牛津大学（University of Oxford），简称“牛津”（Oxford），位于英国牛津，世界顶尖的公立研究型大学，采用书院联邦制。其与剑桥大学并称为牛剑，是罗素大学集团成员，被誉为“金三角名校”和“G5超级精英大学”。牛津大学的具体建校时间已不可考，但有档案明确记载的最早的授课时间为1096年，之后在1167年因得到了英国王室的大力支持而快速发展。牛津大学是英语世界中最古老的大学，也是世界上现存第二古老的高等教育机构。 该校涌现了一批引领时代的科学巨匠，培养了大量开创纪元的艺术大师、国家元首，其中包括28位英国首相及数十位世界各国元首、政商界领袖。 牛津大学在数学、物理、医学、法学、商学等多个领域拥有崇高的学术地位及广泛的影响力，被公认为是当今世界最顶尖的高等教育机构之一。从1902年起，牛津大学还设立了面向全世界本科生的“罗德奖学金”。 截止至2019年3月，牛津大学的校友、教授及研究人员中，共有72位诺贝尔奖得主（世界第九）、3位菲尔兹奖得主（世界第二十）、6位图灵奖得主（世界第九）。2020-2021年度，牛津大学位列2021泰晤士高等教育世界大学排名第1，2022QS世界大学排名第2，2021U.S. News世界大学排名第5，2020软科世界大学学术排名第9 。特别的是，牛津大学于2017-2021年连续五年在泰晤士高等教育世界大学排名荣膺世界第1。
Photocatalytic overall water splitting (POWS) is a developing method of green hydrogen production that utilises light as an energy source. Lanthanide tantalum perovskite oxynitrides (LnTaON2, Ln = lanthanide) are promising photocatalyst candidates as their valence and conduction band positions are suitable for POWS and their small bandgaps (1.8-2.4 eV) allow for utilisation of visible light. For lanthanides other than lanthanum, LnTaON2 perovskites remain underexplored for photocatalytic water splitting, with little work undertaken to optimise their photocatalytic activity and efficiency. In contrast, many successful synthetic variations have been explored to improve the photocatalytic activity of LaTaON2, including solid solution formation, morphology control and heterojunction formation. An overview of current work into non-lanthanum LnTaON2 photocatalysts is provided, along with the successful strategies utilised to improve the photocatalytic performance of LaTaON2, providing a toolbox of techniques to apply to non-lanthanum LnTaON2 photocatalysts to improve their photocatalytic performance in the future.
We give an exact characterization of monotone graph classes G with bounded average degree that satisfy the following property: The domination number of every graph from G is bounded by a linear function of its 2-independence number. In particular, we show that the class of all graphs of maximum average degree at most d has this property exactly if d<6. We also show that for any monotone class of graphs C with bounded average degree that does not contain the 1-subdivisions of all graphs, the domination number of every graph from C is bounded by a function of its 2-independence number, but there are such classes where this function cannot be linear.
In liquid-fuel detonation systems, such as rotating detonation engines (RDEs), the propagation of detonation waves often occurs in an inhomogeneous two-phase environment where fuel droplets and air coexist. High-resolution numerical simulations are performed to investigate the one-dimensional propagation dynamics of liquid-fuel detonation across air layers in n-heptane droplet-air mixtures. A coupled Euler-Lagrange framework is employed, in which the gaseous phase is solved using the reactive Navier-Stokes equations with a skeletal mechanism, while droplet dynamics, evaporation, and phase interactions are tracked in a Lagrangian manner. The influences of air-layer length, fuel-section length, and droplet diameter on detonation structure are systematically examined. Four distinct propagation modes are identified as the detonation wave traverses the air layer: high-intensity shock-detonation, low-intensity shock-detonation, low-intensity shock-deflagration, and low-intensity shock-delayed deflagration. The transitions between these regimes are primarily associated with the strength of the detonation wave across the air layer, the induction time and detonation initiation time of the n-heptane droplets/air mixture. A theoretical model is developed to predict the regime transitions, demonstrating that the propagation behavior is primarily controlled by the phase-velocity differential that determines the droplet evaporation rate. This study provides fundamental insights into the dynamics of liquid-fuel detonation propagation in non-uniform media and offers a framework for modelling detonation-phase interactions in multiphase reactive flows.
The increasing integration of distributed energy resources (DERs) is crucial for power system decarbonization, yet unlocking DERs' flexibility is challenged by their inherent uncertainties and modelling complexity. As traditional optimization methods struggle with such uncertainty and complexity of DERs, reinforcement learning (RL) has emerged as a promising alternative for DER management. However, standard RL methods suffer from sample inefficiency and sub-optimality when trained from scratch. Inspired by the training paradigms in large language models, this paper proposes a Supervised Reinforcement Learning (SRL) framework for learning DER coordination policies. This framework first pre-trains a policy on demonstration data in a supervised-learning fashion, which is then further fine-tuned using RL. Furthermore, we propose a two-step fine-tuning process: offline fine-tuning for enhancing policy performance and online fine-tuning for adapting it to the real-world dynamics. Experiments demonstrate that RL implementations based on the proposed framework significantly outperform all benchmarks, achieving high cost efficiency even under low-quality demonstration data.
In concluding the second series of eight articles in the double special issue on 'Engaging with communication and language in relation to mathematics and its education', we take the opportunity to reflect across the articles to prompt us all to consider potential and emerging directions for future research. This final article is overall a commentary inspired by the communication-centred and language-centred contributions to mathematics education across the entire collection of articles, that is, all those comments that situate mathematics and its education as language and communication praxis. We explore this relationship between mathematics and its education with communication and language through three themes: time, theories and methodologies, and the nuances of the term 'relation' in the title of this double special issue.
