Large loads are expanding rapidly across North America, led by data centers, cryptocurrency mining, hydrogen production facilities, and heavy-duty charging stations. Each class presents distinct electrical characteristics, but data centers are drawing particular attention as AI deployment drives unprecedented capacity growth. Their scale, duty cycles, and converter-dominated interfaces introduce new challenges for transmission interconnections, especially regarding disturbance behavior, steady-state performance, and operational visibility. This paper reviews best practices for large-load interconnections across North America, synthesizing utility and system operator guidelines into a coherent set of technical requirements. The approach combines handbook and manual analysis with cross-utility comparisons and an outlook on European directions. The review highlights requirements on power quality, telemetry, commissioning tests, and protection coordination, while noting gaps in ride-through specifications, load-variation management, and post-disturbance recovery targets. Building on these findings, the paper proposes practical guidance for developers and utilities.
Hallucination in large language models (LLMs) remains a significant challenge for their safe deployment, particularly due to its potential to spread misinformation. Most existing solutions address this challenge by focusing on aligning the models with credible sources or by improving how models communicate their confidence (or lack thereof) in their outputs. While these measures may be effective in most contexts, they may fall short in scenarios requiring more nuanced approaches, especially in situations where access to accurate data is limited or determining credible sources is challenging. In this study, we take North Korea - a country characterised by an extreme lack of reliable sources and the prevalence of sensationalist falsehoods - as a case study. We explore and evaluate how some of the best-performing multilingual LLMs and specific language-based models generate information about North Korea in three languages spoken in countries with significant geo-political interests: English (United States, United Kingdom), Korean (South Korea), and Mandarin Chinese (China). Our findings reveal significant differences, suggesting that the choice of model and language can lead to vastly different understandings of North Korea, which has important implications given the global security challenges the country poses.
Aasif Ahmad Mir, Nina Smirnova, Jeyshankar Ramalingam
et al.
The study aims to highlight the growth and development of Indo-German collaborative research over the past three decades. Moreover, this study encompasses an in-depth examination of funding acknowledgements to gain valuable insights into the financial support that underpins these collaborative endeavors. Together with this paper, we provide an openly accessible dataset of Indo-German research articles for further and reproducible research activities (the "Indo-German Literature Dataset"). The data were retrieved from the Web of Science (WoS) database from the year 1990 till the 30th of November 2022. A total of 36,999 records were retrieved against the employed query. Acknowledged entities were extracted using a NER model specifically trained for this task. Interrelations between the extracted entities and scientific domains, lengths of acknowledgement texts, number of authors and affiliations, number of citations, and gender of the first author, as well as collaboration patterns between Indian and German funders were examined. The study brings to light that Physics, Chemistry, Materials Science, Astronomy and Astrophysics, and Engineering prominently dominate the Indo-German collaborative research. The United States, followed by England and France, are the most active collaborators in Indian and German research. Additionally, relations between entity, entity type, and scientific domain, were discovered. The study highlights a deeper understanding of the composition of the Indo-German collaborative research landscape of the last 30 years and its significance in advancing scientific knowledge and fostering international partnerships. Furthermore, we provide an open version of the original WoS dataset. The Indo-German Literature Dataset consists of 22,844 articles from OpenAlex and is available for related studies like literature studies and Scientometrics.
Marcos Zampieri, Kai North, Tommi Jauhiainen
et al.
Language identification is an important first step in many IR and NLP applications. Most publicly available language identification datasets, however, are compiled under the assumption that the gold label of each instance is determined by where texts are retrieved from. Research has shown that this is a problematic assumption, particularly in the case of very similar languages (e.g., Croatian and Serbian) and national language varieties (e.g., Brazilian and European Portuguese), where texts may contain no distinctive marker of the particular language or variety. To overcome this important limitation, this paper presents DSL True Labels (DSL-TL), the first human-annotated multilingual dataset for language variety identification. DSL-TL contains a total of 12,900 instances in Portuguese, split between European Portuguese and Brazilian Portuguese; Spanish, split between Argentine Spanish and Castilian Spanish; and English, split between American English and British English. We trained multiple models to discriminate between these language varieties, and we present the results in detail. The data and models presented in this paper provide a reliable benchmark toward the development of robust and fairer language variety identification systems. We make DSL-TL freely available to the research community.
This study is devoted to the full-fledged introduction into scientific circulation of a unique early medieval iron war hatchet inlayed with silver originating from the southern steppe zone of the Dniester-Prut interfluve of present-day territory of Republic of Moldova. The ornamented battle axe of 10th–11th centuries was found near the village of Răzeni. The type of form of war axe from Răzeni is similar to type of war axes originated in the Middle Volga region, where there was a tradition of making similar ornamented war axes of this type. The main feature of the war ax from Răzeni is the Scandinavian ornament in Ringerike style. The semantics analysis of the ornamented motif and its interpretation — is one of the main tasks of this research work of the description and introduction of the unique find into scientific circulation. The image is interpreted as the tree of life Yggdrasil sprouting from the deck of the boat, which is worshiped by an anthropomorphic figure interpreted as a Viking. The present scientific work has widely direct and indirect analogies of Scandinavian origin from the Bronze to the Viking Age. The methodology of this research includes a deep and extensive comparative analysis of the motif elements on the surface of the war hatchet. The work makes a deep structural and semantic analysis in the broad context of ancient Germanic sources and previous scientific research on the worldview of the Vikings, reflected in their mythopoetic and mythological concepts and traditions. The ceremonial-battle axe from Răzeni was presumably made and ornamented in the Volga Bulgaria by order of a Viking-man, and then got in the borderland of Ancient Rus in the area of the Bugeac steppe of the Dniester-Prut interfluve of territory of present-day Republic of Moldova. The ceremonial-battle axe could have been brought in Bugeac steppe of the Dniester-Prut interfluve by means of circulation the international water arteries “from the Varangians to the Persians” along the Volga river, and “from the Varangians to the Greeks” along the Dniester river. The uniqueness of the ornamented war hatchet from Republic of Moldova is due to the fact that it is the first find of this type of axes in the Dniester-Prut interfluve of modern Republic of Moldova, and this axe no have direct analogies in eastern Europe. The early medieval ceremonial battle axe decorated with silver, contain the image of Old Norse cosmogonic representations in a stylized motif form on the surface of the hatchet. The unique item that captures the combination and interaction of different cultures at the junction of their intersection in the Medieval world.
Abstract. We propose a new concept of the Weichselian ice dynamics in the south-western sector of the Baltic Sea depression. The review of existing geochronological data from Germany, Denmark and southernmost Sweden in combination with new optically stimulated luminescence (OSL) data from the German Oder Lobe area is the basis for a reassessment and an improvement of previous ice dynamic models. Factors like the pre-existing topography, glaciotectonic features and the occurrence of till beds and inter-till deposits of varying origin are also taken into consideration for our process-based reconstruction of the sedimentary environments close to the ice margin and hence the ice dynamics of the Scandinavian Ice Sheet (SIS). During the early MIS 3 (marine isotope stage), the late MIS 3 and MIS 2, the SIS advanced into present-day terrestrial areas around the south-western Baltic Sea Basin. The first ice advance during the warming phase in early MIS 3 is poorly documented as the Ellund–Warnow Advance in Germany but may be correlated with the numerically dated Ristinge Advance in Denmark and Sweden. The late MIS 3 advance in contrast is reliably documented. It shaped the landforms of the Brandenburg Advance and the maximum Weichselian ice extent in the Oder Lobe area in north-eastern Germany and occurred contemporaneously with the Klintholm Advance in southern Sweden and Denmark. The lack of a corresponding till in various cliff profiles along the Baltic Sea coastline between southern Schleswig-Holstein and the island of Rügen can be explained by the distinct lobate structure of this ice advance, which was strongly guided by the pre-existing low-lying topography. We propose the horst of Bornholm, Denmark, acting as an ice divide, with ice-dammed lakes existing on the lee side between two glacier lobes. This lobate structure had not been considered in previous conceptual models, which led to seemingly conflicting chronological and stratigraphical interpretations. Our introduction of the lobate structure for the first time resolves these contradictions and integrates the data in a coherent model. The dynamics of the MIS 2 readvance to the Last Glacial Maximum (LGM) extent were clearly different to the previous advance and were most likely characterized by a more uniformly advancing ice front with a less lobate structure which also overrode the horst of Bornholm and the island of Rügen. This advance reached the maximum Weichselian ice extent in some parts of the south-western SIS, but, in the Oder Lobe area, it is proven to have terminated at a lesser extent than the early MIS 3 advance, but it did shape the most prominent morphological landform record of the last glacial cycle. In order to advance the reconstruction of Weichselian ice dynamics in the future, we strongly suggest using both an MIS-based terminology and a process-based approach in the interpretation of geochronological data to live up to the dynamic nature of continental ice sheets.
ABSTRACT The history and etymology of Old Scandinavian hinn is a disputed matter. One question concerns whether hinn as a contrastive demonstrative indicating ‘the other (one)/the former (one)’ and hinn as a pre-adjectival article, both of which to some extent are still found in present-day Icelandic, are related or not. Another issue concerns the fact that hinn has no immediate parallel in Germanic outside Scandinavia, which has led scholars to assume that it is a Proto-Scandinavian innovation. This paper argues that Old Scandinavian possessed two hinn words with separate backgrounds, one stemming directly from an anciently inherited distal demonstrative, and one from an innovated proximal demonstrative. However, the innovation was no more founded on common Germanic material than the former hinn was. Instead, it arose from the reinforcement of an ancient precursor. This precursor is traceable in early Icelandic enn, which was used as a pre-adjectival article and as a primitive post-nominal definiteness marker.
Michel Plüss, Lukas Neukom, Christian Scheller
et al.
We present the Swiss Parliaments Corpus (SPC), an automatically aligned Swiss German speech to Standard German text corpus. This first version of the corpus is based on publicly available data of the Bernese cantonal parliament and consists of 293 hours of data. It was created using a novel forced sentence alignment procedure and an alignment quality estimator, which can be used to trade off corpus size and quality. We trained Automatic Speech Recognition (ASR) models as baselines on different subsets of the data and achieved a Word Error Rate (WER) of 0.278 and a BLEU score of 0.586 on the SPC test set. The corpus is freely available for download.
Abstract The form fauþr ‘father’ on the Malt stone is normally understood as a carving error for faþur, but could very well be read at face value as a one-syllable form fǫðr, an archaic accusative singular. In a wider Proto-Germanic context, I propose that this form is part of an early levelling process of the kinship terms to one-syllable stem forms, an alternative paradigm co-existing with the classical hysterodynamic paradigm documented in the Gothic singular. This levelling takes place not only in the plural, but also in the oblique cases of the singular. In a Scandinavian context, this reading sheds light on a handful of seemingly aberrant forms.
This paper presents TextComplexityDE, a dataset consisting of 1000 sentences in German language taken from 23 Wikipedia articles in 3 different article-genres to be used for developing text-complexity predictor models and automatic text simplification in German language. The dataset includes subjective assessment of different text-complexity aspects provided by German learners in level A and B. In addition, it contains manual simplification of 250 of those sentences provided by native speakers and subjective assessment of the simplified sentences by participants from the target group. The subjective ratings were collected using both laboratory studies and crowdsourcing approach.
David Skillicorn, Olivier Walther, Quan Zheng
et al.
This paper explores the spatial and temporal diffusion of political violence in North and West Africa. It does so by endeavoring to represent the mental landscape that lives in the back of a group leader's mind as he contemplates strategic targeting. We assume that this representation is a combination of the physical geography of the target environment, and the mental and physical cost of following a seemingly random pattern of attacks. Focusing on the distance and time between attacks and taking into consideration the transaction costs that state boundaries impose, we wish to understand what constrains a group leader to attack at a location other than the one that would seem to yield the greatest overt payoff. By its very nature, the research problem defies the collection of a full set of structural data. Instead, we leverage functional data from the Armed Conflict Location and Event Data project (ACLED) dataset that, inter alia, meticulously catalogues violent extremist incidents in North and West Africa since 1997, to generate a network whose nodes are administrative regions. These nodes are connected by edges of qualitatively different types: undirected edges representing geographic distance, undirected edges representing borders, and directed edges representing consecutive attacks by the same group at the two endpoints. We analyze the resulting network using novel spectral embedding techniques that are able to account fully for the different types of edges. The result is a map of North and West Africa that depicts the permeability to violence. A better understanding of how location, time, and borders condition attacks enables planning, prepositioning, and response.
We have recently published observations of significant flux density variations at 1.3 cm in HII regions in the star forming regions Sgr B2 Main and North (De Pree et al. 2014). To further study these variations, we have made new 7 mm continuum and recombination line observations of Sgr B2 at the highest possible angular resolution of the Karl G. Jansky Very Large Array (VLA). We have observed Sgr B2 Main and North at 42.9 GHz and at 45.4 GHz in the BnA configuration (Main) and the A configuration (North). We compare these new data to archival VLA 7 mm continuum data of Sgr B2 Main observed in 2003 and Sgr B2 North observed in 2001. We find that one of the 41 known ultracompact and hypercompact HII regions in Sgr B2 (K2-North) has decreased $\sim$27% in flux density from 142$\pm$14 mJy to 103$\pm$10 mJy (2.3$σ$) between 2001 and 2012. A second source, F3c-Main has increased $\sim$30% in flux density from 82$\pm$8 mJy to 107 $\pm$11 mJy (1.8$σ$) between 2003 and 2012. F3c-Main was previously observed to increase in flux density at 1.3 cm over a longer time period between 1989 and 2012 (De Pree et al. 2014). An observation of decreasing flux density, such as that observed in K2-North, is particularly significant since such a change is not predicted by the classical hypothesis of steady expansion of HII regions during massive star accretion. Our new observations at 7 mm, along with others in the literature, suggest that the formation of massive stars occurs through time-variable and violent accretion.
We describe the final product of the North Celestial Cap Survey (NCC Survey, NCCS) - the optical-UV-IR merged catalogue for the region within 10 deg of the North Celestial Pole. The NCC region at DEC > 80 deg is poorly covered by modern CCD-based surveys. The optical part of the survey was observed in V, R and I with the Wise Observatory telescopes and was merged with GALEX UV and WISE IR data, producing the catalogue. More than four million objects were observed in at least one optical band. The final catalogue contains ~1.6 million sources observed in all three optical bands, of which some 1.4 million have WISE counterparts and ~300,000 have GALEX counterparts. The astrometric accuracy of the optical NCCS data, derived from a comparison with the UCAC3 catalogue, is better than 0.2 arcsec and the photometry, when compared with SDSS, is good to ~0.15 mag for sources brighter than V = 20.3, R = 21.0 and I = 19.2 mag. The SDSS point-extended source separation is reproduced with >92% efficiency.