Paula Harder, Luca Schmidt, Francis Pelletier
et al.
Earth System Models (ESM) are our main tool for projecting the impacts of climate change. However, running these models at sufficient resolution for local-scale risk-assessments is not computationally feasible. Deep learning-based super-resolution models offer a promising solution to downscale ESM outputs to higher resolutions by learning from data. Yet, due to regional variations in climatic processes, these models typically require retraining for each geographical area-demanding high-resolution observational data, which is unevenly available across the globe. This highlights the need to assess how well these models generalize across geographic regions. To address this, we introduce RainShift, a dataset and benchmark for evaluating downscaling under geographic distribution shifts. We evaluate state-of-the-art downscaling approaches including GANs and diffusion models in generalizing across data gaps between the Global North and Global South. Our findings reveal substantial performance drops in out-of-distribution regions, depending on model and geographic area. While expanding the training domain generally improves generalization, it is insufficient to overcome shifts between geographically distinct regions. We show that addressing these shifts through, for example, data alignment can improve spatial generalization. Our work advances the global applicability of downscaling methods and represents a step toward reducing inequities in access to high-resolution climate information.
Recent interest in geospatial artificial intelligence (GeoAI) has fostered a wide range of applications using artificial intelligence (AI), especially deep learning, for geospatial problem solving. However, major challenges such as a lack of training data and the neglect of spatial principles and spatial effects in AI model design remain, significantly hindering the in-depth integration of AI with geospatial research. This paper reports our work in developing a deep learning model that enables object detection, particularly of natural features, in a weakly supervised manner. Our work makes three contributions: First, we present a method of object detection using only weak labels. This is achieved by developing a spatially explicit model based on Tobler's first law of geography. Second, we incorporate attention maps into the object detection pipeline and develop a multistage training strategy to improve performance. Third, we apply this model to detect impact craters on Mars, a task that previously required extensive manual effort. The model generalizes to both natural and human-made features on the surfaces of Earth and other planets. This research advances the theoretical and methodological foundations of GeoAI.
Geo-localization is the task of identifying the location of an image using visual cues alone. It has beneficial applications, such as improving disaster response, enhancing navigation, and geography education. Recently, Vision-Language Models (VLMs) are increasingly demonstrating capabilities as accurate image geo-locators. This brings significant privacy risks, including those related to stalking and surveillance, considering the widespread uses of AI models and sharing of photos on social media. The precision of these models is likely to improve in the future. Despite these risks, there is little work on systematically evaluating the geolocation precision of Generative VLMs, their limits and potential for unintended inferences. To bridge this gap, we conduct a comprehensive assessment of the geolocation capabilities of 25 state-of-the-art VLMs on four benchmark image datasets captured in diverse environments. Our results offer insight into the internal reasoning of VLMs and highlight their strengths, limitations, and potential societal risks. Our findings indicate that current VLMs perform poorly on generic street-level images yet achieve notably high accuracy (61\%) on images resembling social media content, raising significant and urgent privacy concerns.
Urban-rural integration is a necessary way to realize the strategy of rural revitalization in the new era, and exploring the characteristics of the spatial correlation network of China's urban-rural integration and its formation mechanism is important for developing a comprehensive understanding of the spatial transmission mechanism of interregional urban-rural integration and provides a new policy perspective for the synergistic enhancement of the urban-rural integration level in each province. Based on China's interprovincial panel data from 2001 to 2021, the entropy method was applied to measure China's interprovincial urban-rural integration level, and the modified gravity model and social network analysis were used to explore the structural characteristics of China's spatial correlation network of urban-rural integration and its influencing factors. The results revealed the following: 1) During the study period, the level of China's interprovincial urban-rural integration showed an increasing trend, but the process of improvement was slow. Spatially, it showed a decreasing gradient from east to west, but the gap between the interprovincial urban-rural integration level gradually narrowed, with obvious spatial non-equilibrium. 2) China's urban-rural integration spatial association network became increasingly dense, complex, and close, and the main linkage flows of urban-rural integration occurred between geographically neighboring provinces and cities, such as Shanghai-Jiangsu, Shanghai-Zhejiang, and Beijing-Tianjin. The network connection was heterogeneous, and the network showed the characteristic of growth. However, the spatial connection of urban-rural integration did not reach the best level, and there is still much room for improvement. 3) Geospatial proximity, differences in the level of economic development, and urbanization had a significant positive effect on the optimization and evolution of the spatial linkage network of urban-rural integration, while differences in agricultural modernization and differences in advanced industrial structure showed a stage-by-stage effect over time. Differences in the scale of financial services did not have a significant effect. Accordingly, this study proposes countermeasures to optimize the spatial correlation network of urban-rural integration in China. Theoretically, it helps to deepen knowledge of the spatial effect and regional transmission of urban-rural integration in the geographical perspective. In practice, optimizing interprovincial urban-rural interaction and promoting cross-regional urban-rural integration development in a scientific and rational way are highly significant.
Economic growth results from countries' accumulation of organizational and technological capabilities. The Economic and Product Complexity Indices, introduced as an attempt to measure these capabilities from a country's basket of exported products, have become popular to study economic development, the geography of innovation, and industrial policies. Despite this reception, the interpretation of these indicators proved difficult. Although the original Method of Reflections suggested a direct interconnection between country and product metrics, it has been proved that the Economic and Product Complexity Indices result from a spectral clustering algorithm that separately groups similar countries or similar products, respectively. This recent approach to economic and product complexity conflicts with the original one and treats separately countries and products. However, building on previous interpretations of the indices and the recent evolution in spectral clustering, we show that these indices simultaneously identify two co-clusters of similar countries and products. This viewpoint reconciles the spectral clustering interpretation of the indices with the original Method of Reflections interpretation. By proving the often neglected intimate relationship between country and product complexity, this approach emphasizes the role of a selected set of products in determining economic development while extending the range of applications of these indicators in economics.
Massive Open Online Courses (MOOCs) have significantly enhanced educational accessibility by offering a wide variety of courses and breaking down traditional barriers related to geography, finance, and time. However, students often face difficulties navigating the vast selection of courses, especially when exploring new fields of study. Driven by this challenge, researchers have been exploring course recommender systems to offer tailored guidance that aligns with individual learning preferences and career aspirations. These systems face particular challenges in effectively addressing the ``cold start'' problem for new users. Recent advancements in recommender systems suggest integrating large language models (LLMs) into the recommendation process to enhance personalized recommendations and address the ``cold start'' problem. Motivated by these advancements, our study introduces RAMO (Retrieval-Augmented Generation for MOOCs), a system specifically designed to overcome the ``cold start'' challenges of traditional course recommender systems. The RAMO system leverages the capabilities of LLMs, along with Retrieval-Augmented Generation (RAG)-facilitated contextual understanding, to provide course recommendations through a conversational interface, aiming to enhance the e-learning experience.
Given the importance of ancient Chinese in capturing the essence of rich historical and cultural heritage, the rapid advancements in Large Language Models (LLMs) necessitate benchmarks that can effectively evaluate their understanding of ancient contexts. To meet this need, we present AC-EVAL, an innovative benchmark designed to assess the advanced knowledge and reasoning capabilities of LLMs within the context of ancient Chinese. AC-EVAL is structured across three levels of difficulty reflecting different facets of language comprehension: general historical knowledge, short text understanding, and long text comprehension. The benchmark comprises 13 tasks, spanning historical facts, geography, social customs, art, philosophy, classical poetry and prose, providing a comprehensive assessment framework. Our extensive evaluation of top-performing LLMs, tailored for both English and Chinese, reveals a substantial potential for enhancing ancient text comprehension. By highlighting the strengths and weaknesses of LLMs, AC-EVAL aims to promote their development and application forward in the realms of ancient Chinese language education and scholarly research. The AC-EVAL data and evaluation code are available at https://github.com/yuting-wei/AC-EVAL.
Philip Wootaek Shin, Jihyun Janice Ahn, Wenpeng Yin
et al.
It has been shown that many generative models inherit and amplify societal biases. To date, there is no uniform/systematic agreed standard to control/adjust for these biases. This study examines the presence and manipulation of societal biases in leading text-to-image models: Stable Diffusion, DALL-E 3, and Adobe Firefly. Through a comprehensive analysis combining base prompts with modifiers and their sequencing, we uncover the nuanced ways these AI technologies encode biases across gender, race, geography, and region/culture. Our findings reveal the challenges and potential of prompt engineering in controlling biases, highlighting the critical need for ethical AI development promoting diversity and inclusivity. This work advances AI ethics by not only revealing the nuanced dynamics of bias in text-to-image generation models but also by offering a novel framework for future research in controlling bias. Our contributions-panning comparative analyses, the strategic use of prompt modifiers, the exploration of prompt sequencing effects, and the introduction of a bias sensitivity taxonomy-lay the groundwork for the development of common metrics and standard analyses for evaluating whether and how future AI models exhibit and respond to requests to adjust for inherent biases.
Hyunuk Kim, Marcus J. Hamilton, Woo-Sung Jung
et al.
All human societies present unique narratives that shape their customs and beliefs. Despite cultural differences, some symbolic elements (e.g., heroes and tricksters) are common across many cultures. Here, we reconcile these seemingly contradictory aspects by analyzing mythological themes and traditions at various scales. Our analysis revealed that global mythologies exhibit both geographic and thematic nesting across different scales, manifesting in a layered structure. The largest geographic clusters correspond to the New and Old Worlds, which further divide into smaller bioregions. This hierarchical manifestation closely aligns with historical human migration patterns at a large scale, suggesting that narrative themes were carried through deep history. At smaller scales, the correspondence with bioregions indicates that these themes are locally adapted and diffused into variations across cultures over time. Our approach, which treats myths and traditions as random variables without considering factors like geography, history, or story lineage, suggests that the manifestation of mythology has been well-preserved over time and thus opens exciting research avenues to reconstruct historical patterns and provide insight into human cultural narratives.
Hasanuddin Chaer, Zulkarnaen Zulkarnaen, Mari'i Mari'i
et al.
This article aims to interpret two aspects of the Kitab Sawi: its grammatical structure and psychological meaning, uncovering both linguistic and psychological dimensions within this Sufi literary text. The subject of this study is the Sufi manuscript Kitab Sawi. To achieve this goal, the article adopts Schleiermacher's hermeneutic theory with a descriptive-analytical approach. The research follows four stages: data collection, data processing, interpretation, and drawing conclusions. The study reveals that the Kitab Sawi manuscript focuses on two main aspects: grammatical structure and psychological meaning. Key insights from the Kitab Sawi, which serves as the primary source for this article, include the following: a) Source of Ethical-Religious Literary Knowledge: The Kitab Sawi manuscript contains strong elements of Sufi literature, aimed at conveying moral and ethical values rooted in religious teachings. b) The Importance of Cultural Heritage: Kitab Sawi is not only valuable as a literary work but also as a representation of past intellectual and spiritual wealth. This manuscript plays a vital role in preserving and communicating cultural and spiritual teachings across generations. c) Psychological and Structural Significance: The analysis demonstrates that Kitab Sawi features a profound and intricate grammatical structure, characteristic of Sufi thought, which frequently employs symbolic and metaphorical language. This structure adds a psychological depth, allowing readers to undergo spiritual transformation through contemplation and engagement with the text. In conclusion, this article offers valuable contributions to the understanding of Sufi literature and Sasak cultural heritage.
Johnathan D. Holman, Dorivar A. Ruiz Diaz, Augustine K. Obour
et al.
Abstract Identifying the limiting nutrient, fertilizer source, rate, placement, additives, and timing of application are critical components of fertilizer management. The objective of this study was to quantify the impact of nitrogen (N) fertilizer source, rate, placement method, additives, application timing, and environment on yields of grain sorghum, forage sorghum [Sorghum bicolor (L.) Moench], and corn (for grain, Zea mays L.). Independent field experiments were conducted at 13 different environments in Kansas from 2008 through 2013 on grain sorghum, forage sorghum, and corn. Treatments were an incomplete factorial combination of four fertilizer placement methods, three fertilizer types, five fertilizer additives, three fertilizer application times, and six fertilizer rates that varied by location and across years. Results showed grain and forage sorghum yields responded to N fertilizer in environments that were not extremely dry (<136 mm) or wet (>651 mm). Corn responded to N fertilizer application only in high‐precipitation environments. For grain sorghum, where rate × placement × source × additive interaction was significant, broadcast application of urea (source) at high rates (67–134 kg N ha−1), with summer application timing, or with additive in winter (with environmentally smart nitrogen [ESN]) resulted in up to 43% greater yield compared with application of urea‐ammonium nitrate (UAN; source) and surface band (placement) at 67 kg ha−1 without additives. In the one site‐year where forage sorghum responded to fertilizer application, forage yields with preplant application of UAN at 56–140 kg ha−1 were 164% greater than the control. For corn, application of either urea or UAN fertilizer, UAN in coulter or surface band, with ESN blend, applied at planting, and at highest rates (160 kg ha−1) resulted in best yields compared with the alternatives and 110% greater yield compared with the unfertilized control. We concluded that fertilizer rate is an important management component as it consistently affected yield regardless of crop considered. Fertilizer placement and timing have crop‐specific importance as they were significant for only corn, but the main effect of additives (N stabilizers) was not significant for any of the crops. Environment and crop type influenced crop response to N fertilizer rate, timing, placement, and additives.
A. V. Shevchenko, A. A. Seleznev, G. P. Malinovsky
et al.
The aim of this study is to provide a tool to assess sediment production in an urban area. The urban environment is affected by a variety of anthropogenic and natural factors that, in particular, lead to the sediment production. The storage of sediments in the urban landscape negatively affects the quality of the urban environment. The model was developed on the basis of landscape studies conducted in residential areas of six Russian cities. The model takes into account (1) the influence of precipitation, spring snowmelt, and vehicles, (2) the influence of erosion factors for two seasons: warm (t>5°C) and cold (t<5°C), and (3) the presence of disturbed surfaces. The application of the developed model to Ekaterinburg city conditions returned sediment production equal to 1.2 kg/m2/y. A comparison of seasonal values shows that sediment production in cold season is 2.5 times higher than in the warm season. In the absence of the disturbed surfaces, sediment production decreases to 0.44 kg/m2/y. Modeling showed a correlation between sediment production in Russian cities and duration of the cold season. The efficiency of various urban area maintenance practices and cleaning measures were evaluated in terms of sediment production and storage. The developed model presented in this paper is based on research in Russian cities, but can be applied to assess the formation of sediment and measures to reduce the value of its accumulation in the urban environment in different regions of the world.
Underwater imagery often exhibits distorted coloration as a result of light-water interactions, which complicates the study of benthic environments in marine biology and geography. In this research, we propose an algorithm to restore the true color (albedo) in underwater imagery by jointly learning the effects of the medium and neural scene representations. Our approach models water effects as a combination of light attenuation with distance and backscattered light. The proposed neural scene representation is based on a neural reflectance field model, which learns albedos, normals, and volume densities of the underwater environment. We introduce a logistic regression model to separate water from the scene and apply distinct light physics during training. Our method avoids the need to estimate complex backscatter effects in water by employing several approximations, enhancing sampling efficiency and numerical stability during training. The proposed technique integrates underwater light effects into a volume rendering framework with end-to-end differentiability. Experimental results on both synthetic and real-world data demonstrate that our method effectively restores true color from underwater imagery, outperforming existing approaches in terms of color consistency.
Twisted bilayer graphene (TBLG) subject to a sequence of commensurate external periodic potentials reveals the formation of moiré fractals (MF) that share striking similarities with the central place theory (CPT) of economic geography, thus uncovering a remarkable connection between twistronics and the geometry of economic zones. MFs arise from the self-similarity of the emergent hierarchy of Brillouin zones (BZ), forming a nested subband structure within the bandwidth of the original moiré bands. We derive the fractal generators (FG) for TBLG under these external potentials and explore their impact on the hierarchy of the BZ edges and the wavefunctions at the Dirac point. By examining realistic super-moiré structures (SMS) and demonstrating their equivalence to MFs with periodic perturbations under specific conditions, we establish MFs as a general description for such systems. Furthermore, we uncover parallels between the modification of the BZ hierarchy and magnetic BZ formation in Hofstadter's butterfly (HB), allowing us to construct an incommensurability measure for MFs \textit{vs.} twist angle. The resulting bandstructure hierarchy bolsters correlation effects, pushing more bands within the same energy window for both commensurate and incommensurate TBLG.
Mirella Kanerva, Nguyen Minh Tue, Tatsuya Kunisue
et al.
The Atlantic salmon (Salmo salar) population in the Baltic Sea consists of wild and hatchery-reared fish that have been released into the sea to support salmon stocks. During feeding migration, salmon migrate to different parts of the Baltic Sea and are exposed to various biotic and abiotic stressors, such as organohalogen compounds (OHCs). The effects of salmon origin (wild or hatchery-reared), feeding area (Baltic Main Basin, Bothnian Sea, and Gulf of Finland), and OHC concentration on the differences in hepatic proteome of salmon were investigated. Multi-level analysis of the OHC concentration, transcriptome, proteome, and oxidative stress biomarkers measured from the same salmon individuals were performed to find the key variables (origin, feeding area, OHC concentrations, and oxidative stress) that best account for the differences in the transcriptome and proteome between the salmon groups. When comparing wild and hatchery-reared salmon, differences were found in xenobiotic and amino acid metabolism-related pathways. When comparing salmon from different feeding areas, the amino acid and carbohydrate metabolic pathways were notably different. Several proteins found in these pathways are correlated with the concentrations of polychlorinated biphenyls (PCBs). The multi-level analysis also revealed amino acid metabolic pathways in connection with PCBs and oxidative stress variables related to glutathione metabolism. Other pathways found in the multi-level analysis included genetic information processes related to ribosomes, signaling and cellular processes related to the cytoskeleton, and the immune system, which were connected mainly to the concentrations of Polychlorinated biphenyls and Dichlorodiphenyltrichloroethane and their metabolites. These results suggest that the hepatic proteome of salmon in the Baltic Sea, together with the transcriptome, is more affected by the OHC concentrations and oxidative stress of the feeding area than the origin of the salmon.
Abstract The recent polymorphism of state intervention and attendant political geographies have been interpreted as a return of state capitalism. While commentators across the social sciences have offered competing characterizations of the new state capitalism, little attention has been dedicated to how narratives and geographical imaginaries of the new state capitalism operate as a form of geopolitical knowledge and practice. Drawing upon critical geopolitics, we make three main arguments. First, we examine the context of wider geopolitical and geo-economic shifts in which the social construction of the geo-category has happened. We contend that the emerging new spatiality of the global economy has prompted the need for new discursive frames and geopolitical lines of reasoning. Second, we argue that this need is fulfilled by the geo-category state capitalism, which acts as a powerful tool in categorizing and hierarchizing the spaces of world politics. It does so by reinstituting a simple narrative of competition between two easily identifiable protagonists – (Western) democratic free-market capitalism and its deviant ‘other’, (Eastern) authoritarian state capitalism – and by reactivating older geopolitical grand narratives. Third, the geo-category state capitalism discursively enables Western business and state actors to justify tougher policy stances in three areas: foreign policy; trade, technology, and investment regulation; and international development.
The concept of nimbers--a.k.a. Grundy-values or nim-values--is fundamental to combinatorial game theory. Nimbers provide a complete characterization of strategic interactions among impartial games in their disjunctive sums as well as the winnability. In this paper, we initiate a study of nimber-preserving reductions among impartial games. These reductions enhance the winnability-preserving reductions in traditional computational characterizations of combinatorial games. We prove that Generalized Geography is complete for the natural class, $\cal{I}^P$ , of polynomially-short impartial rulesets under nimber-preserving reductions, a property we refer to as Sprague-Grundy-complete. In contrast, we also show that not every PSPACE-complete ruleset in $\cal{I}^P$ is Sprague-Grundy-complete for $\cal{I}^P$ . By considering every impartial game as an encoding of its nimber, our technical result establishes the following striking cryptography-inspired homomorphic theorem: Despite the PSPACE-completeness of nimber computation for $\cal{I}^P$ , there exists a polynomial-time algorithm to construct, for any pair of games $G_1$, $G_2$ of $\cal{I}^P$ , a prime game (i.e. a game that cannot be written as a sum) $H$ of $\cal{I}^P$ , satisfying: nimber($H$) = nimber($G_1$) $\oplus$ nimber($G_2$).