Searcharxiv⌕ Search

arXiv subjects

Bardh Prenkaj

Publications and source records attributed to Bardh Prenkaj.

At least 19 recordsLinked to original sources
arXiv2026

Beyond Binary Rewards: A Comparative Study of Reward Design for Reinforcement Unlearning

Machine unlearning seeks to selectively remove specific knowledge from trained language models without full retraining, a growing necessity under privacy regulations such as GDPR and the EU AI Act. Recent work has reformulated unlearning as a Reinforcement Learning with Verifiable Rewards (RLVR) problem, where models are optimized against verifiable rewards computed directly from their outputs. However, existing methods rely on sparse binary rewards that provide minimal learning signal, indicating only whether forbidden content was avoided, and limiting convergence speed. In this paper, we study how reward design affects unlearning efficiency within the Reinforcement Unlearning (RUL) framework. We introduce a principled reward decomposition framework that decouples verifiability from sparsity, and propose two new reward functions: an exponential reward that provides graded penalties based on the count of forbidden-concept occurrences, and a PageRank inspired reward that weights penalties by semantic importance. We conduct experiments on the Real World Knowledge Unlearning (RWKU) benchmark, demonstrating that both rewards consistently outperform the binary setting, while reaching similar forgetting performance up to $3\times$ faster and preserving general model utility. Our results show that reward design is a key driver of unlearning efficiency offering a practical path toward scalable and efficient machine unlearning.

cs.LG↗
arXiv2026

Active Tabular Augmentation via Policy-Guided Diffusion Inpainting

Generative tabular augmentation is appealing in data-scarce domains, yet the prevailing focus on distributional fidelity does not reliably translate into better downstream models. We formalize a fidelity-utility gap: common generative objectives prioritize distributional plausibility, whereas augmentation succeeds only when injected samples reduce the current learner's held-out evaluation loss. This gap motivates learning not just how to generate, but what to generate and when to inject as training evolves. We propose TAP (Tabular Augmentation Policy), which couples diffusion inpainting with a lightweight, learner-conditioned policy to steer generation toward high-utility regions and controls safe injection via explicit gating and conservative windowed commitment. Under severe data scarcity, TAP consistently outperforms strong generative baselines on seven real-world datasets, improving classification accuracy by up to 15.6 percentage points and reducing regression RMSE by up to 32%.

cs.LG↗
arXiv2026

SAGE: Sparse Adaptive Guidance for Dependency-Aware Tabular Data Generation

Generating high-fidelity synthetic tabular data remains a critical challenge for enhancing data availability in privacy-sensitive and low-resource 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.

cs.LG↗
arXiv2026

TabSCM: A practical Framework for Generating Realistic Tabular Data

Most tabular-data generators match marginal statistics yet ignore causal structure, leading downstream models to learn spurious or unfair patterns. We present TabSCM, a mixed-type generator that preserves those causal dependencies. Starting from a Completed Partially Directed Acyclic Graph (CPDAG) found by any causal structure discovery algorithm, TabSCM (i) orients edges to a DAG, (ii) fits root-node marginals with KDE or categorical frequencies, and (iii) learns topologically ordered structural assignments. Such assignments are achieved using conditional diffusion models for continuous variables as child nodes and gradient-boosted trees for categorical ones. Ancestral sampling yields semantically valid records and enables exact counterfactual queries. On seven public datasets, encompassing healthcare, finance, housing, environment, TabSCM matches or surpasses state-of-the-art GAN, diffusion, and LLM baselines in statistical fidelity, downstream utility, and privacy risk, while also cutting rule-violation rates and providing causally meaningful and robust conditional interventions. Because generation is decomposed into explicit equations, it runs up to 583$\times$ faster than diffusion-only models and exposes interpretable knobs for fairness auditing and policy simulation, making TabSCM a practical choice for realism, explainability, and causal soundness.

cs.LG↗
arXiv2026

Analysing the Safety Pitfalls of Steering Vectors

Activation steering has emerged as a powerful tool to shape LLM behavior without the need for weight updates. While its inherent brittleness and unreliability are well-documented, its safety implications remain underexplored. In this work, we present a systematic safety audit of steering vectors obtained with Contrastive Activation Addition (CAA), a widely used steering approach, under a unified evaluation protocol. Using JailbreakBench as benchmark, we show that steering vectors consistently influence the success rate of jailbreak attacks, with stronger amplification under simple template-based attacks. Across LLM families and sizes, steering the model in specific directions can drastically increase (up to 57%) or decrease (up to 50%) its attack success rate (ASR), depending on the targeted behavior. We attribute this phenomenon to the overlap between the steering vectors and the latent directions of refusal behavior. Thus, we offer a traceable explanation for this discovery. Together, our findings reveal the previously unobserved origin of this safety gap in LLMs, highlighting a trade-off between controllability and safety.

cs.CR↗
arXiv2026

Injecting Falsehoods: Adversarial Man-in-the-Middle Attacks Undermining Factual Recall in LLMs

LLMs are now an integral part of information retrieval. As such, their role as question answering chatbots raises significant concerns due to their shown vulnerability to adversarial man-in-the-middle (MitM) attacks. Here, we propose the first principled attack evaluation on LLM factual memory under prompt injection via Xmera, our novel, theory-grounded MitM framework. By perturbing the input given to "victim" LLMs in three closed-book and fact-based QA settings, we undermine the correctness of the responses and assess the uncertainty of their generation process. Surprisingly, trivial instruction-based attacks report the highest success rate (up to ~85.3%) while simultaneously having a high uncertainty for incorrectly answered questions. To provide a simple defense mechanism against Xmera, we train Random Forest classifiers on the response uncertainty levels to distinguish between attacked and unattacked queries (average AUC of up to ~94.8%). We believe that signaling users to be cautious about the answers they receive from black-box and potentially corrupt LLMs is a first checkpoint toward user cyberspace safety.

cs.CR↗
arXiv2026

Reinforcement Unlearning via Group Relative Policy Optimization

During pretraining, LLMs inadvertently memorize sensitive or copyrighted data, posing significant compliance challenges under legal frameworks like the GDPR and the EU AI Act. Fulfilling these mandates demands techniques that can remove information from a deployed model without retraining from scratch. Existing unlearning approaches attempt to address this need, but often leak the very data they aim to erase, sacrifice fluency and robustness, or depend on costly external reward models. We introduce PURGE (Policy Unlearning through Relative Group Erasure), a novel method grounded in the Group Relative Policy Optimization framework that formulates unlearning as a verifiable problem. PURGE uses an intrinsic reward signal that penalizes any mention of forbidden concepts, allowing safe and consistent unlearning. Our approach achieves up to x46 lower token usage per target than state-of-the-art methods, while improving fluency by +5.48% and adversarial robustness by +12.02% over the base model. Extensive evaluation on the Real World Knowledge Unlearning (RWKU) benchmark shows that PURGE reaches 11% unlearning effectiveness while preserving 98% of original utility. PURGE shows that framing LLM unlearning as a verifiable task enables more reliable, efficient, and scalable forgetting, suggesting a promising new direction for unlearning research that combines theoretical guarantees, improved safety, and practical deployment efficiency.

cs.LG↗
arXiv2026

A Multi-Agent Framework for Interpreting Multivariate Physiological Time Series

Continuous physiological monitoring is central to emergency care, yet deploying trustworthy AI is challenging. While LLMs can translate complex physiological signals into clinical narratives, it is unclear how agentic systems perform relative to zero-shot inference. To address these questions, we present Vivaldi, a role-structured multi-agent system that explains multivariate physiological time series. Due to regulatory constraints that preclude live deployment, we instantiate Vivaldi in a controlled, clinical pilot to a small, highly qualified cohort of emergency medicine experts, whose evaluations reveal a context-dependent picture that contrasts with prevailing assumptions that agentic reasoning uniformly improves performance. Our experiments show that agentic pipelines substantially benefit non-thinking and medically fine-tuned models, improving expert-rated explanation justification and relevance by +6.9 and +9.7 points, respectively. Contrarily, for thinking models, agentic orchestration often degrades explanation quality, including a 14-point drop in relevance, while improving diagnostic precision (ESI F1 +3.6). We also find that explicit tool-based computation is decisive for codifiable clinical metrics, whereas subjective targets, such as pain scores and length of stay, show limited or inconsistent changes. Expert evaluation further indicates that gains in clinical utility depend on visualization conventions, with medically specialized models achieving the most favorable trade-offs between utility and clarity. Together, these findings show that the value of agentic AI lies in the selective externalization of computation and structure rather than in maximal reasoning complexity, and highlight concrete design trade-offs and learned lessons, broadly applicable to explainable AI in safety-critical healthcare settings.

cs.LG↗
arXiv2026

Beyond Edge Deletion: A Comprehensive Approach to Counterfactual Explanation in Graph Neural Networks

Graph Neural Networks (GNNs) are increasingly adopted across domains such as molecular biology and social network analysis, yet their black-box nature hinders interpretability and trust. This is especially problematic in high-stakes applications, such as predicting molecule toxicity, drug discovery, or guiding financial fraud detections, where transparent explanations are essential. Counterfactual explanations - minimal changes that flip a model's prediction - offer a transparent lens into GNNs' behavior. In this work, we introduce XPlore, a novel technique that significantly broadens the counterfactual search space. It consists of gradient-guided perturbations to adjacency and node feature matrices. Unlike most prior methods, which focus solely on edge deletions, our approach belongs to the growing class of techniques that optimize edge insertions and node-feature perturbations, here jointly performed under a unified gradient-based framework, enabling a richer and more nuanced exploration of counterfactuals. To quantify both structural and semantic fidelity, we introduce a cosine similarity metric for learned graph embeddings that addresses a key limitation of traditional distance-based metrics, and demonstrate that XPlore produces more coherent and minimal counterfactuals. Empirical results on 13 real-world and 5 synthetic benchmarks show up to +56.3% improvement in validity and +52.8% in fidelity over state-of-the-art baselines, while retaining competitive runtime.

cs.LG↗
arXiv2026

Probabilistic Aggregation and Targeted Embedding Optimization for Collective Moral Reasoning in Large Language Models

Large Language Models (LLMs) have shown impressive moral reasoning abilities. Yet they often diverge when confronted with complex, multi-factor moral dilemmas. To address these discrepancies, we propose a framework that synthesizes multiple LLMs' moral judgments into a collectively formulated moral judgment, realigning models that deviate significantly from this consensus. Our aggregation mechanism fuses continuous moral acceptability scores (beyond binary labels) into a collective probability, weighting contributions by model reliability. For misaligned models, a targeted embedding-optimization procedure fine-tunes token embeddings for moral philosophical theories, minimizing JS divergence to the consensus while preserving semantic integrity. Experiments on a large-scale social moral dilemma dataset show our approach builds robust consensus and improves individual model fidelity. These findings highlight the value of data-driven moral alignment across multiple models and its potential for safer, more consistent AI systems.

cs.CL↗
arXiv2026

Moral Lenses, Political Coordinates: Towards Ideological Positioning of Morally Conditioned LLMs

While recent research has systematically documented political orientation in large language models (LLMs), existing evaluations rely primarily on direct probing or demographic persona engineering to surface ideological biases. In social psychology, however, political ideology is also understood as a downstream consequence of fundamental moral intuitions. In this work, we investigate the causal relationship between moral values and political positioning by treating moral orientation as a controllable condition. Rather than simply assigning a demographic persona, we condition models to endorse or reject specific moral values and evaluate the resulting shifts on their political orientations, using the Political Compass Test. By treating moral values as lenses, we observe how moral conditioning actively steers model trajectories across economic and social dimensions. Our findings show that such conditioning induces pronounced, value-specific shifts in models' political coordinates. We further notice that these effects are systematically modulated by role framing and model scale, and are robust across alternative assessment instruments instantiating the same moral value. This highlights that effective alignment requires anchoring political assessments within the context of broader social values including morality, paving the way for more socially grounded alignment techniques.

cs.CL↗
arXiv2025

From Confidence to Collapse in LLM Factual Robustness

Ensuring the robustness of factual knowledge in LLMs is critical for reliable applications in tasks such as question answering and reasoning. However, existing evaluation methods predominantly focus on performance-based metrics, often investigating from the perspective of prompt perturbations, which captures only the externally triggered side of knowledge robustness. To bridge this gap, we introduce a principled approach to measure factual robustness from the perspective of the generation process by analyzing token distribution entropy in combination with temperature scaling sensitivity. These two factors build the Factual Robustness Score (FRS), a novel metric which quantifies the stability of a fact against perturbations in decoding conditions, given its initial uncertainty. To validate our approach, we conduct extensive experiments on 5 LLMs across 3 closed-book QA datasets (SQuAD, TriviaQA, and HotpotQA). We show that factual robustness varies significantly -- smaller models report an FRS of $0.76$, larger ones $0.93$ -- with accuracy degrading by ~$60\%$ under increased uncertainty. These insights demonstrate how entropy and temperature scaling impact factual accuracy, and lay a foundation for developing more robust knowledge retention and retrieval in future models.

cs.CL↗
arXiv2025

TRADES: Generating Realistic Market Simulations with Diffusion Models

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.

q-fin.TR↗
arXiv2025

CURE: Controlled Unlearning for Robust Embeddings -- Mitigating Conceptual Shortcuts in Pre-Trained Language Models

Pre-trained language models have achieved remarkable success across diverse applications but remain susceptible to spurious, concept-driven correlations that impair robustness and fairness. In this work, we introduce CURE, a novel and lightweight framework that systematically disentangles and suppresses conceptual shortcuts while preserving essential content information. Our method first extracts concept-irrelevant representations via a dedicated content extractor reinforced by a reversal network, ensuring minimal loss of task-relevant information. A subsequent controllable debiasing module employs contrastive learning to finely adjust the influence of residual conceptual cues, enabling the model to either diminish harmful biases or harness beneficial correlations as appropriate for the target task. Evaluated on the IMDB and Yelp datasets using three pre-trained architectures, CURE achieves an absolute improvement of +10 points in F1 score on IMDB and +2 points on Yelp, while introducing minimal computational overhead. Our approach establishes a flexible, unsupervised blueprint for combating conceptual biases, paving the way for more reliable and fair language understanding systems.

cs.CL↗
arXiv2025

Not All Features Deserve Attention: Graph-Guided Dependency Learning for Tabular Data Generation with Language Models

Large Language Models (LLMs) have shown strong potential for tabular data generation by modeling textualized feature-value pairs. However, tabular data inherently exhibits sparse feature-level dependencies, where many feature interactions are structurally insignificant. This creates a fundamental mismatch as LLMs' self-attention mechanism inevitably distributes focus across all pairs, diluting attention on critical relationships, particularly in datasets with complex dependencies or semantically ambiguous features. To address this limitation, we propose GraDe (Graph-Guided Dependency Learning), a novel method that explicitly integrates sparse dependency graphs into LLMs' attention mechanism. GraDe employs a lightweight dynamic graph learning module guided by externally extracted functional dependencies, prioritizing key feature interactions while suppressing irrelevant ones. Our experiments across diverse real-world datasets demonstrate that GraDe outperforms existing LLM-based approaches by up to 12% on complex datasets while achieving competitive results with state-of-the-art approaches in synthetic data quality. Our method is minimally intrusive yet effective, offering a practical solution for structure-aware tabular data modeling with LLMs.

cs.CL↗
arXiv2025

AI on the Pulse: Real-Time Health Anomaly Detection with Wearable and Ambient Intelligence

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.

cs.LG↗
arXiv2025

Doubling Your Data in Minutes: Ultra-fast Tabular Data Generation via LLM-Induced Dependency Graphs

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 Dependency-driven 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 times over LLM-based baselines.

cs.LG↗
arXiv2025

Graph Inverse Style Transfer for Counterfactual Explainability

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.

cs.LG↗