Journals

Predictive simulations of complex systems are essential for applications ranging from weather forecasting to drug design. The veracity of these predictions hinges on their capacity to capture effective system dynamics. Massively parallel simulations predict the system dynamics by resolving all spatiotemporal scales, often at a cost that prevents experimentation, while their findings may not allow for generalization. On the other hand, reduced-order models are fast but limited by the frequently adopted linearization of the system dynamics and the utilization of heuristic closures. Here we present a novel systematic framework that bridges large-scale simulations and reduced-order models to learn the effective dynamics of diverse, complex systems. The framework forms algorithmic alloys between nonlinear machine learning algorithms and the equation-free approach for modelling complex systems. Learning the effective dynamics deploys autoencoders to formulate a mapping between fine- and coarse-grained representations and evolves the latent space dynamics using recurrent neural networks. The algorithm is validated on benchmark problems, and we find that it outperforms state-of-the-art reduced-order models in terms of predictability, and large-scale simulations in terms of cost. Learning the effective dynamics is applicable to systems ranging from chemistry to fluid mechanics and reduces the computational effort by up to two orders of magnitude while maintaining the prediction accuracy of the full system dynamics. We argue that learning the effective dynamics provides a potent novel modality for accurately predicting complex systems.

Increased intracranial pressure is the source of most critical symptoms in patients with glioma, and often the main cause of death. Clinical interventions could benefit from non-invasive estimates of the pressure distribution in the patient's parenchyma provided by computational models. However, existing glioma models do not simulate the pressure distribution and they rely on a large number of model parameters, which complicates their calibration from available patient data. Here we present a novel model for glioma growth, pressure distribution and corresponding brain deformation. The distinct feature of our approach is that the pressure is directly derived from tumour dynamics and patient-specific anatomy, providing non-invasive insights into the patient's state. The model predictions allow estimation of critical conditions such as intracranial hypertension, brain midline shift or neurological and cognitive impairments. A diffuse-domain formalism is employed to allow for efficient numerical implementation of the model in the patient-specific brain anatomy. The model is tested on synthetic and clinical cases. To facilitate clinical deployment, a high-performance computing implementation of the model has been publicly released.

The predictive capabilities of turbulent flow simulations, critical for aerodynamic design and weather prediction, hinge on the choice of turbulence models. The abundance of data from experiments and simulations and the advent of machine learning have provided a boost to turbulence modeling efforts. However, simulations of turbulent flows remain hindered by the inability of heuristics and supervised learning to model the near-wall dynamics. We address this challenge by introducing scientific multi-agent reinforcement learning (SciMARL) for the discovery of wall models for large-eddy simulations (LES). In SciMARL, discretization points act also as cooperating agents that learn to supply the LES closure model. The agents self- learn using limited data and generalize to extreme Reynolds numbers and previously unseen geometries. The present simulations reduce by several orders of magnitude the computa- tional cost over fully-resolved simulations while reproducing key flow quantities. We believe that SciMARL creates unprecedented capabilities for the simulation of turbulent flows.

Simulations are vital for understanding and predicting the evolution of complex molecular systems. However, despite advances in algorithms and special purpose hardware, accessing the time scales necessary to capture the structural evolution of biomolecules remains a daunting task. In this work, we present a novel framework to advance simulation time scales by up to 3 orders of magnitude by learning the effective dynamics (LED) of molecular systems. LED augments the equation-free methodology by employing a probabilistic mapping between coarse and fine scales using mixture density network (MDN) autoencoders and evolves the non-Markovian latent dynamics using long short-term memory MDNs. We demonstrate the effectiveness of LED in the Müller–Brown potential, the Trp cage protein, and the alanine dipeptide. LED identifies explainable reduced-order representations, i.e., collective variables, and can generate, at any instant, all-atom molecular trajectories consistent with the collective variables. We believe that the proposed framework provides a dramatic increase to simulation capabilities and opens new horizons for the effective modeling of complex molecular systems.

In January 2021, the Hellenic Institute of Advanced Study (HIAS) assembled a panel including world leading roboticists from the Hellenic diaspora, who volunteered their scientific expertise to provide a vision for Robotics in Greece. This monograph, entitled “Robotics in the Artificial Intelligence (AI) era,” will hopefully trigger a dialogue towards the development of a national robotics strategy. Our vision is that Robotics in the AI era will be an essential technology of the future for the safety and security of the Hellenic nation, its environment and its citizens, for modernizing its economy towards Industry 4.0, and for inspiring and educating the next generation workforce for the challenges of the 21st century. To contribute towards making this vision a reality, after reviewing global trends in robotics and assessing the Greek robotics ecosystem, we arrived at the following key findings and recommendations: Firstly, we think that Greece should develop a national Hellenic Robotics Initiative that serves as the nation’s long-term vision and strategy across the entire Greek robotics ecosystem. Also, certain societal drivers should be key in the areas of focus. Safety and security is an area of national importance necessitating a national initiative, while agrifood, maritime and logistics provide opportunities for internationally leading innovation.

We recommend the establishment of a mission-driven, government-funded, organization advancing unmanned vehicles in societal drivers of national importance, and Greece should leverage its unique geography and become a living testbed of robotics innovation turning the country into a development site for exportable technologies. In our opinion, universities should create Centers of Excellence in robotics and AI as well as consider innovation-leading research institutes such as the Italian Institute of Technology. We recommend investing in robotics education using Maker Spaces in order to prepare the workforce with 21st century skills to become Industry 4.0 innovators. Furthermore, we believe that the government should collect, measure, and analyze data on the robotics industry, robotics uses, labor shifts, and brain gain and promote awareness via a Hellenic Robotics Day. And finally, the government should regulate robot safety without stifling innovation, provide safe experimentation areas and mechanisms for certifying safety of locally developed robots. This monograph has many additional suggestions that enhance the above main recommendations. As authors, we advocate bringing the robotics ecosystem together in order to sharpen and expand these findings to an ambitious, long-term, and detailed national strategy and roadmap for robotics in the AI era.

Artificial bacteria flagella (ABFs) are magnetic helical microswimmers that can be remotely controlled via a uniform, rotating magnetic field. Previous studies have used the heterogeneous response of microswimmers to external magnetic fields for achieving independent control. Herein, analytical and reinforcement learning control strategies for path planning to a target by multiple swimmers using a uniform magnetic field are introduced. The comparison of the two algorithms shows the superiority of reinforcement learning in achieving minimal travel time to a target. The results demonstrate, for the first time, the effective independent navigation of realistic microswimmers with a uniform magnetic field in a viscous flow field.

Efficient point-to-point navigation in the presence of a background flow field is important for robotic applications such as ocean surveying. In such applications, robots may only have knowledge of their immediate surroundings or be faced with time-varying currents, which limits the use of optimal control techniques. Here, we apply a recently introduced Reinforcement Learning algorithm to discover time-efficient navigation policies to steer a fixed- speed swimmer through unsteady two-dimensional flow fields. The algorithm entails input- ting environmental cues into a deep neural network that determines the swimmer’s actions, and deploying Remember and Forget Experience Replay. We find that the resulting swimmers successfully exploit the background flow to reach the target, but that this success depends on the sensed environmental cue. Surprisingly, a velocity sensing approach significantly out- performed a bio-mimetic vorticity sensing approach, and achieved a near 100% success rate in reaching the target locations while approaching the time-efficiency of optimal navigation trajectories. 

 

We present Korali, an open-source framework for large-scale Bayesian uncertainty quantification and stochastic optimization. The framework relies on non-intrusive sampling of complex multiphysics models and enables their exploitation for optimization and decision-making. In addition, its distributed sampling engine makes efficient use of massively-parallel architectures while introducing novel fault tolerance and load balancing mechanisms. We demonstrate these features by interfacing Korali with existing high-performance software such as APHROS, LAMMPS (CPU-based), and MIRHEO (GPU-based) and show efficient scaling for up to 512 nodes of the CSCS Piz Daint supercomputer. Finally, we present benchmarks demonstrating that Korali outperforms related state-of-the-art software frameworks.

The extreme liquid transport properties of carbon nanotubes present new opportunities for surpassing conventional technologies in water filtration and purification. We demonstrate that carbon nanotubes with wettability surface patterns act as nanopumps for the ultrafast transport of picoliter water droplets without requiring externally imposed pressure gradients. Large-scale molecular dynamics simulations evidence unprecedented speeds and accelerations on the order of 10¹⁰ g of droplet propulsion caused by interfacial energy gradients. This phenomenon is persistent for nanotubes of varying sizes, stepwise pattern configurations, and initial conditions. We present a scaling law for water transport as a function of wettability gradients through simple models for the droplet dynamic contact angle and friction coefficient. Our results show that patterned nanotubes are energy-efficient nanopumps offering a realistic path toward ultrafast water nanofiltration and precision drug delivery.

Swimming organisms can escape their predators by creating and harnessing unsteady flow fields through their body motions. Stochastic optimization and flow simulations have identified escape patterns that are consistent with those observed in natural larval swimmers. However, these patterns have been limited by the specification of a particular cost function and depend on a prescribed functional form of the body motion. Here, we deploy reinforcement learning to discover swimmer escape patterns for larval fish under energy constraints. The identified patterns include the C-start mechanism, in addition to more energetically efficient escapes. We find that maximizing distance with limited energy requires swimming via short bursts of accelerating motion interlinked with phases of gliding. The present, data efficient, reinforcement learning algorithm results in an array of patterns that reveal practical flow optimization principles for efficient swimming and the methodology can be transferred to the control of aquatic robotic devices operating under energy constraints.

Effective intervention strategies for epidemics rely on the identification of their origin and on the robustness of the predictions made by network disease models. We introduce a Bayesian uncertainty quantification framework to infer model parameters for a disease spreading on a network of communities from limited, noisy observations; the state-of-the-art computational framework compensates for the model complexity by exploiting massively parallel computing architectures. Using noisy, synthetic data, we show the potential of the approach to perform robust model fitting and additionally demonstrate that we can effectively identify the disease origin via Bayesian model selection. As disease-related data are increasingly available, the proposed framework has broad practical relevance for the prediction and management of epidemics.

We present high-fidelity numerical simulations of expiratory biosol transport during normal breathing under indoor, stagnant air conditions with and without a facile mask. We investigate mask efficacy to suppress the spread of saliva particles that is underpinnings existing social distancing recommendations. The present simulations incorporate the effect of human anatomy and consider a spectrum of saliva particulate sizes that range from 0.1 to 10 μm while also accounting for their evaporation. The simulations elucidate the vorticity dynamics of human breathing and show that without a facile mask, saliva particulates could travel over 2.2 m away from the person. However, a non-medical grade face mask can drastically reduce saliva particulate propagation to 0.72 m away from the person. This study provides new quantitative evidence that facile masks can successfully suppress the spreading of saliva particulates due to normal breathing in indoor environments.

Simulations of blood flows in microfluidic devices and physiological systems are gaining importance in complementing experimental and clinical studies. The predictive capabilities of these simulations hinge on the parameters of the red blood cell (RBC) model that are usually calibrated from experimental data. However, these parameter values may vary drastically when calibrated using different experimental quan- tities or experimental settings. In turn, the results of existing blood flow simulations largely depend on the utilized parameters that have been chosen to validate a particular experiment. We suggest a revision to this type of model calibration to properly integrate experimental data in the computational models and accordingly inform their predictions. In this context, we introduce the calibration of a popular RBC model using data-driven, hierarchical Bayesian inference. We employ data from classical experiments of RBC stretching by optical tweezers and tank treading in shear flows, and distinguish the calibration of the model parameters through single-level and hierarchical Bayesian uncertainty quantification. We find that the optimal model parameters depend not only on the data used for the inference but also on the way the data are used in the inference process. Single-level Bayesian models predict well the data used in their calibration, but are inferior to the hierarchical Bayesian model at predicting previously unseen data. This work demonstrates that the proper integration of experimental data is essential for the development of a robust and transferable RBC model. We believe that the present study can serve as a prototype across scientific fields, in revising the integration of computational models and heterogeneous experimental data.

Turbulent flow models are critical for applications such as aircraft design, weather forecasting and climate prediction. Existing
models are largely based on physical insight and engineering intuition. More recently, machine learning has been contributing
to this endeavour with promising results. However, all efforts have focused on supervised learning, which is difficult to generalize beyond training data. Here we introduce multi-agent reinforcement learning as an automated discovery tool of turbulence
models. We demonstrate the potential of this approach on large-eddy simulations of isotropic turbulence, using the recovery of
statistical properties of direct numerical simulations as a reward. The closure model is a control policy enacted by cooperating
agents, which detect critical spatio-temporal patterns in the flow field to estimate the unresolved subgrid-scale physics. Results
obtained with multi-agent reinforcement learning algorithms based on experience replay compare favourably with established
modelling approaches. Moreover, we show that the learned turbulence models generalize across grid sizes and flow conditions.

This paper is associated with a video winner of a 2019 American Physical Society's Division of Fluid Dynamics (DFD) Gallery of Fluid Motion Award for work presented at the DFD Gallery of Fluid Motion. The original video is available online at the Gallery of Fluid Motion, https://doi.org/10.1103/APS.DFD.2019.GFM.V0018.

The systematic identification of infected individuals is critical for the containment of the COVID-19 pandemic. Currently, the spread of the disease is mostly quantified by the reported numbers of infections, hospitalisations, recoveries and deaths; these quantities inform epidemiology models that provide forecasts for the spread of the epidemic and guide policy making. The veracity of these forecasts depends on the discrepancy between the numbers of reported, and unreported yet infectious, individuals. We combine Bayesian experimental design with an epidemiology model and propose a methodology for the optimal allocation of limited testing resources in space and time, which maximises the information gain for such unreported infections. The proposed approach is applicable at the onset and spread of the epidemic and can forewarn of a possible recurrence of the disease after relaxation of interventions. We examine its application in Switzerland; the open source software is, however, readily adaptable to countries around the world. We find that following the proposed methodology can lead to vastly less uncertain predictions for the spread of the disease, thus improving estimates of the effective reproduction number and the future number of unreported infections. This information can provide timely and systematic guidance for the effective identification of infectious individuals and for decision-making regarding lockdown measures and the distribution of vaccines.

We examine the efficiency of Recurrent Neural Networks in forecasting the spatiotemporal dynamics of high dimensional and reduced order complex systems using Reservoir Computing (RC) and Backpropagation through time (BPTT) for gated network architectures. We highlight advantages and limitations of each method and discuss their implementation for parallel computing architectures. We quantify the relative prediction accuracy of these algorithms for the long-term forecasting of chaotic systems using as benchmarks the Lorenz-96 and the Kuramoto–Sivashinsky (KS) equations. We find that, when the full state dynamics are available for training, RC outperforms BPTT approaches in terms of predictive performance and in capturing of the long-term statistics, while at the same time requiring much less training time. However, in the case of reduced order data, large scale RC models can be unstable and more likely than the BPTT algorithms to diverge. In contrast, RNNs trained via BPTT show superior forecasting abilities and capture well the dynamics of reduced order systems. Furthermore, the present study quantifies for the first time the Lyapunov Spectrum of the KS equation with BPTT, achieving similar accuracy as RC. This study establishes that RNNs are a potent computational framework for the learning and forecasting of complex spatiotemporal systems.

We present data driven kinematic models for the motion of bubbles in high-Re turbulent fluid flows based on recurrent neural networks with long-short term memory enhancements. The models extend empirical relations, such as Maxey-Riley (MR) and its variants, whose applicability is limited when either the bubble size is large or the flow is very complex. The recurrent neural networks are trained on the trajectories of bubbles obtained by Direct Numerical Simulations (DNS) of the Navier Stokes equations for a two-component incompressible flow model. Long short term memory components exploit the time history of the flow field that the bubbles have encountered along their trajectories and the networks are further augmented by imposing rotational invariance to their structure. We first train and validate the formulated model using DNS data for a turbulent Taylor-Green vortex. Then we examine the model pre- dictive capabilities and its generalization to Reynolds numbers that are different from those of the train- ing data on benchmark problems, including a steady (Hill’s spherical vortex) and an unsteady (Gaussian vortex ring) flow field. We find that the predictions of the developed model are significantly improved compared with those obtained by the MR equation. Our results indicate that data-driven models with history terms are well suited in capturing the trajectories of bubbles in turbulent flows.