Applied Dynamics Archives for Fall 2014 to Spring 2015


Extreme events in excitable systems

When: Thu, September 12, 2013 - 12:30pm
Where: IREAP Large Conference Room, ERF 1207
Speaker: Professor Ulrike Feudel (Carl von Ossietzky Universität Oldenburg) - http://www.komplsyst.icbm.de/
Abstract: We consider the dynamics of rare events at which an observable of a complex dynamical system is extreme in amplitude. Since our goal is to contribute to the understanding of the emergence and the termination of epileptic seizures and of harmful algal blooms, we focus on excitable systems which are useful models to describe those two phenomena. Therefore, we study deterministic networks of excitable elements of FitzHugh–Nagumo type. We demonstrate that these networks are capable of self-generating such events at irregular times, i.e., without a change of control parameters or without an input to the system. To get a better understanding of this behavior, we study a reduced system consisting of two coupled FitzHugh–Nagumo oscillators, which mimics the extreme event dynamics observed for the networks. We discuss the properties of the networks as well as of the reduced system that may be involved in the generation and termination of such extreme events.

Quantum and Dispersive wave analysis via Hydrodynamics

When: Thu, September 26, 2013 - 12:30pm
Where: IREAP Large Conference Room, ERF 1207
Speaker: Prof. Avy Soffer (Math Department, Rutgers University) - http://www.math.rutgers.edu/~soffer/
Abstract: Quantum phenomena like tunneling and resonances are unique to dispersive equations. They play a fundamental role in applications since the early days of Quantum mechanics almost hundred years ago.
I will describe a new way to analize such problems for general class of linear and nonlinear dispersive equations, which are related to macroscopic Quantum systems.
The method I'll present is based on the hydrodynamic formalism, which is used to study the problem of soliton and vortex tunneling from wells, and related phenomena.
In particular, I show how this approach uncovers a new structure of tunneling processes, for both the LINEAR and nonlinear Schroedinger equations. It relates known phenomena in fluid flow to dispersive wave equations.
These new effects, in which the tunneled waves appear as blips or jets outside the well, has subsequently been observed experimentally.

Adaptive Brain Changes in Attention

When: Thu, October 3, 2013 - 12:30pm
Where: IREAP Large Conference Room, ERF 1207
Speaker: Jonathan Fritz (UMD, Institute for Systems Research) - http://www.isr.umd.edu/Labs/NSL/nsl.html
Abstract: Rapid, adaptive task-related receptive field plasticity alters the tuning properties of neurons in the primary auditory cortex in a task-specific fashion that can enhance their ability to encode the attended spectral and temporal features of the relevant acoustic stimuli in the current behavioral task they are engaged in. Such dynamic modulation of sensory filter properties in the brain is a key component of attention. We compared task-related plasticity in primary auditory cortex (A1) vs secondary auditory cortical (AC) areas, in a simple auditory discrimination task. In secondary AC areas, we found relatively heightened responses to targets vs. background stimuli during behavior, compared to responses observed in passive listening. The secondary AC is thus likely to play a key role in the transformations of sound from spectrotemporal pattern to recognized auditory object with associated behavioral meaning. We propose that the PFC controls the flow of sensory inputs via top-down feedback to sensory cortical areas during behavior, reflecting attentional focus and task objectives and rules. Stimulation of PFC, paired with pure tone presentation, elicits changes in A1 receptive fields that are similar to those observed in attention-driven behavior. These results support the idea that PFC and AC are part of a larger auditory attention network that uses neuromodulators and an array of plasticity mechanisms to enhance responses to attended foreground stimuli and suppress responses to background stimuli. I shall discuss convergent results from animal research and human neuroimaging experiments that support the view of the brain as a highly adaptive system.

Symmetries, Cluster Synchronization, and Isolated Desynchronization in Complex Networks

When: Thu, October 10, 2013 - 12:30pm
Where: IREAP Large Conference Room, ERF 1207
Speaker: Lou Pecora and Aaron Hagerstrom (NRL and UMD) -
Abstract: Synchronization is of central importance in power distribution, telecommunication, neuronal, and biological networks. Many networks are observed to produce patterns of synchronized clusters, but it has been difficult to predict these clusters or understand the conditions under which they form, except for in the simplest of networks. In this article, we shed light on the intimate connection between network symmetry and cluster synchronization. We introduce general group theory techniques that use network symmetries to reveal the patterns of synchronized clusters and determine the conditions under which they persist. The connection between symmetry and cluster synchronization is experimentally explored using an electro-optic network. We experimentally observe and theoretically predict a surprising phenomenon in which some clusters lose synchrony while leaving others synchronized. The results could guide the design of new power grid systems or lead to new understanding of the dynamical behavior of networks ranging from neural to social.

Strong Control in Cavity QED

When: Thu, October 17, 2013 - 12:30pm
Where: IREAP Large Conference Room, ERF 1207
Speaker: Dr. Luis Orozco (UMD) - http://www.physics.umd.edu/rgroups/amo/orozco/index.html
Abstract: Cavity QED in the optical regime has two avenues for dissipation: spontaneous emission of the atoms and escape of the light from the cavity. Strong control of the dynamics in the cavity QED relies on both channels and permits the preservation of a conditional quantum state a long time after the expected decay of the system. This talk presents experiments with Rb atoms in a cavity QED system where the single atom coupling and the decay rates are comparable. Single mode and two mode monitoring offer feedback possibilities to preserve conditional atom-cavity oscillatory exchange and ground state superpositions.

Ensemble methods for Lyapunov exponents and data assimilation

When: Thu, October 31, 2013 - 12:30pm
Where: IREAP Large Conference Room, ERF 1207
Speaker: Brian Hunt (UMD, Department of Mathematics and IPST) - http://terpconnect.umd.edu/~bhunt/

Stability of Boolean networks: The joint effects of topology and update rules

When: Thu, November 21, 2013 - 12:30pm
Where: IREAP Large Conference Room, ERF 1207
Speaker: Shane Squires (UMD) -
Abstract: Boolean networks, a widely used model of gene regulatory networks, are known to exhibit a phase transition between a stable regime, in which small perturbations die out, and an unstable regime, in which small perturbations grow exponentially. This talk will address the stability of orbits in large Boolean networks with given complex topology. Recent past work has addressed the separate effects of network topology and certain special classes of update rules on stability, but only crude results exist about how these effects interact. We present a widely applicable solution to this problem, imposing no restrictions on the form of the update rules, which may be correlated with local topological properties of the network. Numerical experiments confirm our theory and show that local correlations between topology and update rules can have profound effects on the qualitative behavior of these systems.

Quantifying Network Topology Robustness in Adversarial Environment: A Game Theoretic Approach

When: Thu, December 5, 2013 - 12:30pm
Where: IREAP Large Conference Room, ERF 1207
Speaker: Assane Gueye (NIST) - http://www.nist.gov/itl/math/cctg/assane.cfm
Abstract: Designing network topologies that are robust and resilient to attacks has been and continues to be an important and challenging topic in the area of communication networks. One of the main difficulties resides in quantifying the robustness of a network in the presence of an intelligent attacker who might exploit the structure of the network topology to design harmful attacks. To capture the strategic nature of the interactions between a defender and an adversary, game theoretic models have been gaining a lot of interest in the study of the security of communication networks. In a recent line of research, network blocking games have been introduced and applied to the analysis of the robustness of network topologies. A network blocking game takes as input the communication model and the topology of a network and models the strategic interactions between an adversary and the network operator as a two-player game. The Nash equilibrium strategies are then used to predict the most likely attacker’s actions and the attacker’s Nash equilibrium payoff serves as a quantification of the vulnerability of the network. In this talk, I will present the notion of network blocking games and show how they can be used to derive network vulnerability metrics by using a series of examples of communication models. I will also show how these metrics can be used to design networks that are robust against attacks and/or strengthen the robustness of existing networks. I will also show how the metrics can be used to identify the most critical links in a network. This is joint work with Prof. Jean C. Walrand, Prof. Venkat Anantharam (UC Berkeley), Dr. Vladimir Marbukh (NIST), and Aron Lazska (Budapest University of Technology and Economics).

Stabilization of incoherence in the disordered Hamiltonian Mean Field model

When: Thu, December 12, 2013 - 12:30pm
Where: IREAP Large Conference Room, ERF 1207
Speaker: Juan Restrepo (University of Colorado, Boulder) - http://amath.colorado.edu/faculty/juanga/Juan_G._Restrepo.html
Abstract: We study the Hamiltonian Mean Field model of coupled Hamiltonian rotors with a heterogeneous distribution of moments of inertia and coupling strengths. We show that when the parameters of the rotors are heterogeneous finite size fluctuations can greatly modify the coupling strength at which the incoherent state loses stability by inducing correlations between the momentum and parameters of the rotors. For unimodal parameter distributions, we find an analytical expression for the modified critical coupling strength in terms of statistical properties of the parameter distributions and confirm our results with numerical simulations. We find numerically that these effects disappear for strongly bimodal parameter distributions.

Rogue Wave Statistics in the Ocean and Beyond

When: Thu, February 6, 2014 - 12:30pm
Where: IREAP Large Conference Room, ERF 1207
Speaker: Prof. Lev Kaplan (Tulane Univesity) - http://www.tulane.edu/~lkaplan/
Abstract: I will discuss several examples of wave dynamics in which the wave statistics may be analyzed from the viewpoint of random, weak semiclassical scattering. (i) Imaging of electron flow in a two-dimensional electron gas exhibits branched behavior, which is associated with caustics in the classical dynamics. The statistical properties of these branches may be computed in the limit where the random potential associated with impurities is weak and is correlated on a scale large compared with the wavelength. (ii) A related approach is applied to calculate the formation probability of rogue waves in the ocean, due to refraction by weak random currents. When the wave steepness is small, the probability distribution is governed by a single “freak index” parameter, which describes the strength of (linear) wave scattering by random currents relative to the angular spread of the incoming random sea. When the average steepness is large, the wave height distribution takes a very similar functional form, but the key variables determining the probability distribution are the steepness, and the angular and frequency spread of the incoming waves. (iii) A quasi-two-dimensional microwave resonator with correlated disorder has been used to demonstrate the mechanism of rogue wave formation by linear refraction in the laboratory. Large deviations from Rayleigh’s law for the wave height distribution are associated with “hot spots”, which again may be interpreted as remnants of singularities in the ray dynamics.

A broken-contact order parameter for inferring inter-cellular communication from patterns of externally guided migrating cells

When: Thu, February 13, 2014 - 12:30pm
Where: IREAP Large Conference Room, ERF 1207
Speaker: Joshua Parker (UMD) -
Abstract: We propose an order parameter for investigating the local spatial organization often seen in migrating cell systems. Using a phenomenological model for the group dynamics of Dictyostelium discoideum, we observe that monitoring the fraction of broken cell-cell contacts over time resolves the characteristic structures seen in simulations that are not captured by other global order parameters. We discuss the application of this order parameter to future experimental and theoretical work on D. discoideum migration as well as extensions to other biological and physical systems.

3D dynamics of granular materials: shear, segregation, and rotation

When: Thu, February 13, 2014 - 1:00pm
Where: IREAP Large Conference Room, ERF 1207
Speaker: Mat (UMD) -
Abstract: A granular material is generally defined as a large collection of discrete, macroscopic particles which primarily interact via repulsive contact forces. Despite the ubiquity of such materials in nature and industry, many characteristics and phenomena of granular media remain difficult to predict and fully describe. The athermal nature of grains percludes these systems from the scope of typical statistical mechanics and thermodynamics, while continuum fluid models may neglect vital microscopic information. One method of experimentally tackling open problems in granular physics is to implement a three-dimensional (3D) imaging technique, which allows one to directly probe individual grain motion and interactions. Microscopic dynamics may lead to mesoscopic rearrangements within and across neighborhoods, and even to bulk phenomena observable at the system scale. One such technique, refractive index-matched scanning, allows non-intrusive imaging of a 3D granular pile under quasistatic circular-shear, and has led to advances in understanding irreversibility and fracture. In this talk, I will describe how this technique has been expanded to characterize individual particle rotation and shear-driven segregation—the tendency of a granular mixture to separate under driving.

Sidewinding and swimming in sand

When: Thu, February 20, 2014 - 12:30pm
Where: IREAP Large Conference Room, ERF 1207
Speaker: Prof. Dan Goldman (Georgia Institute of Technology) - http://crablab.gatech.edu/
Abstract: Many limbless locomotors inhabit sandy deserts, and the bio and neuromechanical principles by which they move effectively in dry granular media are largely unknown. I will describe progress we have made in understanding locomotion both on and below the surface of granular media. We study the above-surface locomotion of the Sidewinder rattlesnake, Crotalus cerastes, a snake that uses a peculiar form of locomotion called sidewinding. In this gait, the animal translates through successive lifting of body segments while other segments remain in static (rolling/peeling) contact with the ground. We find that snake employs an anti-slip strategy to negotiate granular inclines, and can ascend surfaces even close to the angle of repose. Implementing this control strategy in a limbless robot enables the robot to ascend sandy inclines with comparable ability to the animal. I will also describe our study of the subsurface locomotion of the long and slender Mojave shovel-nosed snake, Chionactis occipitalis. Our high speed x-ray imaging reveals that this snake “swims” within sand, and we use a granular resistive force theory (inspired by theory used to describe low Reynolds number swimmers) to model the snake’s locomotion as it swims in a self-generated granular frictional fluid. Our model reveals that the shovel-nosed snake uses a traveling wave pattern which maximizes forward speed and minimizes energy consumption. Our biological studies also reveal new physics of localized intrusion and drag in dry granular media.

Periodic windows within windows within windows

When: Thu, February 27, 2014 - 12:30pm
Where: IREAP Large Conference Room, ERF 1207
Speaker: Madhura Joglekar (UMD) -
Abstract: For a dynamical system xn+1 = fC(xn), there are often infinitely many periodic windows, that is, intervals in the parameter C in which there is stable periodic behavior, followed by period-doubling to chaos, followed by the merging of chaotic attractors, followed by a boundary crisis. The windows display a fractal structure, wherein each window has windows of higher orders. Fortunately there is one situation where these nested windows can be studied effectively, the quadratic map fC(x) = C-x2. For a given small e > 0, we say C is e-uncertain if there is a periodic attractor for exactly one of the values C and C+e. The other presumably has chaotic behavior. We find that for e very small, for the great majority of e-uncertain pairs, one of the two parameters is in a very high order window, that is, a window within a window within a window ... N times for large N.

Avoiding extremes in chaotic systems

When: Thu, February 27, 2014 - 1:00pm
Where: IREAP Large Conference Room, ERF 1207
Speaker: Suddhasattwa Das (UMD) -
Abstract: In chaotic dynamical systems subjected to disturbances which are stronger than the the available control, it is not possible to follow a particular trajectory but the system can be partially controlled to avoid extreme values by staying confi?ned within some compact set. Economic models can be prone to crashes due to the presence of chaotic dynamics. This paper explains for a toy economic model how a general 1-dimensional system can be regulated by the application of relatively weak control, even in the presence of strong external disturbances, thereby avoiding severe downturns in the economy. This partial controllability is defi?ned through the concepts of safe sets. We describe how the safe set varies with parameters, sometimes continuously or discontinuously.

Understanding Complex Adaptive Systems: An Evolutionary Agent-based Approach

When: Thu, March 6, 2014 - 12:30pm
Where: IREAP Large Conference Room, ERF 1207
Speaker: Prof. Kenneth De Jong (GMU) - http://www.cs.gmu.edu/~kdejong/
Abstract: Understanding the properties and the behavior of complex adaptive systems continues to be an important but difficult issue. While some progress can be made using formal analysis tools, the complex non-linear dynamics of many systems of interest often require a significant number of simplifying assumptions in order to produce formal results. We have been exploring an alternative approach that combines the power of two methodologies: agent-based modeling and evolutionary computation. In this talk I will describe the architecture of this hybrid approach and illustrate its application to a number of problem areas including computer network security, homeland defense, and inhalation anthrax studies.

A network measure for the analysis of large-scale graphs

When: Thu, March 27, 2014 - 12:30pm
Where: IREAP Large Conference Room, ERF 1207
Speaker: Roldan Pozo (NIST) -
Abstract: Given an undirected graph, we describe a two-dimensional integer-valued measure, the Q-matrix, based on the connected component size distribution of its degree-limited subgraphs. This generalization of the degree distribution yields a small sparse matrix whose values the number of connected components of a particular size during a prescribed percolation process. When viewed as a two-dimensional histogram, this matrix yields a canonic visual representation, i.e. an identification portrait, revealing important network properties. It can be used as a baseline classification tool, as networks from particular application areas (e.g. co-authorship networks, citation networks, peer-to-peer networks) look similar under this transformation. Finally, this network measure can be used as an analysis tool to help distinguish between real-data networks and random graphs with the same degree-distribution (configuration model) or joint-degree distribution.

Charting Collections of Connections in Social Media: Creating Maps and Measures with NodeXL

When: Thu, April 3, 2014 - 12:30pm
Where: IREAP Large Conference Room, ERF 1207
Speaker: Dr. Marc A. Smith (Connection Action Consulting Group) - http://www.connectedaction.net/marc-smith/
Abstract: Abstract:

Networks are a data structure commonly found across all social media services that allow populations to author collections of connections. The Social Media Research Foundation's NodeXL project makes analysis of social media networks accessible to most users of the Excel spreadsheet application. With NodeXL, Networks become as easy to create as pie charts. Applying the tool to a range of social media networks has already revealed the variations present in online social spaces. A review of the tool and images of Twitter, flickr, YouTube, and email networks will be presented.



Description:

We now live in a sea of tweets, posts, blogs, and updates coming from a significant fraction of the people in the connected world. Our personal and professional relationships are now made up as much of texts, emails, phone calls, photos, videos, documents, slides, and game play as by face-to-face interactions. Social media can be a bewildering stream of comments, a daunting fire hose of content. With better tools and a few key concepts from the social sciences, the social media swarm of favorites, comments, tags, likes, ratings, and links can be brought into clearer focus to reveal key people, topics and sub-communities. As more social interactions move through machine-readable data sets new insights and illustrations of human relationships and organizations become possible. But new forms of data require new tools to collect, analyze, and communicate insights.



Speaker Bio:

Marc Smith is a sociologist specializing in the social organization of online communities and computer mediated interaction. Smith leads the Connected Action consulting group and lives and works in Silicon Valley, California. Smith co-founded the Social Media Research Foundation, a non-profit devoted to open tools, data, and scholarship related to social media research. Smith received a B.S. in International Area Studies from Drexel University in Philadelphia in 1988, an M.Phil. in social theory from Cambridge University in 1990, and a Ph.D. in Sociology from UCLA in 2001. He is an adjunct lecturer at the College of Information Studies at the University of Maryland. Smith is also a Distinguished Visiting Scholar at the Media-X Program at Stanford University.

Best strategies for traversing complex landscapes, or pitfalls of simulated annealing and other "blind" Monte-Carlo techniques

When: Thu, April 10, 2014 - 12:30pm
Where: IREAP Large Conference Room, ERF 1207
Speaker: Dr. Alexandre Morozov (Rutgers University) - http://www.physics.rutgers.edu/~morozov/
Abstract: I will describe a novel strategy for Monte Carlo (MC) optimization on rugged multidimensional landscapes.
The strategy is based on querying the statistical properties of the landscape in order to find the temperature
or set of temperatures which minimize the mean first passage time across the landscape. Thus, in
contrast to other MC algorithms such as simulated annealing, the temperature schedule is explicitly matched to
the statistics of landscape irregularities. In cases where these statistics are approximately invariant over the
entire landscape (e.g. if non-local moves couple distant parts of the landscape), a single-temperature MC
scheme outperforms more "advanced" MC algorithms. Furthermore, we find that in strongly
anisotropic problems such as Coulomb spin glass and traveling salesman, the only relevant statistics
are those of irregularities in low-energy funnels. Our results are applicable to a wide range of
optimization scenarios in science and engineering, and may provide insights into efficiency of protein folding at
room temperatures.

References:

1. Tolkunov and Morozov "Single Temperature for Monte Carlo Optimization on Complex Landscapes", Phys. Rev. Lett. 108, 250602 (2012)

3D dynamics of granular materials: shear, segregation, and rotation

When: Thu, April 17, 2014 - 12:30pm
Where: IREAP Large Conference Room, ERF 1207
Speaker: Matt Harrington (UMD) -

A broken-contact order parameter for inferring inter-cellular communication from patterns of externally guided migrating cells

When: Thu, April 17, 2014 - 1:00pm
Where: IREAP Large Conference Room, ERF 1207
Speaker: Joshua Parker (UMD) -

Unbiased Learning from Big Data: Multidimensional Approaches for Data-Driven Biomedical Research

When: Thu, April 24, 2014 - 12:30pm
Where: IREAP Large Conference Room, ERF 1207
Speaker: Dr. Julian Candia (Department of Physics, UMD) - http://www.juliancandia.com/

Minimal human nature models

When: Thu, May 1, 2014 - 12:30pm
Where: IREAP Large Conference Room, ERF 1207
Speaker: Safa Motesharrei (UMD) -

Feedback control as framework for understanding tradeoffs in biology

When: Thu, May 8, 2014 - 12:30pm
Where: IREAP Large Conference Room, ERF 1207
Speaker: Noah Cowan (Johns Hopkins University) - http://limbs.lcsr.jhu.edu/