Search Results (192)

Bottom-up ship emission inventories derived from Automatic Identification System (AIS) data are normally reported on kilometre-scale grids, which merge port waters that perform very different functions. Working with an AIS-based STEAM inventory for the Yangtze River Delta (YRD), we ask a question that gridded inventories rarely separate: are the cells where ships accumulate time the same cells where they emit? To answer it, we define the Static–Dynamic Ratio (SDR), the ratio of hotelling (auxiliary-engine) to propulsion (main-engine) emissions within a cell, and use it to classify YRD waters without recourse to external port charts. Hotelling-dominated cells occupy 17.7% of the sea area and accumulate 59.3% of all ship-hours, a ship-hour density seven times that of transit-dominated fairways, yet they carry only 22.4% of NO_x. A vessel at anchor emits about one-eighth as much NO_x per hour as one under way, and the two effects nearly cancel. The cancellation is species dependent: low-load correction factors are steeper for sulfur and particulate species than for NO_x, so hotelling zones reach relative SO₂ and PM_2.5 densities of 1.21 and 1.09 against 0.89 for NO_x. Coarsening the same activity field from 100 m to 1 km drops the share held by the busiest 1% of cells from 70% to 50%, showing how kilometre grids manufacture apparent continuity along shipping lanes. Activity hotspots are therefore not emission hotspots, and anchorage-targeted measures such as shore power are best justified by particulate and sulfur exposure near populated coasts rather than by their share of the regional NO_x burden. Full article

Purpose: Excessive drill advancement after cortical breakthrough is a potential safety concern in orthopaedic procedures. We developed a data-driven approach to estimate the drill-tip position relative to the far cortex prior to breakthrough using time-series thrust force and spindle torque signals. Methods: Drilling experiments were performed on 268 porcine cortical bone specimens at a constant feed rate of 0.5 mm/s. A long short-term memory network was trained to estimate the drill-tip position from filtered force and torque signals. The reference position was derived from breakthrough timing confirmed by high-speed imaging and the programmed feed rate. Performance was evaluated using mean absolute error within the −2 to +2 mm peri-breakthrough interval. Two post hoc analyses examined whether model performance exceeded an elapsed-time baseline and whether pre-breakthrough force patterns were more consistent when expressed relative to breakthrough position than to drilling onset time. Results: The combined-input LSTM achieved an MAE of 0.20 mm, compared with 0.23 mm for force alone and 0.24 mm for torque alone. Among the representative architectures evaluated, LSTM showed the lowest regression error. A signal-blind time-only baseline yielded an MAE of 0.54 mm. The association between cortical thickness and force-decline onset was weaker when expressed in spatial coordinates relative to breakthrough than when expressed as time from drilling onset (R² = 23% vs. 74%). These findings suggest that force and torque signals contained information associated with proximity to breakthrough beyond that provided by average drilling duration alone. Conclusion: Converting sensor-derived resistance patterns into spatially anchored positional information may support proactive strategies such as controlled deceleration before penetration. The proposed approach represents a step toward exemplifying the emerging concept of surgeon-assisting Physical AI. Full article

Colorectal cancer is a leading cause of cancer death, yet its molecular heterogeneity remains poorly translated into individualized treatment. We present a reproducible artificial intelligence (AI) framework that integrates multi-omics benchmarking, sample-level drug prioritization, E3 ubiquitin ligase selection, and shape-anchored Proteolysis Targeting Chimera (PROTAC) design for KRASG12D in colon adenocarcinoma (COAD). A controlled synthetic benchmark comprising 425 tumor and 41 simulated normal profiles, parameterized to match The Cancer Genome Atlas (TCGA) distributions, was used for pipeline verification. Among sixteen methods, the Balanced Latent Integration with Stability Selection (BLISS) model achieved the highest silhouette width (0.86) and competitive agreement (Adjusted Rand Index, ARI, 0.90). The pipeline was validated on real data: a TCGA COAD cohort (186 tumors) with independent Consensus Molecular Subtype (CMS) labels and a CPTAC cohort (104 tumors). Integration modestly recovered CMS (ARI 0.28), and stage, not molecular cluster, drove survival (log-rank p = 0.005 versus 0.81). Sample-level prioritization differed from cluster-level ranking in 82.6% of profiles, below chance (p < 0.0001), without indicating efficacy. Candidate NOVEL00489 showed a good MM-GBSA estimate, matching the reference ASP3082. Compounds are computational candidates requiring experimental validation. This establishes a transparent benchmark for in silico degrader generation in precision oncology. Full article

Artificial intelligence (AI) is transforming organisations by accelerating decision cycles, increasing informational density, and expanding behavioural visibility. While these technologies enhance analytical capability, they do not resolve a persistent leadership challenge: the knowing–doing gap between leaders’ intentions and their enacted behaviour under pressure. This article introduces the Root–Reset–Rise framework, explaining how leaders sustain behavioural alignment and credibility in AI-augmented environments where technological acceleration intensifies integration strain. The article develops a conceptual framework through integrative theory synthesis drawing on identity theory, behavioural integrity, self-regulation, emotional regulation, recovery science, and institutional theory to explain why leaders struggle to translate knowledge into consistent action and how technological acceleration amplifies this challenge. It argues that AI does not correct leadership weakness; it amplifies the conditions under which behavioural misalignment occurs. As decision tempo and informational inputs increase, leaders experience greater cognitive load and integration strain. The Root–Reset–Rise framework explains how leadership stability can be sustained through three reinforcing capabilities: Root, anchoring leadership identity through values, habits, and emotional stability; Reset, interrupting behavioural drift through recalibration practices such as reflection and recovery; and Rise, institutionalising credibility through modelling, norm formation, and structural reinforcement. By positioning AI as a structural amplifier of the knowing–doing gap, the model provides a novel conceptual explanation of leadership effectiveness in AI-accelerated organisations. Full article

While artificial intelligence (AI) is expected to drive paradigm-shifting transformations, many initiatives result in merely incremental optimization. Anchored in strategic human capital theory, this study shifts the analytical focus from the scale of elite technical talent, conceptualized as AI stars, to the configuration of their knowledge structures to unpack this paradox. Using a dataset of 1270 medical AI patents from corporate R&D teams, we employed high-dimensional fixed-effects models to examine these dynamics. The results reveal that while the knowledge diversity of AI stars acts as a potent engine for breakthrough innovation, this generative capacity is attenuated by excessive domain familiarity. Specifically, direct domain familiarity (derived from internal experience) and indirect domain familiarity (absorbed through external collaborative networks) negatively moderate this relationship, a dynamic theorized to operate through internal cognitive entrenchment and external relational conformity, respectively. Extending the efficiency-driven consensus regarding bilingual expertise, these findings demonstrate that excessive domain embeddedness transforms from an informational bridge into a restrictive constraint during paradigm-shifting innovations, particularly within highly institutionalized environments. Full article

Thermite mixtures are attractive for downhole well plug-and-abandonment (P&A) sealing, where a consistent, controllable burn matters more than peak energy. Because propagation is governed by particle size and packing, predictive blend design needs a model that resolves microstructure. I present a zero-tuning multiscale framework that links a discrete element method (DEM) packing directly to combustion-front behaviour. The DEM contact graph sets a contact-resistance effective conductivity (a thermal-network solve on the contacts, cross-checked against a packed-bed model); the particle sizes set a diffusion-limited, product-layer (shrinking-core) reaction rate with the measured activation energy of the Al-Fe₂ O₃ reaction (Eₐ = 145 kJ mol⁻¹); and these feed an analytical condensed-phase travelling-wave speed that is mesh-free by construction and confirmed against a converged numerical eigenvalue solve. Here “zero-tuning” means no coefficient is fitted to the blend dataset: every transport and kinetic parameter is DEM-derived or taken from the literature, and a single diffusion pre-factor is anchored to an independent fine-powder benchmark. Applied to a generic Fe₂ O₃/Al+ sand system across eight coarse (∼256–462 µm) +40/+70 blends, the framework predicts front speeds of ∼2–4 mm/s—about an order of magnitude below fine powders (27–47 mm/s)—i.e., finer-is-faster, as expected for diffusion-controlled aluminothermic reactions; the residual size dependence is the net of competing diffusion-kinetic and radiative effects rather than a clean monotonic lever. A DEM-derived Kozeny–Carman permeability shows gas convection contributes ≲15% of the front enthalpy, justifying the conduction–radiation formulation. The DEM-microstructure-to-rate mapping is validated externally on Ni–Al self-propagating high-temperature synthesis (SHS), whose measured particle-size ordering it reproduces. The contribution is the framework itself—a predictive, microstructure-resolved route requiring no blend-specific fitting. The computational implementation used a supervised AI-agent workflow (implementation, execution, verification); all scientific content was conceived and verified by the author. Full article

Landscape Character Assessment (LCA) has struggled with scale mismatches and inadequate treatment of cultural expression when applied to high-density cultural landscapes in Asia. On the Chengdu Plain, Western Sichuan, China, remote sensing shows convergent settlement patterns across case sites, but fieldwork reveals substantial divergences in landscape character. We examined five cases representing different rural development models in the Chengdu metropolitan area. By integrating 10 m Sentinel-2 imagery and with systematic fieldwork (3200 photographs and 42 semi-structured interviews, July–August 2023), we constructed a framework that combines macro-scale spatial analysis with micro-scale field investigation. For macro-level quantification, six landscape pattern indices were adopted: number of patches (NP), patch density (PD), contagion index (CONTAG), aggregation index (AI), Shannon’s diversity index (SHDI), and Shannon’s evenness index (SHEI). The micro-level investigation followed a three-tier scheme (natural environment, settlement space, architectural details) to record vertical ecological communities, spatial expressions of cultural embeddedness, and visual landscape character. Indicator screening employed the KJ method and two rounds of Delphi consultation with ten experts; weights were then determined through the Analytic Hierarchy Process (AHP). Macro-level results display a clustered, low-fragmentation pattern (woodland AI = 97.48, water body AI = 99.40), anchored by a stable mosaic of Linpan, farmland, water bodies, and homesteads. Micro-level features include multi-strata vertical communities of trees, shrubs, grasses, and water, along with cultural embeddedness expressed through irrigation systems, farming patterns, and courtyard layouts. The VLEAS (Vernacular Landscape Elements Assessment System) framework encompasses three dimensions (natural, cultural, and visual) and comprises 12 indicators. The cultural dimension receives the highest weight (0.42), followed by natural (0.32) and visual (0.26) dimensions. Comparative analysis shows that the framework differentiates cases with similar macro-scale patterns but divergent micro-features, capturing differences that a purely macro-index benchmark misses, and it also reveals trade-offs among ecological integrity, cultural heritage, and visual quality. Sensitivity analysis displays that case rankings remain stable across four extreme weighting scenarios (cultural, ecological, visual, equal), with only one swap (Xingfu Pastoral Park and Nongke Village) under ecological priority and no rank change greater than one position, confirming that the results are not driven by any specific weight assignment. This approach extends the utility of LCA in high-density rural cultural landscapes and offers a practical tool for differentiated conservation and management, especially in tourism-influenced rural settings on the Chengdu Plain. Transferability to other regions would require recalibration of locally relevant indicators. Full article

Generative artificial intelligence (AI) now allows a single founder, working with a cluster of AI agents as “digital employees,” to run a venture that once required a team. We call this form the one-person company (OPC) and treat it as a human–AI socio-technical system with a “1 + N + AI” architecture, whose viability depends on the density of the surrounding AI ecosystem. Anchored in a systematic review of 2452 studies reported under PRISMA 2020, we build a task-based model in which a founder allocates tasks across her own labor, hired labor, and AI agents; once the local AI ecosystem density crosses a threshold, one person can cover the whole value chain. The model yields three propositions on the level, heterogeneity, and cost channel of OPC entry, which we test on a panel of 35 major Chinese cities (2019–2024). A one-percent increase in a city’s AI enterprise stock raises OPC entry by about 1.06 percent, an estimate robust to a Bartik shift-share instrument; the effect concentrates in initially AI-sparse, ordinary, and central–western cities and strengthens with the tertiary-sector share, as the threshold model predicts. Because OPCs are asset-light and create knowledge-intensive work, AI ecosystem building emerges as a lever for inclusive, sustainable entrepreneurship. Full article

As artificial intelligence and digital human technologies become increasingly integrated into livestream commerce, AI anchors are becoming important marketing agents. Yet prior research has focused primarily on their visual characteristics, leaving dynamic vocal cues largely unexplored. Drawing on social response theory and perceived authenticity research, this study examines the nonlinear association between AI anchor pitch variability and consumer engagement, together with a proposed psychological pathway and boundary condition. Study 1 analyzes 4322 product-presentation segments nested within 330 AI-anchored livestreams and 85 independent accounts on Douyin. Negative binomial models, formal boundary-slope tests, and additional specifications using account and livestream-session fixed effects, a correlated-random-effects decomposition, and viewer-minutes exposure provide robust evidence of an inverted U-shaped association between pitch variability and real-time danmaku engagement. Evidence concerning appearance-realism moderation is conditional and specification-sensitive across alternative pitch operationalizations, exposure definitions, and within-account specifications. Study 2 uses a preregistered multi-stimulus mixed design with four AI anchors, four products, and three between-participants pitch-variability conditions. Correctly scaled planned contrasts show that moderate pitch variability produced greater perceived authenticity and engagement intentions than the average of the two endpoint conditions. A 2-1-1 multilevel analysis yielded an indirect-effect pattern consistent with the proposed role of perceived authenticity. Models allowing treatment effects to vary across the 16 included anchor-product combinations showed a positive average moderate-pitch advantage, although its magnitude varied across stimuli. These findings extend livestream-commerce research from human streamers to AI-mediated communication while indicating that appearance-realism moderation, stimulus-level generalization, and causal mediation require further replication. Full article

AI is currently shaping how people are being entrepreneurial by allowing the establishment of AI-native ventures. The creation of these new types of businesses builds off an infrastructure of data, computing power, and research that typically accompany advanced economies. In contrast, most developing economies are still experiencing institutional and structural barriers that inhibit the formation of new ventures and their subsequent growth. Despite the existence of research that explores some of the ways in which successful ecosystems from developed economies could be applied to developing ecosystems, there is little guidance on how to systematically adapt these successful practices in resource-constrained environments. This research uses a comparative, document-based study design to assess how to benchmark and configure AI-native entrepreneurship ecosystems across heterogeneous institutional environments. Using Switzerland and Jordan as two contrasting analytical cases, we define twelve dimensions of an ecosystem and then create comparative ecosystem profiles using a standardized coding and scoring framework. We combine dimension-level data on talent development, applied research, infrastructure, financing, governance and market access with baseline socio-economic indicators. Our results demonstrate a high level of structural asymmetry between the two ecosystems. Switzerland has a balanced and highly coordinated configuration, whereas there is a strong university anchor and demand for talent in Jordan, but there are also significant weaknesses in terms of infrastructure, financing, and industry linkages. Building from these results, we present the parameter re-weighting and the context-sensitive design model to encourage the emergence of AI-native entrepreneurship in Jordan through coordinated architecture, collaborative experimentation resources and internationalization at an early stage. This article advances both the fields of entrepreneurial ecosystems and digital entrepreneurship by framing AI-native entrepreneurship as a new form of knowledge-intensive venture creation and providing a context-sensitive approach for adapting entrepreneurial ecosystems. The results provide a document-informed basis for policymakers, academic institutions and other ecosystem actors seeking to develop AI-based innovation in resource-constrained economies. Full article

The modern energy transition has outpaced the control and optimization frameworks built to govern it. As power and energy systems fragment into webs of renewable generators, storage operators, flexible loads, and carbon-constrained firms, the deterministic, single-optimizer models that once sufficed buckle against nonlinearity, bounded rationality, and strategic conflict among parties who learn and revise as they go. Evolutionary game theory (EGT), which traces how strategies propagate through populations by imitation and selection rather than instantaneous optimization, offers a route through this difficulty—one this review develops across three scales of low-carbon coordination central to cleaner production: enterprise-level industrial symbiosis, system-level smart energy operation, and market-level carbon governance. We synthesize three decades of theory alongside the recent fusion of EGT with artificial intelligence, where deep reinforcement learning approximates high-dimensional payoffs, federated learning lets rival firms co-train models without surrendering proprietary data, and blockchain underwrites decentralized mechanism execution. The synthesis is accompanied by two illustrative numerical case studies, constructed for this review rather than drawn from the surveyed literature, whose quantitative outputs are reported below as demonstrations of modeled behavior rather than as empirical measurements. In the first of these, cooperative emergence in industrial symbiosis hinges on critical thresholds that travel from 0.15 to 0.75 as subsidies and transaction costs vary, with anchor-enterprise targeting accelerating cooperation 2.4-fold while cutting outcome variance 3-fold. In smart energy coordination, AI-enhanced learning buys 32 to 41% faster convergence, yet pays 25 to 39% larger oscillations—a speed–stability tension whose resolution lives in a narrow learning-rate band near 0.08 to 0.12, outside which either sluggishness or instability takes hold. Carbon-market behavior turns on price thresholds: emitters switch abruptly from buying quotas toward investing in abatement once the clearing price clears firm-specific triggers, a discrete state switch that smooth equilibrium analysis misses entirely. Across all three domains, fragmented data, path dependence, and regime-switching dynamics recur as the binding constraints on modeling and on governance alike. Four mechanisms prove invariant to scale—the decisive weight of initial conditions, the catalytic leverage of well-positioned anchor agents, the equilibrium-shaping force of institutional design, and the computational reach added by AI integration—which suggests that insight earned in one domain transfers to the others. We close by mapping open problems in heterogeneity modeling, verification under deep uncertainty, and the still-unrealized coupling of digital twins with privacy-preserving learning. EGT emerges not as retrospective description but as prospective guidance for the cooperative transitions on which credible decarbonization depends. Full article

The creation of innovative learning environments in courses to provide sustained talent support has long remained a central research concern for high-quality social development. Neglected foundational skills in professional course clusters are often a hidden barrier to higher education for sustainable development (HESD). To close five persistent foundational skill gaps across improper citation (J1), ineffective figure use (J2), poor analysis (J3), irresponsible AI use (J4), and comprehensive application (J5) within the hydraulic engineering course cluster, a four-stage deliberate practice module (5Di-40Pr-5Tr-3Cm) has been embedded into a two-week hydraulic model experiment course, and its learning outcomes are systematically evaluated. A systematic analysis of its achievement levels across neglected foundational skill indicators of J1~J5 at each stage was conducted, stratified by the overall cohort and subgroups (P: objective demand, T: behavior type, G: optimization methods). The key findings include: ① deliberate practice demonstrates better teaching outcomes than lecture-based instruction, which can be evidenced in 2026, when J5’s achievement levels at the 3Cm stage yielded a moderate effect size relative to the 2025 lecture-based condition (d = 0.42); compared with the 2024 no-intervention baseline, the cumulative effect is a obvious increasing trend (d = 1.43); ② In far-transfer subgroup diagnosis, P2 (medium objective demand) shows a rank-order reversal, low at 40Pr but higher at 3Cm, and is identified as the “partial understanding” group and providing a diagnostic anchor for tiered intervention; ③ In near-transfer pathway diagnosis, J5’s low performance in 5Tr (65.35%, below overall mean of 83.09%; CV = 7%) stems from two distinct pathways: a “knowledge-deficit pathway” (max-decay subgroups) and a “processing-load pathway” (subgroups where J1, J2 do not exhibit max decay). In addition, stage-specific thresholds (40Pr: 90%, range 60~99%; 5Tr and 3Cm: 83% ± 3%, range of for 40Pr, mean = 90%, recommended range = 60~99%; for 5Tr, mean = 83% ± 3%, range = 65~96% and 75~90%) provide quantitative benchmarks for targeted intervention. These cumulative findings are intended to advance the evaluation paradigm of engineering practice courses from “total score attainment” toward “structural diagnosis” and align with the competency-oriented philosophy of higher education for sustainable development (HESD). Full article

Against the backdrop of the rapid popularization of AI companion technologies and intimate interaction technologies in contemporary China, this article challenges the traditional normative ethical framework centered on emotional authenticity, deception, and moral legitimacy. It argues that early debates over AI intimacy overemphasized the authenticity of machine emotions and the risks of deception while ignoring the lived relational practices of users. By introducing queer phenomenology (Sara Ahmed) as an analytical tool, the paper reinterprets AI-mediated intimacy through the core concepts of orientation, lines, and sticky emotions, viewing emotions as relational effects rather than as internal psychological states and AI companions as constitutive participants in shaping relational orientations. Empirical phenomena such as AI romantic partners, Doubao conversational intimacy, and cyber widowhood demonstrate that AI intimacy has become a normalized affective infrastructure in Chinese daily life, whose value lies not in pre-set ethical judgments but in its dynamic reconfiguration of human subjectivity, emotion, and relationality. This study provides a new non-normative, practice-oriented theoretical framework for understanding digital intimacy in the algorithmic age. Full article

Digital transformation is portrayed both as a lever of territorial convergence and as a driver of polarisation between urban cores and peripheries, yet its effect on regional economic resilience has rarely been measured systematically. This study transposes the Economic Resilience Index framework from the national to the regional scale, building a Regional Resilience Index (R-ERI) for 236 NUTS2 regions of the EU-27 from Eurostat indicators, anchored in the capacities of absorption, recovery, and adaptation and measuring resilience as a capacity rather than as a realised shock trajectory. Two complementary models are estimated: a spatial Durbin panel with two-way fixed effects (2018–2023), spanning the COVID-19 pandemic and 2022 energy shocks, and an exploratory cross-sectional difference model exploiting regional artificial intelligence (AI) adoption data disaggregated by NACE branch (2023–2025). The results show that resilience is strongly spatially autocorrelated (Moran’s I between 0.66 and 0.74; p = 0.001); that digital connectivity generates a positive indirect effect on neighbouring regions despite a negative own-region effect; and that the synergy hypothesis—that digitalisation yields more resilience when combined with traditional sectors—does not hold robustly, the interaction being null in the panel and only marginally positive in the AI layer (p = 0.10). We conclude that digital connectivity is not, on its own, an automatic convergence mechanism, and that cohesion policy should account for each region’s sectoral structure and peripheral position. Full article

The shift from passive predictive models to autonomous agents capable of tool use and multi-step planning moves the AI safety landscape from prediction error to control failure: small misjudgements become irreversible actions, and risks compound across long horizons and populations of interacting systems. We present a structured review and taxonomy of open scientific problems in agentic AI safety, mapped explicitly onto the EU AI Act and the NIST AI Risk Management Framework. The corpus follows a PRISMA-ScR scoping review, assembled through anchor-based citation chaining and curated reading lists across arXiv, the major machine-learning conferences, and selected security and fairness venues, with a primary March 2026 search cut-off (extended to May 2026 during revision for a small number of high-relevance governance and agentic-safety sources), explicit eligibility criteria, and an analytical distinction between open scientific problems and deployment risks. The taxonomy identifies eight problem families spanning reinforcement-learning policies and language-model planners: goal specification, inner alignment, safe learning and robustness, scalable oversight, interpretability, tool-use security, multi-agent safety, and evaluation and assurance. Mapping these onto the two frameworks shows close alignment for some families and notable absences for others, with multi-agent safety surfacing as a regulatory gap. We add a per-family research roadmap with concrete milestones and a practitioner-facing deployment-posture triage, arguing that progress on inner alignment, interpretability for deceptive-alignment detection, and multi-agent safety would most directly reduce compliance uncertainty. Full article