SearcharxivSearch

arXiv subjects

Kang-Hun Ahn

Publications and source records attributed to Kang-Hun Ahn.

At least 19 recordsLinked to original sources

Alternating Approach-Putt Models for Multi-Stage Speech Enhancement

Speech enhancement using artificial neural networks aims to remove noise from noisy speech signals while preserving the speech content. However, speech enhancement networks often introduce distortions to the speech signal, referred to as artifacts, which can degrade audio quality. In this work, we propose a post-processing neural network designed to mitigate artifacts introduced by speech enhancement models. Inspired by the analogy of making a `Putt' after an `Approach' in golf, we name our model PuttNet. We demonstrate that alternating between a speech enhancement model and the proposed Putt model leads to improved speech quality, as measured by perceptual quality scores (PESQ), objective intelligibility (STOI), and background noise intrusiveness (CBAK) scores. Furthermore, we illustrate with graphical analysis why this alternating Approach outperforms repeated application of either model alone.

cs.SD

Full simulation on the dynamics of auditory synaptic fusion: Strong clustering of calcium channel might be the origin of the coherent release in the auditory hair cells

The precise timing of synaptic transmission in auditory hair cells is important to hearing and speech recognition. Neurotransmitter release is an underlying step in translating sound. Thus, understanding nature of the synaptic fusion is key to understand the hearing mechanism. Extraordinary large excitatory postsynaptic currents (EPSCs) have been observed in the auditory hair cell synapse, and its origin has been controversial. It is not known yet whether the size and shape of the EPSCs are results of a big vesicle or many small vesicles. We report our numerical simulation of the vesicular fusion process from calcium channel process to the generation of EPSC currents. Our numerical experiments indicate that the origin of the large EPSC with its mysterious form is close to the scenario of the multivesicular release. The large EPSCs might be triggered by strong calcium channeling of the calcium channel clusters.

q-bio.NC

Fast frequency discrimination and phoneme recognition using a biomimetic membrane coupled to a neural network

In the human ear, the basilar membrane plays a central role in sound recognition. When excited by sound, this membrane responds with a frequency-dependent displacement pattern that is detected and identified by the auditory hair cells combined with the human neural system. Inspired by this structure, we designed and fabricated an artificial membrane that produces a spatial displacement pattern in response to an audible signal, which we used to train a convolutional neural network (CNN). When trained with single frequency tones, this system can unambiguously distinguish tones closely spaced in frequency. When instead trained to recognize spoken vowels, this system outperforms existing methods for phoneme recognition, including the discrete Fourier transform (DFT), zoom FFT and chirp z-transform, especially when tested in short time windows. This sound recognition scheme therefore promises significant benefits in fast and accurate sound identification compared to existing methods.

eess.AS

Language and Noise Transfer in Speech Enhancement Generative Adversarial Network

Speech enhancement deep learning systems usually require large amounts of training data to operate in broad conditions or real applications. This makes the adaptability of those systems into new, low resource environments an important topic. In this work, we present the results of adapting a speech enhancement generative adversarial network by finetuning the generator with small amounts of data. We investigate the minimum requirements to obtain a stable behavior in terms of several objective metrics in two very different languages: Catalan and Korean. We also study the variability of test performance to unseen noise as a function of the amount of different types of noise available for training. Results show that adapting a pre-trained English model with 10 min of data already achieves a comparable performance to having two orders of magnitude more data. They also demonstrate the relative stability in test performance with respect to the number of training noise types.

cs.SD

Amplitude death of coupled hair bundles with stochastic channel noise

Hair cells conduct auditory transduction in vertebrates. In lower vertebrates such as frogs and turtles, due to the active mechanism in hair cells, hair bundles(stereocilia) can be spontaneously oscillating or quiescent. Recently, the amplitude death phenomenon has been proposed [K.-H. Ahn, J. R. Soc. Interface, {\bf 10}, 20130525 (2013)] as a mechanism for auditory transduction in frog hair-cell bundles, where sudden cessation of the oscillations arises due to the coupling between non-identical hair bundles. The gating of the ion channel is intrinsically stochastic due to the stochastic nature of the configuration change of the channel. The strength of the noise due to the channel gating can be comparable to the thermal Brownian noise of hair bundles. Thus, we perform stochastic simulations of the elastically coupled hair bundles. In spite of stray noisy fluctuations due to its stochastic dynamics, our simulation shows the transition from collective oscillation to amplitude death as inter-bundle coupling strength increases. In its stochastic dynamics, the formation of the amplitude death state of coupled hair bundles can be seen as a sudden suppression of the displacement fluctuation of the hair bundles as the coupling strength increases. The enhancement of the signal-to-noise ratio through the amplitude death phenomenon is clearly seen in the stochastic dynamics. Our numerical results demonstrate that the multiple number of transduction channels per hair bundle is an important factor to the amplitude death phenomenon, because the phenomenon may disappear for a small number of transduction channels due to strong gating noise.

physics.bio-ph

Enhanced signal-to-noise ratios in frog hearing can be achieved through amplitude death

In the ear, hair cells transform mechanical stimuli into neuronal signals with great sensitivity relying on certain active processes. Individual hair cell bundles of non-mammals such as frogs and turtles are known to show spontaneous oscillation. However hair bundles in vivo must be quiet in the absence of stimuli, otherwise, the signal is drowned in intrinsic noise. Thus, a certain mechanism is needed to exist in order to suppress intrinsic noise. Here, through a model study of elastically coupled hair bundles of bullfrog sacculi, we show that a low stimulus threshold and a high signal-to-noise ratio (SNR) can be achieved through the amplitude death phenomenon (the cessation of spontaneous oscillations by coupling). This phenomenon occurs only when the coupled hair bundles have inhomogeneous distribution, which is likely to be the case in biological systems. We show that the SNR has non-monotonic dependence on the mass of the overlying membrane, and find out that the SNR has maximum value in the region of the amplitude death. The low threshold of stimulus through amplitude death may account for the experimentally observed high sensitivity of frog sacculi in detecting vibration. The hair bundles' amplitude death mechanism provides a smart engineering design for low-noise amplification.

physics.bio-ph

Amplitude death state for hearing

We propose amplitude death phenomenon as an underlying mechanism of auditory transduction. When non-identical auditory hair bundles are elastically coupled, their spontaneous oscillations can be quenched to form an amplitude death state. We show, in this state, the hair cells are quiet and ready to detect oscillatory stimulus with coupling-strength dependent amplification. Numerical demonstration of the mechanism suggests that the non-uniformity of coupled hair cells can contribute to noise-robust auditory transduction.

physics.bio-ph

The interplay between real and pseudo magnetic field in graphene with strain

We investigate electric and magnetic properties of graphene with rotationally symmetric strain. The strain generates large pseudo magnetic field with alternating sign in space, which forms a strongly confined quantum dot connected to six chiral channels. The orbital magnetism, degeneracy, and channel opening can be understood from the interplay between real and pseudo magnetic field which have different parities under time reversal and mirror reflection. While the orbital magnetic response of the confined state is diamagnetic, it can be paramagnetic if there is an accidental degeneracy with opposite mirror reflection parity.

cond-mat.mes-hall

Ferromagnetically coupled magnetic impurities in a quantum point contact

We investigate the ground and excited states of interacting electrons in a quantum point contact using exact diagonalization method. We find that strongly localized states in the point contact appear when a new conductance channel opens due to momentum mismatch. These localized states form magnetic impurity states which are stable in a finite regime of chemical potential and excitation energy. Interestingly, these magnetic impurities have ferromagnetic coupling, which shed light on the experimentally observed puzzling coexistence of Kondo correlation and spin filtering in a quantum point contact.

cond-mat.mes-hall

Dynamical electron transport through a nanoelectromechanical wire in a magnetic field

We investigate dynamical transport properties of interacting electrons moving in a vibrating nanoelectromechanical wire in a magnetic field. We have built an exactly solvable model in which electric current and mechanical oscillation are treated fully quantum mechanically on an equal footing. Quantum mechanically fluctuating Aharonov-Bohm phases obtained by the electrons cause nontrivial contribution to mechanical vibration and electrical conduction of the wire. We demonstrate our theory by calculating the admittance of the wire which are influenced by the multiple interplay between the mechanical and the electrical energy scales, magnetic field strength, and the electron-electron interaction.

cond-mat.mes-hall

Admittance and noise in an electrically driven nano-structure: Interplay between quantum coherence and statistics

We investigate the interplay between the quantum coherence and statistics in electrically driven nano-structures. We obtain expression for the admittance and the current noise for a driven nano-capacitor in terms of the Floquet scattering matrix and derive a non-equilibrium fluctuation-dissipation relation. As an interplay between the quantum phase coherence and the many-body correlation, the admittance has peak values whenever the noise power shows a step as a function of near-by gate voltage. Our theory is demonstrated by calculating the admittance and noise of driven double quantum dots.

cond-mat.mes-hall

Spontaneously Symmetry-Broken Current in Coupled Nanomechanical Shuttles

We investigate the transport and the dynamical properties of tunnel-coupled double charge shuttles. The oscillation frequencies of two shuttles are mode-locked to integer multiples of the applied voltage frequency $ω$. We show that left/right-symmetric double shuttles may generate direct net current due to bistable motions caused by parametric instability. The symmetry-broken direct current appears near $ω=Ω_{0}/(2j-1)$, ($j=1,2,...$), where $Ω_{0}$ is the dressed resonance frequency of the relative motion of the two shuttles.

cond-mat.mes-hall

Electrical driving single barrier spin cell

We propose a spin cell based on photon-assisted tunneling through a conventional semiconductor barrier. The Dresselhaus spin-orbit interaction is included to break the spin rotation symmetry. Due to the in-plane electric field induced asymmetric momentum distribution in one lead, continuous flows of spin currents are driven through a barrier by a AC field. The net charge current remains zero. The spin current via photon-assisted tunneling can be readily adjusted via tuning the AC frequency or the in-plane electric field. This device may function as an ideal spin cell to supply spin currents in the spintronics circuit.

cond-mat.mes-hall

Floquet scattering theory of negative magneto-conductance of 2DEG under microwave radiation

We develop a theory of magneto-transport properties of two-dimensional electron gas in cylindrical geometry where finite bias and microwave are locally applied. By invesitgating the Landauer conductance based on Floquet theory, we obtain the conductance without perturbation analysis. We show that the dc conductance becomes negative when dc electric field is applied inside the 2DEG with sufficient microwave power. A positive sign of the current is restored by increasing dc bias. We analyze the radiation induced non-equilibrium distribution function with respect to the frequency of the applied microwave.

cond-mat.mes-hall

Transport through a Strongly Correlated Quantum-Dot with Fano Interference

We present the transport properties of a strongly correlated quantum dot attached to two leads with a side coupled non-interacting quantum dot. Transport properties are analyzed using the slave boson mean field theory which is reliable in the zero temperature and low bias regime. It is found that the transport properties are determined by the interplay of two fundamental physical phenomena,i.e. the Kondo effects and the Fano interference. The linear conductance will depart from the unitary limit and the zero bias anomaly will be suppressed in the presence of interdot coupling. The zero bias shot noise Fano factor increases with the interdot coupling and tends to the Poisson value. The shot noise Fano factor shows a non-monotonic behavior as a function of the interdot coupling for various side dot energy levels.

cond-mat.mes-hall

Elementary Excitations in One-Dimensional Electromechanical Systems; Transport with Back-Reaction

Using an exactly solvable model, we study low-energy properties of a one-dimensional spinless electron fluid contained in a quantum-mechanically moving wire located in a static magnetic field. The phonon and electric current are coupled via Lorentz force and the eigenmodes are described by two independent boson fluids. At low energies, the two boson modes are charged while one of them has excitation gap due to back-reaction of the Lorentz force. The theory is illustrated by evaluating optical absorption spectra. Our results are exact and show a non-perturbative regime of electron transport.

cond-mat.mes-hall

Spectral Correlation in Incommensurate Multi-Walled Carbon Nanotubes

We investigate the energy spectra of clean incommensurate double-walled carbon nanotubes, and find that the overall spectral properties are described by the so-called critical statistics of Anderson metal-insulator transition. In the energy spectra, there exist three different regimes characterized by Wigner-Dyson, Poisson, and semi-Poisson distributions. This feature implies that the electron transport in incommensurate multi-walled nanotubes can be either diffusive, ballistic, or intermediate between them, depending on the position of the Fermi energy.

cond-mat.mes-hall

Quantum Friction of Micromechanical Resonators at Low Temperatures

Dissipation of micro- and nano-scale mechanical structures is dominated by quantum-mechanical tunneling of two-level defects intrinsically present in the system. We find that at high frequencies--usually, for smaller, micron-scale structures--a novel mechanism of phonon pumping of two-level defects gives rise to weakly temperature-dependent internal friction, $Q^{-1}$, concomitant to the effects observed in recent experiments. Due to their size, comparable to or shorter than the emitted phonon wavelength, these structures suffer from superradiance-enhanced dissipation by the collective relaxation of a large number of two-level defects contained within the wavelength.

cond-mat.mes-hall