arXiv subjects

Sang Eon Park

Publications and source records attributed to Sang Eon Park.

16 recordsLinked to original sources

arXiv2026

Strontium ${}^{1}S_{0}\!\rightarrow\!{}^{1}P_{1}$ transition frequency measurements assisted by a photonic grating chip

We measure the absolute frequency of the ${}^{1}S_{0}\!\rightarrow\!{}^{1}P_{1}$ transition in strontium using two methods: fluorescence spectroscopy of a thermal atomic beam source from a compact low-power oven and velocity measurements of a slow atomic beam from a two-dimensional grating magneto-optical trap (2D gMOT). The measurements for both methods are performed in the same ultra-high vacuum chamber containing a diffraction grating chip which is placed below the strontium atoms that are being interrogated. The first method uses a probe laser beam incident on the grating chip such that the grating acts as an end mirror, with the first-order diffracted beam providing a retro-reflected probe beam. The counter-propagating laser beams traverse an atomic beam emitted from an oven, enabling spatially resolved fluorescence spectroscopy through CCD imaging and hyperfine-constrained multi-isotope fitting. The second method relies on a large profile cooling laser beam normally incident onto the grating chip which laser cools strontium atoms for a slow atomic beam source. The velocity of the atoms exiting the 2D gMOT is measured as a function of the laser detuning and intensity from which the resonance frequency can be estimated. The two methods are consistent within their quoted uncertainties. Using three datasets based on retro-beam spectroscopy measurements, and one dataset using slow atom beam velocity measurements, we determine the ${}^{1}S_{0}\!\rightarrow\!{}^{1}P_{1}$ transition frequency to be $650.503\,815(5)~\mathrm{THz}$. Our result provides a re-evaluation of this $461$ nm transition demonstrated on a compact laser cooling apparatus based on a diffraction grating platform.

physics.atom-ph↗

arXiv2026

FPGA-based disturbance-observer servo for broadband noise suppression in laser frequency stabilization

We have demonstrated broadband frequency-noise suppression in a laser stabilization system by augmenting a conventional proportional-integral-derivative (PID) controller with a digital disturbance observer (DOB) implemented on a field-programmable gate array (FPGA). The DOB employs a first-order exponential moving average filter as its Q-filter, replacing multi-parameter frequency-domain plant identification with a single one-dimensional gain sweep. Using modulation transfer spectroscopy on the 87Rb D2 line at 780.24 nm, we have measured the frequency-noise power spectral density and the Allan deviation of the beat note between two independently stabilized lasers. The integrated rms frequency noise below 40 kHz decreased by 16.9 dB compared with PID alone, corresponding to a reduction from approximately 140 kHz to 20 kHz. The short-term fractional frequency instability improved from sigma_y(1 ms) = $7.9 \times 10^{-12}$ to $4.6 \times 10^{-12}$, while the long-term stability at tau > 1 s remained within statistical uncertainty. This DOB-augmented architecture offers a simple and effective route to enhanced noise rejection in FPGA-based servo systems for atomic physics experiments.

physics.atom-ph↗

arXiv2026

High-stability offset-frequency locking of two lasers using a balanced filter discriminator

We demonstrate a high-stability laser offset-frequency locking technique based on a balanced filter discriminator. The beat note between two 852 nm external-cavity diode lasers is down-converted in two parallel arms using local-oscillator frequencies placed symmetrically around the desired offset frequency. After low-pass filtering and RMS detection, differential subtraction of the two detector outputs produces a dispersive frequency-error signal with a zero crossing primarily defined by the reference local-oscillator frequencies. This balanced configuration reduces sensitivity to common beat-power fluctuations and can improve the effective error-signal signal-to-noise ratio. The system was implemented for an 8.653 GHz offset corresponding to the cesium repumping frequency difference used in our laser-cooling setup. Measurements with different low-pass filters reveal a trade-off between discrimination sensitivity and feedback bandwidth. With an SLP-1.9+ filter, the locked beat frequency reached a fractional instability of $4\times10^{-15}$ at 10 s when referred to the 852 nm optical carrier. The residual dependence on photodetector optical power was also characterized, showing that amplitude-to-frequency conversion remains small in the optimized differential configuration. This approach provides a practical frequency-only offset-locking method for atomic-physics experiments requiring stable and tunable microwave-scale laser frequency offsets.

physics.atom-ph↗

arXiv2026

AutoSciDACT: Automated Scientific Discovery through Contrastive Embedding and Hypothesis Testing

Novelty detection in large scientific datasets faces two key challenges: the noisy and high-dimensional nature of experimental data, and the necessity of making statistically robust statements about any observed outliers. While there is a wealth of literature on anomaly detection via dimensionality reduction, most methods do not produce outputs compatible with quantifiable claims of scientific discovery. In this work we directly address these challenges, presenting the first step towards a unified pipeline for novelty detection adapted for the rigorous statistical demands of science. We introduce AutoSciDACT (Automated Scientific Discovery with Anomalous Contrastive Testing), a general-purpose pipeline for detecting novelty in scientific data. AutoSciDACT begins by creating expressive low-dimensional data representations using a contrastive pre-training, leveraging the abundance of high-quality simulated data in many scientific domains alongside expertise that can guide principled data augmentation strategies. These compact embeddings then enable an extremely sensitive machine learning-based two-sample test using the New Physics Learning Machine (NPLM) framework, which identifies and statistically quantifies deviations in observed data relative to a reference distribution (null hypothesis). We perform experiments across a range of astronomical, physical, biological, image, and synthetic datasets, demonstrating strong sensitivity to small injections of anomalous data across all domains.

cs.LG↗

arXiv2025

Probing Geometry of Next Token Prediction Using Cumulant Expansion of the Softmax Entropy

We introduce a cumulant-expansion framework for quantifying how large language models (LLMs) internalize higher-order statistical structure during next-token prediction. By treating the softmax entropy of each layer's logit distribution as a perturbation around its "center" distribution, we derive closed-form cumulant observables that isolate successively higher-order correlations. Empirically, we track these cumulants in GPT-2 and Pythia models on Pile-10K prompts. (i) Structured prompts exhibit a characteristic rise-and-plateau profile across layers, whereas token-shuffled prompts remain flat, revealing the dependence of the cumulant profile on meaningful context. (ii) During training, all cumulants increase monotonically before saturating, directly visualizing the model's progression from capturing variance to learning skew, kurtosis, and higher-order statistical structures. (iii) Mathematical prompts show distinct cumulant signatures compared to general text, quantifying how models employ fundamentally different processing mechanisms for mathematical versus linguistic content. Together, these results establish cumulant analysis as a lightweight, mathematically grounded probe of feature-learning dynamics in high-dimensional neural networks.

cs.CL↗

arXiv2024

Product Manifold Machine Learning for Physics

Physical data are representations of the fundamental laws governing the Universe, hiding complex compositional structures often well captured by hierarchical graphs. Hyperbolic spaces are endowed with a non-Euclidean geometry that naturally embeds those structures. To leverage the benefits of non-Euclidean geometries in representing natural data we develop machine learning on $\mathcal P \mathcal M$ spaces, Cartesian products of constant curvature Riemannian manifolds. As a use case we consider the classification of "jets", sprays of hadrons and other subatomic particles produced by the hadronization of quarks and gluons in collider experiments. We compare the performance of $\mathcal P \mathcal M$-MLP and $\mathcal P \mathcal M$-Transformer models across several possible representations. Our experiments show that $\mathcal P \mathcal M$ representations generally perform equal or better to fully Euclidean models of similar size, with the most significant gains found for highly hierarchical jets and small models. We discover significant correlation between the degree of hierarchical structure at a per-jet level and classification performance with the $\mathcal P \mathcal M$-Transformer in top tagging benchmarks. This is a promising result highlighting a potential direction for further improving machine learning model performance through tailoring geometric representation at a per-sample level in hierarchical datasets. These results reinforce the view of geometric representation as a key parameter in maximizing both performance and efficiency of machine learning on natural data.

hep-ph↗

arXiv2023

FAIR AI Models in High Energy Physics

The findable, accessible, interoperable, and reusable (FAIR) data principles provide a framework for examining, evaluating, and improving how data is shared to facilitate scientific discovery. Generalizing these principles to research software and other digital products is an active area of research. Machine learning (ML) models -- algorithms that have been trained on data without being explicitly programmed -- and more generally, artificial intelligence (AI) models, are an important target for this because of the ever-increasing pace with which AI is transforming scientific domains, such as experimental high energy physics (HEP). In this paper, we propose a practical definition of FAIR principles for AI models in HEP and describe a template for the application of these principles. We demonstrate the template's use with an example AI model applied to HEP, in which a graph neural network is used to identify Higgs bosons decaying to two bottom quarks. We report on the robustness of this FAIR AI model, its portability across hardware architectures and software frameworks, and its interpretability.

hep-ex↗

arXiv2022

Neural Embedding: Learning the Embedding of the Manifold of Physics Data

In this paper, we present a method of embedding physics data manifolds with metric structure into lower dimensional spaces with simpler metrics, such as Euclidean and Hyperbolic spaces. We then demonstrate that it can be a powerful step in the data analysis pipeline for many applications. Using progressively more realistic simulated collisions at the Large Hadron Collider, we show that this embedding approach learns the underlying latent structure. With the notion of volume in Euclidean spaces, we provide for the first time a viable solution to quantifying the true search capability of model agnostic search algorithms in collider physics (i.e. anomaly detection). Finally, we discuss how the ideas presented in this paper can be employed to solve many practical challenges that require the extraction of physically meaningful representations from information in complex high dimensional datasets.

hep-ph↗

arXiv2022

A FAIR and AI-ready Higgs boson decay dataset

To enable the reusability of massive scientific datasets by humans and machines, researchers aim to adhere to the principles of findability, accessibility, interoperability, and reusability (FAIR) for data and artificial intelligence (AI) models. This article provides a domain-agnostic, step-by-step assessment guide to evaluate whether or not a given dataset meets these principles. We demonstrate how to use this guide to evaluate the FAIRness of an open simulated dataset produced by the CMS Collaboration at the CERN Large Hadron Collider. This dataset consists of Higgs boson decays and quark and gluon background, and is available through the CERN Open Data Portal. We use additional available tools to assess the FAIRness of this dataset, and incorporate feedback from members of the FAIR community to validate our results. This article is accompanied by a Jupyter notebook to visualize and explore this dataset. This study marks the first in a planned series of articles that will guide scientists in the creation of FAIR AI models and datasets in high energy particle physics.

hep-ex↗

arXiv2021

Applications and Techniques for Fast Machine Learning in Science

In this community review report, we discuss applications and techniques for fast machine learning (ML) in science -- the concept of integrating power ML methods into the real-time experimental data processing loop to accelerate scientific discovery. The material for the report builds on two workshops held by the Fast ML for Science community and covers three main areas: applications for fast ML across a number of scientific domains; techniques for training and implementing performant and resource-efficient ML algorithms; and computing architectures, platforms, and technologies for deploying these algorithms. We also present overlapping challenges across the multiple scientific domains where common solutions can be found. This community report is intended to give plenty of examples and inspiration for scientific discovery through integrated and accelerated ML solutions. This is followed by a high-level overview and organization of technical advances, including an abundance of pointers to source material, which can enable these breakthroughs.

cs.LG↗

arXiv2021

Quasi Anomalous Knowledge: Searching for new physics with embedded knowledge

Discoveries of new phenomena often involve a dedicated search for a hypothetical physics signature. Recently, novel deep learning techniques have emerged for anomaly detection in the absence of a signal prior. However, by ignoring signal priors, the sensitivity of these approaches is significantly reduced. We present a new strategy dubbed Quasi Anomalous Knowledge (QUAK), whereby we introduce alternative signal priors that capture some of the salient features of new physics signatures, allowing for the recovery of sensitivity even when the alternative signal is incorrect. This approach can be applied to a broad range of physics models and neural network architectures. In this paper, we apply QUAK to anomaly detection of new physics events at the CERN Large Hadron Collider utilizing variational autoencoders with normalizing flow.

hep-ph↗

arXiv2021

The LHC Olympics 2020: A Community Challenge for Anomaly Detection in High Energy Physics

A new paradigm for data-driven, model-agnostic new physics searches at colliders is emerging, and aims to leverage recent breakthroughs in anomaly detection and machine learning. In order to develop and benchmark new anomaly detection methods within this framework, it is essential to have standard datasets. To this end, we have created the LHC Olympics 2020, a community challenge accompanied by a set of simulated collider events. Participants in these Olympics have developed their methods using an R&D dataset and then tested them on black boxes: datasets with an unknown anomaly (or not). This paper will review the LHC Olympics 2020 challenge, including an overview of the competition, a description of methods deployed in the competition, lessons learned from the experience, and implications for data analyses with future datasets as well as future colliders.

hep-ph↗

arXiv2019

Support-Area Dependence of Vibration-Insensitive Optical Cavities

The vibration sensitivities of optical cavities depending on the support-area were investigated both numerically and experimentally. We performed the numerical simulation with two models; one with total constraint over the support area, and the other with only vertical constraint. A support-area-size insensitive optimal support condition could be found by the numerical simulation. The support-area was determined in the experiment by a Viton rubber pad. The vertical, transverse, and longitudinal vibration sensitivities were measured experimentally. The experimental result agreed with the numerical simulation with a sliding model (only vertical constraint).

physics.ins-det↗

arXiv2017

Charge asymmetry dependence of anisotropic flow in pPb and PbPb collisions with CMS experiment

In nucleus-nucleus collisions, the linear dependence found for the elliptic flow harmonic of both positive or negative charged particles as a function of event charge asymmetry is predicted by the phenomenon known as the Chiral Magnetic Wave (CMW) due to its induced electric quadrupole moment. Here, the event charge asymmetry $A_{\rm ch}$ is defined as $\frac{N_{+}-N_{-}}{N_{+}+N_{-}}$, where $N_{+}$ and $N_{-}$ are the number of positive and negative charged particles, respectively. However, other scenarios are also possible and may provide alternative explanations for the experimental results. New measurements of elliptic ($v_{\rm 2}$) and triangular ($v_{\rm 3}$) flow for positive and negative charged particles as a function of $A_{\rm ch}$ in pPb and PbPb collisions at $\sqrt{s_{_{NN}}} = 5.02~\mathrm{TeV}$ are presented, using data collected by the CMS experiment during the LHC runs 1 and 2. The slopes and intercepts of the charged-dependent $v_{n}$ harmonics vs. $A_{\rm ch}$ are directly compared for pPb and PbPb collisions with similar charged-particle multiplicities, where a strong CMW effect is not expected in very high multiplicity pPb events. Moreover, a comparison is made of the slope parameters between $v_{2}$ and $v_{3}$ harmonics normalized by the inclusive charge particle $v_{n}$ in PbPb collisions as a function of centrality. These results provide a means to discriminate between the CMW and other scenarios such as local charge conservation as possible explanations for the observed charge dependent behavior.

nucl-ex↗

arXiv2016

Drift-compensated Low-noise Frequency Synthesis Based on a cryoCSO for the KRISS-F1

In this paper we report on the implementation and stability analysis of a drift-compensated frequency synthesizer from a cryogenic sapphire oscillator (CSO) designed for a Cs/Rb atomic fountain clock. The synthesizer has two microwave outputs of 7 GHz and 9 GHz for Rb and Cs atom interrogation, respectively. The short-term stability of these microwave signals, measured using an optical frequency comb locked to an ultra-stable laser, is better than $5\times10^{-15}$ at an averaging time of 1 s. We demonstrate that the short-term stability of the synthesizer is lower than the quantum projection noise limit of the Cs fountain clock, KRISS-F1(Cs) by measuring the short-term stability of the fountain with varying trapped atom number. The stability of the fountain at 1-s averaging time reaches $2.5\times10^{-14}$ at the highest atom number in the experiment when the synthesizer is used as an interrogation oscillator of the fountain. In order to compensate the frequency drift of the CSO, the output frequency of a waveform generator in the synthesis chain is ramped linearly. By doing this, the stability of the synthesizer at an average time of one hour reaches a level of $10^{-16}$ which is measured with the fountain clock.

physics.ins-det↗

arXiv2013

Absolute frequency measurement of 1S0 (F = 1/2) - 3P0 (F = 1/2) transition of 171Yb atoms in a one-dimensional optical lattice at KRISS

We measured the absolute frequency of the optical clock transition 1S0 (F = 1/2) - 3P0 (F = 1/2) of 171Yb atoms confined in a one-dimensional optical lattice and it was determined to be 518 295 836 590 863.5(8.1) Hz. The frequency was measured against Terrestrial Time (TT; the SI second on the geoid) by using an optical frequency comb of which the frequency was phase-locked to an H-maser as a flywheel oscillator traceable to TT. The magic wavelength was also measured as 394 798.48(79) GHz. The results are in good agreement with two previous measurements of other institutes within the specified uncertainty of this work.

physics.atom-ph↗