arXiv subjects

Jungwon Kim

Publications and source records attributed to Jungwon Kim.

At least 19 recordsLinked to original sources

arXiv2026

From Simulated Citizens to Simulated Deliberation: Challenges in Representation and Interaction

Multi-agent LLM deliberation has been explored as a scalable way to simulate public deliberation. For such simulations to be informative, persona agents should reflect population opinion patterns and interaction should shape their conclusions. We evaluate whether LLM-based deliberation can meet these two conditions using census-grounded Korean personas debating real policy questions benchmarked against national surveys. Persona agents do not reliably reproduce population opinion patterns: responses are often far more concentrated and frequently reverse demographic differences in the human data. Deliberations nonetheless produce reasoned, reciprocal, and varied arguments alongside substantial stance movement. Yet much of this movement does not require peer exchange: sealed-monologue agents change position at similar rates and reach nearly the same final balance as full debates, while groups initialized with very different positions often converge to similar endpoints. Anchoring population-informed starting positions, meanwhile, sharply suppresses updating. Thus, population representation, argument generation, and interaction-driven opinion change do not necessarily go together. The simulations readily surface arguments on both sides, though whether they capture the diversity of human perspectives remains untested, leaving open a promising role for argument surfacing even as population simulation requires further validation.

cs.AI↗

arXiv2026

PolicyGuard: Prompt-Configurable Semantic DLP for LLM Coding Agents

AI coding agents accept free-form natural language prompts that may inadvertently contain credentials, personally identifiable information (PII), or proprietary business data. Existing data loss prevention (DLP) solutions rely on rigid regex patterns, model fine-tuning, or vendor-managed classifiers with limited customizability. We present PolicyGuard, a pre-model interception framework that classifies user prompts using an LLM guided by a natural language policy file. Our key contributions are: (1) the policy-as-prompt paradigm, where DLP classification criteria are defined entirely in a plaintext policy document editable by non-engineers without code changes or model retraining; (2) a sealed evaluation protocol with template-family-level data splits, hidden holdouts, and frozen test sets to rigorously assess generalization; and (3) a comprehensive empirical evaluation across 2,000 multilingual prompts demonstrating 96.5% effective block rate (EBR) with only 3.0% false positive rate (FPR) on a frozen test set of 927 prompts, and perfect 100% accuracy on a 217-prompt hidden holdout. Information-matched baseline experiments show that PolicyGuard's natural language format significantly outperforms equivalent content in JSON format (McNemar chi-squared = 31.58, p < 0.001) and dramatically outperforms zero-shot classification (Cohen's h = 0.915). Cross-model portability experiments demonstrate that the same policy achieves 86.4-96.5% EBR across four different LLMs without modification.

cs.CR↗

arXiv2026

Self-stabilization of microcombs

Optical frequency combs form phase-locked spectral lines arranged on an equidistant grid fully defined by two degrees of freedom, i.e., the repetition rate and frequency offset. Stabilizing these parameters to a common frequency reference results in a coherent frequency ruler, central for modern precision metrology. However, extending this level of stability to chip-scale microcombs remains an outstanding challenge. Here, we demonstrate a self-stabilizing mechanism based on self-injection locking of a selected comb line via an external feedback loop. This process establishes a second anchor point in addition to the pump, thereby constraining the comb's frequency noise dynamics. We show that, with an appropriate choice between pump frequency noise and feedback strength, collective fluctuations of the repetition rate are strongly suppressed. The result is a microcomb exhibiting ultralow phase noise and dramatically reduced timing jitter. In a 100 GHz silicon nitride soliton microcomb, we achieve an unprecedented combination of high-conversion efficiency, sub-Hertz intrinsic linewidth across the entire C band, and an integrated timing jitter of 1 fs. This approach enables chip-scale microcombs with remarkable noise performance and fs-level pulse stability, surpassing conventional noise limits and opening new avenues for precision metrology at the chip scale.

physics.optics↗

arXiv2026

Atomic-referenced Hz-linewidth lasers via fiber interferometric stabilization

Narrow-linewidth lasers with absolute frequency anchoring are essential for precision metrology, coherent sensing, and emerging quantum technologies beyond laboratory environments. Optical cavities and interferometers provide exceptional short-term spectral purity but lack intrinsic absolute frequency references. Atomic transitions, in contrast, provide stable frequency anchors but offer limited discrimination sensitivity. Recent hybrid approaches have demonstrated the combination of compact optical resonators with atomic references, yet achieving the Hz-level regime remains challenging. Here, we present a hybrid architecture that enables simultaneous realization of Hz-level linewidth and atomic-referenced frequency stability. An external-cavity diode laser is first stabilized to a fiber interferometer to achieve Hz-level spectral purity, while the interferometer is subsequently anchored to an 87Rb D2 transition via modulation transfer spectroscopy to suppress long-term drift and define the laser frequency relative to the atomic transition. This dual-stabilization scheme realizes a compact atomic-referenced laser with a 3.4-Hz linewidth (1-rad integrated-phase method), a minimum fractional frequency stability of 3.4x10-14 at 0.56 s, and 9x10-13 at 100 s. This architecture establishes a practical and scalable route toward compact and field-deployable atomic-referenced narrow-linewidth lasers for precision metrology and quantum technologies.

physics.optics↗

arXiv2026

Ideological Bias in LLMs' Economic Causal Reasoning

Do large language models (LLMs) exhibit systematic ideological bias when reasoning about economic causal effects? As LLMs are increasingly used in policy analysis and economic reporting, where directionally correct causal judgments are essential, this question has direct practical stakes. We present a systematic evaluation by extending the EconCausal benchmark with ideology-contested cases - instances where intervention-oriented (pro-government) and market-oriented (pro-market) perspectives predict divergent causal signs. From 10,490 causal triplets (treatment-outcome pairs with empirically verified effect directions) derived from top-tier economics and finance journals, we identify 1,056 ideology-contested instances and evaluate 20 state-of-the-art LLMs on their ability to predict empirically supported causal directions. We find that ideology-contested items are consistently harder than non-contested ones, and that across 18 of 20 models, accuracy is systematically higher when the empirically verified causal sign aligns with intervention-oriented expectations than with market-oriented ones. Moreover, when models err, their incorrect predictions disproportionately lean intervention-oriented, and this directional skew is not eliminated by one-shot in-context prompting. These results highlight that LLMs are not only less accurate on ideologically contested economic questions, but systematically less reliable in one ideological direction than the other, underscoring the need for direction-aware evaluation in high-stakes economic and policy settings.

cs.AI↗

arXiv2025

GeoReg: Weight-Constrained Few-Shot Regression for Socio-Economic Estimation using LLM

Socio-economic indicators like regional GDP, population, and education levels, are crucial to shaping policy decisions and fostering sustainable development. This research introduces GeoReg a regression model that integrates diverse data sources, including satellite imagery and web-based geospatial information, to estimate these indicators even for data-scarce regions such as developing countries. Our approach leverages the prior knowledge of large language model to address the scarcity of labeled data, with the language model functioning as a data engineer by extracting informative features to enable effective estimation in few-shot settings. Specifically, our model obtains contextual relationships between data features and the target indicator, categorizing their correlations as positive, negative, mixed, or irrelevant. These features are then fed into the linear estimator with tailored weight constraints for each category. To capture nonlinear patterns, the model also identifies meaningful feature interactions and integrates them, along with nonlinear transformations. Experiments across three countries at different stages of development demonstrate that our model outperforms baselines in estimating socio-economic indicators, even for low-income countries with limited data availability.

cs.LG↗

arXiv2025

Optical frequency comb integration in radio telescopes: advancing signal generation and phase calibration

Very long baseline interferometry (VLBI) enables high-angular-resolution observations in astronomy and geodesy by synthesizing a virtual telescope with baselines spanning hundreds to thousands of kilometres. Achieving high instrumental phase stability in VLBI relies on the generation of high-quality, atomic-referenced RF local oscillator (LO) and RF-comb signals for the effective downconversion of celestial RF signals and precise phase calibration, respectively. As observing frequencies move into higher ranges with wider bandwidth, conventional electronic methods face significant challenges in maintaining the quality of these signals. Here, we demonstrate that an optical frequency comb (OFC) can be used as a versatile tool to generate and distribute low-noise and atomic-referenced RF-comb and RF-LO signals in the VLBI telescope. Hydrogen maser-stabilized optical pulses are transmitted over a timing-stabilized fibre link from the observatory building to the VLBI receiver system at the telescope, where photodetection converts them into the required RF-comb and RF-LO signals. In VLBI test observation, we successfully detected VLBI fringes and extracted the RF-combs characteristics in a format suitable for VLBI instrumental phase calibration. These results highlight the high potential of OFC-based technology for enhancing next-generation broadband VLBI measurements, advancing astrophysical research and facilitating intercontinental clock comparison.

astro-ph.IM↗

arXiv2024

10$^{-15}$-level laser stabilization down to fiber thermal noise limit using self-homodyne detection

We demonstrate a self-homodyne detection method to stabilize a continuous-wave 1550-nm laser to a 1-km optical fiber delay line, achieving a frequency instability of 6.3x10 -15 at a 16-ms averaging time. This result, limited by fiber thermal noise, is achieved without the need for a vacuum system, highlighting the potential of our approach for ultra-stable laser systems in non-laboratory environments. The system utilizes only a few passive fiber optic components and a single balanced photodetector, significantly simplifying the laser stabilization process while maintaining high performance. The entire optical setup is compactly packaged in a portable metal air-tight case.

physics.optics↗

arXiv2023

Palm-sized, vibration-insensitive and vacuum-free all-fiber-photonic module for 10-14-level stabilization of CW lasers and frequency combs

Compact and robust frequency-stabilized laser sources are critical for a variety of fields that require stable frequency standards, including field spectroscopy, radio astronomy, microwave generation, and geophysical monitoring. In this work, we applied a simple and compact fiber ring-resonator configuration that can stabilize both a continuous-wave laser and a self-referenced optical frequency comb to a vibration-insensitive optical fiber delay-line. We could achieve a thermal-noise-limited frequency noise level in the 10 Hz - 1 kHz offset frequency range for both the continuous-wave laser and the optical frequency comb with the minimal frequency instability of 2.7x10-14 at 0.03-s and 2.6x10-14 at 0.01-s averaging time, respectively, in non-vacuum condition. The optical fiber spool, working as a delay reference, is designed to be insensitive to external vibration, with a vibration sensitivity of sub-10-10 [1/g] and volume of 32 mL. Finally, the ring-resonator setup is packaged in a palm-sized aluminum case with 171-mL volume with a vibration-insensitive spool, as well as an even smaller 97-mL-volume case with an ultra-compact 9-mL miniaturized fiber spool.

physics.optics↗

arXiv2020

Photonic flywheel in a monolithic fiber resonator

We demonstrate the first compact photonic flywheel with sub-fs time jitter (averaging times up to 10 μs) at the quantum-noise limit of a monolithic fiber resonator. Such quantum-limited performance is accessed through novel two-step pumping scheme for dissipative Kerr soliton (DKS) generation. Controllable interaction between stimulated Brillouin lasing and Kerr nonlinearity enhances the DKS coherence and mitigate the thermal instability challenge, achieving a remarkable 22-Hz intrinsic comb linewidth and an unprecedented phase noise of -180 dBc/Hz at 945 MHz carrier at free running. The scheme can be generalized to various device platforms for field-deployable precision metrology.

physics.optics↗

arXiv2018

Highly tunable repetition-rate multiplication of mode-locked lasers using all-fibre harmonic injection locking

Higher repetition-rate optical pulse trains have been desired for various applications such as high-bit-rate optical communication, photonic analogue-to-digital conversion, and multi- photon imaging. Generation of multi GHz and higher repetition-rate optical pulse trains directly from mode-locked oscillators is often challenging. As an alternative, harmonic injection locking can be applied for extra-cavity repetition-rate multiplication (RRM). Here we have investigated the operation conditions and achievable performances of all-fibre, highly tunable harmonic injection locking-based pulse RRM. We show that, with slight tuning of slave laser length, highly tunable RRM is possible from a multiplication factor of 2 to >100. The resulting maximum SMSR is 41 dB when multiplied by a factor of two. We further characterize the noise properties of the multiplied signal in terms of phase noise and relative intensity noise. The resulting absolute rms timing jitter of the multiplied signal is in the range of 20 fs to 60 fs (10 kHz - 1 MHz) for different multiplication factors. With its high tunability, simple and robust all-fibre implementation, and low excess noise, the demonstrated RRM system may find diverse applications in microwave photonics, optical communications, photonic analogue-to-digital conversion, and clock distribution networks.

physics.optics↗

arXiv2017

Ultrasensitive, high-dynamic-range and broadband strain sensing by time-of-flight detection with femtosecond-laser frequency combs

Ultrahigh-resolution optical strain sensors provide powerful tools in various scientific and engineering fields, ranging from long-baseline interferometers to civil and aerospace industries. Here we demonstrate an ultrahigh-resolution fibre strain sensing method by directly detecting the time-of-flight (TOF) change of the optical pulse train generated from a free-running passively mode-locked laser (MLL) frequency comb. We achieved a local strain resolution of 18 pε/Hz1/2 and 1.9 pε/Hz1/2 at 1 Hz and 3 kHz, respectively, with largedynamic range of >154 dB at 3 kHz. For remote-point sensing at 1-km distance, 80 pε/Hz1/2 (at 1 Hz) and 2.2 pε/Hz1/2 (at 3 kHz) resolution is demonstrated. While attaining both ultrahigh resolution and large dynamic range, the demonstrated method can be readily extended for multiple-point sensing as well by taking advantage of the broad optical comb spectra. These advantages may allow various applications of this sensor in geophysical science, structural health monitoring, and underwater science.

physics.optics↗

arXiv2014

Characterization of timing jitter spectra in free-running mode-locked lasers with 340 dB dynamic range over 10 decades of Fourier frequency

We demonstrate a method that enables accurate timing jitter spectral density characterization of free-running mode-locked laser oscillators over more than 10-decade of Fourier frequency from mHz to tens MHz range. The method is based on analyzing both the input voltage noise to the slave laser and the output voltage noise from the balanced optical cross- correlator (BOC), when two mode-locked lasers are synchronized in repetition rate by the BOC. As a demonstration experiment, timing jitter spectrum of a free-running mode-locked Er-fiber laser with a dynamic range of >340 dB is measured over Fourier frequency ranging from 1 mHz to 38.5 MHz (Nyquist frequency). The demonstrated method can resolve different noise mechanisms that cause specific jitter characteristics in free-running mode-locked laser oscillators for a vast range of time scales from <100-ns to >1000-s.

physics.optics↗

arXiv2014

Sub-20-Attosecond Timing Jitter Mode-Locked Fiber Lasers

We demonstrate 14.3-attosecond timing jitter [integrated from 10 kHz to 94 MHz offset frequency] optical pulse trains from 188-MHz repetition-rate mode-locked Yb-fiber lasers. In order to minimize the timing jitter, we shorten the non-gain fiber length to shorten the pulsewidth and reduce excessive higher-order nonlinearity and nonlinear chirp in the fiber laser. The measured jitter spectrum is limited by the amplified spontaneous emission limited quantum noise in the 100 kHz - 1 MHz offset frequency range, while it was limited by the relative intensity noise-converted jitter in the lower offset frequency range. This intrinsically low timing jitter enables sub-100-attosecond synchronization between the two mode-locked Yb-fiber lasers over the full Nyquist frequency with a modest 10-kHz locking bandwidth. The demonstrated performance is the lowest timing jitter measured from any free-running mode-locked fiber lasers, comparable to the performance of the lowest-jitter Ti:sapphire solid-state lasers.

physics.optics↗

arXiv2013

Frequency comb-based microwave transfer over fiber with $7 \times 10^{-19}$ instability using fiber-loop optical-microwave phase detectors

We demonstrate a remote microwave/radio-frequency (RF) transfer technique based on the stabilization of a fiber link using a fiber-loop optical-microwave phase detector (FLOM-PD). This method compensates for the excess phase fluctuations introduced in fiber transfer by direct phase comparison between the optical pulse train reflected from the remote site and the local microwave/RF signal using the FLOM-PD. This enables sub-fs resolution and long-term stable link stabilization while having wide timing detection range and less demand in fiber dispersion compensation. The demonstrated relative frequency instability between 2.856-GHz RF oscillators separated by a 2.3-km fiber link is $7.6 \times 10^{-18}$ and $6.5 \times 10^{-19}$ at 1000 s and 82500 s averaging time, respectively.

physics.optics↗

arXiv2013

Gigahertz repetition rate, sub-femtosecond timing jitter optical pulse train directly generated from a mode-locked Yb:KYW laser

We show that a 1.13-GHz repetition rate optical pulse train with 0.70 fs high-frequency timing jitter (integration bandwidth of 17.5 kHz - 10 MHz, where the measurement instrument-limited noise floor contributes 0.41 fs in 10 MHz bandwidth) can be directly generated from a free-running, single-mode diode-pumped Yb:KYW laser mode-locked by single-walled carbon nanotube (SWCNT)-coated mirrors. To our knowledge, this is the lowest timing jitter optical pulse train with the GHz repetition rate ever measured. If this pulse train is used for direct sampling of 565-MHz signals (Nyquist frequency of the pulse train), the demonstrated jitter level corresponds to the projected effective-number-of-bit (ENOB) of 17.8, which is much higher than the thermal noise limit of 50-ohm load resistance (~14 bits).

physics.optics↗

arXiv2013

Ultralow Phase Noise Microwave Generation From Mode-Locked Er-Fiber Lasers With Subfemtosecond Integrated Timing Jitter

We demonstrate ultralow phase noise 10-GHz microwave signal generation from a free-running mode-locked Er-fiber laser with -142 and -157 dBc/Hz single-sideband absolute phase noise at 10- and 100-kHz offset frequencies, respectively. The absolute RMS timing jitter is 1.5 fs when integrated from 1-kHz to 5-GHz (Nyquist frequency) offset frequency. In the 10-kHz to 10-MHz integration bandwidth typically used for microwave generators, the RMS integrated jitter is 0.49 fs. The Er-fiber laser is operated in the stretched-pulse regime at close-to-zero dispersion to minimize the phase noise of extracted microwaves. In order to suppress the excess phase noise in the optical-to-electronic conversion process, we synchronize a low-noise voltage-controlled oscillator to the fiber laser using a fiber Sagnac-loop-based optical-microwave phase detector.

physics.optics↗

arXiv2012

Sub-femtosecond synchronization of microwave oscillators with mode-locked Er-fiber lasers

We synchronize a 8.06-GHz microwave signal from a voltage-controlled oscillator with an optical pulse train from a 77.5- MHz mode-locked Er-fiber laser using a fiber-based optical-microwave phase detector. The residual phase noise between the optical pulse train and the synchronized microwave signal is -133 dBc/Hz (-154 dBc/Hz) at 1 Hz (5 kHz) offset frequency, which results in 838 as integrated rms timing jitter [1 Hz - 1 MHz]. The long-term residual phase drift is 847 as (rms) measured over 2 hours, which reaches 4\times10^{-19} fractional frequency instability at 1800 s averaging time. This method has a potential to provide both sub-fs-level short-term phase noise and long-term phase stability in microwave extraction from mode-locked fiber lasers.

physics.optics↗