67471 research outputs found
Sort by
Learning to Balance: Equitable Districting and Routing in Last-Mile Logistics via Graph Neural Networks
No full textThis work provides a data-driven, deep learning-based solution to the districting and routing problem. Related previous solution approaches focus on cost minimization and face limitations by yielding highly imbalanced districts. This imbalance can cause practical problems such as excessive service times, low customer satisfaction, and unfair workload distribution among deliverers. We propose a deep learning-based solution architecture based on Graph Neural Networks that integrates balance-awareness into the learning process. Evaluation on a large set of real-world cities demonstrates that our approach achieves a significant improvement in workload balance
IT Attacks, OT Panic: Mischaracterizations of Cyberattacks on Critical Infrastructure in Public Discourse
Get PDFRecent high-profile cyberattacks have raised concerns about the security of operational technology (OT) in critical infrastructure. However, many incidents attributed to compromised OT were actually attacks on traditional information technology (IT) systems. This mischaracterization distorts public perception of cyber threats and can lead to uninformed policy responses. Through case studies on Colonial Pipeline, Saudi Aramco, and Norsk Hydro, this paper examines how incidents that primarily affected IT systems can be easily mischaracterized as attacks on OT environments, exaggerating the threat to physical infrastructure. Our analysis found substantial mischaracterization of all three cases in public discourse, despite forensic evidence showing that disruptions stemmed from IT compromises rather than direct OT breaches. These misattributions have practical consequences for digital government, influencing how agencies allocate funds, coordinate responses, and build resilience. This work advocates for clearer distinctions between IT and OT, improved reporting standards, and stronger collaboration among experts, media, and government institutions
FA-RAG: Financial Advice System Using RAG Based on LLMs
Get PDFThis study presents FA-RAG, a Retrieval Augmented Generation (RAG) framework leveraging large language models (LLMs) to improve the quality of automated financial advice. The system is constructed from 940 expert-authored financial consultation cases, which include household income and expenditure data, demographic attributes of the consulter, and professional advice from financial planners. Both textual and numerical data are transformed into vector representations and stored in a Pinecone vector database, enabling efficient retrieval of relevant cases. When new consultations are processed, the system queries this database to supply the LLM with contextually relevant information, thereby producing more accurate and personalized advice. For evaluation, FA-RAG was applied to 100 consultation cases, with multiple insurance premium levels examined to analyze how recommended reduction rates vary under different financial conditions. The results demonstrate the potential of integrating domain-specific data and retrieval mechanisms with LLMs to support reliable and tailored financial planning assistance
Detecting Synthetic Text Profiles: Human Discernment Versus AI Analytics
Get PDFThis study evaluated human vulnerability to synthetic text profiles generated by artificial intelligence. As large language models become more sophisticated, distinguishing between human and AI-generated content grows increasingly difficult. Our findings show that humans consistently struggled to detect synthetic profiles, which highlights the importance of developing AI tools not only for detection but also for supporting human discernment. By establishing a benchmark of current human capabilities, this research offers a snapshot in time - a necessary step toward tracking how human judgment evolves alongside advancing AI systems through a direct comparison of human performance vs analytic detection approaches. Without such records, it becomes difficult to support users in keeping pace with synthetic media and maintaining control over how credibility is evaluated in digital environments
Introduction to the Minitrack on Adversarial Influences: Erosion of Societal Norms and Institutions from Influences in Digital and social Media
No full textPolarized Distinctiveness: How Platform Designs and Superstar Connections Shape Crowdfunding Success
Get PDFCrowdfunding platforms often see a few top campaigns succeed while most struggle. Optimal distinctiveness theory (ODT) suggests campaigns need to balance fitting in and standing out, but platform designs can interfere with this balance. This study examines how Indiegogo's “keep-what-you-raise” model changes optimal distinctiveness levels. We find a U-shaped relationship between distinctiveness and crowdfunding success. In contrast to the “all-or-nothing” model that favors moderate distinctiveness, the “keep-what-you-raise” model demands that campaigns either closely conform or be exceptionally distinct to succeed. Collaborations with superstars steepen this U-shape, amplifying penalties for moderate distinctiveness while boosting gains for extremes through knowledge transfer and legitimacy spillovers. We extend ODT theoretically to demonstrate how platform dynamics disrupt conventional balancing acts by incentivizing strategic extremism. Methodologically, we advance the distinctiveness measurement through a contextualized heterogeneous graph neural network. Practically, our findings guide campaigners to adopt platform-specific distinctiveness strategies and advise platforms to design mechanisms that support niches
Prompting for the Unknown: Leveraging In-Context-Learning for Few-Shot Open Set Classification
Get PDFRecognising customer intent is crucial for applications such as chatbots and virtual assistants, requiring accurate interpretation of user inputs. While traditional intent recognition systems depend on large datasets and complex machine learning pipelines, large language models (LLMs) offer competitive performance with significantly less training data through in-context learning (ICL). In this work, we assess the effectiveness of ICL for intent recognition, with a particular focus on detecting out-of-distribution (OOD) inputs. We explore prompting strategies to improve OOD detection and systematically evaluate few-shot classifiers under varying OOD proportions. Our results show that implicit prompting strategies yield better precision for OOD detection, while explicit strategies excel at recall. Moreover, we confirm that LLMs perform comparably to conventional classifiers on in-distribution data. However, a significant fraction of OOD errors are non-overlapping between LLMs and traditional models, highlighting limitations in LLM robustness and suggesting new directions for enhancing generalisation in intent recognition systems
Educational Games to Teach Fundamental Principles of Cybersecurity
Get PDFThe benefit of using serious games in education has been known for more than a decade [Dicerbo, 2012; Guillen-Nieto, 2015]. Recently, the use of games to introduce cybersecurity principles to students in grades 5-12 has been conducted. Results have been very positive and the potential to replicate what was done in cybersecurity for other STEM disciplines has been proposed. This paper will address the use of the Cyber Threat Defender (CTD) Collectible Card Game to introduce cybersecurity principles to students. Of importance, especially to rural and Title I schools, is the use of the game along with lesson plans which have been developed to be used by teachers who have no background in cybersecurity. A second game, the Community Cybersecurity Game, was also briefly introduced as a companion that addresses a separate aspect of cybersecurity