179566 research outputs found
Sort by
Introduction:The visual and material worlds of James VI and I
No full textThis introduction/editorial challenges the assumption that James VI and I was indifferent to art, showing instead how visual and material display underpinned authority, favour, diplomacy, and identity throughout his reign
A Langevin sampling algorithm inspired by the Adam optimizer
No full textWe present a framework for adaptive-stepsize MCMC sampling based on time-rescaled Langevin dynamics, in which the stepsize variation is dynamically driven by an additional degree of freedom. Our approach augments the phase space by an additional variable which in turn defines a time reparameterization. The use of an auxiliary relaxation equation allows accumulation of a moving average of a local monitor function and provides for precise control of the timestep while circumventing the need to modify the drift term in the physical system. Our algorithm is straightforward to implement and can be readily combined with any off-the-peg fixed-stepsize Langevin integrator. As a particular example, we consider control of the stepsize by monitoring the norm of the log-posterior gradient, which takes inspiration from the Adam optimizer, the stepsize being automatically reduced in regions of steep change of the log posterior and increased on plateaus, improving numerical stability and convergence speed. As in Adam, the stepsize variation depends on the recent history of the gradient norm, which enhances stability and improves accuracy compared to more immediate control approaches. We demonstrate the potential benefit of this method--both in accuracy and in stability--in numerical experiments including Neal's funnel and a Bayesian neural network for classification of MNIST data
Cluster Formation in Diffusive Systems
No full textIn this paper, we study the formation of clusters for stochastic interacting particle systems (SIPS) that interact through short-range attractive potentials in a periodic domain. We consider kinetic (underdamped) Langevin dynamics and focus on the low-friction regime. Employing a linear stability analysis for the kinetic McKean-Vlasov equation, we show that, at sufficiently low temperatures, and for sufficiently short-ranged interactions, the particles form clusters that correspond to metastable states of the mean-field dynamics. We derive the friction and particle-count dependent cluster-formation time and numerically measure the friction-dependent times to reach a stationary state (given by a state in which all particles are bound in a single cluster). By providing both theory and numerical methods in the inertial stochastic setting, this work acts as a bridge between cluster formation studies in overdamped Langevin dynamics and the Hamiltonian (microcanonical) limit
Anakinra restores immunological misfiring that drives influenza-associated pulmonary aspergillosis
No full textValid Conformal Prediction for Dynamic GNNs
Get PDFDynamic graphs provide a flexible data abstraction for modelling many sorts of real-world systems, such as transport, trade, and social networks. Graph neural networks (GNNs) are powerful tools allowing for different kinds of prediction and inference on these systems, but getting a handle on uncertainty, especially in dynamic settings, is a challenging problem. In this work we propose to use a dynamic graph representation known in the tensor literature as the unfolding, to achieve valid prediction sets via conformal prediction. This representation, a simple graph, can be input to any standard GNN and does not require any modification to existing GNN architectures or conformal prediction routines. One of our key contributions is a careful mathematical consideration of the different inference scenarios which can arise in a dynamic graph modelling context. For a range of practically relevant cases, we obtain valid prediction sets with almost no assumptions, even dispensing with exchangeability. In a more challenging scenario, which we call the semi-inductive regime, we achieve valid prediction under stronger assumptions, akin to stationarity. We provide real data examples demonstrating validity, showing improved accuracy over baselines, and sign-posting different failure modes which can occur when those assumptions are violated
Oscillatory integral operators and variable Schrödinger propagators: beyond the universal estimates
Get PDFWe consider a class of Hörmander-type oscillatory integral operators in \mathbb{R}^n for n \geq 3 odd with real analytic phase. We derive weak conditions on the phase which ensure L^p bounds beyond the universal p \geq 2 \cdot \frac{n+1}{n-1} range guaranteed by Stein's oscillatory integral theorem. This expands and elucidates pioneering work of Bourgain from the early 1990s. We also consider a closely related class of variable coefficient Schrödinger propagator-type operators, and show that the corresponding theory differs significantly from that of the Hörmander-type operators. The main ingredient in the proof is a curved Kakeya/Nikodym maximal function estimate. This is established by combining the polynomial method with certain uniform sublevel set estimates for real analytic functions. The sublevel set estimates are the main novelty in the argument and can be interpreted as a form of quantification of linear independence in the C^{\omega} category
Polycomb Repressive Complexes 1 and 2 are recruited independently to pericentromeric heterochromatin in response to hypomethylation in mouse embryonic stem cells
Get PDFPericentromeric heterochromatin (PCH) is delineated by the enrichment of repressive epigenetic modifications, specifically trimethylated histone H3 at lysine 9 (H3K9me3) and DNA methylation (5-methylcytosine), which establish and maintain a condensed, transcriptionally silenced chromatin state. Depletion of either H3K9me3 or DNA methylation in mouse embryonic stem cells (mESCs) induces a permissive chromatin configuration that permits de novo recruitment and deposition of normally excluded Polycomb Repressive Complexes 1 and 2 (PRC1 and PRC2), characterized by H2AK119ub1 and H3K27me3 modifications, respectively, at PCH. Here, we demonstrate that H2AK119ub1 and H3K27me3 are independently recruited to hypomethylated PCH using a doxycycline-inducible mESC model allowing modulation of Dnmt1 expression levels and catalytic activity. We further investigate the roles of proposed mediators of PRC1/2 targeting, including SCML2, BEND3, KDM2b, and TET enzymes, in this context, our findings indicate that neither PRC1 nor PRC2 recruitment at hypomethylated PCH depends on these factors. Additionally, our data suggest that the permissive chromatin environment resulting from DNA hypomethylation is the principal facilitator of Polycomb complex spreading, offering novel insights into the mechanisms governing epigenetic modifier dynamics and interactions during periods of DNA methylation reprogramming
Climate science for 2050
Get PDFKnowledge of the functioning of the climate system, including the physical, dynamical and biogeochemical feedback processes expected to occur in response to anthropogenic climate forcing, has increased substantially over recent decades. Today, climate science is at a crossroads, with new and urgent demands arising from the needs of society to deal with future climate change, and the need for the climate science community to refine its strategic goals to meet these demands rapidly. All possible—but currently unknown—worlds in 2050, with a larger global population, unprecedented climate conditions with higher temperatures, more frequent extreme weather events, sea level rise, disrupted ecosystems, changes in habitability and increased climate-induced displacement and migration, and the emergence of new geopolitical tensions, will require limiting society’s vulnerability both through mitigation measures to minimize further warming and through the implementation of innovative adaptation initiatives. The development of a skillful climate information system, based on the most advanced Earth system science, will be required to inform decision-makers and the public around the world about the local and remote impacts of climate change, and guide them in optimizing their adaptation and mitigation agendas. This information will also help manage renewable resources in a warmer world and strengthen resilience to the expected interconnected impacts of climate change. In this paper, we summarize the major advances needed to understand the multiscale dynamics of the Earth system. We highlight the need to develop an integrated information system accessible to decision-makers and citizens in all parts of the world, and present some of the key scientific questions that need to be addressed to inform decisions on mitigation and adaptation. Finally, we speculate about the values and ethics of climate science and the nature of climate research in a world that will be increasingly affected by global warming in a geopolitical context very different from that of recent decades.</p
The influence of metal cations on the formation of phthalocyanine/porphyrin heterodimers and nanoporous cocrystals
Get PDFPorphyrin-containing Phthalocyanine Nanoporous Crystals (PNCs) have been prepared previously by the cocrystallisation of octa(2′,6′-di-iso-propyphenoxy)phthalocyanine [(dipPhO)8Pc] and its metalated derivatives with metal-free tetraphenylporphyrin (TPP). Here we report the effect of the central cations (M = H2, Co, Cu, Zn) on heterodimer formation, both in solution and in cocrystallisations, for all sixteen possible combinations of (dipPhO)8PcM with TPPM. UV/visible spectroscopy shows a clear trend of Cu2+ > 2H+ > Co2+ > Zn2+ for encouraging or discouraging heterodimer formation, which can be rationalised by the propensity of the metal cation to bind to axial ligands when incorporated into porphyrin or phthalocyanine macrocycles. Attempted cocrystallisations of all 16 (dipPhO)8PcM/TPPM combinations yielded 13 heterodimer-based PNCs, with only (dipPhO)8PcZn failing to provide a cocrystal with TPPM (Co, Cu, Zn). This study further demonstrates the remarkable efficiency of (dipPhO)8PcM in directing the structure of PNC cocrystallisations. It suggests a simple method for placing two metal cations of choice within van der Waals distance of one another within a well-defined, predictable crystal structure.</p
A matter of perspective? Differences between adolescent-parent and parent-teacher pairs in responses to the Strengths and Difficulties Questionnaire using a Scottish national cohort study
Get PDFAlthough multiple-respondent scoring methods are increasingly recommended for youth mental health questionnaires, utilisation of parent-only responses remains common in survey research. The substantive, epistemological and methodological ramifications of this perspective gap remain under-explored despite the widespread adoption of youth psychometrics in social survey datasets. Modelling the impact of respondent pair identities on inter-respondent discrepancies in youth mental health questionnaires reveals “whose” responses differ and how measurement error may be patterned in single-respondent models. Comparing Goodman’s Strengths and Difficulties Questionnaire (SDQ) responses from parents, teachers and adolescents themselves, we apply latent difference score modelling to parent-adolescent (age 14, n=2,943) and parent teacher (age 10, n=1,833) pairs from the Growing Up in Scotland birth cohort study and present significant inter-respondent differences in behavioural perceptions between these groups. Higher levels of difficulties are associated with larger inter-respondent discrepancy levels. The impact of gender, housing tenure, finances,family composition, maternal mental health and education on score discrepancies vary in direction, magnitude and significance between SDQ behavioural components. Therefore, discrepancies depend upon characteristics of the each measured behaviour, not a global propensity to dis/agreement. Evaluating the implications of these findings, we advocate for the inclusion of youth self-reporting in survey datasets and discuss how to caveat research with the potential impact of respondent identities on missing perspectives when multiple respondent data are unavailable
Access Repository Dashboard
Do you manage Edinburgh Research Explorer? Access insider analytics, issue reports and manage access to outputs from your repository in the CORE Repository Dashboard!