IA Scholar Query: Formally verified asymptotic consensus in robust networks.
https://scholar.archive.org/
Internet Archive Scholar query results feedeninfo@archive.orgTue, 29 Nov 2022 00:00:00 GMTfatcat-scholarhttps://scholar.archive.org/help1440PurdueThesis_XuejunZhao
https://scholar.archive.org/work/j4izygnldbgxfhthvie3jdiony
This study examines data-driven contract design in the small data regime and large data regime respectively, and the implications from contract pricing in the pharmaceutical supply chain.Xuejun Zhaowork_j4izygnldbgxfhthvie3jdionyTue, 29 Nov 2022 00:00:00 GMTConditional Gradient Methods
https://scholar.archive.org/work/b2imrksvmfclhaik7ghfh6bcte
The purpose of this survey is to serve both as a gentle introduction and a coherent overview of state-of-the-art Frank--Wolfe algorithms, also called conditional gradient algorithms, for function minimization. These algorithms are especially useful in convex optimization when linear optimization is cheaper than projections. The selection of the material has been guided by the principle of highlighting crucial ideas as well as presenting new approaches that we believe might become important in the future, with ample citations even of old works imperative in the development of newer methods. Yet, our selection is sometimes biased, and need not reflect consensus of the research community, and we have certainly missed recent important contributions. After all the research area of Frank--Wolfe is very active, making it a moving target. We apologize sincerely in advance for any such distortions and we fully acknowledge: We stand on the shoulder of giants.Gábor Braun, Alejandro Carderera, Cyrille W. Combettes, Hamed Hassani, Amin Karbasi, Aryan Mokhtari, Sebastian Pokuttawork_b2imrksvmfclhaik7ghfh6bcteFri, 25 Nov 2022 00:00:00 GMTExploring the Nonlinear Relationship between Renewable Energy Consumption and Economic Growth in the Context of Global Climate Change
https://scholar.archive.org/work/mfakyprkrjhxfe4c5tphuhy7ca
In recent years, the impact of global climate change has increasingly revealed that energy transformation has become an indispensable part of achieving carbon neutrality. Thus, the relationship between energy transformation and economic growth has become the focus of academic attention. This study examines energy transition issues by using the panel threshold method. It explores the nonlinear impact of renewable energy consumption on economic growth, identifies various factors that lead to this nonlinear impact, and verifies its threshold effect. A comprehensive analysis reveals the following. (1) Overall, renewable energy consumption inhibits real gross domestic product (GDP) growth, but, in the long run, the negative impact becomes positive. (2) The threshold effect of energy consumption intensity (EI) is significant, with a threshold value of approximately 3.213. This means that when EI ≤ 3.213, renewable energy consumption promotes economic growth. However, EI > 3.213 indicates that this impact is significantly negative, which means that advancing the energy transition at this time may occur at the expense of real GDP growth. (3) There is also a significant threshold effect in energy transformation, with a threshold value of approximately 6.456. Similarly, when energy consumption transition (ET) ≤ 6.456, renewable energy consumption dampens real economic growth, and the economic cost of promoting renewable energy consumption is greater at this time. Alternatively, when ET > 6.456, this impact is significant at the 1 percent level and significantly positive. (4) There is also a significant threshold effect for emerging technologies, with a threshold value of approximately 1.367. When ET ≤ 1.367, renewable energy consumption dampens real economic growth, and the economic cost of promoting renewable energy consumption is greater. When ET > 1.367, the impact is significantly positive at the 1% level. To promote the positive development of economic growth, climate change, and energy transition, the nonlinear relationship studied in this paper can fill the gaps in existing research in theory and provide a theoretical basis for the government to adopt different policies at different stages of the energy transition to lay the foundation for improving global climate change in practice.Yuting Feng, Tong Zhaowork_mfakyprkrjhxfe4c5tphuhy7caThu, 24 Nov 2022 00:00:00 GMTDelay-constrained wireless multi-hop networks in the tactile internet
https://scholar.archive.org/work/mm6k7adcpjcj7arm6zbxfr3kpu
Das Taktile Internet verspricht weltweite haptische Kommunikation, also physische Interaktion zwischen Endpunkten, als weitere Modalität zusätzlich zu Audio- und Videosignalen. Haptische Kommunikation zeichnet sich imWesentlichen durch hohe Latenzanforderungen im Bereich von einer Millisekunde aus (Round-Trip-Time). Weiterhin bestehen hohe Ansprüche an die Zuverlässigkeit im Sinne einer niedrigen Paketverlustrate im Bereich um 0:001%. Der Trend wird häufig unter dem Begriff URLLC (engl. Ultra Reliable and Low-Latency Communication) zusammengefasst. URLLC ist durch physikalische Grenzen in der Signalübertragung auf eine maximale Kommunikationsreichweite von etwa 100–150km beschränkt. Zur Zeit lässt sich ein genereller Trend hin zu drahtloser Kommunikation beobachten. Latenzarme drahtlose Multi-Hop-Netze können somit ein Fundament für das Taktile Internet bilden. Das drahtlose Medium birgt jedoch noch zahlreiche Herausforderungen, besonders hinsichtlich unseres Verständnisses über das Latenzverhalten. Es hat sich in der Forschungsliteratur gezeigt, dass das gesamte Netz nicht einfach anhand des Verhaltens einzelner Knoten modelliert werden kann. Bisher konnte für drahtlose Multi-Hop-Netze nur der Durchsatz zufriedenstellend modelliert werden, was aber für das Taktile Internet nicht mehr ausreicht. Zudem müssen sich Applikationen im Taktilen Internet aber auch mit der Reichweitenbeschränkung von URLLC arrangieren, was noch viele Forschungsfragen aufwirft. Mit dieser Arbeit führen wir zunächst den Begriff der Tactile Internet Application ein, um eine Abgrenzung zwischen reichweitenbeschränkter URLLC und potenziell weltweit ausgedehnter haptischer Kommunikation im Taktilen Internet zu schaffen. Über hohe Distanzen muss vor allem der propagation Delay mit Software-Mitteln beseitigt werden. In unserer Vision vom Taktilen Internet fungieren Digitale Zwilinge als Stellvertreter für reale Objekte, welche frei im Internet insatanziiert werden und so zumindest virtuell Distanzen verringern können. Unser Forschungsbeitrag ist [...]Frank Engelhardt, Universitäts- Und Landesbibliothek Sachsen-Anhalt, Martin-Luther Universität, Mesut Güneşwork_mm6k7adcpjcj7arm6zbxfr3kpuWed, 23 Nov 2022 00:00:00 GMTOnline Federated Learning via Non-Stationary Detection and Adaptation amidst Concept Drift
https://scholar.archive.org/work/e7lp76vorbauhaznj3kpmqcd54
Federated Learning (FL) is an emerging domain in the broader context of artificial intelligence research. Methodologies pertaining to FL assume distributed model training, consisting of a collection of clients and a server, with the main goal of achieving optimal global model with restrictions on data sharing due to privacy concerns. It is worth highlighting that the diverse existing literature in FL mostly assume stationary data generation processes; such an assumption is unrealistic in real-world conditions where concept drift occurs due to, for instance, seasonal or period observations, faults in sensor measurements. In this paper, we introduce a multiscale algorithmic framework which combines theoretical guarantees of FedAvg and FedOMD algorithms in near stationary settings with a non-stationary detection and adaptation technique to ameliorate FL generalization performance in the presence of model/concept drifts. We present a multi-scale algorithmic framework leading to 𝒪 ( min{√(LT) , Δ^1/3T^2/3 + √(T)}) dynamic regret for T rounds with an underlying general convex loss function, where L is the number of times non-stationary drifts occured and Δ is the cumulative magnitude of drift experienced within T rounds.Bhargav Ganguly, Vaneet Aggarwalwork_e7lp76vorbauhaznj3kpmqcd54Tue, 22 Nov 2022 00:00:00 GMTUnivariate and multivariate statistical approaches for the analyses of omics data: sample classification and two-block integration
https://scholar.archive.org/work/mc3pdonidnh2fc64ocq3jawkhu
The wealth of information generated by high-throughput omics technologies in the context of large-scale epidemiological studies has made a significant contribution to the identification of factors influencing the onset and progression of common diseases. Advanced computational and statistical modelling techniques are required to manipulate and extract meaningful biological information from these omics data as several layers of complexity are associated with them. Recent research efforts have concentrated in the development of novel statistical and bioinformatic tools; however, studies thoroughly investigating the applicability and suitability of these novel methods in real data have often fallen behind. This thesis focuses in the analyses of proteomics and transcriptomics data from the EnviroGenoMarker project with the purpose of addressing two main research objectives: i) to critically appraise established and recently developed statistical approaches in their ability to appropriately accommodate the inherently complex nature of real-world omics data and ii) to improve the current understanding of a prevalent condition by identifying biological markers predictive of disease as well as possible biological mechanisms leading to its onset. The specific disease endpoint of interest corresponds to B-cell Lymphoma, a common haematological malignancy for which many challenges related to its aetiology remain unanswered. The seven chapters comprising this thesis are structured in the following manner: the first two correspond to introductory chapters where I describe the main omics technologies and statistical methods employed for their analyses. The third chapter provides a description of the epidemiological project giving rise to the study population and the disease outcome of interest. These are followed by three results chapters that address the research aims described above by applying univariate and multivariate statistical approaches for sample classification and data integration purposes. A summary of findings, c [...]Javiera Garrido Manriquez, Marc Chadeau, Paolo Vineis, Paul Elliott, Agencia Nacional De Investigación Y Desarrollo, Chile (ANID)work_mc3pdonidnh2fc64ocq3jawkhuTue, 22 Nov 2022 00:00:00 GMTStock Return Volatility in Southeast Asian Markets
https://scholar.archive.org/work/gqzvhzsiizcjtbi6dx3foomfla
This thesis explores determinants of stock return volatility in Southeast Asian markets. Firstly, we examine the effects of fundamental factors (macroeconomic and corporate variables) and behavioural factors (index composition and political risk) on stock market volatility. The factors are studied based on two theories in finance: neoclassical finance and behavioural finance. The results show that behavioural factors affect stock market volatility more significantly than fundamental factors in the countries with underdeveloped financial systems and unstable economies. This implies that the concept of behavioural finance is more reliable. However, the implication of the results presented in this thesis contributes to numerous finance researches regarding the theory of neoclassical finance in the countries with developed financial systems and stable economies. Next, we investigate evidence of primary factors influencing stock return volatility at industry level. Our results indicate that behavioural factors influence industry return volatility more significantly compared to those on aggregate market level. In other words, noise trading tends to be found in industry level, while informed traders become more dominant in the market, and thus the impact of noise trading on aggregate market volatility is reduced. Finally, we further evaluate the significance of stock return volatility by investigating the dynamics of volatility for the implications of real economic activity and crisis. A number of tests are conducted to examine whether the characteristics of return volatility involve a crisis and whether information in time-varying return volatilities explains macroeconomic fluctuations. As expected, we have found some characteristics of return volatility related to a crisis. There are the same results in both aggregate market and industry levels. The positive and relatively large value of β coefficient, which indicates high persistence in volatility, has been found in the time of pre-crisis. In addition, the results su [...]Natthinee Inphongwork_gqzvhzsiizcjtbi6dx3foomflaTue, 22 Nov 2022 00:00:00 GMTAnalysing global Taenia solium transmission dynamics to refine transmission modelling and support identification of optimal strategies for its control and elimination in low and middle-income countries
https://scholar.archive.org/work/727xpv2g3zac7hw2jd54wrr4ru
Taenia solium taeniasis/cysticercosis is a neglected zoonotic disease caused by the Taenia solium pork tapeworm, responsible for a significant global public health and economic burden. Infection of the central nervous system with T. solium cysts, leading to neurocysticercosis, is the single most important risk factor for acquired (infectious) epilepsy worldwide. Economic consequences pertain not only to the human health sector, by also to the animal sector, resulting from pig infection in the food-value chain. The 2012 World Health Organization (WHO), first, roadmap and the London Declaration on Neglected Tropical Diseases (NTDs) called for a validated strategy towards T. solium control and elimination by 2015, and for interventions to be scaled up in selected countries by 2020. These targets have not yet been met. More recently, the second, post-2020 NTD WHO roadmap (published in April, 2020) reformulated the milestones for T. solium, as hyperendemic settings/countries having adopted intensified control by 2030. Mathematical transmission models can support efforts aimed to setting, achieving, and evaluating the proposed 2030 goals. Although a number of transmission dynamics and control frameworks for T. solium taeniasis/cysticercosis already exist, refinement of some of these will be required for the purposes of increasing their usefulness to inform public health policy and practice. Taenia solium has a complex multi-host transmission cycle, including the adult tapeworm stage in the human definitive host, larval stages in the intermediate porcine host (with humans also acting as accidental intermediate hosts), and contamination of the environment with eggs/proglottids. Various T. solium models have been published, capturing the transmission cycle to varying degrees of complexity. Therefore, this thesis first reviews the current state of the T. solium transmission modelling field, including assessing models for the wider Taeniidae family of tapeworms to identify important gaps and understand how these may be addr [...]Matthew Andrew Stephen Dixon, Maria-Gloria Basáñez, Peter Winskill, Wendy Harrison, Lesong Conteh, Medical Research Council (Great Britain)work_727xpv2g3zac7hw2jd54wrr4ruMon, 21 Nov 2022 00:00:00 GMTHarmonic-Copuled Riccati Equations and its Applications in Distributed Filtering
https://scholar.archive.org/work/wa57ro3uxzdkphstkwbdxvot3m
The coupled Riccati equations are cosisted of multiple Riccati-like equations with solutions coupled with each other, which can be applied to depict the properties of more complex systems such as markovian systems or multi-agent systems. This paper manages to formulate and investigate a new kind of coupled Riccati equations, called harmonic-coupled Riccati equations (HCRE), from the matrix iterative law of the consensus on information-based distributed filtering (CIDF) algortihm proposed in [1], where the solutions of the equations are coupled with harmonic means. Firstly, mild conditions of the existence and uniqueness of the solution to HCRE are induced with collective observability and primitiviness of weighting matrix. Then, it is proved that the matrix iterative law of CIDF will converge to the unique solution of the corresponding HCRE, hence can be used to obtain the solution to HCRE. Moreover, through applying the novel theory of HCRE, it is pointed out that the real estimation error covariance of CIDF will also become steady-state and the convergent value is simplified as the solution to a discrete time Lyapunov equation (DLE). Altogether, these new results develop the theory of the coupled Riccati equations, and provide a novel perspective on the performance analysis of CIDF algorithm, which sufficiently reduces the conservativeness of the evaluation techniques in the literature. Finally, the theoretical results are verified with numerical experiments.Jiachen Qian, Peihu Duan, Zhisheng Duanwork_wa57ro3uxzdkphstkwbdxvot3mMon, 21 Nov 2022 00:00:00 GMTEnergy dynamics, heat production and heat-work conversion with qubits: towards the development of quantum machines
https://scholar.archive.org/work/5jk4gl2uz5fzdoognhj6fpcqiy
We present an overview of recent advances in the study of energy dynamics and mechanisms for energy conversion in qubit systems with special focus on realizations in superconducting quantum circuits. We briefly introduce the relevant theoretical framework to analyze heat generation, energy transport and energy conversion in these systems with and without time-dependent driving considering the effect of equilibrium and non-equilibrium environments. We analyze specific problems and mechanisms under current investigation in the context of qubit systems. These include the problem of energy dissipation and possible routes for its control, energy pumping between driving sources and heat pumping between reservoirs, implementation of thermal machines and mechanisms for energy storage. We highlight the underlying fundamental phenomena related to geometrical and topological properties, as well as many-body correlations. We also present an overview of recent experimental activity in this field.Liliana Arracheawork_5jk4gl2uz5fzdoognhj6fpcqiySun, 20 Nov 2022 00:00:00 GMTMaking Scientific Peer Review Scientific
https://scholar.archive.org/work/5waxaog6mzbpde3hdsrzol3ote
Nowadays many important applications such as hiring, university admissions, and scientific peer review rely on the collective efforts of a large number of individuals. These applications often operate at an extremely large scale which creates both opportunities and challenges. On the opportunity side, the large amount of data generated in these applications enables a novel data science perspective on the classical problem of decision-making. On the challenge side, in many of these applications, human decision-makers need to interact with various interfaces and algorithms, and follow various policies. When not carefully designed, such interfaces, algorithms, and policies may lead to unintended consequences. Identifying and overcoming such unintended consequences is an important research problem. In this thesis, we explore these opportunities and tackle these challenges with a general goal of understanding and improving distributed human decision-making in a principled manner. One application where the need for improvement is especially strong is scientific peer review. On the one hand, peer review is the backbone of academia, and scientific community agrees on the importance of improvement of the system. On the other hand, peer review is a microcosm of distributed decision-making that features a complex interplay between noise, bias, and incentives. Thus, insights learned from this specific domain apply to many other areas where similar problems arise. All in all, in this thesis, we aim at developing a principled approach towards scientific peer review—an important prerequisite for fair, equitable, and efficient progression of science. The three broad challenges that arise in peer review are noise, bias, and incentives. In this thesis, we work on each of these challenges: Noise and reviewer assignment. A suitable choice of reviewers is a cornerstone of peer review: poor assignment of reviewers to submissions may resul [...]Ivan Stelmakhwork_5waxaog6mzbpde3hdsrzol3oteFri, 18 Nov 2022 00:00:00 GMTStrategies for Black-Box and Multi-Objective Optimization
https://scholar.archive.org/work/h3k3elcd6rgwxg2jo53ia2er6a
Black-box optimization (BBO) problems occur frequently in many engineering and scientific disciplines, where one has access to zeroth-order evaluations of a function (black-box), that has to be optimized over a specified domain. In many situations, the function is expensive to evaluate, and hence the number of evaluations is limited by a budget. A popular class of algorithms known as Bayesian Optimization model the black-box function via surrogates, and proceed by evaluating points that are most likely to lead to the optimum. Multiobjective optimization (MOO) is another topic in optimization where the goal is to simultaneously optimize for multiple objectives defined over a common domain. Typically, these objectives do not achieve their optima for the same inputs. In such scenarios, rather than searching for a single best solution, a set of Pareto optimal solutions is desired. In this thesis, we study several optimization strategies for BBO and MOO and their applications. The first half of this thesis is about BBO for expensive functions. First, we propose a simple and flexible approach for multi-objective black-box optimization (MOBO) based on the idea of random scalarizations. We introduce a notion of multi-objective regret and show that our strategy achieves zero regret as the budget grows. Next, we study the effectiveness of neural networks for expensive BBO. We show that a simple greedy approach can achieve a performance close to that of Gaussian process Bayesian optimization. Using recently studied connections between Gaussian processes and training dynamics of very wide neural networks, we prove upper bounds on the regret of our proposed algorithm. Lastly, we propose a cost-aware Bayesian optimization framework that takes into account the cost of each evaluation. This approach is useful in settings where the evaluation cost varies across the input domain and low cost evaluations can provide a large amount of information about the maximum. The second half of this thesis is about the application of MOO to tw [...]Biswajit Pariawork_h3k3elcd6rgwxg2jo53ia2er6aFri, 18 Nov 2022 00:00:00 GMTMin-max optimization in two-team zero-sum games
https://scholar.archive.org/work/jhu6g25wqresbbcmfhdb6egcsq
Στην παρούσα εργασία εξετάζουμε τη (μη-)σύγκλιση μίας σειράς γνωστών αλγορίθμων βελτιστοποίησης για τον υπολογισμό σημείων ισορροπίας Nash σε παίγνια δύο ομάδων μηδενικού αθροίσματος. Τα παίγνια δύο ομάδων μηδενικού αθροίσματος μπορούν να μοντελοποιήσουν τη δυναμική της σύγκρουσης μεταξύ δύο αντιτιθέμενων μερών χωρίς να καταφεύγουν σε απλοϊκοποίηση του μοντέλου ως μία σύγκρουση μεταξύ δύο μετα-παικτών. Από άποψη υπολογιστικής πολυπλοκότητας, δείχνουμε ότι το πρόβλημα υπολογισμού σημείων ισορροπίας Nash είναι CLSδύσκολο. Στη συνέχεια, αποδεικνύουμε ότι για μία οικογένεια παιγνίων δύο ομάδων, μία σειρά αλγορίθμων πρώτου βαθμού (GDA, OGDA, EG, OMWU) αποτυγχάνουν να συγκλίνουν. Στον αντίποδα, συνεισφέρουμε τον σχεδιασμό ενός νέου αλγορίθμου πρώτου βαθμού που κάτω από ικανές συνθήκες συγκλίνει σε σημείο ισορροπίας Nash τόσο στη συγκεκριμένη οικόγενεια παιγνίων όσο και σε οποιοδήποτε παίγνιο (πιθανά μη κυρτό-μη κοίλο). Τέλος, παρουσιάζουμε έναν αριθμό πειραμάτων σε αρχιτεκτονικές νευρωνικών δικτύων (GANs) όπου η μοντελοποίηση τους ως σύγκρουση δύο ομάδων έχει προνομιακό πεδίο εφαρμογής.Fivos Kalogiannis, National Technological University Of Athenswork_jhu6g25wqresbbcmfhdb6egcsqFri, 18 Nov 2022 00:00:00 GMTA Time-Varying Network for Cryptocurrencies
https://scholar.archive.org/work/4seqihr2ozerpnh4qhs6zubhym
Cryptocurrencies return cross-predictability and technological similarity yield information on risk propagation and market segmentation. To investigate these effects, we build a time-varying network for cryptocurrencies, based on the evolution of return cross-predictability and technological similarities. We develop a dynamic covariate-assisted spectral clustering method to consistently estimate the latent community structure of cryptocurrencies network that accounts for both sets of information. We demonstrate that investors can achieve better risk diversification by investing in cryptocurrencies from different communities. A cross-sectional portfolio that implements an inter-crypto momentum trading strategy earns a 1.08% daily return. By dissecting the portfolio returns on behavioral factors, we confirm that our results are not driven by behavioral mechanisms.Li Guo, Wolfgang Karl Härdle, Yubo Taowork_4seqihr2ozerpnh4qhs6zubhymThu, 17 Nov 2022 00:00:00 GMTDistributed Random Reshuffling over Networks
https://scholar.archive.org/work/3mbgyifajnhg5ot5jyenslkzi4
In this paper, we consider distributed optimization problems where n agents, each possessing a local cost function, collaboratively minimize the average of the local cost functions over a connected network. To solve the problem, we propose a distributed random reshuffling (D-RR) algorithm that invokes the random reshuffling (RR) update in each agent. We show that D-RR inherits favorable characteristics of RR for both smooth strongly convex and smooth nonconvex objective functions. In particular, for smooth strongly convex objective functions, D-RR achieves 𝒪(1/T^2) rate of convergence (where T counts epoch number) in terms of the squared distance between the iterate and the global minimizer. When the objective function is assumed to be smooth nonconvex, we show that D-RR drives the squared norm of gradient to 0 at a rate of 𝒪(1/T^2/3). These convergence results match those of centralized RR (up to constant factors) and outperform the distributed stochastic gradient descent (DSGD) algorithm if we run a relatively large number of epochs. Finally, we conduct a set of numerical experiments to illustrate the efficiency of the proposed D-RR method on both strongly convex and nonconvex distributed optimization problems.Kun Huang, Xiao Li, Andre Milzarek, Shi Pu, Junwen Qiuwork_3mbgyifajnhg5ot5jyenslkzi4Thu, 17 Nov 2022 00:00:00 GMTDesign and training of deep reinforcement learning agents
https://scholar.archive.org/work/v4bnexxtdbgkvdgyu4jub7gulm
Deep reinforcement learning is a field of research at the intersection of reinforcement learning and deep learning. On one side, the problem that researchers address is the one of reinforcement learning: to act efficiently. A large number of algorithms were developed decades ago in this field to update value functions and policies, explore, and plan. On the other side, deep learning methods provide powerful function approximators to address the problem of representing functions such as policies, value functions, and models. The combination of ideas from these two fields offers exciting new perspectives. However, building successful deep reinforcement learning experiments is particularly difficult due to the large number of elements that must be combined and adjusted appropriately. This thesis proposes a broad overview of the organization of these elements around three main axes: agent design, environment design, and infrastructure design. Arguably, the success of deep reinforcement learning research is due to the tremendous amount of effort that went into each of them, both from a scientific and engineering perspective, and their diffusion via open source repositories. For each of these three axes, a dedicated part of the thesis describes a number of related works that were carried out during the doctoral research. The first part, devoted to the design of agents, presents two works. The first one addresses the problem of applying discrete action methods to large multidimensional action spaces. A general method called action branching is proposed, and its effectiveness is demonstrated with a novel agent, named BDQ, applied to discretized continuous action spaces. The second work deals with the problem of maximizing the utility of a single transition when learning to achieve a large number of goals. In particular, it focuses on learning to reach spatial locations in games and proposes a new method called Q-map to do so efficiently. An exploration mechanism based on this method is then used to demonstrate the effect [...]Fabio Pardo, Petar Kormushev, Andrew Davison, Dyson Technology Limited (Firm)work_v4bnexxtdbgkvdgyu4jub7gulmTue, 15 Nov 2022 00:00:00 GMTDagstuhl Reports, Volume 12, Issue 3, March 2022, Complete Issue
https://scholar.archive.org/work/atzdjkohs5hsdlk3uxxlt2ydry
Dagstuhl Reports, Volume 12, Issue 3, March 2022, Complete Issuework_atzdjkohs5hsdlk3uxxlt2ydryMon, 14 Nov 2022 00:00:00 GMTConvergence and variance reduction for stochastic differential equations in sampling and optimisation
https://scholar.archive.org/work/js6325m75rgw5ny6ykf5rzzilq
Three problems that are linked by way of motivation are addressed in this work. In the first part of the thesis, we study the generalised Langevin equation for simulated annealing with the underlying goal of improving continuous-time dynamics for the problem of global optimisation of nonconvex functions. The main result in this part is on the convergence to the global optimum, which is shown using techniques from hypocoercivity given suitable assumptions on the nonconvex function. Alongside, we investigate numerically the problem of parameter tuning in the continuous-time equation. In the second part of the thesis, this last problem is addressed rigorously for the underdamped Langevin dynamics. In particular, a systematic procedure for finding the optimal friction matrix in the sampling problem is presented. We give an expression for the gradient of the asymptotic variance in terms of solutions to Poisson equations and present a working algorithm for approximating its value. Lastly, regularity of an associated semigroup, twice differentiable-in-space solutions to the Kolmogorov equation and weak numerical convergence rates of order one are shown for a class of stochastic differential equations with superlinearly growing, non-globally monotone coefficients. In the relation to the previous part, the results allow the use of Poisson equations for variations of Langevin dynamics not permissible before.Martin Chak, Grigorios Pavliotis, Nikolas Kantas, Engineering And Physical Sciences Research Councilwork_js6325m75rgw5ny6ykf5rzzilqFri, 11 Nov 2022 00:00:00 GMTIV Estimation of Spatial Dynamic Panels with Interactive Effects: Large Sample
https://scholar.archive.org/work/36unbcvbzzbmfpxrj3cc5pnpnu
The present paper develops a new Instrumental Variables (IV) estimator for spatial, dynamic panel data models with interactive effects under large N and T asymptotics. For this class of models, the only approaches available in the literature are based on quasi-maximum likelihood estimation. The approach put forward in this paper is appealing from both a theoretical and a practical point of view for a number of reasons. Firstly, the proposed IV estimator is linear in the parameters of interest and it is computationally inexpensive. Secondly, the IV estimator is free from asymptotic bias. In contrast, existing ML estimators suffer from incidental parameter bias, depending on the magnitude of unknown parameters. Thirdly, the IV estimator retains the attractive feature of Method of Moments estimation in that it can accommodate endogenous regressors, so long as external exogenous instruments are available. The IV estimator is consistent and asymptotically normal as N, T → ∞, with N/T^2 → 0 and T /N^2 → 0. The proposed methodology is employed to study the determinants of risk attitude of banking institutions. The results of our analysis provide evidence that the more risksensitive capital regulation that was introduced by the Basel III framework in 2011 has succeeded in influencing banks' behaviour in a substantial manner.Guowei Cui, Vasilis Sarafidis, Takashi Yamagatawork_36unbcvbzzbmfpxrj3cc5pnpnuThu, 10 Nov 2022 00:00:00 GMTHidden Symmetries in Gravity : Black holes and other minisuperspaces
https://scholar.archive.org/work/zxr24noinzcjtouttlzs4ebcny
This thesis is dedicated to the study of symmetries in reduced models of gravity, with some frozen degrees of freedom. We focus on the minisuperspace reduction whith a finite number of degrees of freedom. Minisuperspaces are treated as mechanical models, evolving in one spacetime direction. This evolution parameter represents the orthogonal coordinate to the homogeneous foliation of the spacetime. I investigate their classical symmetries and the algebra of the corresponding Noether charges. After presenting the formalism allowing us to describe the reduced models in terms of an action principle, we discuss the condition for having an (extended) conformal symmetry. In particular, the black hole model enlightens the subtle role of the spacelike boundary of the homogeneous slice. The latter interplays with the conformal symmetry, being associated with a conserved quantity from the mechanical point of view. The absence of the infinite tower of charges, characteristic of the full theory, is due here to a symmetry-breaking mechanism. This is made explicit by looking at the infinite-dimensional extension of the symmetry group. This allows to look at the equation of motion of the mechanical system in terms of the infinite-dimensional group, who in turn has the effect of rescaling the coupling constants of the theory. Finally, the presence of the finite symmetry group allows defining a quantum model in terms of the corresponding representation theory. At the level of the effective theory, accounting for the quantum effects, the request that the symmetry is protected provides a powerful tool to discriminate between different modifications. In the end, the conformal invariance of the black hole background opens the door to its holographic properties and might have important consequences in the corresponding perturbation theory.Francesco Sartiniwork_zxr24noinzcjtouttlzs4ebcnyWed, 09 Nov 2022 00:00:00 GMT