Bulletin of the American Physical Society
APS March Meeting 2023
Volume 68, Number 3
Las Vegas, Nevada (March 5-10)
Virtual (March 20-22); Time Zone: Pacific Time
Session D01: GSNP Student and Postdoc Speaker Awards |
Hide Abstracts |
Sponsoring Units: GSNP Chair: Daphne Klotsa, UNC Chapel Hill Room: Room 124 |
Monday, March 6, 2023 3:00PM - 3:12PM |
D01.00001: Session Introduction: Rules for the Awards Finalists Daphne Klotsa Session Introduction: Rules for the Awards Finalists |
Monday, March 6, 2023 3:12PM - 3:29PM |
D01.00002: The Fluctuations of Small Elastic Objects in Fluid with Linear and Nonlinear Restoring Forces Johnathon R Barbish, Hagen Gress, Kamil L Ekinci, Mark Paul As an elastic object, such as a beam, is uniformly reduced in size its resonant frequency increases while its stiffness typically decreases. As technology continues to push towards smaller objects to exploit this favorable combination, we will reach a point where Brownian motion will drive the dynamics into the nonlinear regime. For example, a strongly driven beam will yield a geometric nonlinearity where the stiffness increases cubically with displacement. We explore the stochastic dynamics of elastic objects in fluid for a range of restoring forces from linear to strongly nonlinear. For a linear restoring force the fluctuation-dissipation theorem can be used to provide a deterministic theoretical description. We compare theory with experiment for a beam that is under tension and immersed in a fluid where excellent agreement is found. We numerically probe the role of nonlinearity on the dynamics of elastic objects with a Duffing restoring force using a stochastic differential equation description. We use a finite element approach to study the dynamics of 3D elastic objects whose geometry and properties have been tailored to yield a nonlinear response. |
Monday, March 6, 2023 3:29PM - 3:46PM |
D01.00003: A Braess paradox analog in optimal search networks. Georgios Gounaris, Eleni Katifori What is the optimal network architecture to minimize the time it takes for a random walker to find a randomly selected target node? A low first encounter time is the key to successful exploration both for diffusion in real space, like the motion of a protein inside a living cell and for the stochastic transitions in an energy landscape. Intuition suggests that adding a shortcut between the random walk's starting and target nodes will reduce the pair's mean first passage time. Considering the mean first passage time between all pairs, one would assume that a topologically well-connected network would be optimal. Counterintuitively, we find that this is not always the case. We show a Braess paradox analog in the case of diffusive exploration in spatially embedded graphs in which the transit time through an edge scales as the mean squared displacement of the random walk through the edge. For regular diffusion a shortcut longer than the average edge length of the graph can deteriorate the overall search efficiency of the network, although it bridges topologically distant nodes. Ultimately, to investigate the interplay between the graph structure and anomalous diffusion, we propose an optimization scheme according to which each edge can adapt its conductivity until the graph's average pairwise mean first passage time is minimized. The optimization reveals a crossover in the network's architecture: for super-diffusive motion, the optimal graph is small-world, while for sub-diffusive propagation short-range networks are optimal. We believe that this optimization approach might give insights to investigate the mechanisms that highly optimized biological systems employ to solve the problem of efficient exploration in various length scales. |
Monday, March 6, 2023 3:46PM - 4:03PM |
D01.00004: Fluctuation theorems for halting times of computers Gülce Kardes, Édgar Roldán, Gonzalo Manzano Paule, David H Wolpert Computer science theory considers abstract systems not involving thermodynamic variables. However, real world computation is subject to fluctuations and energetic constraints. Developing a physical theory of computation is thus an open challenging task at the interface of non-equilibrium thermodynamics and computer science. An important part of this task requires deriving second laws of thermodynamics for physical processes which can be modeled as implementations of abstract models of computation, such as finite automata or Turing machines. Here we extend the martingale theory of stochastic thermodynamics to derive universal second-law-like inequalities for computational machines whose halting time is a stochastic variable. We explicitly introduce new fluctuation relations for finite automata at halting times. Furthermore, we consider the thermodynamic costs of computers observing a physical system that stop the system when it achieves a pre-defined condition. We show that executing some stopping conditions require higher computational power than others. |
Monday, March 6, 2023 4:03PM - 4:20PM |
D01.00005: Minimal entropy production in anisotropic temperature fields Olga Movilla Miangolarra Anisotropy of temperature fields, chemical potentials and ion concentration gradients provide the fuel that feeds dynamical processes that sustain life. Dynamical flows in respective environments incur losses manifested as entropy production. In this work we consider an overdamped stochastic thermodynamic system in an anisotropic temperature heat bath, and analyze the problem to minimize entropy production while driving the system between thermodynamic states in finite time. Entropy production in a fully isotropic temperature field, can be expressed as the Wasserstein-2 length of the path traversed by the thermodynamic state of the system. In the presence of an anisotropic temperature field, the mechanism of entropy production is substantially more complicated as it entails seepage of energy between the ambient heat sources. We show that, in this case, the entropy production can be expressed as the solution of a suitably constrained and generalized Optimal Mass Transport (OMT) problem. In contrast to the situation in standard OMT, entropy production may not be identically zero, even when the thermodynamic state remains unchanged. Physically, this is due to the fact that maintaining a Non-Equilibrium Steady State (NESS), incurs an intrinsic entropic cost. As already noted, NESSs are the hallmark of life. Thus our problem of minimizing entropy production appears of central importance in understanding biological processes and how they may have evolved to optimize for usage of available resources. |
Monday, March 6, 2023 4:20PM - 4:37PM |
D01.00006: Anomalous relaxation in dissipative quantum chaos Lucas Sá, Antonio M García-García, Tomaz Prosen, Pedro Ribeiro, Jacobus J Verbaarschot, Jie Ping Zheng We study the nonequilibrium dynamics of the Sachdev-Ye-Kitaev (SYK) model, N strongly coupled fermions with q-body random all-to-all interactions, coupled to a Markovian environment through jump operators either linear or quadratic in the Majoranas. We develop a dynamical mean-field theory for the Lindbladian time evolution on the Keldysh contour and numerically solve the saddle-point equations for $q=4$. For strong dissipation, the system relaxes exponentially to its steady state at a rate linear in the coupling, characteristic of a dissipation-driven relaxation. However, for weak coupling, there are oscillatory corrections to the exponential relaxation and we observe an anomalously large decay rate with a finite value even in the absence of an explicit coupling to the environment. Remarkably, close to the steady state, the real-time Lindbladian dynamics of this system is identical to the low-temperature dynamics in Euclidean time of a two-site non-Hermitian SYK with intersite coupling. Its gravity dual is similar to a Euclidean wormhole in a near-AdS2 background but with a wrong-sign Schwarzian. It is this configuration, dubbed a Keldysh wormhole, together with quantum chaos, that causes anomalous relaxation. |
Monday, March 6, 2023 4:37PM - 4:54PM |
D01.00007: Generation and convergence of DNA and RNA structural ensembles Swapnil Baral, Satvik Manjigani, Joseph Robertson, Michael Zwolak DNA and RNA secondary and higher-order structure plays a central role in biology, medicine, bioengineering, and biomolecular nanotechnology, such as RNA therapeutics, Ribocomputing, and DNA origami. However, the study of nucleic acid secondary structure is demanding, as it spans an exponentially large phase space even when employing a high-level representation of structure. In other words, the low free energy manifold can be vast and have many structures contributing to observable or functional characteristics, such as hybridization efficiency, RNA programming errors, or assembly yield. We develop an approach to generate the equilibrium manifold and a conformational classification that enables testing ensemble convergence at varying degrees of fine graining. The procedure starts with a "seed" ensemble from an Ising-like spin model. This ensemble is transferred into a nucleic acid model (such as oxDNA or oxRNA, but a fully atomistic description is also possible for smaller scale structures). Replica exchange anneals this seed ensemble. We then classify, via k-means and hierarchical clustering, structures according to a distance measure on the secondary structure. This permits testing both the convergence of the ensemble with respect to the number of members (i.e., the static ensemble) and the time dynamics (i.e., the dynamic ensemble). Nucleic acid ensembles are intractable despite the large growth in computational resources. Our approach efficiently generates an ensemble and endows it with a natural metric for convergence, an approach that can also be employed inhomogeneously to identify and converge regions of hypervariability. |
Monday, March 6, 2023 4:54PM - 5:11PM |
D01.00008: An information engine that rectifies nonequilibrium fluctuations Jannik Ehrich, Tushar K Saha, Momcilo Gavrilov, Susanne Still, David A Sivak, John Bechhoefer Information engines produce useful output work from heat by using feedback to rectify thermal fluctuations. We report on an experimental realization of such an engine that, when in contact with a bath that is out of equilibrium, can extract orders of magnitude more work than an information engine in contact with an equilibrium bath. We place a micron-scale bead in a harmonic potential that ratchets up, capturing favorable fluctuations. Adding a fluctuating electric field, which generates nonequilibrium fluctuations, drastically increases work extraction. Calculating the minimum thermodynamic costs to achieve feedback control, we estimate the engine's efficiency, illustrating that information engines in nonequilibrium baths can not only do more work but also be more efficient than conventional engines. |
Monday, March 6, 2023 5:11PM - 5:28PM |
D01.00009: Exponential sensitivity by a nonequilibrium cooperativity mechanism Jeremy A Owen, Jordan M Horowitz Sensitivity to a small perturbation is a basic observable in statistical physics, as well as an important figure of merit for many living processes, including chemical sensing, signal transduction, and morphogenesis. At thermodynamic equilibrium, the basic biophysical mechanism for sensitivity is cooperative binding, for which it can be shown that the Hill coefficient, a sensitivity measure, cannot exceed the number of binding sites. This bound is a manifestation of equilibrium laws relating sensitivity to system structure and spontaneous fluctuations, but living things can violate these laws by expending energy. In search of the ultimate limits to sensitivity, we uncover a remarkable nonequilibrium cooperative binding mechanism—nested hysteresis—with sensitivity exponential in the number of binding sites. We show that this dramatic improvement over equilibrium is, in a precise sense, the best possible. Nested hysteresis could in principle be realized in the complex, energy-expending gene regulatory systems of eukaryotes, and has implications for our understanding of the function of biomolecular condensates. |
Monday, March 6, 2023 5:28PM - 5:45PM |
D01.00010: Understanding the punctuated dynamics of scientific and technological frontiers Yian Yin, Dashun Wang Despite its widespread importance for human society, our quantitative understanding of the mechanisms governing scientific and technological frontiers remains limited. Here we collect novel data tracking 5.9M solutions of 5.8K problems across 7 diverse areas in science and technology. We find that, in contrast with predictions of canonical models, the emergence of knowledge frontiers follow a bursty pattern, highlighting an intriguing co-existence between rapid progression (unexpectedly rapid growth of frontiers) and long stasis (power-law waiting time distribution between frontiers). Building on rich literature on innovation, record statistics and cultural evolution, we build a simple yet general mathematical framework integrating incremental and radical innovations, where new solutions can be created through either minor refinements on current frontiers or novel explorations in a random manner. The model not only captures the growth pattern and punctuated nature of frontiers progress, but also offers further rich predictions that are supported across all datasets. Together, these results not only offer a new framework to understand frontiers dynamics in science and technology but also have important implications for a wide range of complex social systems. |
Monday, March 6, 2023 5:45PM - 6:02PM |
D01.00011: A Catch-22 of Reservoir Computing Yuanzhao Zhang, Sean P Cornelius Reservoir Computing (RC) is a simple and efficient model-free framework for data-driven predictions of nonlinear dynamical systems. Recently, Next Generation Reservoir Computing (NGRC) has emerged as an especially attractive variant of RC. By shifting the nonlinearity from the reservoir to the readout layer, NGRC requires less data and has fewer hyperparameters to optimize, making it suitable for challenging tasks such as predicting basins of attraction. Here, using paradigmatic multistable systems including magnetic pendulums and coupled Kuramoto oscillators, we show that the performance of NGRC models can be extremely sensitive to the choice of readout nonlinearity. In particular, by incorporating the exact nonlinearity from the original equations, NGRC trained on a single trajectory can predict pseudo-fractal basins with almost perfect accuracy. However, even a small uncertainty on the exact nonlinearity can completely break NGRC, rendering the prediction accuracy no better than chance. This creates a catch-22 for NGRC since it may not be able to make useful predictions unless a key part of the system being predicted (i.e., its nonlinearity) is already known. Our results highlight the challenges faced by data-driven methods in learning complex dynamical systems. |
Follow Us |
Engage
Become an APS Member |
My APS
Renew Membership |
Information for |
About APSThe American Physical Society (APS) is a non-profit membership organization working to advance the knowledge of physics. |
© 2024 American Physical Society
| All rights reserved | Terms of Use
| Contact Us
Headquarters
1 Physics Ellipse, College Park, MD 20740-3844
(301) 209-3200
Editorial Office
100 Motor Pkwy, Suite 110, Hauppauge, NY 11788
(631) 591-4000
Office of Public Affairs
529 14th St NW, Suite 1050, Washington, D.C. 20045-2001
(202) 662-8700