monitored behaviours and a trajectory generation component, which extracts features from the densest cluster centroids. Finally, we apply an ensemble of divergence tests on sliding reference and detection windows to detect drift periods in the behavioural sequence.

Anomalies are rare and anomaly detection is often therefore framed as One-Class Classification (OCC), i.e. trained solely on normalcy. Leading OCC techniques constrain the latent representations of normal1motions to limited volumes and detect as abnormal anything outside, which accounts satisfactorily for the openset’ness of anomalies. But normalcy shares the same openset’ness property since humans can perform the same action in several ways, which the leading techniques neglect.We propose a novel generative model for video anomaly detection (VAD), which assumes that both normality and abnormality are multimodal. We consider skeletal representations and leverage state-of-the-art diffusion probabilistic models to generate multimodal future human poses. We contribute a novel conditioning on the past motion of people and exploit the improved mode coverage capabilities of diffusion processes to generate different-but-plausible future motions. Upon the statistical aggregation of future modes, an anomaly is detected when the generated set of motions is not pertinent to the actual future. We validate our model on 4 established benchmarks: UBnormal, HR-UBnormal, HR-STC, and HR-Avenue, with extensive experiments surpassing state-of-the-art results.

The black-box nature and the lack of interpretability detract from constant improvements in Graph Neural Networks (GNNs) performance in social network tasks like friendship prediction and community detection. Graph Counterfactual Explanation (GCE) methods aid in understanding the prediction of GNNs by generating counterfactual examples that promote trustworthiness, debiasing, and privacy in social networks. Alas, the literature on GCE lacks standardised definitions, explainers, datasets, and evaluation metrics. To bridge the gap between the performance and interpretability of GNNs in social networks, we discuss GRETEL, a unified framework for GCE methods development and evaluation. We demonstrate how GRETEL comes with fully extensible built-in components that allow users to define ad-hoc explainer methods, generate synthetic datasets, implement custom evaluation metrics, and integrate state-of-the-art prediction models.

Online courses and e-degrees, although present since the mid-1990, have received enormous attention only in the last decade. Moreover, the new Coronavirus disease (COVID-19) outbreak forced many nations (e.g. Italy, the US, and other countries) to massively push their education system towards an online environment. Academics now are also looking at the crisis as an opportunity for universities to adopt digital technologies for teaching more broadly. But they will have to understand what possible ways of evaluating and effectively teaching will be in this new scenario. The depicted overview, in conjunction with the utility and ubiquitous access to the educational platforms of online courses, entails a vast amount of enrolments. Nevertheless, a high enrolment rate usually translates into a significant dropout (or withdrawal) rate of students (40-80% of online students drop out). Student dropout prediction (SDP) consists of modelling and fore-casting student behaviour when interacting with e-learning platforms. It is a significant phenomenon that has repercussions on online institutions, the involved students and professors. Early approaches tended to perform manual analytic examinations to devise retention strategies. Recent research has adopted automated policies to thoroughly exploit the advantages of student activities(hereafter e-tivities) in the e-platforms and identify at-risk students. These approaches include machine learning and deep learning techniques to predict the student dropout status. Therefore, being able to cope with the trend shifting of student interactions with the course platforms in real-time has become of paramount importance. In this tutorial, we comprehensively overview the SDP problem in the literature. We provide mathematical formalisation to the different definitions proposed, and we introduce simple and complex predictive methods adhering to the following: Student dropout definition, Input modelling, Underlying machine and deep learning techniques, Evaluation measures, Datasets, and privacy concerns.

In this paper, we compare several deep and surface state-of-the-art machine learning methods for risk prediction in problems that can be modelled as a trajectory of events separated by irregular time intervals. Trajectories are the abstract representation of many real-life data, such as patient records, student e-tivities, online financial transactions, and many others. Given the continuously increasing number of machine learning methods to predict future high-risk events in these contexts, we aim to provide more insight into reproducibility and applicability of these methods when changing datasets, parameters, and evaluation measures. As an additional contribution, we release to the community the implementations of all compared methods.