260851 research outputs found
Sort by
A critical meta-survey of the lifecycle greenhouse gas emissions of hydrogen energy systems
No full textHydrogen energy systems are central to decarbonization strategies, yet their full climatic footprint remains contested. Lifecycle assessments reveal significant variability influenced by feedstock choices, energy inputs, and production processes. This paper conducts a meta-survey to examine assumptions behind varying estimations and identify improvement areas. We analyzed 653 academic studies, applying rigorous exclusion criteria to extract 906 estimations across ten lifecycle stages from 109 peer-reviewed studies covering 90 % of global hydrogen production over 2000–2024. We calculated emissions intensities based on grams of carbon dioxide equivalent per kilowatt-hour (gCO2e/kWh) and kilograms per kilogram of hydrogen produced (kgCO2e/kgH2), disaggregating estimates by hydrogen “colors” (turquoise, blue, green, etc.) based on different energy inputs. Results indicate median carbon footprints for production alone of 164.3 gCO2e/kWh or 5.5 kgCO2e/kgH2. Full lifecycle accounting—including upstream sourcing, conversion, storage, distribution, end-use, and decommissioning—increases total median footprints to 435.4 gCO2e/kWh or 15.2 kgCO2e/kgH2. Mean values are significantly higher (1038.4 gCO2e/kWh and 34.5 kgCO2e/kgH2), highlighting extreme outliers' impact. Key variability sources include process type, energy inputs, sectoral application, geographic location, leakage rates, and system capacity. We identify gaps in lifecycle methodologies, particularly system boundary definitions, reporting standards, and end-of-life infrastructure treatment. The findings challenge generic assumptions about hydrogen's climate benefits and emphasize the need for granular, pathway-specific analysis in policy, investment, and modeling decisions.</p
Iran’s New Protests Explained - Interview with Kamran Matin
No full textAfter years of economic decline, sanctions, and political repression, protests have once again spread across Iran. What began with demonstrations by bazaar merchants in Tehran over the collapse of the national currency has expanded into a broader wave of unrest across dozens of cities, with reports of deaths, arrests, and growing pressure on the state.In this in-depth interview, Mahtab Mahboub, contributor to The Amargi, speaks with Kamran Matin, Reader in International Relations at the University of Sussex, about the structural forces driving Iran’s latest protests.Drawing on Iran’s political economy, regional geopolitics, and the aftermath of recent military escalation, Matin argues that the current unrest is shaped not only by economic hardship, but by a deeper crisis of legitimacy following intensified sanctions, the suspension of nuclear diplomacy, and the fallout from the June 12-day war.In this video, we explore:Why the protests began in Tehran’s bazaar, and what that signals about the regime’s social baseHow sanctions, inflation, and diplomatic deadlock have closed off prospects for economic reliefThe long-term role of foreign policy in sustaining the Islamic Republic’s internal legitimacyHow the collapse of Iran’s regional power projection has weakened that strategyComparisons between the current protests and earlier waves in 2017, 2019, and the Jin Jiyan Azadi movementThe rise of monarchist narratives around Reza Pahlavi, and their limits inside IranWhether cracks within Iran’s security apparatus resemble early dynamics of the 1979 revolutionAbout the guest:Kamran Matin is an Associate Professor of International Relations at the University of Sussex and the author of Recasting Iranian Modernity: International Relations and Social Change. He writes extensively on Iranian, Kurdish, and Middle Eastern politics, focusing on the intersection of domestic crises and international power structures.Watch the full conversation for a grounded analysis of why Iran’s current protest wave reflects a more profound crisis within the post-revolutionary order and why its outcome remains deeply uncertain.</p
Factors influencing digital transformation adoption: a multi-stakeholder quantitative study in Saudi Arabian Universities
No full textThis study investigates the factors influencing digital transformation (DT) adoption in Saudi Arabian universities through a stakeholder-specific lens. Drawing on the Technology Acceptance Model (TAM) and the Technology–Organization–Environment (TOE) framework, three models were developed one for Top Managers, one for IT Staff, and one for Academic Staff and Students. Data were collected from 2,317 participants across Saudi universities via four structured questionnaires: 408 top managers, 420 IT staff, 665 academic staff and 824 students. Structural Equation Modeling (SEM) was employed to test the proposed relationships. The results reveal that determinants of DT adoption vary across stakeholder groups, while certain shared factors were identified. Leader’s support, cost-effectiveness, compatibility, technology readiness and managers' digital skills were significant for both top managers and IT staff, while leaders’ creativity was insignificant for either. IT staff technical skills were valued by top managers but not by IT staff themselves, while maintenance was uniquely important for IT staff. For students and academic staff, ease of use, usefulness and trust were significant drivers of DT adoption, while resistance to change was insignificant for either group. Awareness influenced students’ adoption but was insignificant among IT and academic staff. Notably, security and quality of services emerged as influential factors across all groups, underscoring their foundational role. This study contributes theoretically by extending TAM and TOE with stakeholder-specific dimensions, and practically by offering tailored insights for universities seeking to advance digital transformation. The results emphasize the need for differentiated strategies aligned with each stakeholder group’s role and priorities.</p
Hallucination detection and reclassification using graph co-training
No full textThe proliferation of large language models (LLMs)
has brought forth unprecedented capabilities in text generation, yet it is often accompanied by the pervasive issue of ”hallucinations,” meaning plausible but factually incorrect or nonsensical outputs. This paper proposes a novel approach for the identification and conceptual reclassification (correction) of such hallucinations
using Graph Co-training, a semi-supervised learning
paradigm that leverages the inherent relationships within textual data. We present a methodology that constructs a graph representation of text, where nodes represent entities or concepts, and edges signify their semantic relationships. This graph structure facilitates the propagation of labels and the discovery of latent
patterns indicative of hallucinatory content. The proposed Graph Co-training framework integrates multiple views of the text (e.g., lexical, semantic, and factual consistency) and iteratively refines a classifier’s understanding of hallucinations. Experimental results demonstrate that the proposed method significantly outperforms state-of-the-art machine learning, deep learning, and even existing co-training and Graph Neural Network (GNN) approaches in accurately identifying hallucinatory text segments. Furthermore, we illustrate a conceptual pipeline for reclassifying (correcting) these identified hallucinations, offering a robust solution to enhance the reliability of LLM-generated content.</p
The Ludic Unconscious: towards a symptomatic reading of play and narrative in Watch Dogs 2
No full textThis essay proposes that videogames are necessary objects of ideological critique. It does so by drawing on the lineage of ideological critique advanced by Louis Althusser via “symptomatic reading” and Fredric Jameson’s notion of “the political unconscious.” In such a framework, I propose the notion of the ludic unconscious, in which both gameplay mechanics and narrative sensibilities converge. To illustrate this, I examine a specific game -- Ubisoft’s Watch Dogs 2 (2016). I seek to provide both a micro- and macro-level analysis of the narrative and gameplay representations of surveillance. I do so with an emphasis on the military-industrial complex, ideology, technology and race. Ultimately, I posit the notion of ludic unconscious, which encapsulates videogames’ ludic and narrative dimensions, their reliance on repetition and displacement of real-life political contradictions.</p
Burning Barns: existential uncertainty and the destructive self
No full textThe work of Erich Fromm has provided valuable insights into why people engage in violent, aggressive, and destructive behaviour. Fromm is known for his analysis of sadism, masochism, conformity and authoritarianism and he was particularly interested in how people maintain a strong sense of security and belonging. In Escape from Freedom ([1941] 1969), he examined the psychological strategies people employ to help alleviate feelings of existential uncertainty and ontological insecurity. These include authoritarianism (involving a sacrifice of freedom and personal autonomy), destructiveness (a defence that aims to eliminate and control), and automaton conformity (the uncritical internalisation of prevailing norms, values, desires, and expectations). Fromm posits that all escape mechanisms are rooted in the need to escape feelings of isolation, self-doubt, uncertainty, and powerlessness. This article specifically focuses on his work on destructiveness by examining Haruki Murakami’s ([1993] 2003) mysterious story Barn Burning, where one of the protagonists confesses his enjoyment of secretly burning down barns. The overall aim of the article is to explore Fromm’s notion that destructiveness results from a suppressed and unlived life.</p
Understanding the 2024 summer riots in the UK: three case studies
No full textThe wave of riots in England in summer 2024 constituted the biggest wave of disorder in the country for more than a decade. These were followed by swift policy responses, based on assumptions about the events and the participants, before any detailed empirical investigation had been carried out. There is a need for detailed description of events as a solid basis for both social psychological theory and policy . This paper therefore presents case studies of the disorders in Bristol, Hanley, and Tamworth, using interviews and multiple secondary sources, to understand what happened and who was involved. Our analysis suggests that it is inaccurate to see the events as ‘protests’, since they consisted of collective attacks (on asylum seekers’ accommodation and on mosques). Protagonists were ethnically white but not homogeneous. At least four different parties were involved – anti-immigrant participants, police, counter-protesters, the targets of the actions (asylum seekers and Muslims), and on one occasion ‘community defenders’. We compare these events to the 2011 English riots, and we specify remaining ‘unknowns’ that future research should address.</p
Constructive community race: full-density spiking neural network model drives neuromorphic computing
No full textThe local circuitry of the mammalian brain is a focus of the search for generic computational principles because it is largely conserved across species and modalities. In 2014 a model was proposed representing all neurons and synapses of the stereotypical cortical microcircuit below 1mm2 of brain surface. The model reproduces fundamental features of brain activity but its impact remained limited because of its computational demands. For theory and simulation, however, the model was a breakthrough because it is full-scale, therefore free of uncertainties of downscaling, and larger models are less densely connected. This sparked a race in the neuromorphic computing community and the model became a de facto standard benchmark. Within a few years real-time performance was reached and surpassed at significantly reduced energy consumption. We review how the computational challenge was tackled by different simulation technologies and derive guidelines for the next generation of benchmarks and other domains of science.</p
A comparative analysis of long-term effective population sizes across eukaryotes
No full textThe effective population size (Ne) is a fundamental parameter in population genetics. Despite its central importance, there are relatively few estimates of Ne available and there have been limited attempts to compare values across eukaryotes. Here, we estimate long-term effective population sizes for 120 species, broadly distributed across the eukaryotic tree of life, using nucleotide diversity and direct mutation rate estimates. We find that Ne varies by nearly 4 orders of magnitude and that it shows strong phylogenetic structure across broad taxonomic scales but not within individual lineages. Phylogenetically controlled regressions reveal that Nₑ correlates with key life history traits, including generation time and propagule size, and that nucleotide diversity serves as a useful proxy for Nₑ. Finally, we show that small Nₑ is generally associated with a reduction in the efficacy of natural selection, as indicated by an elevated ratio of non-synonymous to synonymous diversity (πN/πS), but not with an increase in genome size after accounting for phylogenetic non-independence. These results provide a broad comparative perspective on the factors driving variation in Nₑ and its evolutionary consequences across eukaryotes.</p
Systematic evaluation of commercially available pain-management mHealth apps for chronic pain in the United Kingdom
No full textOBJECTIVES: Self-management is central in chronic pain care, and mobile health (mHealth) applications (apps) offer scalable tools to support symptom monitoring and management. Although promising, these apps vary in quality, adaptability, and integration of evidence-based behaviour change techniques (BCTs). Many remain unregulated and under-evaluated, leaving their benefits for pain management unclear. We systematically evaluated the quality of commercially available pain management apps in the United Kingdom and examined the prevalence of pain-related BCTs and adaptive features. DESIGN AND METHODS: Freely available English-language apps from the 'Health and Fitness' or 'Medical' categories in the Apple® and Google Play® stores were screened and assessed for quality using the Mobile App Rating Scale (MARS; 1 = inadequate, 5 = excellent) and coded for BCTs and adaptive features. RESULTS: Twenty-three apps were included, with a mean MARS score of 3.03 (range = 1.8-4.6). Five scored >4.0, while 39% scored 3.0-3.9, indicating moderate quality. Apps included a mean of 3.3 BCTs, most commonly self-monitoring (87%), instruction (61%), and behaviour-health links (52%). Social support (13%) and goal setting (17%) were rare. An average of 2.3 adaptive mechanisms were identified, with proximal outcomes in all apps and intervention options in 70%, but decision points and tailoring variables were infrequent. CONCLUSION: Commercially available pain apps in the United Kingdom are generally of moderate quality, with limited integration of social, goal-setting, and adaptive features. Greater personalization is needed to strengthen engagement and clinical impact in digital pain self-management.</p