Research Papers

Large Language Models are fundamental actors in the modern IT landscape dominated by AI solutions. However, security threats associated with them might prevent their reliable adoption in critical application scenarios such as government organizations and medical institutions. For this reason, commercial LLMs typically undergo a sophisticated censoring mechanism to eliminate any harmful output they could possibly produce. These mechanisms maintain the integrity of LLM alignment by guaranteeing that the models respond safely and ethically. In response to this, attacks on LLMs are a significant threat to such protections, and many previous approaches have already demonstrated their effectiveness across diverse domains. Existing LLM attacks mostly adopt a generate-and-test strategy to craft malicious input. To improve the comprehension of censoring mechanisms and design a targeted attack, we propose an Explainable-AI solution that comparatively analyzes the behavior of censored and uncensored models to derive unique exploitable alignment patterns. Then, we propose XBreaking, a novel approach that exploits these unique patterns to break the security and alignment constraints of LLMs by targeted noise injection. Our thorough experimental campaign returns important insights about the censoring mechanisms and demonstrates the effectiveness and performance of our approach.

Non-muscle-invasive bladder cancer (NMIBC) is a relentless challenge in oncology, with recurrence rates soaring as high as 70-80%. Each recurrence triggers a cascade of invasive procedures, lifelong surveillance, and escalating healthcare costs - affecting 460,000 individuals worldwide. However, existing clinical prediction tools remain fundamentally flawed, often overestimating recurrence risk and failing to provide personalized insights for patient management. In this work, we propose an interpretable deep learning framework that integrates vector embeddings and attention mechanisms to improve NMIBC recurrence prediction performance. We incorporate vector embeddings for categorical variables such as smoking status and intravesical treatments, allowing the model to capture complex relationships between patient attributes and recurrence risk. These embeddings provide a richer representation of the data, enabling improved feature interactions and enhancing prediction performance. Our approach not only enhances performance but also provides clinicians with patient-specific insights by highlighting the most influential features contributing to recurrence risk for each patient. Our model achieves accuracy of 70% with tabular data, outperforming conventional statistical methods while providing clinician-friendly patient-level explanations through feature attention. Unlike previous studies, our approach identifies new important factors influencing recurrence, such as surgical duration and hospital stay, which had not been considered in existing NMIBC prediction models.

The integration of unmanned aerial vehicles (UAVs) into cellular networks presents significant mobility management challenges, primarily due to frequent handovers caused by probabilistic line-of-sight conditions with multiple ground base stations (BSs). To tackle these challenges, reinforcement learning (RL)-based methods, particularly deep Q-networks (DQN), have been employed to optimize handover decisions dynamically. However, a major drawback of these learning-based approaches is their black-box nature, which limits interpretability in the decision-making process. This paper introduces an explainable AI (XAI) framework that incorporates Shapley Additive Explanations (SHAP) to provide deeper insights into how various state parameters influence handover decisions in a DQN-based mobility management system. By quantifying the impact of key features such as reference signal received power (RSRP), reference signal received quality (RSRQ), buffer status, and UAV position, our approach enhances the interpretability and reliability of RL-based handover solutions. To validate and compare our framework, we utilize real-world network performance data collected from UAV flight trials. Simulation results show that our method provides intuitive explanations for policy decisions, effectively bridging the gap between AI-driven models and human decision-makers.

AI Awareness

Xiaojian Li, Haoyuan Shi, Rongwu Xu +1

Recent breakthroughs in artificial intelligence (AI) have brought about increasingly capable systems that demonstrate remarkable abilities in reasoning, language understanding, and problem-solving. These advancements have prompted a renewed examination of AI awareness not as a philosophical question of consciousness, but as a measurable, functional capacity. AI awareness is a double-edged sword: it improves general capabilities, i.e., reasoning, safety, while also raising concerns around misalignment and societal risks, demanding careful oversight as AI capabilities grow. In this review, we explore the emerging landscape of AI awareness, which includes metacognition (the ability to represent and reason about its own cognitive state), self-awareness (recognizing its own identity, knowledge, limitations, inter alia), social awareness (modeling the knowledge, intentions, and behaviors of other agents and social norms), and situational awareness (assessing and responding to the context in which it operates). First, we draw on insights from cognitive science, psychology, and computational theory to trace the theoretical foundations of awareness and examine how the four distinct forms of AI awareness manifest in state-of-the-art AI. Next, we systematically analyze current evaluation methods and empirical findings to better understand these manifestations. Building on this, we explore how AI awareness is closely linked to AI capabilities, demonstrating that more aware AI agents tend to exhibit higher levels of intelligent behaviors. Finally, we discuss the risks associated with AI awareness, including key topics in AI safety, alignment, and broader ethical concerns.

This paper investigates the integration of graph neural networks (GNNs) with Qualitative Explainable Graphs (QXGs) for scene understanding in automated driving. Scene understanding is the basis for any further reactive or proactive decision-making. Scene understanding and related reasoning is inherently an explanation task: why is another traffic participant doing something, what or who caused their actions? While previous work demonstrated QXGs' effectiveness using shallow machine learning models, these approaches were limited to analysing single relation chains between object pairs, disregarding the broader scene context. We propose a novel GNN architecture that processes entire graph structures to identify relevant objects in traffic scenes. We evaluate our method on the nuScenes dataset enriched with DriveLM's human-annotated relevance labels. Experimental results show that our GNN-based approach achieves superior performance compared to baseline methods. The model effectively handles the inherent class imbalance in relevant object identification tasks while considering the complete spatial-temporal relationships between all objects in the scene. Our work demonstrates the potential of combining qualitative representations with deep learning approaches for explainable scene understanding in autonomous driving systems.

Hallucinations in Large Language Models (LLMs) are widely regarded as errors - outputs that deviate from factual accuracy. However, in creative or exploratory contexts, these "mistakes" may represent unexpected avenues for innovation. We introduce Purposefully Induced Psychosis (PIP), a novel approach that amplifies LLM hallucinations for imaginative tasks such as speculative fiction, interactive storytelling, and mixed-reality simulations. Drawing on Herman Melville's Moby-Dick, where Pip's "madness" reveals profound insight, we reframe hallucinations as a source of computational imagination rather than a flaw. Our method fine-tunes LLMs to encourage speculative, metaphorical, and surreal outputs - hallucinations that are useful when factual accuracy is not the chief objective. Inspired by the consensual illusions of theater and stage magic, PIP situates these creative missteps in contexts where users willingly suspend disbelief, thereby transforming "errors" into catalysts for new ways of thinking. We discuss potential applications, design principles for ensuring user consent, preliminary observations, and implications for broader AI ethics and human-AI collaboration.

High-stakes domains like cyber operations need responsible and trustworthy AI methods. While large language models (LLMs) are becoming increasingly popular in these domains, they still suffer from hallucinations. This research paper provides learning outcomes from a case study with LinkQ, an open-source natural language interface that was developed to combat hallucinations by forcing an LLM to query a knowledge graph (KG) for ground-truth data during question-answering (QA). We conduct a quantitative evaluation of LinkQ using a well-known KGQA dataset, showing that the system outperforms GPT-4 but still struggles with certain question categories - suggesting that alternative query construction strategies will need to be investigated in future LLM querying systems. We discuss a qualitative study of LinkQ with two domain experts using a real-world cybersecurity KG, outlining these experts' feedback, suggestions, perceived limitations, and future opportunities for systems like LinkQ.

As artificial intelligence (AI) systems rapidly gain autonomy, the need for robust responsible AI frameworks becomes paramount. This paper investigates how organizations perceive and adapt such frameworks amidst the emerging landscape of increasingly sophisticated agentic AI. Employing an interpretive qualitative approach, the study explores the lived experiences of AI professionals. Findings highlight that the inherent complexity of agentic AI systems and their responsible implementation, rooted in the intricate interconnectedness of responsible AI dimensions and the thematic framework (an analytical structure developed from the data), combined with the novelty of agentic AI, contribute to significant challenges in organizational adaptation, characterized by knowledge gaps, a limited emphasis on stakeholder engagement, and a strong focus on control. These factors, by hindering effective adaptation and implementation, ultimately compromise the potential for responsible AI and the realization of ROI.

Evaluation and Incident Prevention in an Enterprise AI Assistant

Akash V. Maharaj, David Arbour, Daniel Lee +6

Enterprise AI Assistants are increasingly deployed in domains where accuracy is paramount, making each erroneous output a potentially significant incident. This paper presents a comprehensive framework for monitoring, benchmarking, and continuously improving such complex, multi-component systems under active development by multiple teams. Our approach encompasses three key elements: (1) a hierarchical ``severity'' framework for incident detection that identifies and categorizes errors while attributing component-specific error rates, facilitating targeted improvements; (2) a scalable and principled methodology for benchmark construction, evaluation, and deployment, designed to accommodate multiple development teams, mitigate overfitting risks, and assess the downstream impact of system modifications; and (3) a continual improvement strategy leveraging multidimensional evaluation, enabling the identification and implementation of diverse enhancement opportunities. By adopting this holistic framework, organizations can systematically enhance the reliability and performance of their AI Assistants, ensuring their efficacy in critical enterprise environments. We conclude by discussing how this multifaceted evaluation approach opens avenues for various classes of enhancements, paving the way for more robust and trustworthy AI systems.

Trustworthy AI encompasses many aspirational aspects for aligning AI systems with human values, including fairness, privacy, robustness, explainability, and uncertainty quantification. The ultimate goal of Trustworthy AI research is to achieve all aspects simultaneously. However, efforts to enhance one aspect often introduce unintended trade-offs that negatively impact others. In this position paper, we review notable approaches to five aspects and systematically consider every pair, detailing the negative interactions that can arise. For example, applying differential privacy to model training can amplify biases, undermining fairness. Drawing on these findings, we take the position that current research practices of improving one or two aspects in isolation are insufficient. Instead, research on Trustworthy AI must account for interactions between aspects and adopt a holistic view across all relevant axes at once. We offer recommendations for how researchers can work towards integrated trust, and provide guidance for practitioners to manage interactions today.

Cancer remains a leading global health challenge and a major cause of mortality. This study leverages machine learning (ML) to predict the survivability of cancer patients with metastatic patterns using the comprehensive MSK-MET dataset, which includes genomic and clinical data from 25,775 patients across 27 cancer types. We evaluated five ML models-XGBoost, Naïve Bayes, Decision Tree, Logistic Regression, and Random Fores using hyperparameter tuning and grid search. XGBoost emerged as the best performer with an area under the curve (AUC) of 0.82. To enhance model interpretability, SHapley Additive exPlanations (SHAP) were applied, revealing key predictors such as metastatic site count, tumor mutation burden, fraction of genome altered, and organ-specific metastases. Further survival analysis using Kaplan-Meier curves, Cox Proportional Hazards models, and XGBoost Survival Analysis identified significant predictors of patient outcomes, offering actionable insights for clinicians. These findings could aid in personalized prognosis and treatment planning, ultimately improving patient care.

Deep learning has been successfully applied to medical image segmentation, enabling accurate identification of regions of interest such as organs and lesions. This approach works effectively across diverse datasets, including those with single-image contrast, multi-contrast, and multimodal imaging data. To improve human understanding of these black-box models, there is a growing need for Explainable AI (XAI) techniques for model transparency and accountability. Previous research has primarily focused on post hoc pixel-level explanations, using methods gradient-based and perturbation-based apporaches. These methods rely on gradients or perturbations to explain model predictions. However, these pixel-level explanations often struggle with the complexity inherent in multi-contrast magnetic resonance imaging (MRI) segmentation tasks, and the sparsely distributed explanations have limited clinical relevance. In this study, we propose using contrast-level Shapley values to explain state-of-the-art models trained on standard metrics used in brain tumor segmentation. Our results demonstrate that Shapley analysis provides valuable insights into different models' behavior used for tumor segmentation. We demonstrated a bias for U-Net towards over-weighing T1-contrast and FLAIR, while Swin-UNETR provided a cross-contrast understanding with balanced Shapley distribution.