Geonhee Oh

As the application range of light-emitting diode (LED) has rapidly expanded in various fields and functional LED-implanted electronic devices have deeply permeated human daily lives, research on the electrically off-grid power supply method for LEDs is being widely conducted around the world in order to further strengthen and utilize the strong points of LEDs. In this study, a shaft-shared triboelectric-electromagnetic hybrid portable energy harvester (STEP) considering the electromotive force-based hybridization is proposed to drive various functional LEDs with a wide range of wavelengths using biomechanical energy. The respective roles and functionalities of triboelectric nanogenerator (TENG) and electromagnetic generator (EMG) in a hybrid energy harvester are experimentally and visually demonstrated by comparative investigation in parallel and series circuits. Then, based on the experimental result, the necessity of hybridization of TENG and EMG is described from the perspective of both electromotive force and current for the LED-involved devices. When the STEP operation speed is 500 rpm, the average output voltage and current of the TENG and EMG parts are experimentally characterized, and the average output voltage from TENG is 238 V and the average output current from EMG is 18 mA. The maximum peak powers of TENG and EMG are 1.92 and 26.7 mW when the external load resistances are 30 MΩ and 400 Ω, respectively. STEP is portable in size and weight, and it can be operated by human force thanks to an optimally designed gearbox. When STEP is driven by human power, it can power a variety of functional LEDs ranging from 275 to 850 nm of wavelengths, including 250 visible, 50 infrared, and dozens of UV LEDs. Given that the superior functionalities of LEDs have been validated, and that the synergistic effect through convergence with electrical grid-independent power supply technology is thought to have great potential in a variety of fields, STEP is expected to serve as a guideline for effective energy supply strategy to various functional LEDs in related fields such as lighting, engineering, security, and healthcare industries.

Abstract Electrical signatures for quantum magnetic phases of α‐RuCl 3 sensed by an adjacent graphene field‐effect transistor (FET) are reported. The gate‐voltage dependence of the graphene FET reveals p ‐type doping beneath α‐RuCl 3 caused by the charge transfer due to the work function difference. Furthermore, the gate‐voltage hysteresis exhibits a marked change in the transport behavior with the temperature variation. The high‐ T simple paramagnetic phase and the low‐ T zigzag antiferromagnetic phase of α‐RuCl 3 exhibit negligible hysteresis. In sharp contrast, a huge hysteresis of the graphene FET at intermediate temperatures is observed. The possibility that the observed conductance hysteresis is associated with the intriguing continuum excitations of the thermally induced Kitaev paramagnet state is discussed. The results show that the proximity effect‐based electric approach can be utilized for investigating charge‐neutral quasiparticles in quantum magnetic insulators.

Light Sensing and Ranging (LiDAR) is a widely used technique for reconstructing three-dimensional (3D) scenes in a variety of applications, including augmented reality and virtual reality, automotive, industrial machine vision, earth mapping, planetary science, etc. Recent progresses in 3D stacking technologies provided an important step forward in SPAD or SiPM array development, allowing to reach smaller pitch, higher pixel count and more complex processing electronics. However, these also have certain disadvantages that should be considered in specific applications that as limited dynamic range, afterpulsing, crosstalk, and noise. For example, SiPM exhibit dark counts, which are spurious signals generated in the absence of incident photons. Dark counts contribute to the noise floor of SiPMs and can limit their sensitivity, especially in low-light conditions. Efforts are made to reduce dark counts, but they still exist to some extent. Therefore, we have quantitatively analyzed the limits of SiPM compared to APD in a noisy environment in this paper. For example, when the target size is constant, and the beam size is larger than the object, the SNR (Signal-to-Noise Ratio) of the pulsed signal due to ambient light can be analyzed mathematically.

The unstructured-textual crash descriptions recorded by police officers is rarely utilized, despite containing detailed information on traffic situations. This lack of utilization is mainly due to the difficulty in analyzing text data, as there is currently no innovative methodology for extracting meaningful information from it. Given limitations and challenges in analyzing traffic crash descriptions, this study developed a methodology to classify significant words in unstructured data that describe traffic crash scenarios into standardized data. Ultimately, a natural language processing technique, specifically a bidirectional encoder representation from transformer (BERT), was used to extract meaningful information from crash descriptions. This BERT-based model effectively extracts information on the exact collision point and the pre-crash vehicle maneuver from crash descriptions. Its practical approach allows for the interpretation of traffic crash descriptions and outperforms other natural language processing models. Importantly, this method of extracting crash scene information from traffic crash descriptions can aid in better comprehending the unique characteristics of traffic crashes. This comprehension can ultimately aid in the development of appropriate countermeasures, leading to the prevention of future traffic crashes.

The analog front-end for the Low Gain Avalanche Detector (LGAD) based precision timing application in the CMS Endcap Timing Layer (ETL) has been prototyped in a 65 nm CMOS mini-ASIC named ETROC0. Serving as the very first prototype of ETL readout chip (ETROC), ETROC0 aims to study and demonstrate the performance of the analog frontend, with the goal to achieve 40 to 50 ps time resolution per hit with LGAD (therefore reach about 30ps per track with two detector-layer hits per track). ETROC0 consists of preamplifier and discriminator stages, which amplifies the LGAD signal and generates digital pulses containing time of arrival and time over threshold information. This paper will focus on the design considerations that lead to the ETROC front-end architecture choice, the key design features of the building blocks, the methodology of using the LGAD simulation data to evaluate and optimize the front-end design. The ETROC0 prototype chips have been extensively tested using charge injection and the measured performance agrees well with simulation. The initial beam test results are also presented, with time resolution of around 33 ps observed from the preamplifier waveform analysis and around 41 ps from the discriminator pulses analysis. A subset of ETROC0 chips have also been tested to a total ionizing dose of 100 MRad with X-ray and no performance degradation been observed.

<p>In light of recent advancements in deep and machine learning, federated learning has been proposed as a means to prevent privacy invasion. However, a reconstruction attack that exploits gradients to leak learning data has recently been developed. With increasing research into federated learning and the importance of data usage, it is crucial to prepare for such attacks. Specifically, when face data are used in federated learning, the damage caused by privacy infringement can be significant. Therefore, attack studies are necessary to develop effective defense strategies against these attacks. In this study, we propose a new attack method that uses labels to achieve faster and more accurate reconstruction performance than previous reconstruction attacks. We demonstrate the effectiveness of our proposed method on the Yale Face Database B, MNIST, and CIFAR-10 datasets, as well as under non-IID conditions, similar to real federated learning. The results show that our proposed method outperforms random labeling in terms of reconstruction performance in all evaluations for MNIST and CIFAR-10 datasets in round 1.</p> <p> </p>

As the application range of light-emitting diode (LED) has rapidly expanded in various fields and functional LED-implanted electronic devices have deeply permeated human daily lives, research on the electrical grid-independent power supply method for LEDs is being widely conducted around the world in order to further strengthen and utilize the strong points of LEDs. In this study, a shaft-shared triboelectric-electromagnetic hybrid portable energy harvester (STEP) is proposed to drive various functional LEDs with a wide range of wavelengths using biomechanical energy. When the STEP operation speed is 500 rpm, the average output voltage and current of the TENG and EMG parts are experimentally characterized, and the average output voltage from TENG is 238 V and the average output current from EMG is 18 mA. STEP is portable in size and weight, and it can be operated by human force thanks to an optimally designed gearbox. When STEP is driven by human power, it can power a variety of functional LEDs ranging from 275 to 850 nm of wavelengths, including 250 visible, 50 infrared, and dozens of UV LEDs. Given that the superior functionalities of LEDs have been validated, and that the synergistic effect through convergence with electrical grid-independent power supply technology is thought to have great potential in a variety of fields, STEP is expected to serve as a guideline for effective energy supply strategy to various functional LEDs in related fields such as lighting, engineering, security, and healthcare industries.

Text-to-image generative models have garnered immense attention for their ability to produce high-fidelity images from text prompts. Among these, Stable Diffusion distinguishes itself as a leading open-source model in this fast-growing field. However, the intricacies of fine-tuning these models pose multiple challenges from new methodology integration to systematic evaluation. Addressing these issues, this paper introduces LyCORIS (Lora beYond Conventional methods, Other Rank adaptation Implementations for Stable diffusion) [https://github.com/KohakuBlueleaf/LyCORIS], an open-source library that offers a wide selection of fine-tuning methodologies for Stable Diffusion. Furthermore, we present a thorough framework for the systematic assessment of varied fine-tuning techniques. This framework employs a diverse suite of metrics and delves into multiple facets of fine-tuning, including hyperparameter adjustments and the evaluation with different prompt types across various concept categories. Through this comprehensive approach, our work provides essential insights into the nuanced effects of fine-tuning parameters, bridging the gap between state-of-the-art research and practical application.

기존의 센서 네트워크 연구들은 센서 자체가 가지는 에너지 측면만 강조하였다. 그러나 실제 센서 네트워크를 구성하였을 경우 특정 센서의 많은 활용으로 인해 센서 네트워크의 부분 단절을 초래한다. 이는 결국 센서네트워크가 오랜 시간 효율적으로 운영되지 못하는 단점이 되어 오히려 특정 센서 에너지 효율성이 센서 네트워크의 효율성을 저하시키는 결과를 초래하였다. 센서 네트워크들이 클러스터로 구성되었거나 하나의 큰 네트워크로 구성되어 있는 경우에도 센서의 에너지 효율성을 강조하기 때문에 결국 센서 네트워크의 단절을 회피할 수 없다. 따라서 센서 네트워크를 구성하는 모든 센서들을 고루 사용함으로써 센서 네트워크의 센서들이 단절을 회피하도록 하여 센서 네트워크의 수명을 연장할 수 있도록 한다. 본 논문은 유비쿼터스 환경에서 센서네트워크를 구성하는 프로토콜로 구성된 센서 네트워크의 에너지를 효율적으로 관리하여 센서 네트워크의 단절을 방지함으로써 구성된 센서 네트워크가 오랜 시간 유지되는 프로토콜을 제안한다. Existing sensor network studies have only emphasized energy aspects that sensors themselves had. But when an actual sensor network is established, biased use of a specific sensor may cause a partial disconnection of the sensor network. It becomes an disadvantage to fail efficient operation of the sensor network for a long time and energy efficiency of specific sensor energy causes to drop the efficiency of the sensor network. When a sensor network is composed of many clusters or made up of a large network, sensor network's disconnection cannot be avoided because they emphasize sensor's energy efficiency. Therefore, it was tried to lengthen the lifespan of the sensor network by making sensors in the sensor network avoid disconnection through even use of all the sensors composing of the sensor network. This article proposes a protocol to maintain a sensor network for a long time by preventing a sensor networks' disconnection through efficient management of sensor network energy composed of the protocols composing of the sensor network in ubiquitous environments.

A multi-dimensional analysis of prompt charmonia in pp and pPb collisions at sNN=5.02TeV with the CMS detector is presented. The pPb differential cross-sections of prompt J/ψ are shown in a wide kinematic region, for transverse momentum pT spanning from 2 to 30 GeV/c and a rapidity interval between -2.4 to 1.93 in the center of mass of the collision. The final results on prompt ψ(2S) meson production cross section in pp and pPb collisions at 5.02 TeV are also reported as a function of pT and rapidity, for pT from 4 to 30 GeV/c. The nuclear modification factor is found to be smaller than that of prompt J/ψ in all measured bins, especially at low pT and at backward rapidity. Such a different behaviour between the ground and excited states cannot be reproduced considering nPDF effects alone.

Bottomonia are important probes of the quark-gluon plasma since they are produced at early times and propagate through the medium. The production cross sections of the three $\Upsilon$ states (1S, 2S, 3S) were measured by CMS in pp and PbPb collisions at $\sqrt{s_{\mathrm{NN}}\space}$ = 5.02 ${\mathrm{TeV}}\space$. The nuclear modification factors, ${R_\mathrm{AA}}$, derived from the ${\text{PbPb}}$ to pp ratio of yields for each state, are studied as functions of meson rapidity between 0 to 2.4 and transverse momentum ${p_\mathrm{T}}$ spanning from 0 to 30 ${\mathrm{GeV/\it{c}}}$, as well as PbPb collision centrality 0 to 100\%. A strong suppression of \raa is observed in PbPb collisions but the $\Upsilon$(3S) was not observed clearly in PbPb collisions. The upper limit on the ${R_\mathrm{AA}}$ of $\Upsilon$(3S) integrated over ${p_\mathrm{T}}$, rapidity and centrality is 0.096 at 95\% confidence level, which is the strongest suppression observed for a quarkonium state in heavy ion collisions to date. The suppression of $\Upsilon$(1S) is larger than that seen at $\sqrt{s_{\mathrm{NN}}\space}$ = 2.76 ${\mathrm{TeV}}$, although the two are still compatible within uncertainties.

The Compact Muon Solenoid (CMS) detector at the CERN Large Hadron Collider (LHC) is undergoing an extensive Phase II upgrade program to prepare for the challenging conditions of the High-Luminosity LHC (HL-LHC). A new timing layer is designed to measure minimum ionizing particles (MIP) with a time resolution of $\sim$ 30 ps and hermetic coverage up to a pseudo-rapidity of $|\eta|$ = 3. The precise time information from the MIP timing detector (MTD) will serve as an excellent time-of-flight detector for particle identification in QCD and heavy-ion physics. Together with the wide coverage of tracker and calorimetry, the MTD will enable a broad range of new and unique opportunities in heavy-ion physics at CMS. We present the current status and ongoing R$\&$D of the MTD and performance of extending heavy-ion physics program at CMS with particle identification, focusing on measurements involving hard probes such as heavy flavor hadron reconstruction over wide rapidity down to a very low transverse momentum, correlations of jets and identified hadrons.