arXiv subjects

Chang Sub Kim

Publications and source records attributed to Chang Sub Kim.

12 recordsLinked to original sources

arXiv2024

Bayesian Mechanics of Synaptic Learning under the Free Energy Principle

The brain is a biological system comprising nerve cells and orchestrates its embodied agent's perception, behavior, and learning in the dynamic environment. The free energy principle (FEP) advocated by Karl Friston explicates the local, recurrent, and self-supervised neurodynamics of the brain's higher-order functions. In this paper, we continue to finesse the FEP through the physics-guided formulation; specifically, we apply our theory to synaptic learning by considering it an inference problem under the FEP and derive the governing equations, called Bayesian mechanics. Our study uncovers how the brain infers weight change and postsynaptic activity, conditioned on the presynaptic input, by deploying the generative models of the likelihood and prior belief. Consequently, we exemplify the synaptic plasticity in the brain with a simple model: we illustrate that the brain organizes an optimal trajectory in neural phase space during synaptic learning in continuous time, which variationally minimizes synaptic surprisal.

q-bio.NC↗

arXiv2022

Free energy and inference in living systems

Organisms are nonequilibrium, stationary systems self-organized via spontaneous symmetry breaking and undergoing metabolic cycles with broken detailed balance in the environment. The thermodynamic free-energy principle describes an organism's homeostasis as the regulation of biochemical work constrained by the physical free-energy cost. In contrast, recent research in neuroscience and theoretical biology explains a higher organism's homeostasis and allostasis as Bayesian inference facilitated by the informational free energy. As an integrated approach to living systems, this study presents a free-energy minimization theory overarching the essential features of both the thermodynamic and neuroscientific free-energy principles. Our results reveal that the perception and action of animals result from active inference entailed by free-energy minimization in the brain, and the brain operates as Schr{ö}dinger's machine conducting the neural mechanics of minimizing sensory uncertainty. A parsimonious model suggests that the Bayesian brain develops the optimal trajectories in neural manifolds and induces a dynamic bifurcation between neural attractors in the process of active inference.

q-bio.NC↗

arXiv2021

Bayesian mechanics of perceptual inference and motor control in the brain

The free energy principle (FEP) in the neurosciences stipulates that all viable agents induce and minimize informational free energy in the brain to fit their environmental niche. In this study, we continue our effort to make the FEP a more physically principled formalism by implementing free energy minimization based on the principle of least action. We build a Bayesian mechanics (BM) by casting the formulation reported in the earlier publication (Kim 2018) to considering active inference beyond passive perception. The BM is a neural implementation of variational Bayes under the FEP in continuous time. The resulting BM is provided as an effective Hamilton's equation of motion and subject to the control signal arising from the brain's prediction errors at the proprioceptive level. To demonstrate the utility of our approach, we adopt a simple agent-based model and present a concrete numerical illustration of the brain performing recognition dynamics by integrating BM in neural phase space. Furthermore, we recapitulate the major theoretical architectures in the FEP by comparing our approach with the common state-space formulations.

q-bio.NC↗

arXiv2020

Deformation-induced charge redistribution in Ceria thin film at room temperature

Tuning electronic properties through strain engineering of metal oxides is an important step toward understanding electrochemical and catalytic reactions in energy storage and conversion devices. Traditionally, strain engineering studies focused on movement of oxygen ions at high temperatures (500 °C and above), complicating electrical properties by introducing mixed electronic and ionic conductivity. In this study, we demonstrate room temperature charge redistribution in a CeO2 thin film as a result of phase transformation in a localized region by mechanical deformation. Mechanical indentation of the CeO2 thin film at room temperature results in irreversible deformation. Conductive-tip atomic force microscopy (C-AFM) analysis shows increased current passing through the locally deformed area of the CeO2 thin film. Electron energy loss spectroscopy (EELS) analysis equipped with transmission electron microscopy (TEM) suggests that the increase in current contrast in the deformed region arises from an increased concentration of Ce3+ ions. We herein discuss the fundamental reason behind the increased amount of Ce3+ ions in the deformed area, based on the atomic scale computational works performed by molecular dynamics (MD) simulations and first-principles density functional theory (DFT) calculations. Plastic deformation induces a phase transformation of cubic fluorite CeO2 into a newly formed T-CeO2 structure. This phase transformation occurs mainly by oxygen ions moving closer to cerium ions to release mechanical energy absorbed in the CeO2 thin film, followed by charge redistribution from the initial CeO2 to the newly created T-CeO2 structure.

cond-mat.mtrl-sci↗

arXiv2019

Classical and quantum dissipative dynamics in Josephson junctions: an Arnold problem, bifurcation and capture into resonance

We theoretically study the phase dynamics in Josephson junctions, which maps onto the oscillatory motion of a point-like particle in the washboard potential. Under appropriate driving and damping conditions, the Josephson phase undergoes intriguing bistable dynamics near a saddle point in the quasienergy landscape. The bifurcation mechanism plays a critical role in superconducting quantum circuits with relevance to non-demolition measurements such as high-fidelity readout of qubit states. We address the question `what is the probability of capture into either basin of attraction' and answer it concerning both classical and quantum dynamics. Consequently, we derive the Arnold probability and numerically analyze its implementation of the controlled dynamical switching between two steady states under the various nonequilibrium conditions.

cond-mat.stat-mech↗

arXiv2018

Recognition Dynamics in the Brain under the Free Energy Principle

We formulate the computational processes of perception in the framework of the principle of least action by postulating the theoretical action as a time integral of the free energy in the brain sciences. The free energy principle is accordingly rephrased as that for autopoietic grounds all viable organisms attempt to minimize the sensory uncertainty about the unpredictable environment over a temporal horizon. By varying the informational action, we derive the brain's recognition dynamics (RD) which conducts Bayesian filtering of the external causes from noisy sensory inputs. Consequently, we effectively cast the gradient-descent scheme of minimizing the free energy into Hamiltonian mechanics by addressing only positions and momenta of the organisms' representations of the causal environment. To manifest the utility of our theory, we show how the RD may be implemented in a neuronally based biophysical model at a single-cell level and subsequently in a coarse-grained, hierarchical architecture of the brain. We also present formal solutions to the RD for a model brain in linear regime and analyze the perceptual trajectories around attractors in neural state space.

q-bio.NC↗

arXiv2017

The free energy principle for action and perception: A mathematical review

The 'free energy principle' (FEP) has been suggested to provide a unified theory of the brain, integrating data and theory relating to action, perception, and learning. The theory and implementation of the FEP combines insights from Helmholtzian 'perception as inference', machine learning theory, and statistical thermodynamics. Here, we provide a detailed mathematical evaluation of a suggested biologically plausible implementation of the FEP that has been widely used to develop the theory. Our objectives are (i) to describe within a single article the mathematical structure of this implementation of the FEP; (ii) provide a simple but complete agent-based model utilising the FEP; (iii) disclose the assumption structure of this implementation of the FEP to help elucidate its significance for the brain sciences.

q-bio.NC↗

arXiv2015

Statistical work-energy theorems in deterministic dynamics

We theoretically explore the Bochkov-Kuzovlev-Jarzynski-Crooks work theorems in a finite system subject to external control, which is coupled to a heat reservoir. We first elaborate the mechanical energy-balance between the system and the surrounding reservoir and proceed to formulate the statistical counterpart under the general nonequilibrium conditions. Consequently, a consistency condition is derived, underpinning the nonequilibrium equalities, both in the framework of the system-centric and nonautonomous Hamiltonian pictures and its utility is examined in a few examples. Also, we elucidate that the symmetric fluctuation associated with forward and backward manipulation of the nonequilibrium work is contingent on time-reversal invariance of the underlying mesoscopic dynamics.

cond-mat.stat-mech↗

arXiv2005

Condensate wave function of neutral Bose atoms in power-law traps

The macroscopic quantum states of a condensed neutral Bose gas in one-dimensional power-law traps are obtained by solving the Gross-Pitaevskii equation numerically. A suitable candidate for a trial wave function for the variational calculation of ground state energy is suggested, and shown that the suggested function produces a lower ground state energy than the conventional Gaussian trial wave function.

cond-mat.stat-mech↗

arXiv2004

Intra-Landau-level collective excitations in a bilayer disordered electronic system

We investigate intra-Landau-level collective excitations in a bilayer disordered two-dimensional electron system exposed to a perpendicular magnetic field. The energy spectrum is calculated within the random phase approximation by taking into account electron-impurity scattering in the self-consistent Born approximation which includes consistent vertex corrections. Signatures of these bilayer excitations in drag and collective excitation measurements are identified.

cond-mat.mes-hall↗

arXiv2003

Ultrafast spectroscopy of propagating coherent acoustic phonons in GaN/InGaN heterostructures

We show that large amplitude, coherent acoustic phonon wavepackets can be generated and detected in In$_x$Ga$_{1-x}$N/GaN epilayers and heterostructures in femtosecond pump-probe differential reflectivity experiments. The amplitude of the coherent phonon increases with increasing Indium fraction $x$ and unlike other coherent phonon oscillations, both \textit{amplitude} and \textit{period} are strong functions of the laser probe energy. The amplitude of the oscillation is substantially and almost instantaneously reduced when the wavepacket reaches a GaN-sapphire interface below the surface indicating that the phonon wavepackets are useful for imaging below the surface. A theoretical model is proposed which fits the experiments well and helps to deduce the strength of the phonon wavepackets. Our model shows that localized coherent phonon wavepackets are generated by the femtosecond pump laser in the epilayer near the surface. The wavepackets then propagate through a GaN layer changing the local index of refraction, primarily through the Franz-Keldysh effect, and as a result, modulate the reflectivity of the probe beam. Our model correctly predicts the experimental dependence on probe-wavelength as well as epilayer thickness.

cond-mat.mes-hall↗

arXiv2003

Frictional magnetodrag between spatially separated two-dimensional electron systems: Coulomb versus phonon mediated electron-electron interaction

We study the frictional drag due to Coulomb and phonon mediated electron-electron interaction in a double layer electron system exposed to a perpendicular magnetic field. Within the random phase approximation we calculate the dispersion relation of the intra Landau level magnetoplasmons at finite temperatures and distinguish their contribution to the magnetodrag. We calculate the transresistivity $ρ_{Drag}$ as a function of magnetic field $B$, temperature $T$, and interlayer spacing $Λ$ for a matched electron density. For $Λ=200$ nm we find that $ρ_{Drag}$ is solely due to phonon exchange and shows no double-peak structure as a function of $B$. For $Λ=30$ nm, $ρ_{Drag}$ shows the double-peak structure and is mainly due to Coulomb interaction. The value of $ρ_{Drag}$ is about 0.3 $Ω$ at T=2 K and for the half-filled second Lanadau level, which is about 13 times larger than the value for $Λ=200$ nm. At lower edge of the temperature interval from 0.1 to 8 K, $ρ_{Drag}/ T^{2}$ remains finite for $Λ=30$ nm while it tends to zero for $Λ=200$ nm. Near the upper edge of this interval, $ρ_{Drag}$ for $Λ=30$ nm is approximately linear in $T$ while for $Λ=200$ nm it decreases slowly in $T$. Therefore, the peak of $ρ_{Drag}/ T^{2}$ is very sharp for $Λ=200$ nm. This strikingly different magnetic field and temperature dependence of $ρ_{Drag}$ ascribe we mainly to the weak screening effect at large interlayer separations.

cond-mat.mes-hall↗