arXiv subjects

Chenghui Yu

Publications and source records attributed to Chenghui Yu.

13 recordsLinked to original sources

arXiv2026

Status report towards implementation of a Compton polarimeter at BEPCII

Precision beam polarization measurements based on Compton polarimeters are essential for the physics program of future high-energy colliders. In order to prepare for these and to extend the scope of physics measurements of the BESIII experiment at the BEPCII, a diagnostic of electron beam transverse polarization at BEPCII is of interest. The design and status report of the commissioning, until July 2025, of this device is reported in this paper. We report unambiguous observation of Compton interaction, discuss current limitations of the experimental setup and draw prospects for improvements and actual measurement of electron beam polarization in the near future.

physics.acc-ph↗

arXiv2026

When Rules Fall Short: Agent-Driven Discovery of Emerging Content Issues in Short Video Platforms

Trends on short-video platforms evolve at a rapid pace, with new content issues emerging every day that fall outside the coverage of existing annotation policies. However, traditional human-driven discovery of emerging issues is too slow, which leads to delayed updates of annotation policies and poses a major challenge for effective content governance. In this work, we propose an automatic issue discovery method based on multimodal LLM agents. Our approach automatically recalls short videos containing potential new issues and applies a two-stage clustering strategy to group them, with each cluster corresponding to a newly discovered issue. The agent then generates updated annotation policies from these clusters, thereby extending coverage to these emerging issues. Our agent has been deployed in the real system. Both offline and online experiments demonstrate that this agent-based method significantly improves the effectiveness of emerging-issue discovery (with an F1 score improvement of over 20%) and enhances the performance of subsequent issue governance (reducing the view count of problematic videos by approximately 15%). More importantly, compared to manual issue discovery, it greatly reduces time costs and substantially accelerates the iteration of annotation policies.

cs.CV↗

arXiv2025

Physics Briefing Book: Input for the 2026 update of the European Strategy for Particle Physics

The European Strategy for Particle Physics (ESPP) reflects the vision and presents concrete plans of the European particle physics community for advancing human knowledge in fundamental physics. The ESPP is updated every five-to-six years through a community-driven process. It commences with the submission of specific proposals and other input from the community at large, outlining projects envisioned for the near-, mid-, and long-term future. All submitted contributions are evaluated by the Physics Preparatory Group (PPG), and a preliminary analysis is presented at a Symposium meant to foster a broad community discussion on the scientific value and feasibility of the various ideas proposed. The outcomes of the analysis and the deliberations at the Symposium are synthesized in the current Briefing Book, which provides an important input in the deliberations of the Strategy recommendations by the European Strategy Group (ESG).

hep-ex↗

arXiv2024

USM: Unbiased Survey Modeling for Limiting Negative User Experiences in Recommendation Systems

Reducing negative user experiences is essential for the success of recommendation platforms. Exposing users to inappropriate content could not only adversely affect users' psychological well-beings, but also potentially drive users away from the platform, sabotaging the platform's long-term success. However, recommendation algorithms tend to weigh more heavily on positive feedback signals due to the scarcity of negative ones, which may result in the neglect of valuable negative user feedback. In this paper, we propose an approach aimed at limiting negative user experiences. Our method primarily relies on distributing in-feed surveys to the users, modeling the users' feedback collected from the survey, and integrating the model predictions into the recommendation system. We further enhance the baseline survey model by integrating the Learning Hidden Unit Contributions module and the Squeeze-and-Excitation module. In addition, we strive to resolve the problem of response Bias by applying a survey-submit model; The A/B testing results indicate a reduction in survey sexual rate and survey inappropriate rate, ranging from -1.44\% to -3.9\%. Additionally, we compared our methods against an online baseline that does not incorporate our approach. The results indicate that our approach significantly reduces the report rate and dislike rate by 1\% to 2.27\% compared to the baseline, confirming the effectiveness of our methods in enhancing user experience. After we launched the survey model based our approach on our platform, the model is able to bring reductions of 1.75\%, 2.57\%, 2.06\% on reports, dislikes, survey inappropriate rate, respectively.

cs.IR↗

arXiv2023

Benchmarking Robustness and Generalization in Multi-Agent Systems: A Case Study on Neural MMO

We present the results of the second Neural MMO challenge, hosted at IJCAI 2022, which received 1600+ submissions. This competition targets robustness and generalization in multi-agent systems: participants train teams of agents to complete a multi-task objective against opponents not seen during training. The competition combines relatively complex environment design with large numbers of agents in the environment. The top submissions demonstrate strong success on this task using mostly standard reinforcement learning (RL) methods combined with domain-specific engineering. We summarize the competition design and results and suggest that, as an academic community, competitions may be a powerful approach to solving hard problems and establishing a solid benchmark for algorithms. We will open-source our benchmark including the environment wrapper, baselines, a visualization tool, and selected policies for further research.

cs.AI↗

arXiv2023

Emergent collective intelligence from massive-agent cooperation and competition

Inspired by organisms evolving through cooperation and competition between different populations on Earth, we study the emergence of artificial collective intelligence through massive-agent reinforcement learning. To this end, We propose a new massive-agent reinforcement learning environment, Lux, where dynamic and massive agents in two teams scramble for limited resources and fight off the darkness. In Lux, we build our agents through the standard reinforcement learning algorithm in curriculum learning phases and leverage centralized control via a pixel-to-pixel policy network. As agents co-evolve through self-play, we observe several stages of intelligence, from the acquisition of atomic skills to the development of group strategies. Since these learned group strategies arise from individual decisions without an explicit coordination mechanism, we claim that artificial collective intelligence emerges from massive-agent cooperation and competition. We further analyze the emergence of various learned strategies through metrics and ablation studies, aiming to provide insights for reinforcement learning implementations in massive-agent environments.

cs.AI↗

arXiv2022

Seeking Common Ground While Reserving Differences: Multiple Anatomy Collaborative Framework for Undersampled MRI Reconstruction

Recently, deep neural networks have greatly advanced undersampled Magnetic Resonance Image (MRI) reconstruction, wherein most studies follow the one-anatomy-one-network fashion, i.e., each expert network is trained and evaluated for a specific anatomy. Apart from inefficiency in training multiple independent models, such convention ignores the shared de-aliasing knowledge across various anatomies which can benefit each other. To explore the shared knowledge, one naive way is to combine all the data from various anatomies to train an all-round network. Unfortunately, despite the existence of the shared de-aliasing knowledge, we reveal that the exclusive knowledge across different anatomies can deteriorate specific reconstruction targets, yielding overall performance degradation. Observing this, in this study, we present a novel deep MRI reconstruction framework with both anatomy-shared and anatomy-specific parameterized learners, aiming to "seek common ground while reserving differences" across different anatomies.Particularly, the primary anatomy-shared learners are exposed to different anatomies to model flourishing shared knowledge, while the efficient anatomy-specific learners are trained with their target anatomy for exclusive knowledge. Four different implementations of anatomy-specific learners are presented and explored on the top of our framework in two MRI reconstruction networks. Comprehensive experiments on brain, knee and cardiac MRI datasets demonstrate that three of these learners are able to enhance reconstruction performance via multiple anatomy collaborative learning.

eess.IV↗

arXiv2019

Offset Simultaneous Conjugate Atom Interferometers

Correlating the signals from simultaneous atom interferometers has enabled some of the most precise determinations of fundamental constants. Here, we show that multiple interferometers with strategically chosen initial conditions ("offset simultaneous conjugate interferometers" or OSCIs) can provide multi-channel readouts that amplify or suppress specific effects. This allows us to measure the photon recoil, and thus the fine structure constant, while being insensitive to gravity gradients, general acceleration gradients, and unwanted diffraction phases - these effects can be simultaneously monitored in other channels. An expected 4-fold reduction of sensitivity to spatial variations of gravity (due to higher-order gradients) and a 6-fold suppression of diffraction phases paves the way to measurements of the fine structure constant below the 0.1-ppb level, or to simultaneous sensing of gravity, the gravity gradient, and rotations.

physics.atom-ph↗

arXiv2018

Measurement of the fine-structure constant as a test of the Standard Model

Measurements of the fine-structure constant alpha require methods from across subfields and are thus powerful tests of the consistency of theory and experiment in physics. Using the recoil frequency of cesium-133 atoms in a matter-wave interferometer, we recorded the most accurate measurement of the fine-structure constant to date: alpha = 1/137.035999046(27) at 2.0 x 10^-10 accuracy. Using multiphoton interactions (Bragg diffraction and Bloch oscillations), we demonstrate the largest phase (12 million radians) of any Ramsey-Borde interferometer and control systematic effects at a level of 0.12 parts per billion. Comparison with Penning trap measurements of the electron gyromagnetic anomaly ge-2 via the Standard Model of particle physics is now limited by the uncertainty in ge-2; a 2.5 sigma tension rejects dark photons as the reason for the unexplained part of the muon's magnetic moment at a 99 percent confidence level. Implications for dark-sector candidates and electron substructure may be a sign of physics beyond the Standard Model that warrants further investigation.

physics.atom-ph↗

arXiv2017

Limits on Lorentz violation in gravity from worldwide superconducting gravimeters

We investigated Lorentz violation through anisotropy of gravity using a worldwide array of 12 superconducting gravimeters. The Lorentz-violating signal is extracted from the difference between measured gravity and a tidal model. At the level of sensitivity we reach, ocean tides start to play an important role. However, most models available that include ocean tides are empirically based on measured gravity data, which may contain Lorentz-violating signal. In this work we used an ocean tides included tidal model derived from first principles to extract Lorentz-violating signal for the first time. We have bounded space-space components of gravitational Lorentz violation in the minimal standard model extension (SME) up to the order of $10^{-10}$, one order of magnitude improved relative to previous atom-interferometer tests.

gr-qc↗

arXiv2017

Laser frequency stabilization by combining modulation transfer and frequency modulation spectroscopy

We present a hybrid laser frequency stabilization method combining modulation transfer spectroscopy (MTS) and frequency modulation spectroscopy (FMS) for the cesium D2 transition. In a typical pump-probe setup, the error signal is a combination of the DC-coupled MTS error signal and the AC-coupled FMS error signal. This combines the long-term stability of the former with the high signal-to-noise ratio of the latter. In addition, we enhance the long-term frequency stability with laser intensity stabilization. By measuring the frequency difference between two independent hybrid spectroscopies, we investigate the short-term and long-term stability. We find a long-term stability of 7.8 kHz characterized by a standard deviation of the beating frequency drift over the course of 10 hours, and a short-term stability of 1.9 kHz characterized by an Allan deviation of that at 2 seconds of integration time.

physics.atom-ph↗

arXiv2016

Controlling the Multiport Nature of Bragg Diffraction in Atom Interferometry

Bragg diffraction has been used in atom interferometers because it allows signal enhancement through multiphoton momentum transfer and suppression of systematics by not changing the internal state of atoms. Its multi-port nature, however, can lead to parasitic interferometers, allows for intensity-dependent phase shifts in the primary interferometers, and distorts the ellipses used for phase extraction. We study and suppress these unwanted effects. Specifically, phase extraction by ellipse fitting and the resulting systematic phase shifts are calculated by Monte Carlo simulations. Phase shifts arising from the thermal motion of the atoms are controlled by spatial selection of atoms and an appropriate choice of Bragg intensity. In these simulations, we found that Gaussian Bragg pulse shapes yield the smallest systematic shifts. Parasitic interferometers are suppressed by a "magic" Bragg pulse duration. The sensitivity of the apparatus was improved by the addition of AC Stark shift compensation, which permits direct experimental study of sub-part-per-billion (ppb) systematics. This upgrade allows for a 310\,$\hbar k$ momentum transfer, giving an unprecedented 6.6\,Mrad measured in a Ramsey-Bord\'e interferometer.

physics.atom-ph↗

arXiv2014

High-resolution atom interferometers with suppressed diffraction phases

We experimentally and theoretically study the diffraction phase of large-momentum transfer beam splitters in atom interferometers based on Bragg diffraction. We null the diffraction phase and increase the sensitivity of the interferometer by combining Bragg diffraction with Bloch oscillations. We demonstrate agreement between experiment and theory, and a 1500-fold reduction of the diffraction phase, limited by measurement noise. In addition to reduced systematic effects, our interferometer has high contrast with up to 4.4 million radians of phase difference, and a resolution in the fine structure constant of $\delta \alpha/\alpha=0.25\,$ppb in 25 hours of integration time.

physics.atom-ph↗