arXiv subjects

Sunwoo Lee

Publications and source records attributed to Sunwoo Lee.

50 records · Page 3Linked to original sources

arXiv2022

A Case Study on Parallel HDF5 Dataset Concatenation for High Energy Physics Data Analysis

In High Energy Physics (HEP), experimentalists generate large volumes of data that, when analyzed, helps us better understand the fundamental particles and their interactions. This data is often captured in many files of small size, creating a data management challenge for scientists. In order to better facilitate data management, transfer, and analysis on large scale platforms, it is advantageous to aggregate data further into a smaller number of larger files. However, this translation process can consume significant time and resources, and if performed incorrectly the resulting aggregated files can be inefficient for highly parallel access during analysis on large scale platforms. In this paper, we present our case study on parallel I/O strategies and HDF5 features for reducing data aggregation time, making effective use of compression, and ensuring efficient access to the resulting data during analysis at scale. We focus on NOvA detector data in this case study, a large-scale HEP experiment generating many terabytes of data. The lessons learned from our case study inform the handling of similar datasets, thus expanding community knowledge related to this common data management task.

cs.DC↗

arXiv2022

Partial Model Averaging in Federated Learning: Performance Guarantees and Benefits

Local Stochastic Gradient Descent (SGD) with periodic model averaging (FedAvg) is a foundational algorithm in Federated Learning. The algorithm independently runs SGD on multiple workers and periodically averages the model across all the workers. When local SGD runs with many workers, however, the periodic averaging causes a significant model discrepancy across the workers making the global loss converge slowly. While recent advanced optimization methods tackle the issue focused on non-IID settings, there still exists the model discrepancy issue due to the underlying periodic model averaging. We propose a partial model averaging framework that mitigates the model discrepancy issue in Federated Learning. The partial averaging encourages the local models to stay close to each other on parameter space, and it enables to more effectively minimize the global loss. Given a fixed number of iterations and a large number of workers (128), the partial averaging achieves up to 2.2% higher validation accuracy than the periodic full averaging.

cs.LG↗

arXiv2022

Layer-wise Adaptive Model Aggregation for Scalable Federated Learning

In Federated Learning, a common approach for aggregating local models across clients is periodic averaging of the full model parameters. It is, however, known that different layers of neural networks can have a different degree of model discrepancy across the clients. The conventional full aggregation scheme does not consider such a difference and synchronizes the whole model parameters at once, resulting in inefficient network bandwidth consumption. Aggregating the parameters that are similar across the clients does not make meaningful training progress while increasing the communication cost. We propose FedLAMA, a layer-wise model aggregation scheme for scalable Federated Learning. FedLAMA adaptively adjusts the aggregation interval in a layer-wise manner, jointly considering the model discrepancy and the communication cost. The layer-wise aggregation method enables to finely control the aggregation interval to relax the aggregation frequency without a significant impact on the model accuracy. Our empirical study shows that FedLAMA reduces the communication cost by up to 60% for IID data and 70% for non-IID data while achieving a comparable accuracy to FedAvg.

cs.LG↗

arXiv2021

SSFL: Tackling Label Deficiency in Federated Learning via Personalized Self-Supervision

Federated Learning (FL) is transforming the ML training ecosystem from a centralized over-the-cloud setting to distributed training over edge devices in order to strengthen data privacy. An essential but rarely studied challenge in FL is label deficiency at the edge. This problem is even more pronounced in FL compared to centralized training due to the fact that FL users are often reluctant to label their private data. Furthermore, due to the heterogeneous nature of the data at edge devices, it is crucial to develop personalized models. In this paper we propose self-supervised federated learning (SSFL), a unified self-supervised and personalized federated learning framework, and a series of algorithms under this framework which work towards addressing these challenges. First, under the SSFL framework, we demonstrate that the standard FedAvg algorithm is compatible with recent breakthroughs in centralized self-supervised learning such as SimSiam networks. Moreover, to deal with data heterogeneity at the edge devices in this framework, we have innovated a series of algorithms that broaden existing supervised personalization algorithms into the setting of self-supervised learning. We further propose a novel personalized federated self-supervised learning algorithm, Per-SSFL, which balances personalization and consensus by carefully regulating the distance between the local and global representations of data. To provide a comprehensive comparative analysis of all proposed algorithms, we also develop a distributed training system and related evaluation protocol for SSFL. Our findings show that the gap of evaluation accuracy between supervised learning and unsupervised learning in FL is both small and reasonable. The performance comparison indicates the representation regularization-based personalization method is able to outperform other variants.

cs.LG↗

arXiv2020

Transformable Reflective Telescope for optical testing and education

We propose and experimentally demonstrate the Transformable Reflective Telescope (TRT) Kit for educational purposes and for performing various optical tests with a single kit. The TRT Kit is a portable optical bench setup suitable for interferometry, spectroscopy, measuring stray light, and developing adaptive optics, among other uses. Supplementary modules may be integrated easily thanks to the modular design of the TRT Kit. The Kit consists of five units; a primary mirror module, a secondary mirror module, a mounting base module, a baffle module, and an alignment module. Precise alignment and focusing are achieved using a precision optical rail on the alignment module. The TRT Kit transforms into three telescope configurations: Newtonian, Cassegrain, and Gregorian. Students change telescope configurations by exchanging the secondary mirror. The portable design and the aluminum primary mirror of the TRT Kit enable students to perform experiments in various environments. The minimized baffle design utilizes commercial telescope tubes, allowing users to look directly into the optical system while suppressing stray light down to $\sim$10$^{-8}$ point source transmittance (PST). The TRT Kit was tested using a point source and field images. Point source measurement of the Newtonian telescope configuration resulted in an 80\% encircled energy diameter (EED) of 23.8 $μ$m.

physics.ed-ph↗

arXiv2020

Development of Linear Astigmatism Free -- Three Mirror System (LAF-TMS)

We present the development of Linear Astigmatism Free - Three Mirror System (LAF-TMS). This is a prototype of an off-axis telescope that enables very wide field of view (FoV) infrared satellites that can observe Paschen-$α$ emission, zodiacal light, integrated star light, and other infrared sources. It has the entrance pupil diameter of 150 mm, the focal length of 500 mm, and the FoV of 5.5$^\circ$ $\times$ 4.1$^\circ$. LAF-TMS is an obscuration-free off-axis system with minimal out-of-field baffling and no optical support structure diffraction. This optical design is analytically optimized to remove linear astigmatism and to reduce high-order aberrations. Sensitivity analysis and Monte-Carlo simulation reveal that tilt errors are the most sensitive alignment parameters that allow $\sim$1$^\prime$. Optomechanical structure accurately mounts aluminum mirrors, and withstands satellite-level vibration environments. LAF-TMS shows optical performance with 37 $μ$m FWHM of the point source image satisfying Nyquist sampling requirements for typical 18 $μ$m pixel Infrared array detectors. The surface figure errors of mirrors and scattered light from the tertiary mirror with 4.9 nm surface micro roughness may affect the measured point spread function (PSF). Optical tests successfully demonstrate constant optical performance over wide FoV, indicating that LAF-TMS suppresses linear astigmatism and high-order aberrations.

astro-ph.IM↗

arXiv2019

Improving MPI Collective I/O Performance With Intra-node Request Aggregation

Two-phase I/O is a well-known strategy for implementing collective MPI-IO functions. It redistributes I/O requests among the calling processes into a form that minimizes the file access costs. As modern parallel computers continue to grow into the exascale era, the communication cost of such request redistribution can quickly overwhelm collective I/O performance. This effect has been observed from parallel jobs that run on multiple compute nodes with a high count of MPI processes on each node. To reduce the communication cost, we present a new design for collective I/O by adding an extra communication layer that performs request aggregation among processes within the same compute nodes. This approach can significantly reduce inter-node communication congestion when redistributing the I/O requests. We evaluate the performance and compare with the original two-phase I/O on a Cray XC40 parallel computer with Intel KNL processors. Using I/O patterns from two large-scale production applications and an I/O benchmark, we show the performance improvement of up to 29 times when running 16384 MPI processes on 256 compute nodes.

cs.DC↗

arXiv2017

Bright visible light emission from graphene

Graphene and related two-dimensional materials are promising candidates for atomically thin, flexible, and transparent optoelectronics. In particular, the strong light-matter interaction in graphene has allowed for the development of state-of-the-art photodetectors, optical modulators, and plasmonic devices. In addition, electrically biased graphene on SiO2 substrates can be used as a low-efficiency emitter in the mid-infrared range. However, emission in the visible range has remained elusive. Here we report the observation of bright visible-light emission from electrically biased suspended graphenes. In these devices, heat transport is greatly minimised; thus hot electrons (~ 2800 K) become spatially localised at the centre of graphene layer, resulting in a 1000-fold enhancement in the thermal radiation efficiency. Moreover, strong optical interference between the suspended graphene and substrate can be utilized to tune the emission spectrum. We also demonstrate the scalability of this technique by realizing arrays of chemical-vapour-deposited graphene bright visible-light emitters. These results pave the way towards the realisation of commercially viable large-scale, atomically-thin, flexible and transparent light emitters and displays with low-operation voltage, and graphene-based, on-chip ultrafast optical communications.

cond-mat.mes-hall↗

arXiv2016

Graphene mechanical oscillators with tunable frequency

Oscillators, which produce continuous periodic signals from direct current power, are central to modern communications systems, with versatile applications such as timing references and frequency modulators. However, conventional oscillators typically consist of macroscopic mechanical resonators such as quartz crystals, which require excessive off-chip space. Here we report oscillators built on micron-size, atomically-thin graphene nanomechanical resonators, whose frequencies can be electrostatically tuned by as much as 14\%. The self-sustaining mechanical motion of the oscillators is generated and transduced at room temperature by simple electrical circuitry. The prototype graphene voltage controlled oscillators exhibit frequency stability and modulation bandwidth sufficient for modulation of radio-frequency carrier signals. As a demonstration, we employ a graphene oscillator as the active element for frequency modulated signal generation, and achieve efficient audio signal transmission.

cond-mat.mes-hall↗

arXiv2016

SU-8 clamped CVD graphene drum resonators

Graphene mechanical resonators are the ultimate two-dimensional nanoelectromechanical systems (NEMS) with applications in sensing and signal processing. While initial devices have shown promising results, an ideal graphene NEMS resonator should be strain engineered, clamped at the edge without trapping gas underneath, and electrically integratable. In this letter, we demonstrate fabrication and direct electrical measurement of circular SU-8 polymer-clamped chemical vapor deposition (CVD) graphene drum resonators. The clamping increases device yield and responsivity, while providing a cleaner resonance spectrum from eliminated edge modes. Furthermore, this resonator is highly strained, indicating its potential in strain engineering for performance enhancement.

cond-mat.mes-hall↗

arXiv2016

Modulation of mechanical resonance by chemical potential oscillation in graphene

The classical picture of the force on a capacitor assumes a large density of electronic states, such that the electrochemical potential of charges added to the capacitor is given by the external electrostatic potential and the capacitance is determined purely by geometry. Here we consider capacitively driven motion of a nano-mechanical resonator with a low density of states, in which these assumptions can break down. We find three leading-order corrections to the classical picture: the first of is a modulation in the static force due to variation in the internal chemical potential, the second and third are change in static force and dynamic spring constant due to the rate of change of chemical potential, expressed as the quantum (density of states) capacitance. As a demonstration, we study a capacitively driven graphene mechanical resonators, where the chemical potential is modulated independently of the gate voltage using an applied magnetic field to manipulate the energy of electrons residing in discrete Landau levels. In these devices, we observe large periodic frequency shifts consistent with the three corrections to the classical picture. In devices with extremely low strain and disorder, the first correction term dominates and the resonant frequency closely follows the chemical potential. The theoretical model fits the data with only one adjustable parameter representing disorder-broadening of the Landau levels. The underlying electromechanical coupling mechanism is not limited the particular choice of material, geometry, or mechanism for variation in chemical potential, and can thus be extended to other low-dimensional systems.

cond-mat.mes-hall↗

arXiv2016

Strong interfacial exchange field in the graphene/EuS heterostructure

Exploiting 2D materials for spintronic applications can potentially realize next-generation devices featuring low-power consumption and quantum operation capability. The magnetic exchange field (MEF) induced by an adjacent magnetic insulator enables efficient control of local spin generation and spin modulation in 2D devices without compromising the delicate material structures. Using graphene as a prototypical 2D system, we demonstrate that its coupling to the model magnetic insulator (EuS) produces a substantial MEF (> 14 T) with potential to reach hundreds of Tesla, which leads to orders-of-magnitude enhancement in the spin signal originated from Zeeman spin-Hall effect. Furthermore, the new ferromagnetic ground state of Dirac electrons resulting from the strong MEF may give rise to quantized spin-polarized edge transport. The MEF effect shown in our graphene/EuS devices therefore provides a key functionality for future spin logic and memory devices based on emerging 2D materials in classical and quantum information processing.

cond-mat.mes-hall↗

arXiv2013

Graphene Metallization of High-Stress Silicon Nitride Resonators for Electrical Integration

High stress stoichiometric silicon nitride resonators, whose quality factors exceed one million, have shown promise for applications in sensing and signal processing. Yet, electrical integration of the insulating silicon nitride resonators has been challenging, as depositing even a thin layer of metal degrades the quality factor significantly. In this work, we show that graphene used as a conductive coating for Si3N4 membranes reduces the quality factor by less than 30 % on average, which is minimal when compared to the effect of conventional metallization layers such as chromium or aluminum. The electrical integration of Si3N4-Graphene (SiNG) heterostructure resonators is demonstrated with electrical readout and electro-static tuning of the frequency by up to 1 % per volt. These studies demonstrate the feasibility of hybrid graphene/nitride mechanical resonators in which the electrical properties of graphene are combined with the superior mechanical performance of silicon nitride.

cond-mat.mtrl-sci↗

arXiv2011

Broadband Tuning of Optomechanical Cavities

We demonstrate broadband tuning of an optomechanical microcavity optical resonance by exploring the large optomechanical coupling of a double-wheel microcavity and its uniquely low mechanical stiffness. Using a pump laser with only 13 mW at telecom wavelengths we show tuning of the silicon nitride microcavity resonances over 32 nm. This corresponds to a tuning power efficiency of only 400 $μ$W/nm. By choosing a relatively low optical Q resonance ($\approx$18,000) we prevent the cavity from reaching the regime of regenerative optomechanical oscillations. The static mechanical displacement induced by optical gradient forces is estimated to be as large as 60 nm.

physics.optics↗