arXiv subjects

Peter Sollich

Publications and source records attributed to Peter Sollich.

At least 55 records · Page 3Linked to original sources

arXiv2021

Diversity of self-propulsion speeds reduces motility-induced clustering in confined active matter

Self-propelled swimmers such as bacteria agglomerate into clusters as a result of their persistent motion. In 1D, those clusters do not coalesce macroscopically and the stationary cluster size distribution (CSD) takes an exponential form. We develop a minimal lattice model for active particles in narrow channels to study how clustering is affected by the interplay between self-propulsion speed diversity and confinement. A mixture of run-and-tumble particles with a distribution of self-propulsion speeds is simulated in 1D. Particles can swap positions at rates proportional to their relative self-propulsion speed. Without swapping, we find that the average cluster size $L_\text{c}$ decreases with diversity and follows a non-arithmetic power mean of the single-component $L_\text{c}$'s, unlike the case of tumbling-rate diversity previously studied. Effectively, the mixture is thus equivalent to a system of identical particles whose self-propulsion speed is the harmonic mean self-propulsion speed of the mixture. With swapping, particles escape more quickly from clusters. As a consequence, $L_\text{c}$ decreases with swapping rates and depends less strongly on diversity. We derive a dynamical equilibrium theory for the CSDs of binary and fully polydisperse systems. Similarly to the clustering behaviour of one-component models, our qualitative results for mixtures are expected to be universal across active matter. Using literature experimental values for the self-propulsion speed diversity of unicellular swimmers known as choanoflagellates, which naturally differentiate into slower and faster cells, we predict that the error in estimating their $L_\text{c}$ via one-component models which use the conventional arithmetic mean self-propulsion speed is around $30\%$.

cond-mat.soft↗

arXiv2021

Fragmentation in trader preferences among multiple markets: Market coexistence versus single market dominance

Technological advancement has lead to an increase in number and type of trading venues and diversification of goods traded. These changes have re-emphasized the importance of understanding the effects of market competition: does proliferation of trading venues and increased competition lead to dominance of a single market or coexistence of multiple markets? In this paper, we address these questions in a stylized model of Zero Intelligence traders who make repeated decisions at which of three available markets to trade. We analyse the model numerically and analytically and find that parameters that govern traders' decisions -- memory length and intensity of choice, e.g. how strongly decisions are based on past success -- make the key distinctions between consolidated and fragmented steady states of the population of traders. All three markets coexist with equal shares of traders only when either learning is too weak and traders choose randomly, or when markets are identical. In the latter case, the population of traders is fragmented across the markets. For the more general case of markets with different biases, we note that market dominance is the more typical scenario. These results are interesting because previously either strong differentiation of markets or heterogeneity in the needs of traders was found to be a necessary condition for market coexistence. We show that, in contrast, these states can emerge simply as a consequence of co-adaptation of an initially homogeneous population of traders.

q-fin.TR↗

arXiv2021

Learning Waveform-Based Acoustic Models using Deep Variational Convolutional Neural Networks

We investigate the potential of stochastic neural networks for learning effective waveform-based acoustic models. The waveform-based setting, inherent to fully end-to-end speech recognition systems, is motivated by several comparative studies of automatic and human speech recognition that associate standard non-adaptive feature extraction techniques with information loss which can adversely affect robustness. Stochastic neural networks, on the other hand, are a class of models capable of incorporating rich regularization mechanisms into the learning process. We consider a deep convolutional neural network that first decomposes speech into frequency sub-bands via an adaptive parametric convolutional block where filters are specified by cosine modulations of compactly supported windows. The network then employs standard non-parametric 1D convolutions to extract relevant spectro-temporal patterns while gradually compressing the structured high dimensional representation generated by the parametric block. We rely on a probabilistic parametrization of the proposed neural architecture and learn the model using stochastic variational inference. This requires evaluation of an analytically intractable integral defining the Kullback-Leibler divergence term responsible for regularization, for which we propose an effective approximation based on the Gauss-Hermite quadrature. Our empirical results demonstrate a superior performance of the proposed approach over comparable waveform-based baselines and indicate that it could lead to robustness. Moreover, the approach outperforms a recently proposed deep convolutional neural network for learning of robust acoustic models with standard FBANK features.

stat.ML↗

arXiv2021

Robust Prediction of Force Chains in Jammed Solids using Graph Neural Networks

Force chains, which are quasi-linear self-organised structures carrying large stresses, are ubiquitous in jammed amorphous materials, such as granular materials, foams, emulsions or even assemblies of cells. Predicting where they will form upon mechanical deformation is crucial in order to describe the physical properties of such materials, but remains an open question. Here we demonstrate that graph neural networks (GNN) can accurately infer the location of these force chains in frictionless materials from the local structure prior to deformation, without receiving any information about the inter-particle forces. Once trained on a prototypical system, the GNN prediction accuracy proves to be robust to changes in packing fraction, mixture composition, amount of deformation, and the form of the interaction potential. The GNN is also scalable, as it can make predictions for systems much larger than those it was trained on. Our results and methodology will be of interest for experimental realizations of granular matter and jammed disordered systems, e.g. in cases where direct visualisation of force chains is not possible or contact forces cannot be measured.

cond-mat.soft↗

arXiv2021

Coarse graining of biochemical systems described by discrete stochastic dynamics

Many biological systems can be described by finite Markov models. A general method for simplifying master equations is presented that is based on merging adjacent states. The approach preserves the steady-state probability distribution and all steady-state fluxes except the one between the merged states. Different levels of coarse graining of the underlying microscopic dynamics can be obtained by iteration, with the result being independent of the order in which states are merged. A criterion for the optimal level of coarse graining or resolution of the process is proposed via a tradeoff between the simplicity of the coarse-grained model and the information loss relative to the original model. As a case study, the method is applied to the cycle kinetics of the molecular motor kinesin.

physics.bio-ph↗

arXiv2021

The random first-order transition theory of active glass in the high-activity regime

Dense active matter, in the fluid or amorphous-solid form, has generated intense interest as a model for the dynamics inside living cells and multicellular systems. An extension of the random first-order transition theory (RFOT) to include activity was developed, whereby the activity of the individual particles was added to the free energy of the system in the form of the potential energy of an active particle, trapped by a harmonic potential that describes the effective confinement by the surrounding medium. This active-RFOT model was shown to successfully account for the dependence of the structural relaxation time in the active glass, extracted from simulations, as a function of the activity parameters: the magnitude of the active force ($f_0$) and its persistence time ($τ_p$). However, significant deviations were found in the limit of large activity (large $f_0$ and/or $τ_p$). Here we extend the active-RFOT model to high activity using an activity-dependent harmonic confining potential, which we solve self-consistently. The extended model predicts qualitative changes in the high activity regime, which agree with the results of simulations in both three-dimensional and two-dimensional models of active glass.

cond-mat.soft↗

arXiv2021

Shear Induced Orientational Ordering in Active Glass

Dense assemblies of self propelled particles, also known as active or living glasses are abundantaround us, covering different length and time scales: from the cytoplasm to tissues, from bacterialbio-films to vehicular traffic jams, from Janus colloids to animal herds. Being structurally disorderedas well as strongly out of equilibrium, these systems show fascinating dynamical and mechanicalproperties. Using extensive molecular dynamics simulation and a number of different dynamicaland mechanical order parameters we differentiate three dynamical steady states in a sheared modelactive glassy system: (a) a disordered phase, (b) a propulsion-induced ordered phase, and (c) ashear-induced ordered phase. We supplement these observations with an analytical theory based onan effective single particle Fokker-Planck description to rationalise the existence of the novel shear-induced orientational ordering behaviour in our model active glassy system that has no explicitaligning interactions,e.g.of Vicsek-type. This ordering phenomenon occurs in the large persistencetime limit and is made possible only by the applied steady shear. Using a Fokker-Planck descriptionwe make testable predictions without any fit parameters for the joint distribution of single particleposition and orientation. These predictions match well with the joint distribution measured fromdirect numerical simulation. Our results are of relevance for experiments exploring the rheologicalresponse of dense active colloids and jammed active granular matter systems.

cond-mat.soft↗

arXiv2020

Dissipative failure of adiabatic quantum transport as a dynamical phase transition

Entanglement is the central resource in adiabatic quantum transport. Dephasing affects the availability of that resource by biasing trajectories, driving transitions between success and failure. This depletion of entanglement is important for the practical implementation of quantum technologies. We present a new perspective on the failure of adiabatic computation by understanding the failure of adiabatic transport as a dynamical phase transition. These ideas are demonstrated in a toy model of adiabatic quantum transport in a two spin system.

quant-ph↗

arXiv2020

Aging in a mean field elastoplastic model of amorphous solids

We construct a mean-field elastoplastic description of the dynamics of amorphous solids under arbitrary time-dependent perturbations, building on the work of Lin and Wyart [Phys. Rev. X 6, 011005 (2016)] for steady shear. Local stresses are driven by power-law distributed mechanical noise from yield events throughout the material, in contrast to the well-studied Hébraud-Lequeux model where the noise is Gaussian. We first use a mapping to a mean first passage time problem to study the phase diagram in the absence of shear, which shows a transition between an arrested and a fluid state. We then introduce a boundary layer scaling technique for low yield rate regimes, which we first apply to study the scaling of the steady state yield rate on approaching the arrest transition. These scalings are further developed to study the aging behaviour in the glassy regime, for different values of the exponent $μ$ characterizing the mechanical noise spectrum. We find that the yield rate decays as a power-law for $1<μ<2$, a stretched exponential for $μ=1$ and an exponential for $μ<1$, reflecting the relative importance of far-field and near-field events as the range of the stress propagator is varied. Comparison of the mean-field predictions with aging simulations of a lattice elastoplastic model shows excellent quantitative agreement, up to a simple rescaling of time.

cond-mat.soft↗

arXiv2020

How to Study a Persistent Active Glassy System?

We explore glassy dynamics of dense assemblies of soft particles that are self-propelled by active forces. These forces have a fixed amplitude and a propulsion direction that varies on a timescale tau_p, the persistence timescale. Numerical simulations of such active glasses are computationally challenging when the dynamics is governed by large persistence times. We describe in detail a recently proposed scheme that allows one to study directly the dynamics in the large persistence time limit, on timecales around and well above the persistence time. We discuss the idea behind the proposed scheme, which we call activity-driven dynamics, as well as its numerical implementation. We establish that our prescription faithfully reproduces all dynamical quantities in the appropriate limit tau_p goes to infinity. We deploy the approach to explore in detail the statistics of Eshelby-like plastic events in the steady state dynamics of a dense and intermittent active glass.

cond-mat.soft↗

arXiv2020

Coarse-grained Second Order Response Theory

While linear response theory, manifested by the fluctuation dissipation theorem, can be applied at any level of coarse graining, nonlinear response theory is fundamentally of microscopic nature. For perturbations of equilibrium systems, we develop an exact theoretical framework for analyzing the nonlinear (second order) response of coarse grained observables to time-dependent perturbations, using a path-integral formalism. The resulting expressions involve correlations of the observable with coarse grained path weights. The time symmetric part of these weights depends on the paths and perturbation protocol in a complex manner; in addition, the absence of Markovianity prevents slicing of the coarse-grained path integral. We show that these difficulties can be overcome and the response function can be expressed in terms of path weights corresponding to a single-step perturbation. This formalism thus leads to an extrapolation scheme where measuring linear responses of coarse-grained variables suffices to determine their second order response. We illustrate the validity of the formalism with an exactly solvable four-state model and the near-critical Ising model.

cond-mat.stat-mech↗

arXiv2020

Active mixtures in a narrow channel: Motility diversity changes cluster sizes

The persistent motion of bacteria produces clusters with a stationary cluster size distribution (CSD). Here we develop a minimal model for bacteria in a narrow channel to assess the relative importance of motility diversity (i.e. polydispersity in motility parameters) and confinement. A mixture of run-and-tumble particles with a distribution of tumbling rates (denoted generically by $α$) is considered on a 1D lattice. Particles facing each other cross at constant rate, rendering the lattice quasi-1D. To isolate the role of diversity, the global average $α$ stays fixed. For a binary mixture with no particle crossing, the average cluster size ($L_\text{c}$) increases with the diversity as lower-$α$ particles trap higher-$α$ ones for longer. At finite crossing rate, particles escape from the clusters sooner, making $L_\text{c}$ smaller and the diversity less important, even though crossing can enhance demixing of particle types between the cluster and gas phases. If the crossing rate is increased further, the clusters become controlled by particle crossing. We also consider an experiment-based continuous distribution of tumbling rates, revealing similar physics. Using parameters fitted from experiments with Escherichia coli bacteria, we predict that the error in estimating $L_\text{c}$ without accounting for polydispersity is around $60\%$. We discuss how to find a binary system with the same CSD as the fully polydisperse mixture. An effective theory is developed and shown to give accurate expressions for the CSD, the effective $α$, and the average fraction of mobile particles. We give reasons why our qualitative results are expected to be valid for other active matter models and discuss the changes that would result from polydispersity in the active speed rather than in the tumbling rate.

cond-mat.soft↗

arXiv2020

Tractable nonlinear memory functions as a tool to capture and explain dynamical behaviours

Mathematical approaches from dynamical systems theory are used in a range of fields. This includes biology where they are used to describe processes such as protein-protein interaction and gene regulatory networks. As such networks increase in size and complexity, detailed dynamical models become cumbersome, making them difficult to explore and decipher. This necessitates the application of simplifying and coarse graining techniques in order to derive explanatory insight. Here we demonstrate that Zwanzig-Mori projection methods can be used to arbitrarily reduce the dimensionality of dynamical networks while retaining their dynamical properties. We show that a systematic expansion around the quasi-steady state approximation allows an explicit solution for memory functions without prior knowledge of the dynamics. The approach not only preserves the same steady states but also replicates the transients of the original system. The method also correctly predicts the dynamics of multistable systems as well as networks producing sustained and damped oscillations. Applying the approach to a gene regulatory network from the vertebrate neural tube, a well characterised developmental transcriptional network, identifies features of the regulatory network responsible dfor its characteristic transient behaviour. Taken together, our analysis shows that this method is broadly applicable to multistable dynamical systems and offers a powerful and efficient approach for understanding their behaviour.

math.NA↗

arXiv2020

From entropic to energetic barriers in glassy dynamics: The Barrat-Mézard trap model on sparse networks

Trap models describe glassy dynamics as a stochastic process on a network of configurations representing local energy minima. We study within this class the paradigmatic Barrat-Mézard model, which has Glauber transition rates. Our focus is on the effects of the network connectivity, where we go beyond the usual mean field (fully connected) approximation and consider sparse networks, specifically random regular graphs. We obtain the spectral density of relaxation rates of the master operator using the cavity method, revealing very rich behaviour as a function of network connectivity $c$ and temperature $T$. We trace this back to a crossover from initially entropic barriers, resulting from a paucity of downhill directions, to energy barriers that govern the escape from local minima at long times. The insights gained are used to rationalize the relaxation of the energy after a quench from high $T$, as well as the corresponding correlation and persistence functions.

cond-mat.dis-nn↗

arXiv2020

Wrinkles, folds and ripplocations: unusual deformation structures of confined elastic sheets at non-zero temperatures

We study the deformation of a fluctuating crystalline sheet confined between two flat rigid walls as a simple model for layered solids where bonds among atoms {\it within} the same layer are much stronger than those {\it between} layers. When subjected to sufficiently high loads in an appropriate geometry, these solids deform and fail in unconventional ways. Recent experiments suggest that configurations named {\it ripplocations}, where a layer folds backwards over itself, are involved. These structures are distinct and separated by large free energy barriers from smooth {\it ripples} of the atomic layers that are always present at any non-zero temperature. We use Monte Carlo simulation in combination with an umbrella sampling technique to obtain conditions under which such structures form and study their specific experimental signatures.

cond-mat.soft↗

arXiv2020

Systematic model reduction captures the dynamics of extrinsic noise in biochemical subnetworks

We consider the general problem of describing the dynamics of subnetworks of larger biochemical reaction networks, e.g. protein interaction networks involving complex formation and dissociation reactions. We propose the use of model reduction strategies to understand the 'extrinsic' sources of stochasticity arising from the rest of the network. Our approaches are based on subnetwork dynamical equations derived by projection methods and by path integrals. The results provide a principled derivation of the different components of the extrinsic noise that is observed experimentally in cellular biochemical reactions, over and above the intrinsic noise from the stochasticity of biochemical events in the subnetwork. We explore several intermediate approximations to assess systematically the relative importance of different extrinsic noise components, including initial transients, long-time plateaus, temporal correlations, multiplicative noise terms and nonlinear noise propagation. The best approximations achieve excellent accuracy in quantitative tests on a simple protein network and on the epidermal growth factor receptor signalling network.

q-bio.QM↗

arXiv2019

A nucleation theory for yielding of nearly defect-free crystals: understanding rate dependent yield points

Experiments and simulations show that when an initially defect free rigid crystal is subjected to deformation at a constant rate, irreversible plastic flow commences at the so-called {\em yield point}. The yield point is a weak function of the deformation rate, which is usually expressed as a power law with an extremely small non-universal exponent. We re-analyze a representative set of published data on nanometer sized, mostly defect free, Cu, Ni and Au crystals in the light of a recently proposed theory of yielding based on nucleation of stable stress-free regions inside the metastable rigid solid. The single relation derived here, which is {\em not} a power law, explains data covering {\em fifteen} orders of magnitude in time scales.

cond-mat.mtrl-sci↗

arXiv2019

Multiple Types of Aging in Active Glass

Recent experiments and simulations have revealed glassy features in the cytoplasm, living tissues as well as dense assemblies of self propelled colloids. This leads to a fundamental question: how do these non-equilibrium (active) amorphous materials differ from conventional passive glasses, created either by lowering temperature or by increasing density? To address this we investigate the aging behaviour after a quench to an almost arrested state of a model active glass former, a Kob-Andersen glass in two dimensions. Each constituent particle is driven by a constant propulsion force whose direction diffuses over time. Using extensive molecular dynamics simulations we reveal rich aging behaviour of this dense active matter system: short persistence times of the active forcing lead to effective thermal aging; in the opposite limit we find a two-step aging process with active athermal aging at short times followed by activity-driven aging at late times. We develop a dedicated simulation method that gives access to this long-time scaling regime for highly persistent active forces.

cond-mat.soft↗