This article sets off for an exploration of the still evolving discourse surrounding artificial intelligence (AI) in the wake of the release of ChatGPT. It scrutinizes the pervasive narratives that are shaping the societal engagement with AI, spotlighting key themes such as agency and decision-making, autonomy, truthfulness, knowledge processing, prediction, general purpose, neutrality and objectivity, apolitical optimization, sustainability game-changer, democratization, mass unemployment, and the dualistic portrayal of AI as either a harbinger of societal utopia or dystopia. Those narratives are analysed critically based on insights from critical computer science, critical data and algorithm studies, from STS, data protection theory, as well as from the philosophy of mind and semiotics. To properly analyse the narratives presented, the article first delves into a historical and technical contextualisation of the AI discourse itself. The article then introduces the notion of "Zeitgeist AI" to critique the imprecise and misleading application of the term "AI" across various societal sectors. Then, by discussing common narratives with nuance, the article contextualises and challenges often assumed socio-political implications of AI, uncovering in detail and with examples the inherent political, power infused and value-laden decisions within all AI applications. Concluding with a call for a more grounded engagement with AI, the article carves out acute problems ignored by the narratives discussed and proposes new narratives recognizing AI as a human-directed tool necessarily subject to societal governance.

Climate change simulations project the slowdown of the Atlantic Meridional Overturning Circulation (AMOC) in the 21st century, though the rate and extent of the decline vary across models. In the Euro-Atlantic region, where the AMOC strongly influences sea surface temperature (SST) variability, climate projections also show significant uncertainty in the evolution of large-scale atmospheric circulation. We hypothesise that the decline of the AMOC and the uncertainty in Euro-Atlantic atmospheric circulation are interconnected. To test this, we analyse three coordinated experiments from the CMIP6 archive (SSP2\\($\$\)-\\($\$\)4.5, SSP5\\($\$\)-\\($\$\)8.5 and abrupt-4xCO2) adopting a storyline approach. In particular, we compare groups of models projecting a larger AMOC decline to groups of models projecting a smaller AMOC decline. Our results indicate that a stronger AMOC weakening is associated with an intensification of the North Atlantic storm track and jet stream, as well as an increased frequency of the NAO\\($\$\)+\\($\$\)weather regime. We link these atmospheric changes to the influence of a reduced warming of the subpolar North Atlantic, known as the North Atlantic Warming Hole (NAWH), associated with the AMOC weakening. To further support our findings, we conduct an additional experiment using the EC-Earth3 model, comparing an abrupt-4xCO2 experiment with one where the AMOC is fixed at preindustrial levels. This experiment corroborates the CMIP6 analysis and validates our hypothesis. We conclude that the AMOC plays a critical role in driving atmospheric changes over the Euro-Atlantic region. Improved monitoring and better constraints on future AMOC decline would help reduce uncertainty in predictions of atmospheric circulation and climate impacts over Europe.

Childhood obesity, driven by an obesogenic environment, remains a public health priority. To review the effectiveness of non-pharmacological interventions for treating childhood obesity, with a focus on weight reduction and consideration of social determinants of health (SDOH). The review included randomized and quasi-experimental studies, published between August 2016 and December 2021, that evaluated the effectiveness of interventions aimed at reducing body weight in children aged 2 to 14 years with obesity. (PROSPERO ID: CRD42022301191). A total of 59 interventions were identified, showing wide variability in design, duration, components and settings. Most included physical activity (93%) and nutrition (81%), while fewer incorporated psychological support (46%) or digital tools (25%). Despite the diversity of approaches, 56% of the interventions were effective, particularly those implemented in healthcare or school settings, involving physical activity, and delivered by multidisciplinary teams. However, only six demonstrated sustained effects beyond 6 months. Risk of bias was high in over half of the studies, and only a small proportion (22%) reported participants' social characteristics such as family structure, race, or socioeconomic status. Importantly, very few interventions were designed or evaluated with explicit consideration of SDOH. CONCLUSION: Overall, effective interventions tended to target both children and parents, include multiple components-especially physical activity-and be delivered in accessible settings such as schools or primary care. Yet, the generalizability of findings is limited, as most studies involved children from relatively advantaged backgrounds. Future interventions should integrate SDOH in both design and evaluation to ensure equity and long-term effectiveness in diverse populations. WHAT IS KNOWN: • Childhood obesity remains a critical public health issue, heavily influenced by obesogenic environments and social determinants of health (SDOH). • Non-pharmacological interventions, particularly those incorporating physical activity and nutritional guidance, have been commonly used to address childhood obesity, though with varying degrees of success. WHAT IS NEW: • This review highlights that interventions delivered in healthcare or school settings by multidisciplinary teams, and involving both physical activity and parental engagement, show higher effectiveness in reducing childhood obesity. • The study identifies a significant gap in the integration of social determinants of health in both the design and evaluation of interventions, limiting their long-term impact and equity across diverse populations.

Exploring the outer reaches of the Solar System presents significant propulsion and mission design challenges. This study assesses the feasibility of a mission to Sedna using two advanced propulsion concepts: the Direct Fusion Drive (DFD) rocket engine, based on D-3He thermonuclear fusion, and a solar sail utilizing thermal desorption of its coating for propulsion. Both are evaluated for a one-way Earth-to-Sedna mission; however, due to the different performances the DFD would enable orbit insertion, whereas for the solar sail a flyby is envisioned. The analysis evaluates key mission parameters, including delivered payload capacity, travel time, and potential science return. For the DFD, we assume a 1.6 MW system with constant thrust and specific impulse, while for the solar sail, we consider acceleration via thermal desorption and a gravity-assist maneuver around Jupiter. The mission analysis incorporates four key phases: departure, interplanetary acceleration, interplanetary coasting, and rendezvous. Sedna is expected to pass through the perihelion of its orbit in 2075--2076 and then move again away from the Sun. Considering the distances involved, a mission targeting the object would need to be launched "relatively" soon, especially if using conventional propulsion systems, which could require up to 30 years of deep-space travel. In our study, results indicate that the DFD could reach Sedna in approximately 10 years, with 1.5 years of thrusting, while the solar sail, assisted by Jupiter's gravity, could complete the journey in 7 years. The feasibility of science payload accommodation, power availability, and communication constraints is also considered. These findings provide a comparative foundation for future deep-space mission planning.

The use of VR for educational purposes provides the opportunity for integrating VR applications into assessments or graded examinations. Interacting with an VR environment requires specific human abilities, thus suggesting the existence of a VR competence. With regard to the emerging field of VR-based examinations, this VR competence might influence a candidate's final grade and hence should be taken into account. In this paper, we proposed and developed a VR competence assessment application. The application features eight individual challenges that are based on generic 3D interaction techniques. In a pilot study, we measured the performance of 18 users. By identifying significant correlations between VR competence score, previous VR experience and theoretically-grounded contributing human abilities and characteristics, we provide first evidence that our VR competence assessment is effective. In addition, we provide first data that a specific VR competence exists. Our analyses further revealed that mainly spatial ability but also immersive tendency correlated with VR competence scores. These insights not only allow educators and researchers to assess and potentially equalize the VR competence level of their subjects, but also help designers to provide effective tutorials for first-time VR users.

The integrity of peer review is fundamental to scientific progress, but the rise of large language models (LLMs) has introduced concerns that some reviewers may rely on these tools to generate reviews rather than writing them independently. Although some venues have banned LLM-assisted reviewing, enforcement remains difficult as existing detection tools cannot reliably distinguish between fully generated reviews and those merely polished with AI assistance. In this work, we address the challenge of detecting LLM-generated reviews. We consider the approach of performing indirect prompt injection via the paper's PDF, prompting the LLM to embed a covert watermark in the generated review, and subsequently testing for presence of the watermark in the review. We identify and address several pitfalls in naïve implementations of this approach. Our primary contribution is a rigorous watermarking and detection framework that offers strong statistical guarantees. Specifically, we introduce watermarking schemes and hypothesis tests that control the family-wise error rate across multiple reviews, achieving higher statistical power than standard corrections such as Bonferroni, while making no assumptions about the nature of human-written reviews. We explore multiple indirect prompt injection strategies--including font-based embedding and obfuscated prompts--and evaluate their effectiveness under various reviewer defense scenarios. Our experiments find high success rates in watermark embedding across various LLMs. We also empirically find that our approach is resilient to common reviewer defenses, and that the bounds on error rates in our statistical tests hold in practice. In contrast, we find that Bonferroni-style corrections are too conservative to be useful in this setting.

The use of VR for educational purposes provides the opportunity for integrating VR applications into assessments or graded examinations. Interacting with an VR environment requires specific human abilities, thus suggesting the existence of a VR competence. With regard to the emerging field of VR-based examinations, this VR competence might influence a candidate's final grade and hence should be taken into account. In this paper, we proposed and developed a VR competence assessment application. The application features eight individual challenges that are based on generic 3D interaction techniques. In a pilot study, we measured the performance of 18 users. By identifying significant correlations between VR competence score, previous VR experience and theoretically-grounded contributing human abilities and characteristics, we provide first evidence that our VR competence assessment is effective. In addition, we provide first data that a specific VR competence exists. Our analyses further revealed that mainly spatial ability but also immersive tendency correlated with VR competence scores. These insights not only allow educators and researchers to assess and potentially equalize the VR competence level of their subjects, but also help designers to provide effective tutorials for first-time VR users.

Inflammaging, an age-associated increase in chronic inflammation, is considered a hallmark of aging. However, there is no consensus approach to measuring inflammaging based on circulating cytokines. Here we assessed whether an inflammaging axis detected in the Italian InCHIANTI dataset comprising 19 cytokines could be generalized to a different industrialized population (Singapore Longitudinal Aging Study) or to two indigenous, nonindustrialized populations: the Tsimane from the Bolivian Amazon and the Orang Asli from Peninsular Malaysia. We assessed cytokine axis structure similarity and whether the inflammaging axis replicating the InCHIANTI result increased with age or was associated with health outcomes. The Singapore Longitudinal Aging Study was similar to InCHIANTI except for IL-6 and IL-1RA. The Tsimane and Orang Asli showed markedly different axis structures with little to no association with age and no association with age-related diseases. Inflammaging, as measured in this manner in these cohorts, thus appears to be largely a byproduct of industrialized lifestyles, with major variation across environments and populations.

We introduce FedGVI, a probabilistic Federated Learning (FL) framework that is robust to both prior and likelihood misspecification. FedGVI addresses limitations in both frequentist and Bayesian FL by providing unbiased predictions under model misspecification, with calibrated uncertainty quantification. Our approach generalises previous FL approaches, specifically Partitioned Variational Inference (Ashman et al., 2022), by allowing robust and conjugate updates, decreasing computational complexity at the clients. We offer theoretical analysis in terms of fixed-point convergence, optimality of the cavity distribution, and provable robustness to likelihood misspecification. Further, we empirically demonstrate the effectiveness of FedGVI in terms of improved robustness and predictive performance on multiple synthetic and real world classification data sets.

In fields ranging from business to systems biology, directed graphs with edges labeled by signs are used to model systems in a simple way: the nodes represent entities of some sort, and an edge indicates that one entity directly affects another either positively or negatively. Multiplying the signs along a directed path of edges lets us determine indirect positive or negative effects, and if the path is a loop we call this a positive or negative feedback loop. Here we generalize this to graphs with edges labeled by a monoid, whose elements represent `polarities' possibly more general than simply "positive" or "negative". We study three notions of morphism between graphs with labeled edges, each with its own distinctive application: to refine a simple graph into a complicated one, to transform a complicated graph into a simple one, and to find recurring patterns called "motifs". We construct three corresponding symmetric monoidal double categories of "open" graphs. We study feedback loops using a generalization of the homology of a graph to homology with coefficients in a commutative monoid. In particular, we describe the emergence of new feedback loops when we compose open graphs using a variant of the Mayer-Vietoris exact sequence for homology with coefficients in a commutative monoid.

Ochre, an iron-rich pigment, is widely associated with symbolic communication, but its functional applications in the Middle Stone Age (MSA) remain poorly understood. Experimental and ethnographic evidence suggests ochre being useful for hide tanning, hafting adhesives, and skin protection, although direct archeological evidence is scarce. We address this gap by presenting ochre tools from Blombos Cave, South Africa, found in Still Bay to pre–Still Bay layers dated 90 to 70,000 years ago. Seven ochre pieces were deliberately modified into lithic retouchers, showing clear use-wear patterns and evidence of intentional shaping. Targeted experiments confirm that some were used for pressure flaking and were rejuvenated to maintain function. These findings provide direct evidence of ochre being used to retouch lithic artifacts during the MSA, highlighting its role in technological systems of this period. The results emphasize the multifunctionality of ochre and suggest that such curated tools may have held personal, cultural, or technological significance within early modern human communities. Seven ochre artifacts from Blombos Cave show evidence of lithic retouching and pressure flaking, revealing specialized MSA tools.

Fourier analysis has been an instrumental tool in the development of signal processing. This leads us to wonder whether this framework could similarly benefit generative modelling. In this paper, we explore this question through the scope of time series diffusion models. More specifically, we analyze whether representing time series in the frequency domain is a useful inductive bias for score-based diffusion models. By starting from the canonical SDE formulation of diffusion in the time domain, we show that a dual diffusion process occurs in the frequency domain with an important nuance: Brownian motions are replaced by what we call mirrored Brownian motions, characterized by mirror symmetries among their components. Building on this insight, we show how to adapt the denoising score matching approach to implement diffusion models in the frequency domain. This results in frequency diffusion models, which we compare to canonical time diffusion models. Our empirical evaluation on real-world datasets, covering various domains like healthcare and finance, shows that frequency diffusion models better capture the training distribution than time diffusion models. We explain this observation by showing that time series from these datasets tend to be more localized in the frequency domain than in the time domain, which makes them easier to model in the former case. All our observations point towards impactful synergies between Fourier analysis and diffusion models.

The artificial glow of the night sky is a form of light pollution; its global change over time is not well known. Developments in lighting technology complicate any measurement because of changes in lighting practice and emission spectra. We investigated the change in global sky brightness from 2011 to 2022 using 51,351 citizen scientist observations of naked-eye stellar visibility. The number of visible stars decreased by an amount that can be explained by an increase in sky brightness of 7 to 10% per year in the human visible band. This increase is faster than emissions changes indicated by satellite observations. We ascribe this difference to spectral changes in light emission and to the average angle of light emissions. Artificial lighting that escapes into the sky causes it to glow, preventing humans and animals from seeing the stars. Satellites can measure the light emitted upward, but they are not sensitive to all wavelengths produced by LED lighting or to light emitted horizontally. Kyba et al. used data from citizen scientists to measure how light pollution is affecting human views of the stars worldwide (see the Perspective by Falchi and Bará). Participants were shown maps of the sky at different levels of light pollution and asked which most closely matched their view. Trends in the data showed that the average night sky got brighter by 9.6% per year from 2011 to 2022, which is equivalent to doubling the sky brightness every 8 years. —KTS Observations of the night sky by citizen scientists show that it is rapidly getting brighter due to light pollution.

In industrial environments, having instrumentation able to attain fast, accurate, and autonomous measurements is pivotal to understanding the dynamics of liquid and particles during transport. Ideally, these instruments, consisting of either probes or sensors, should be robust, fast, and unintrusive, i.e., not cause interference on the very flows being monitored, and require minimal maintenance. Beyond monitoring, the process knowledge gained through real time inspection allows teams to make informed technical decisions based on particle behavior, i.e., settling of particles causing pipe wear and clustering or blockages that can damage the unit or cause shutdowns, both of which with economical drawbacks. The purpose of this review is to examine experimental measurement techniques used to characterize physical properties and operational parameters of solid-liquid slurry flows, focusing on non-ionizing radiation methods. With this text the intent is not to provide an exhaustive examination of each individual technique but rather an overview on the most pertinent types of instrumentation, which will be presented, in addition to application examples from the literature, while directing the reader for pertinent seminal and review papers for a more in-depth analysis.

The Busy Beaver function, with its incomprehensibly rapid growth, has captivated generations of computer scientists, mathematicians, and hobbyists. In this survey, I o er a personal view of the BB function 58 years after its introduction, emphasizing lesser-known insights, recent progress, and especially favorite open problems. Examples of such problems include: when does the BB function rst exceed the Ackermann function? Is the value of BB(20) independent of set theory? Can we prove that BB(n + 1) > 2BB(n) for large enough n? Given BB(n), how many advice bits are needed to compute BB(n + 1)? Do all Busy Beavers halt on all inputs, not just the 0 input? Is it decidable, given n, whether BB(n) is even or odd?

This research has been conducted on a main agricultural production region of the NorthEast Germany over the time frame of 2015. Five widely used crop growth models, namely, WOFOST, Aqua Crop, Ceres Wheat, CropSyst and Light Use Efficiency (LUE), are compared to predict the biomass of winter wheat. To obtain the higher spatial and temporal coverage, this study explains the usage of a fusion model in which both Landsat (30m) and Modis (500m) were combined to generate the synthetic remote sensing time series for the accurate evaluation of vegetation on a daily basis.

Micro bore finishing for metal and ceramic materials has been a challenge in the manufacturing industry. Unfortunately, little is understood about how to polish a micro bore and how to assess its inner wall quality because it is difficult to access the micro bore for either polishing or measurement. This article reports on a feasibility study of the abrasive flow polishing of micro bores of 260 ˜ 500-µm diameters and 25 ˜ 50 length/diameter ratios for both metal and ceramic materials. An abrasive flow polishing machine was designed and built with turbulent flow characteristics. Polishing of steel S45C bores of 400- and 500-µm diameters, stainless steel 304 bores of 500-µm diameter, and zirconia bores of 260-µm diameter was conducted. Surface roughness and topography of the polished inner walls of micro bores were characterized using profilometry and optical interferometry from the three-dimensional point of view. Significant reduction in surface roughness of the micro bore inner walls has been made in the polishing processes. The results indicate that it is feasible to apply the abrasive flow polishing technology for metal and ceramic micro bores of diameters of 260 µm or larger and the length/diameter ratios of 25 or higher. It is found that surface roughness of the polished micro bore inner walls decreases with an increase of the abrasive flow passes.

A new finishing method for inner wall of capillaries has been developed, in which the polishing is performed by letting the slurry flow through the capillary at high speed. With the increasing number of slurry passes, the surface roughness of as-received capillaries decreases rapidly in the early stage of the low number of slurry passes, and then tends to saturate to a certain level depending on the diameter of the capillaries. It is also found that the decreasing surface roughness of the inner wall results in the increasing conductance of gas flow through the capillary. Accounting for the effect of the capillary diameter D on the differential pressure which is measured in nitrogen gas flow through the capillary at a constant flow rate, an empirical equation is found between the differential pressure P and the surface roughness Ra, logP · D4 = 0.362 + logRa1/6.

In any model of typed \($\lambda$\)-calculus conianing some basic arithmetic, a functional p - * (procedure--* expression) will be defined which inverts the evaluation functional for typed X-terms, Combined with the evaluation functional, p-e yields an efficient normalization algorithm. The method is extended to X-calculi with constants and is used to normalize (the X-representations of) natural deduction proofs of (higher order) arithmetic. A consequence of theoretical interest is a strong completeness theorem for \($\beta$\)\($\eta$\)-reduction, generalizing results of Friedman [1] and Statman [31: If two Xterms have the same value in some model containing representations of the primitive recursive functions (of level 1) then they are provably equal in the \($\beta$\)\($\eta$\)- calculus.