Results for “Data”

Journal articles (OpenAlex) Yan Huang, Yungang Liu · Territory Politics Governance South China Sea

Tourism as maritime statecraft: Hainan’s policy entrepreneurship in the South China Sea

This article examines how China’s South China Sea (SCS) tourism was formed, enacted and recalibrated over time. Drawing on multi-decade data, we show that SCS tourism was more than a top-down territorialisation strategy but also a socio-economic enterprise, reflecting both Beijing’s evolving foreign-policy calculations and the uneven weight of bureaucratic concerns and local lobbying. The Hainan government acted as a policy entrepreneur in advancing Paracel tourism and…

Read at Journal articles (OpenAlex)

arXiv (research papers) Manar Abouzaid SI

Hybrid Adversarial Defence for Natural Language Understanding Tasks

Large Language Models (LLMs) are vulnerable both to hallucination and adversarial manipulation. Although these problems are closely related, existing defences typically address them separately. We investigate a hybrid defence framework that combines entropy-based models, designed to reduce hallucinations, with uncertainty-based models and geometric-based models, designed to reduce vulnerability. Under in-domain tests on Natural Language Understanding datasets (FEVER…

Read at arXiv (research papers)

Journal articles (OpenAlex) Jihye Lee, Mitra Naseh, Yingying Zeng et al. · Journal of Immigrant & Refugee Studies Geopolitics

Multidimensional Poverty and Perceived Discrimination Among Newly Resettled Afghans in the United States

Refugees experience high poverty rates, yet research on refugees’ economic and social integration remains limited, especially among Afghan refugees resettled in the United States following the 2021 Taliban takeover. This study explores multidimensional poverty and discrimination among 185 newly resettled Afghan refugee households in St. Louis, Missouri. Cross-sectional survey data were collected in Dari, Pashto, and English. Poverty was assessed using an adapted…

Read at Journal articles (OpenAlex)

arXiv (research papers) Yingqi Zhang SI

Agent libOS: A Runtime Substrate for Capability-Controlled Self-Evolving LLM Agents

Large language model (LLM) agents can persist across tasks, acquire memory, activate Skills, synthesize tools, fork child processes, attach remote resources, and commit checkpoints as reusable images. These mechanisms expand the action surface after deployment and create authority-escalation and data-exfiltration risks when visibility is mistaken for permission. We present Agent libOS, an agent-native library OS substrate that separates three planes. Operation admission…

Read at arXiv (research papers)

arXiv (research papers) Ryle Goehausen SI

Gate AI: LLM Security Benchmark Evaluation Methodology and Results

Published evaluations of prompt-injection and jailbreak detectors for Large Language Models often suffer from two systematic weaknesses: per-dataset threshold tuning and undisclosed operating points. We describe an evaluation harness that addresses both. The detector under evaluation is scored across 16 public benchmarks (12,111 samples) using 5-fold cross-validation. StratifiedKFold (by row) is the headline pass; a parallel StratifiedGroupKFold pass over a composite key…

Read at arXiv (research papers)

arXiv (research papers) Hiskias Dingeto SI

AgentRedBench: Dynamic Redteaming and Integration-Aware Defense for LLM Agents over SaaS Integrations

Indirect prompt injection in tool-use agents is a concrete production threat: LLM agents read from integrations (third-party services such as Gmail, Salesforce, or Jira accessed through tool calls) whose response content the user neither writes nor controls. Existing benchmarks under-measure the threat: most cover only a handful of integrations with the same attack payload replayed across runs, and open-source guards are trained on chat-style data rather than tool-response…

Read at arXiv (research papers)

Journal articles (OpenAlex) Bülent Aras, Burcu Fazlıoğlu · World Affairs Geopolitics

The Normative Shield: How the BRICS Loose Alliance Promotes Authoritarian Resilience

ABSTRACT This article investigates whether the BRICS grouping strategically deploys high‐level diplomatic visits to bolster incumbent autocrats facing domestic instability, thereby promoting “authoritarian durability.” Utilizing a novel dataset of official visits by the five core members (Brazil, Russia, India, China, and South Africa) between 1991 and 2024, the study models the relationship between host‐country unrest and diplomatic engagement, revealing a significant…

Read at Journal articles (OpenAlex)

arXiv (research papers) Mohammadreza Rashidi SI

Depth-Dependent Indirect Prompt Injection in Tool-Calling ReAct Agents: Injection Depth, Payload Framing, and Turn-Budget Sensitivity

ReAct agents that interleave chain-of-thought reasoning with tool calls are increasingly deployed for real tasks such as scheduling, file retrieval, and data access. Their tool observation loop creates a direct attack surface: an adversary who controls any tool's return value can embed instructions that redirect the agent away from the user's goal, a threat known as indirect prompt injection. Existing benchmarks evaluate attack success rate (ASR) at a fixed injection…

Read at arXiv (research papers)

Journal articles (OpenAlex) Enze Han · The China Quarterly South China Sea

How the Chinese Public Perceives South-East Asia: Affinity, Annoyance and Apprehension

Abstract This article explores how the Chinese public perceives South-East Asia. Using original survey data collected in China, it suggests three perceptual frames. The “affinity” frame, which is rooted in historical Chinese migration and diaspora ties, generates feelings of kinship and cultural connection. Conversely, the “annoyance” frame stems from contemporary geopolitical tensions, particularly South China Sea disputes, where several South-East Asian countries are…

Read at Journal articles (OpenAlex)

arXiv (research papers) Travis Lelle SI

Token-Level Generalization in LoRA Adapter Backdoors: Attack Characterization and Behavioral Detection

We show that LoRA adapters, the dominant distribution format for fine-tuned LLMs, can be reliably backdoored through training data poisoning while preserving baseline task performance. On a Qwen 2.5 1.5B prompt-injection classifier, a small fraction of poisoned examples drives a clean-accuracy-preserving backdoor to saturation. The resulting backdoor generalizes at the token feature level rather than the structural pattern level: a model trained on one RFC reference…

Read at arXiv (research papers)

Journal articles (OpenAlex) Deimantas Jastramskis · Journalism Geopolitics

Factors of political influence on journalistic activities in a post-communist consolidated democracy

The article examines which factors of journalistic activity, journalists’ roles, and attitudes are significant for journalists experiencing political influence in a post-communist consolidated democracy. The analysis is based on data from a representative survey of 302 Lithuanian journalists. The study found that the key factors of political influence on journalism are external informational influences, most often directly related to political entities. Political influences…

Read at Journal articles (OpenAlex)

Journal articles (OpenAlex) Anilesh Kumar, Daya Kishan Thussu · Media International Australia Geopolitics

Mediated multipolarity: Russia Today's discursive construction of India before and after Russia's invasion of Ukraine

This article examines how Russia Today (RT), a state-funded international broadcaster, reframed its coverage of India before and after the Russian invasion of Ukraine. Drawing on a longitudinal dataset of 635 English-language articles published between 2019 and 2023, the study combines topic modelling with discursive framing analysis to trace a shift in RT's India-related narratives. In the pre-war period, India appeared sporadically and was framed through domestic…

Read at Journal articles (OpenAlex)

Journal articles (OpenAlex) Sandra Eckert, Lucas Seghezzo, Cristian Darío Venencia et al. · Global Environmental Change Advances Geopolitics

Linking land investments, supply chains, and deforestation: Insights from soy land deals in South America

This study investigates large-scale land acquisitions (LSLAs) for soy production in Argentina and Brazil, analysing investor origin, trade destinations, and links to deforestation risk. This analysis combines data from the Land Matrix, TRASE, and ORBIS and applies a spatially explicit approach to link investor, trade, deforestation, and soy cultivation data to LSLAs and analyse their geographic and quantitative relationships. We find that domestic investors account for the…

Read at Journal articles (OpenAlex)

Journal articles (OpenAlex) Ziba Karjoo, Loris Rubini · International Economic Journal Geopolitics

Exporters and Energy Efficiency: Evidence from Chilean Firms

We explore how exporting affects firms' energy consumption. One concern about international trade is its impact on the environment: when a firm exports, it expands its output, increasing pollution. This concern would be mitigated if exporters use “cleaner” technologies than non-exporters.Using Chilean data, we investigate whether exporting drives firms to substitute fuel, a dirty technology, for electricity, a clean technology. We do not find a significant difference between…

Read at Journal articles (OpenAlex)

Simon Willison SI

Microsoft Copilot Cowork Exfiltrates Files

Microsoft Copilot Cowork Exfiltrates Files The biggest challenge in designing agentic systems continues to be preventing them from enabling attackers to exfiltrate data. In this case Microsoft Copilot Cowork (yes, that's a real product name) was allowing agents to send emails to the user's own inbox without approval... but those messages were then displayed in a way that could leak data to an attacker via rendered images: Because these messages can contain external images…

Read at Simon Willison

arXiv (research papers) Zedian Shao SI

Cordyceps: Covert Control Attacks on LLMs via Data Poisoning

Large language models (LLMs) are often fine-tuned on uncurated text datasets that adversaries can poison. Existing poisoning attacks primarily rely on fixed trigger phrases that defenses such as outlier detection, clean-data regularization, or online monitoring can neutralize. In this paper, we propose a data poisoning method that teaches an LLM an information hiding scheme reliably and stealthily through semantic associations between shared knowledge such as facts or…

Read at arXiv (research papers)

arXiv (research papers) Peiran Wang SI

Aligning Provenance with Authorization: A Dual-Graph Defense for LLM Agents

LLM-based agents are increasingly deployed in high-stakes scenarios such as email management, financial transactions, and code execution, where they interact with the external world through tool calling. During execution, these agents must read external data sources (emails, webpages, files) that attackers can control; through indirect prompt injection, attackers embed malicious instructions in this data to manipulate agents into performing unauthorized operations such as…

Read at arXiv (research papers)

Journal articles (OpenAlex) Yuhang Chen, Feng Xu, Yilin Zhong et al. · Applied Economics Geopolitics

Turning crisis into opportunity: how export controls reconfigure corporate human capital structures

With global trade tensions escalating, U.S. export controls on China constitute an important external factor that influences the human capital structure of Chinese corporations. Using panel data on Chinese A-share listed companies from 2010 to 2024, this study applies a two-way fixed effects model to examine the impact of U.S. export controls on the human capital structure of Chinese corporations and the underlying mechanisms. The results indicate that (1) export controls…

Read at Journal articles (OpenAlex)

arXiv (research papers) Faruk Alpay SI

AgentSecBench: Measuring Prompt Injection, Privacy Leakage, and Tool-Use Integrity in LLM Agents

LLM agents process trusted instructions, retrieved records, and tool observations through a common generative channel. This conflates data flow with authority: an untrusted string can affect a secret-bearing response or an action proposal even when no application policy authorizes that influence. We introduce AgentSecBench as an empirical instantiation of a formal security framework for this problem. The framework defines three games-instruction-integrity…

Read at arXiv (research papers)

arXiv (research papers) Zixuan Chen SI

IterInject: Indirect Prompt Injection Against LLM Agents via Feedback-Guided Iterative Optimization

LLM-based agents are increasingly deployed for complex tasks requiring planning, tool use, and interaction with external services. Their reliance on untrusted external content exposes them to indirect prompt injection (IPI), in which adversarial instructions embedded in retrieved data hijack agent behavior. Existing attacks rely on static payloads that cannot adapt to agent-specific defenses; even recent adaptive methods lack structured feedback to guide optimization. We…

Read at arXiv (research papers)

arXiv (research papers) Rohan Pandey SI

Poisoning the Watchtower: Prompt Injection Attacks Against LLM-Augmented Security Operations Through Adversarial Log Content

Large language models (LLMs) are increasingly used as analyst assistants in security operations centers (SOCs), where they ingest log and alert data to produce triage labels, incident summaries, or remediation advice. We study a structural failure mode of this design: many log fields are attacker controlled. User agents, URLs, payloads, DNS queries, and attempted usernames can therefore carry instructions to the model alongside evidence of the intrusion. We call this setting…

Read at arXiv (research papers)

arXiv (research papers) Qingcao Li SI

MixFake: Benchmarking and Enhancing Audio Deepfake Detection in Diverse Real-world Mixed Audio

Speech deepfake detection has achieved remarkable success in clean environments but faces significant challenges in complex, real-world scenarios where speech is often mixed with background music or noise. Current state-of-the-art methods rely on semantic features from self-supervised learning (SSL) models, which often fail when processing non-speech or mixed-source audio. In this paper, we first introduce MixFake, a large-scale benchmark dataset designed to simulate diverse…

Read at arXiv (research papers)

Journal articles (OpenAlex) Ahmad Khan, Melanie M. Tidman · Cureus Geopolitics

Provincial Trends in Childhood Vaccine Coverage in Afghanistan From 2000 to 2024: An Ecological Study of Findings From the Institute for Health Metrics and Evaluation

This study analyzes vaccination coverage trends across 34 provinces in Afghanistan from 2000 to 2024, using data from the Institute for Health Metrics and Evaluation (IHME). Coverage rates for bacillus Calmette-Guérin (BCG), diphtheria-tetanus-pertussis third dose (DTP3), measles-containing vaccine first dose (MCV1), polio vaccine (three doses), pneumococcal conjugate vaccine third dose (PCV3), hepatitis B vaccine third dose (HepB3), Haemophilus influenzae type b vaccine…

Read at Journal articles (OpenAlex)

arXiv (research papers) Sahar Abdelnabi SI

AI Agents May Always Fall for Prompt Injections

Prompt injection is the most critical vulnerability in deployed AI agents. Despite recent progress, we show that the prevailing defense paradigm (data-instruction separation) both fails to detect attacks that operate through contextual manipulation and degrades contextually appropriate behavior. We then recast prompt injection via the lens of Contextual Integrity (CI), a privacy theory that judges information flow compliance with contextual norms. This explains types of…

Read at arXiv (research papers)

Journal articles (OpenAlex) Daniella Bendo, Gabrielle Gooch, Stefania Maggi · Children s Geographies Geopolitics

‘It hurts when they don’t listen to you’: young people’s perceptions, emotions, and experiences engaging in youth-led climate activism

This paper documents the analysis of qualitative data, collected from focus groups and interviews with diverse young people ages 18–24 from across Canada. It analyzes the spectrum of emotions surrounding their perceptions and experiences of what constitutes youth-led climate activism and their understanding of participatory involvement in climate activism. Findings from the qualitative data reveal the complexities associated with the intersectional realities of youth-led…

Read at Journal articles (OpenAlex)

arXiv (research papers) Tsafac Nkombong Regine Cyrille SI

STRIDE-AI: A Threat Modeling Framework for Generative AI Security Assessment

Traditional cybersecurity methodologies target deterministic systems and fail to address the probabilistic nature of AI, leaving systems vulnerable to attack vectors such as model inversion, data poisoning, and prompt injection. Recent industry reports indicate that a majority of organizations deploying AI lack a dedicated security strategy, with adversarial attacks increasing rapidly year-over-year. We present \textit{STRIDE-AI}, a framework that bridges the gap between…

Read at arXiv (research papers)

Journal articles (OpenAlex) Moa Frödin Gruneau, Valeriya Mechkova · Journal of European Public Policy Geopolitics

Becoming Sweden: unpacking women's representation and welfare state building from a historical perspective

How did women's political inclusion shape the early development of the welfare state? We examine Sweden – one of the most egalitarian countries today – to explore the relationship between women's political representation and welfare state formation. Our focus is on the first local election after universal suffrage, analysing over 700 rural municipalities at a time when the welfare state was in its infancy, and women were newcomers in the political arena. Using data on…

Read at Journal articles (OpenAlex)

Journal articles (OpenAlex) Wasiu Olorunlambe, Andreas Jud · Transcultural Psychiatry Geopolitics

The Effects of Boko Haram Terrorism on Adolescents' and Young Adults' Psychosocial Adjustment in the Northeast of Nigeria

Children are a particularly vulnerable group and, in northeast Nigeria, they are severely affected by Boko Haram terrorism. Although previous studies have looked at the effects of terrorism on children’s mental health, there is a lack of peer-reviewed studies on the effects of terrorism on adolescents’ psychosocial adjustment. This study addresses this gap with data on the consequences of Boko Haram terrorism in the northeast of Nigeria. A cross-sectional research design was…

Read at Journal articles (OpenAlex)

Journal articles (OpenAlex) Leah Nann · Ethnos Geopolitics

Learning to Feel Otherwise: Emotional Training and Feminist Counter Tactics Amid Online Misogyny in German-Speaking Disruptive Climate Activism

This article examines the experiences with and consequences of online misogyny towards members of the disruptive climate justice movement ‘Letzte Generation’, foregrounding data collection through online and in-person co-constructed ‘in-depth’ conversations with activists based in Germany and Austria. Firstly, the analysis focuses on the role of ritualised organisational training practices. I conceptualise these as ‘emotive institutions’ whose socially organised practices…

Read at Journal articles (OpenAlex)

arXiv (research papers) Joel Rorseth SI

RUBEN: Rule-Based Explanations for Retrieval-Augmented LLM Systems

This paper demonstrates RUBEN, an interactive tool for discovering minimal rules to explain the outputs of retrieval-augmented large language models (LLMs) in data-driven applications. We leverage novel pruning strategies to efficiently identify a minimal set of rules that subsume all others. We further demonstrate novel applications of these rules for LLM safety, specifically to test the resiliency of safety training and effectiveness of adversarial prompt injections.

Read at arXiv (research papers)

Older stories