Search

Advances in artificial intelligence (AI) have great potential to help address societal challenges that are both collective in nature and present at national or trans-national scale. Pressing challenges in healthcare, finance, infrastructure and sustainability, for instance, might all be productively addressed by leveraging and amplifying AI for national-scale collective intelligence. The development and deployment of this kind of AI faces distinctive challenges, both technical and socio-technical. Here, a research strategy for mobilising inter-disciplinary research to address these challenges is detailed and some of the key issues that must be faced are outlined., Comment: 25 pages, 3 figures, Accepted for publication at Knowledge Engineering Review (KER)

Galaxy properties are known to be affected by their environment. This is well established for the extremes of the density scales, between the high-density cluster environment and the low-density field. It is however not fully understood how the intermediate-density regime of cosmic web filaments affects galaxy evolution. We investigate this environmental effect using a mass complete sample of 23,441 galaxies in the Sloan Digital Sky Survey DR8 Main Galaxy Sample (${M}_{\text{Stellar}} > 10^{9.91} \text{M}_{\odot}$). We define 6 environments, probing different density regimes and representing unique stages in the structure formation process, comparing the differences in star formation activity and morphology between them. We find that galaxies in filaments tend to be less star forming and favour more early-type morphologies than those in the field. These differences persist when considering stellar mass-matched samples, suggesting that this is a consequence of the environment. We further investigate whether these trends are a result of the large scale or local environment through constructing samples matched both in stellar mass and local galaxy density. We find that when also matching in local galaxy density, the differences observed between the filament and field population vanishes, concluding that the environmental effect of filaments can be entirely parameterised by a local galaxy density index. We find that differences can still be seen in comparisons with the interiors of clusters, suggesting these are unique environments which can impart additional physical processes not characterised by local galaxy density., Comment: 19 pages, 10 figures, Accepted for publication in MNRAS

The theoretical basis for the Eddy Damped Markovian Anisotropic Closure (EDMAC) is formulated for two-dimensional anisotropic turbulence interacting with Rossby waves in the presence of advection by a large-scale mean flow. The EDMAC is as computationally efficient as the Eddy Damped Quasi Normal Markovian (EDQNM) closure but in contrast is realizable in the presence of transient waves. The EDMAC is arrived at through systematic simplification of a generalization of the non-Markovian Direct Interaction Approximation (DIA) closure that has its origin in renormalized perturbation theory. Markovian Anisotropic Closures (MACs) are obtained from the DIA by using three variants of the Fluctuation Dissipation Theorem (FDT) with the information in the time history integrals instead carried by Markovian differential equations for two relaxation functions. One of the MACs is simplified to the EDMAC with analytical relaxation functions and high numerical efficiency, like the EDQNM. Sufficient conditions for the EDMAC to be realizable in the presence of Rossby waves are determined. Examples of the numerical integration of the EDMAC compared with the EDQNM are presented for two-dimensional isotropic and anisotropic turbulence, at moderate Reynolds numbers, possibly interacting with Rossby waves and large-scale mean flow. The generalization of the EDMAC for the statistical dynamics of other physical systems, to higher dimension and higher order nonlinearity is considered., Comment: 29 pages, 2 figures

Turbulence is the most important, ubiquitous, and difficult problem of classical physics. Feynman viewed it as essentially unsolved, without a rigorous mathematical basis to describe the statistical dynamics of this most complex of fluid motion. However, the paradigm shift came in 1959, with the formulation of the Eulerian direct interaction approximation (DIA) closure by Kraichnan. It was again based on renormalized perturbation theory, like quantum electrodynamics, and is a bare vertex theory that is manifestly realizable. Here, we review some of the subsequent exciting achievements in closure theory. We also document in some detail the progress that has been made in extending statistical dynamical turbulence theory to the real world of interactions with mean flows, waves and inhomogeneities such as topography. This includes numerically efficient inhomogeneous closures, like the realizable quasi-diagonal direct interaction approximation (QDIA), and even more efficient Markovian inhomogeneous closures (MICs). Recent developments include the formulation and testing of an eddy damped Markovian anisotropic closure (EDMAC) that is realizable in interactions with transient waves but is as efficient as the eddy damped quasi-normal Markovian (EDQNM). As well a similarly efficient closure, the realizable eddy damped Markovian inhomogeneous closure (EDMIC) has been developed. Moreover, we present subgrid models that cater for the complex interactions that occur in geophysical flows. Recent progress includes the determination of complete sets of subgrid terms for skilful large eddy simulations of baroclinic inhomogeneous turbulent atmospheric and oceanic flows interacting with Rossby waves and topography. The success of these inhomogeneous closures has also led to further applications in data assimilation and ensemble prediction and generalization to quantum fields., Comment: 79 pages, 21 figures

Robots incurring component failures ought to adapt their behavior to best realize still-attainable goals under reduced capacity. We formulate the problem of planning with actuators known a priori to be susceptible to failure within the Markov Decision Processes (MDP) framework. The model captures utilization-driven malfunction and state-action dependent likelihoods of actuator failure in order to enable reasoning about potential impairment and the long-term implications of impoverished future control. This leads to behavior differing qualitatively from plans which ignore failure. As actuators malfunction, there are combinatorially many configurations which can arise. We identify opportunities to save computation through re-use, exploiting the observation that differing configurations yield closely related problems. Our results show how strategic solutions are obtained so robots can respond when failures do occur -- for instance, in prudently scheduling utilization in order to keep critical actuators in reserve., Comment: 8 pages, 5 figures, submitted to 2024 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS)

Robotic systems are typically composed of various subsystems, such as localization and navigation, each encompassing numerous configurable components (e.g., selecting different planning algorithms). Once an algorithm has been selected for a component, its associated configuration options must be set to the appropriate values. Configuration options across the system stack interact non-trivially. Finding optimal configurations for highly configurable robots to achieve desired performance poses a significant challenge due to the interactions between configuration options across software and hardware that result in an exponentially large and complex configuration space. These challenges are further compounded by the need for transferability between different environments and robotic platforms. Data efficient optimization algorithms (e.g., Bayesian optimization) have been increasingly employed to automate the tuning of configurable parameters in cyber-physical systems. However, such optimization algorithms converge at later stages, often after exhausting the allocated budget (e.g., optimization steps, allotted time) and lacking transferability. This paper proposes CURE -- a method that identifies causally relevant configuration options, enabling the optimization process to operate in a reduced search space, thereby enabling faster optimization of robot performance. CURE abstracts the causal relationships between various configuration options and robot performance objectives by learning a causal model in the source (a low-cost environment such as the Gazebo simulator) and applying the learned knowledge to perform optimization in the target (e.g., Turtlebot 3 physical robot). We demonstrate the effectiveness and transferability of CURE by conducting experiments that involve varying degrees of deployment changes in both physical robots and simulation., Comment: Revised submission in IEEE Transactions on Robotics (T-RO), 2024

We study the problem of automatically discovering Granger causal relations from observational multivariate time-series data.Vector autoregressive (VAR) models have been time-tested for this problem, including Bayesian variants and more recent developments using deep neural networks. Most existing VAR methods for Granger causality use sparsity-inducing penalties/priors or post-hoc thresholds to interpret their coefficients as Granger causal graphs. Instead, we propose a new Bayesian VAR model with a hierarchical factorised prior distribution over binary Granger causal graphs, separately from the VAR coefficients. We develop an efficient algorithm to infer the posterior over binary Granger causal graphs. Comprehensive experiments on synthetic, semi-synthetic, and climate data show that our method is more uncertainty aware, has less hyperparameters, and achieves better performance than competing approaches, especially in low-data regimes where there are less observations.

Estimating the structure of a Bayesian network, in the form of a directed acyclic graph (DAG), from observational data is a statistically and computationally hard problem with essential applications in areas such as causal discovery. Bayesian approaches are a promising direction for solving this task, as they allow for uncertainty quantification and deal with well-known identifiability issues. From a probabilistic inference perspective, the main challenges are (i) representing distributions over graphs that satisfy the DAG constraint and (ii) estimating a posterior over the underlying combinatorial space. We propose an approach that addresses these challenges by formulating a joint distribution on an augmented space of DAGs and permutations. We carry out posterior estimation via variational inference, where we exploit continuous relaxations of discrete distributions. We show that our approach performs competitively when compared with a wide range of Bayesian and non-Bayesian benchmarks on a range of synthetic and real datasets.

Predicting lithium-ion battery lifetime is one of the greatest unsolved problems in battery research right now. Recent years have witnessed a surge in lifetime prediction papers using physics-based, empirical, or data-driven models, most of which have been validated against the remaining capacity (capacity fade) and sometimes resistance (power fade). However, there are many different combinations of degradation mechanisms in lithium-ion batteries that can result in the same patterns of capacity and power fade, making it impossible to find a unique validated solution. Experimentally, degradation mode analysis involving measuring the loss of lithium inventory, loss of active material at both electrodes, and electrode drift/slippage has emerged as a state-of-the-art requirement for cell degradation studies. In this paper we coupled five degradation mechanisms together for the first time. We also showed how three models with different levels of complexity can all fit the remaining capacity and resistance well, but only the model with five coupled degradation mechanisms could also fit the degradation modes at all temperatures. This work proves that validating only against capacity and power fade is no longer sufficient, and state-of-the-art experimental and modelling degradation studies should include degradation mode analysis for validation in the future.

A model of a lithium-ion battery containing a cosolvent electrolyte is developed and implemented within the open-source PyBaMM platform. Lithium-ion electrolytes are essential to battery operation and normally contain at least two solvents to satisfy performance requirements. The widely used Doyle-Fuller-Newman battery model assumes that the electrolyte comprises a salt dissolved in a single effective solvent, however. This single-solvent approximation has been disproved experimentally and may hinder accurate battery modelling. Here, we present a two-solvent model that resolves the transport of ethylene carbonate (EC) and lithium salt in a background linear carbonate. EC concentration polarization opposes that of Li+ during cycling, affecting local electrolyte properties and cell-level overpotentials. Concentration gradients of Li+ can be affected by cross-diffusion, whereby EC gradients enhance or impede salt flux. A rationally parametrized model that includes EC transport predicts 6% more power loss at 4.5C discharge and ~0.32% more capacity loss after a thousand 1C cycles than its single-solvent equivalent. This work provides a tool to model more transport behaviour in the electrolyte that may affect degradation and enables the transfer of microscopic knowledge about solvation structure-dependent performance to the macroscale.

High-throughput computing (HTC) is a pivotal asset in many scientific fields, such as biology, material science and machine learning. Applying HTC to the complex physics-based degradation models of lithium-ion batteries enables efficient parameter identification and sensitivity analysis, which further leads to optimal battery design and operating conditions. However, running physics-based degradation models comes with pitfalls, as solvers can crash or get stuck in infinite loops due to numerical errors. Also, how to pipeline HTC for degradation models has seldom been discussed. To fill these gaps, we have created ParaSweeper, a Python script tailored for HTC, designed to streamline parameter sweeping by running as many ageing simulations as computational resources allow, each with different parameters. We have demonstrated the capability of ParaSweeper based on the open-source platform PyBaMM, and the approach can also apply to other numerical models which solve partial differential equations. ParaSweeper not only manages common solver errors, but also integrates various methods to accelerate the simulation. Using a high-performance computing platform, ParaSweeper can run millions of charge/discharge cycles within one day. ParaSweeper stands to benefit both academic researchers, through expedited model exploration, and industry professionals, by enabling rapid lifetime design, ultimately contributing to the prolonged lifetime of batteries.

A realizable Eddy Damped Markovian Anisotropic Closure (EDMAC) is presented for the interaction of two dimensional turbulence and transient waves such as Rossby waves. The structure of the EDMAC ensures that it is as computationally efficient as the Eddy Damped Quasi Normal Markovian (EDQNM) closure but unlike the EDQNM is guaranteed to be realizable in the presence of transient waves. Jack Herring's important contributions to laying the foundations of statistical dynamical closure theories of fluid turbulence are briefly reviewed. The topics covered include equilibrium statistical mechanics, Eulerian and Lagrangian statistical dynamical closure theories, and the statistical dynamics of the interaction of turbulence with topography. The impact of Herring's work is described and placed in the context of related developments. Some of the further works that have built on Herring's foundations are discussed. The relationships between theoretical approaches employed in statistical classical and quantum field theories, and their overlap, are outlined. The seminal advances made by the pioneers in strong interaction fluid turbulence are put into perspective by comparing related developments in strong interaction quantum filed theory., Comment: 23 pages, 0 figures

Battery models generally fall into two categories: physics-based models and ECM models. Physics-based Doyle-Fuller-Newman (DFN) models can accurately simulate the battery internal electrochemical processes, but to properly account for thermal effects requires a strong coupling between a DFN model and a 3D thermal model, which is computationally unaffordable. Distributed Equivalent Circuit Network (ECN) models can perform simulations with high speed and reasonable accuracy. However, these models rely heavily on the characterisation experiments for ECN parameter identification, which is resource-intensive and can lead to inaccurate parametrisation outcomes due to internal thermal inhomogeneity. To harness the strengths of both models, we propose a computational framework to integrate electrochemical DFN model and 3D distributed ECN model together. Using this framework, we simulate constant current discharge experiments of Kokam 7.5 Ah pouch cell (Model SLPB75106100) and compare the simulations with the commonly-used lumped DFN-thermal model. The computational model outperforms the lumped DFN model at low-temperature and/or high C-rate scenarios significantly. The largest predicting error of the framework at 3 C-rate &Tam = 25oC and at 1 C-rate &Tam = 0 oC is approximately 1/3 of that for DFN model. At 3 C-rate &Tam = 5oC, the difference between these two can rise to 377 mV. By integrating DFN and 3D-distributed ECN together, the computational framework can simulate the complicated interplay between electrochemistry, thermal process, and electricity within a cell fast and accurately. We anticipate this computational framework to be a valuable toolset to assist researchers and engineers in the design and control of Li-ion batteries.

Gaussian process state-space models (GPSSMs) provide a principled and flexible approach to modeling the dynamics of a latent state, which is observed at discrete-time points via a likelihood model. However, inference in GPSSMs is computationally and statistically challenging due to the large number of latent variables in the model and the strong temporal dependencies between them. In this paper, we propose a new method for inference in Bayesian GPSSMs, which overcomes the drawbacks of previous approaches, namely over-simplified assumptions, and high computational requirements. Our method is based on free-form variational inference via stochastic gradient Hamiltonian Monte Carlo within the inducing-variable formalism. Furthermore, by exploiting our proposed variational distribution, we provide a collapsed extension of our method where the inducing variables are marginalized analytically. We also showcase results when combining our framework with particle MCMC methods. We show that, on six real-world datasets, our approach can learn transition dynamics and latent states more accurately than competing methods., Comment: Updating to final version to appear in the proceedings

Robotic systems have subsystems with a combinatorially large configuration space and hundreds or thousands of possible software and hardware configuration options interacting non-trivially. The configurable parameters are set to target specific objectives, but they can cause functional faults when incorrectly configured. Finding the root cause of such faults is challenging due to the exponentially large configuration space and the dependencies between the robot's configuration settings and performance. This paper proposes CaRE -- a method for diagnosing the root cause of functional faults through the lens of causality. CaRE abstracts the causal relationships between various configuration options and the robot's performance objectives by learning a causal structure and estimating the causal effects of options on robot performance indicators. We demonstrate CaRE's efficacy by finding the root cause of the observed functional faults and validating the diagnosed root cause by conducting experiments in both physical robots (Husky and Turtlebot 3) and in simulation (Gazebo). Furthermore, we demonstrate that the causal models learned from robots in simulation (e.g., Husky in Gazebo) are transferable to physical robots across different platforms (e.g., Husky and Turtlebot 3)., Comment: IEEE Robotics and Automation Letters, 2023 (Accepted)

During its 53-day polar orbit around Jupiter, Juno often crosses the boundaries of the Jovian magnetosphere (namely the magnetopause and bow shock). From the boundary locations, the upstream solar wind dynamic pressure can be inferred, which in turn illustrates the state of compression or relaxation of the system. The aim of this study is to examine Jovian radio emissions during magnetospheric compressions, in order to determine the relationship between the solar wind and Jovian radio emissions. In this paper, we give a complete list of bow shock and magnetopause crossings (from June 2016 to August 2022), along with some extra informations (e.g. solar wind dynamic pressure and position of the standoff distances inferred from Joy et al. (2002)). We then select two compression events that occur in succession (inferred from magnetopause crossings) and we present a case study of the response of the Jovian radio emissions. We demonstrate that magnetospheric compressions lead to the activation of new radio sources. Newly activated broadband kilometric emissions are observed almost simultaneously to compression of the magnetosphere, with sources covering a large range of longitudes. Decametric emission sources are seen to be activated more than one rotation later only at specific longitudes and dusk local times. Finally, the activation of narrowband kilometric radiation is not observed during the compression phase, but when the magnetosphere is in its expansion phase.

Aims. Working towards improved space weather predictions, we aim to quantify how the critical height at which the torus instability drives coronal mass ejections (CMEs) varies over time in a sample of solar active regions. Methods. We model the coronal magnetic fields of 42 active regions and quantify the critical height at their central polarity inversion lines throughout their observed lifetimes. We then compare these heights to the changing magnetic flux at the photospheric boundary and identify CMEs in these regions. Results. In our sample, the rates of CMEs per unit time are twice as high during phases when magnetic flux is increasing than when it is decreasing, and during those phases of increasing flux, the rate of CMEs is 63% higher when the critical height is rising than when it is falling. Furthermore, we support and extend the results of previous studies by demonstrating that the critical height in active regions is generally proportional to the separation of their magnetic polarities through time. When the separation of magnetic polarities in an active region increases, for example during the continuous emergence and expansion of a magnetic bipole, the critical height also tends to increase. Conversely, when the polarity separation decreases, for example due to the emergence of a new, compact bipole at the central inversion line of an existing active region or into a quiet Sun environment, the critical height tends to decrease., Comment: Accepted for publication in Astronomy & Astrophysics. 10 pages, 4 figures, 6 tables