87191 research outputs found
Sort by
Direct Nonparametric Predictive Inference Classification Trees
Get PDFClassification is the task of assigning a new instance to one of a set of predefined categories based on the attributes of the instance. A classification tree is one of the most commonly used techniques in the area of classification. In this paper, we introduce a novel classification tree algorithm which we call Direct Nonparametric Predictive Inference (D-NPI) classification algorithm. The D-NPI algorithm is completely based on the Nonparametric Predictive Inference (NPI) approach, and it does not use any other assumptions. NPI is a statistical methodology which learns from data in the absence of prior knowledge and uses only few modelling assumptions, enabled by the use of lower and upper probabilities to quantify uncertainty. Due to the predictive nature of NPI, it is well suited for classification, as the nature of classification is explicitly predictive as well. The D-NPI algorithm uses a new split criterion called Correct Indication (CI). CI reflects how informative attribute variables are, hence if the attribute variable is very informative, it gives high NPI lower and upper probabilities for CI. In addition, CI reports the strength of the evidence that the attribute variables will indicate regarding the possible class state for future instances, based on the data. To demonstrate its real-world applicability, the D-NPI algorithm is tested on benchmark data sets from various domains obtained from the UCI machine learning repository. The performance of the D-NPI classification algorithm is tested against several other classification algorithms using classification accuracy, in-sample accuracy and tree size. The experimental results indicate that the D-NPI classification algorithm performs well and tends to slightly outperform the other classification algorithms
Decentralised consensus-based hierarchical control of a hybrid AC-DC microgrid robust to communication delays
No full textThe continuous growth of distributed renewable energy resources in medium and low-voltage networks supports the case for a distributed architecture for electricity supply relying on microgrids. Whereas decentralised architectures for the primary and secondary control layers of the hierarchical control of microgrids have already been proposed, the tertiary control (economic dispatch) has been mainly formulated as a centralised control or as a distributed control that relies on a central agent to coordinate the operation between distributed generators. This paper proposes a fully decentralised hierarchical microgrid control based on a consensus-based economic dispatch problem. Unlike the proposals found in the literature, the convergence of the proposed tertiary control layer is robust to communication delays and the secondary control layer applies an incremental formulation for a seamless integration with the tertiary layer. The robustness and performance of the proposed hierarchical control are compared with those of a conventional centralised approach and a previously published decentralised approach. The proposal is validated using detailed non-linear real-time simulations, and its stability is proved by modal analysis and using Lyapunov functions
Sparcs – combining radiation hydrodynamics with non-equilibrium metal chemistry in the Swift astrophysical code
No full textWe present Sparcs, which combines the moment-based radiative transfer method Sph-m1rt with the non-equilibrium metal chemistry solver Chimes in the modern highly parallel astrophysical code Swift. Sparcs enables on-the-fly radiation hydrodynamic simulations, with multifrequency ultraviolet radiative transfer coupled with all ionization states of 11 major elements, in the presence of dust, cosmic ray ionization and heating, and self-gravity. Direct radiation pressure on gas and dust is also accounted for. We validate Sparcs against analytic solutions and standard photoionization codes such as cloudy in idealized tests. As an example application, we simulate an ionization front propagating through an inhomogeneous interstellar medium with solar metallicity. We produce mock optical emission line observations with the level population calculation code HyLight and the diagnostic radiative transfer code Radmc3D. We find that non-equilibrium effects and inhomogeneities can boost the low ion fractions by up to an order of magnitude. Possible applications of Sparcs include studying the dynamical impact of radiation on gas in star-forming regions, and in the interstellar, intergalactic, and circumgalactic medium, as well as interpreting line diagnostics in such environments, and galactic or active galactic nucleus outflows
CFT dual to gravitational non-locality in string theory
No full textWe investigate non-perturbative bulk corrections arising from instantons in string theory and M-theory. By deriving non-local curvature corrections of the form e-γRRμνRμν, we demonstrate how these modifications emerge from wrapped brane instantons and their summation over multi-instanton configurations. Utilizing holographic techniques, we establish a direct connection between these non-perturbative effects and large-N gauge theories, identifying the appropriate holographic dual conformal field theory (CFT). We further analyze this connection through resummations in the large-N expansion. Additionally, we study black hole solutions in AdS backgrounds and show that these instanton-induced corrections significantly modify the near-horizon geometry. Finally, we explore the regularization of curvature singularities via these exponential damping terms, providing a natural resolution mechanism in quantum gravity. Our findings underscore the fundamental role of non-perturbative physics in shaping the structure of spacetime and its holographic duals
Caste and Credit: Discrimination in India’s Credit Sector
No full textThis article examines how caste shapes access to credit in India’s formal and informal lending markets. Using nationally representative data from the India Human Development Survey (2011–2012), we analyse loan application rates and loan amounts and compare outcomes between General Castes (GC) and three lower-caste groups: Other Backward Castes (OBC), Scheduled Castes (SC) and Scheduled Tribes (ST). We find that GC households are more likely to apply for and receive larger loans from formal banks, while lower-caste households rely more heavily on informal moneylenders. A substantial share of these credit gaps—particularly in bank lending—remains unexplained by observable characteristics, pointing to potential caste-based discrimination. In contrast, moneylenders do not appear to penalize lower-caste borrowers to the same extent and, in some cases, lend more than expected to OBC households. These findings suggest that entrenched caste hierarchies continue to influence credit access in India, with formal institutions reinforcing rather than correcting social inequalities
Neural reranking for UK statutory retrieval: Provision-level evaluation and an open distilled model
No full textThis work explores provision-level retrieval and neural reranking for UK primary and secondary legislation. We introduce UK-StatuteCorpus, a corpus of recent UK Acts and statutory instruments from legislation.gov.uk, together with a 100-query evaluation set of practitioner-style questions whose graded relevance judgements distinguish legally operative, supporting and contextual provisions. Using BM25 and an MPNet-based dense retriever to build candidate sets, we evaluate ten neu-ral rerankers, including transformer cross-encoders, a late-interaction reranker, an LLM-based listwise reranker and proprietary APIs. Across both sparse and dense pools, neural reranking consistently improves normalized Discounted Cumulative Gain (nDCG) and Mean Reciprocal Rank (MRR) over first-stage retrieval. We further distil a proprietary Voyage reranker into a ModernBERT-based cross-encoder, Distilled-Voyage-ModernBERT, which approaches the teacher's effectiveness and outperforms other open rerankers on our benchmark. Results are based on 100 expert-validated queries, each linked to three graded provisions from a single UK instrument, so they characterise single-instrument, provision-level retrieval over recent UK legislation
Rebecca Leong, Living More Human(e)ly? Towards Better Approaches to Dating and Couple Relationships in Contemporary Anglo-American Contexts
No full textAssessing the effectiveness of the one paleopathology workshop
No full textLay Summary: This article describes the creation, content, and results of a One Paleopathology workshop held at the University of Durham (UK) in August 2024, prior to the annual meeting of the International Society of Evolutionary Medicine and Public Health (ISEMPH). The workshop goals included extending today’s One Health paradigm to include the past, fostering an appreciation of planetary health—past and present— and forging transdisciplinary relationships to address complex health challenges today. After the workshop, two participant surveys measured the effectiveness of the workshop in meeting these goals. The surveys suggest that researchers have sustained interest in this concept and will continue to extend their health-related research to reflect the integration of humans, animals, and the environment
A comparison of tool-use flexibility between captive chimpanzees and bonobos
No full textDespite chimpanzees and bonobos sharing close phylogenetic ties to humans, chimpanzees are the more common model species in multiple fields of comparative research. One reason for this bias is the variation in tool repertoire size observed between the two species. Previous studies have examined the factors driving this difference, but few have targeted flexibility in how tools are used. We studied bonobos and chimpanzees under similar conditions in captivity, thus excluding any ecological variation present in these species’ natural habitats. We examined whether the species differed in their ability to switch between tools, a trait that may facilitate tool innovation in primates. To do so, we provided the apes with a task that required switching tool type from a rigid stick to a bendable rope to forage successfully. Our data suggest that there are no significant differences in tool-use performance between chimpanzees and bonobos in captivity. However, we found significant differences in the species’ exploration tendencies. While chimpanzees fixed their attention on stick tools, bonobos switched their attention more easily towards the rope, potentially due to less functional fixedness. We also found significant within-species differences between institutions. These findings suggest that future research should disentangle intrinsic flexibility in exploration and account for institution and group level effects
Data sources:
Access Repository Dashboard
Do you manage Open Research Online? Become a CORE Member to access insider analytics, issue reports and manage access to outputs from your repository in the CORE Repository Dashboard! 👇