BZPEERReading, trust and collaboration for research
Menu

Discover

GraphFollow connections around a paper, topic or saved view.

Workspaces

HomePick up where your research flow left off.InboxHandle requests, replies and notifications from one place.CollectionsCurate reading lists, evidence sets and shareable dossiers.ResearchSee your private provenance graph, trail and imports.CollaborationMove from discovery into focused discussion rooms.PublishDraft, preview and publish full-text research articles.

Network

ExpertsFind specialists worth following or asking for help.CommunitiesJoin research circles organized around shared work.PartnersSee external organizations active around the same topics.

Opportunities

ListingsBrowse roles, calls and collaborations with clearer fit signals.

Account

Log inReturn to your reading and collaboration workspace.Create accountStart saving work, following people and joining teams.
Log inCreate account

Researcher profile

Rui Shi

Rui Shi contributes to research discovery and scholarly infrastructure.

ResearcherAffiliation not importedOpen to collaborate
Artificial Intelligenceastro-ph.COastro-ph.GAastro-ph.IMcond-mat.softMachine Learningphysics.chem-phastro-ph.HEBiological PhysicsComputer Visioncond-mat.stat-mecheess.IVgr-qchep-phhep-thmath.OAmath.OCNeural and Evolutionary Computingnucl-exphysics.acc-phphysics.data-anphysics.ins-det

Trust snapshot

Quick read

Trust 21 - EmergingVerification L1Unclaimed author
17works
0followers
22topics
4close collaborators

Actions

Decide how to stay connected

Follow researcher0

Identity and collaboration

How to connect with this researcher

Claiming links this public author record to a researcher profile and unlocks direct collaboration workflows.

Log in to claim

Direct collaboration

Open a focused conversation when the fit is right

Claim this author entity first to unlock direct invitations.

Research graph

See the researcher in context

Open full explorer

Inspect adjacent work, topics, institutions and collaborators without jumping out to a separate graph page.

Building this graph slice

BZPEER is loading the nearby papers, people, topics and institutions for this page.

Published work

17 published item(s)

preprint2026arXiv

Eigen Microstate Condensation and Critical Phenomena in the Lennard-Jones Fluid

Despite extensive study of the liquid-gas phase transition, accurately determining the critical point and the critical exponents in fluid systems through direct simulation remains a challenge. We employ the eigen microstate theory (EMT) to investigate the liquid-gas continuous phase transition in the Lennard-Jones (LJ) fluid within the canonical ensemble. In EMT, the probability amplitudes of eigen microstates serve as the order parameter. Using finite-size scaling of probability amplitudes, we simultaneously determine the critical temperature, $T_c = 1.188(2)$, and critical density, $ρ_c = 0.320(4)$. Furturemore, we obtain critical exponents of the LJ fluid, $β= 0.32(2)$ and $ν= 0.64(3)$, which demonstrate a great agreement with the Ising universality class. This method also reveals the mesoscopic structure of the emergent phase, characterizing the three-dimensional (3D) spatial configuration of the fluid in the critical region. This work also confirms the finite-size scaling behavior of the probability amplitudes of the eigen microstates in the critical region. The EMT provides a powerful tool for studying the critical phenomena of complex fluid system.

0 citations0 reviewsNo code linkedOpen access
preprint2026arXiv

Multi-Scale Dequant: Eliminating Dequantization Bottleneck via Activation Decomposition for Efficient LLM Inference

Quantization is essential for efficient large language model (LLM) inference, yet the dequantization step-converting low-bit weights back to high-precision for matrix multiplication has become a critical bottleneck on modern AI accelerators. On architectures with decoupled compute units (e.g., Ascend NPUs), dequantization operations can consume more cycles than the matrix multiplication itself, leaving the high-throughput tensor cores underutilized. This paper presents Multi-Scale Dequant (MSD), a quantization framework that removes weight/KV dequantization from the GEMM critical path. Instead of lifting low-bit weights to BF16 precision, MSD decomposes high-precision BF16 activations into multiple low-precision components, each of which can be multiplied directly with quantized weights via native hardware-accelerated GEMM. This approach shifts the computational paradigm from precision conversion to multi-scale approximation, avoiding INT8-to-BF16 weight conversion before GEMM. We instantiate MSD for two weight formats and derive tight error bounds for each. For INT8 weights (W4A16), two-pass INT8 decomposition achieves near 16 effective bits. For MXFP4 weights (W4A16), two-pass MXFP4 decomposition yields near 6.6 effective bits with error bound 1/64 per block surpassing single-pass MXFP8(5.24 bits) while maintaining the same effective GEMM compute time. We further derive closed-form latency and HBM traffic models showing that MSD avoids the Vector-Cube pipeline stall caused by dequantization and reduces KV cache HBM traffic by up to 2.5 times in attention. Numerical simulations on matrix multiplication and Flash Attention kernels confirm that MSD does not degrade accuracy compared to dequantization baselines, and in many settings achieves lower L2 error.

0 citations0 reviewsNo code linkedOpen access
preprint2025arXiv

ChatTS: Aligning Time Series with LLMs via Synthetic Data for Enhanced Understanding and Reasoning

Understanding time series is crucial for its application in real-world scenarios. Recently, large language models (LLMs) have been increasingly applied to time series tasks, leveraging their strong language capabilities to enhance various applications. However, research on multimodal LLMs (MLLMs) for time series understanding and reasoning remains limited, primarily due to the scarcity of high-quality datasets that align time series with textual information. This paper introduces ChatTS, a novel MLLM designed for time series analysis. ChatTS treats time series as a modality, similar to how vision MLLMs process images, enabling it to perform both understanding and reasoning with time series. To address the scarcity of training data, we propose an attribute-based method for generating synthetic time series with detailed attribute descriptions. We further introduce Time Series Evol-Instruct, a novel approach that generates diverse time series Q&As, enhancing the model's reasoning capabilities. To the best of our knowledge, ChatTS is the first TS-MLLM that takes multivariate time series as input for understanding and reasoning, which is fine-tuned exclusively on synthetic datasets. We evaluate its performance using benchmark datasets with real-world data, including six alignment tasks and four reasoning tasks. Our results show that ChatTS significantly outperforms existing vision-based MLLMs (e.g., GPT-4o) and text/agent-based LLMs, achieving a 46.0% improvement in alignment tasks and a 25.8% improvement in reasoning tasks. We have open-sourced the source code, model checkpoint and datasets at https://github.com/NetManAIOps/ChatTS.

0 citations0 reviewsNo code linkedOpen access
preprint2023arXiv

Beyond PID Controllers: PPO with Neuralized PID Policy for Proton Beam Intensity Control in Mu2e

We introduce a novel Proximal Policy Optimization (PPO) algorithm aimed at addressing the challenge of maintaining a uniform proton beam intensity delivery in the Muon to Electron Conversion Experiment (Mu2e) at Fermi National Accelerator Laboratory (Fermilab). Our primary objective is to regulate the spill process to ensure a consistent intensity profile, with the ultimate goal of creating an automated controller capable of providing real-time feedback and calibration of the Spill Regulation System (SRS) parameters on a millisecond timescale. We treat the Mu2e accelerator system as a Markov Decision Process suitable for Reinforcement Learning (RL), utilizing PPO to reduce bias and enhance training stability. A key innovation in our approach is the integration of a neuralized Proportional-Integral-Derivative (PID) controller into the policy function, resulting in a significant improvement in the Spill Duty Factor (SDF) by 13.6%, surpassing the performance of the current PID controller baseline by an additional 1.6%. This paper presents the preliminary offline results based on a differentiable simulator of the Mu2e accelerator. It paves the groundwork for real-time implementations and applications, representing a crucial step towards automated proton beam intensity control for the Mu2e experiment.

0 citations0 reviewsNo code linkedOpen access
preprint2023arXiv

Towards Relating Fragile-To-Strong Transition to Fragile Glass

Glass formers are in general classified as strong or fragile depending on whether their relaxation rates follow Arrhenius or super-Arrhenius temperature dependence. There are however notable exceptions such as water, which exhibit a fragile-to-strong (FTS) transition and behave as fragile and strong respectively at high and low temperatures. In this work, the FTS transition is studied using a distinguishable-particle lattice model previously demonstrated to be capable of simulating both strong and fragile glasses [Phys. Rev. Lett. 125, 265703 (2020)]. Starting with a bimodal pair-interaction distribution appropriate for fragile glasses, we show that by narrowing down the energy dispersion in the low-energy component of the distribution, a FTS transition is observed. The transition occurs at a temperature at which the stretching exponent of the relaxation is minimized, in agreement with previous molecular dynamics simulations.

0 citations0 reviewsNo code linkedOpen access
preprint2022arXiv

A machine learning approach to infer the accreted stellar mass fractions of central galaxies in the TNG100 simulation

We propose a random forest (RF) machine learning approach to determine the accreted stellar mass fractions ($f_\mathrm{acc}$) of central galaxies, based on various dark matter halo and galaxy features. The RF is trained and tested using 2,710 galaxies with stellar mass $\log_{10}M_\ast/M_\odot>10.16$ from the TNG100 simulation. Galaxy size is the most important individual feature when calculated in 3-dimensions, which becomes less important after accounting for observational effects. For smaller galaxies, the rankings for features related to merger histories increase. When an entire set of halo and galaxy features are used, the prediction is almost unbiased, with root-mean-square error (RMSE) of $\sim$0.068. A combination of up to three features with different types (galaxy size, merger history and morphology) already saturates the power of prediction. If using observable features, the RMSE increases to $\sim$0.104, and a combined usage of stellar mass, galaxy size plus galaxy concentration achieves similar predictions. Lastly, when using galaxy density, velocity and velocity dispersion profiles as features, which approximately represent the maximum amount of information extracted from galaxy images and velocity maps, the prediction is not improved much. Hence the limiting precision of predicting $f_\mathrm{acc}$ is $\sim$0.1 with observables, and the multi-component decomposition of galaxy images should have similar or larger uncertainties. If the central black hole mass and the spin parameter of galaxies can be accurately measured in future observations, the RMSE is promising to be further decreased by $\sim$20%.

0 citations0 reviewsNo code linkedOpen access
preprint2022arXiv

Construction of a Large Diameter Reflective Half-Wave Plate Modulator for Millimeter Wave Applications

Polarization modulation is a powerful technique to increase the stability of measurements by enabling the distinction of a polarized signal from dominant slow system drifts and unpolarized foregrounds. Furthermore, when placed as close to the sky as possible, modulation can reduce systematic errors from instrument polarization. In this work, we introduce the design and preliminary drive system laboratory performance of a new 60 cm diameter reflective half-wave plate (RHWP) polarization modulator. The wave plate consists of a wire array situated in front of a flat mirror. Using \mbox{50 $μ$m} diameter wires with \mbox{175 $μ$m} spacing, the wave plate will be suitable for operation in the millimeter wavelength range with flatness of the wires and parallelism to the mirror held to a small fraction of a wavelength. The presented design targets the 77--108 GHz range. Modulation is performed by a rotation of the wave plate with a custom rotary drive utilizing an actively controlled servo motor.

0 citations0 reviewsNo code linkedOpen access
preprint2022arXiv

Design and characterization of new 90 GHz detectors for the Cosmology Large Angular Scale Surveyor (CLASS)

The Cosmology Large Angular Scale Surveyor (CLASS) is a polarization-sensitive telescope array located at an altitude of 5,200 m in the Chilean Atacama Desert. CLASS is designed to measure "E-mode" (even parity) and "B-mode" (odd parity) polarization patterns in the Cosmic Microwave Background (CMB) over large angular scales with the aim of improving our understanding of inflation, reionization, and dark matter. CLASS is currently observing with three telescopes covering four frequency bands: one at 40 GHz (Q); one at 90 GHz (W1); and one dichroic system at 150/220 GHz (G). In these proceedings, we discuss the updated design and in-lab characterization of new 90 GHz detectors. The new detectors include design changes to the transition-edge sensor (TES) bolometer architecture, which aim to improve stability and optical efficiency. We assembled and tested four new detector wafers, to replace four modules of the W1 focal plane. These detectors were installed into the W1 telescope, and will achieve first light in the austral winter of 2022. We present electrothermal parameters and bandpass measurements from in-lab dark and optical testing. From in-lab dark tests, we also measure a median NEP of 12.3 $\mathrm{aW\sqrt{s}}$ across all four wafers about the CLASS signal band, which is below the expected photon NEP of 32 $\mathrm{aW\sqrt{s}}$ from the field. We therefore expect the new detectors to be photon noise limited.

0 citations0 reviewsNo code linkedOpen access
preprint2022arXiv

Long-Timescale Stability in CMB Observations at Multiple Frequencies using Front-End Polarization Modulation

The Cosmology Large Angular Scale Surveyor (CLASS) is a telescope array observing the Cosmic Microwave Background (CMB) at frequency bands centered near 40, 90, 150, and 220 GHz. CLASS measures the CMB polarization on the largest angular scales to constrain the inflationary tensor-to-scalar ratio and the optical depth due to reionization. To achieve the long time-scale stability necessary for this measurement from the ground, CLASS utilizes a front-end, variable-delay polarization modulator on each telescope. Here we report on the improvements in stability afforded by front-end modulation using data across all four CLASS frequencies. Across one month of modulated linear polarization data in 2021, CLASS achieved median knee frequencies of 9.1, 29.1, 20.4, and 36.4 mHz for the 40, 90, 150, and 220 GHz observing bands. The knee frequencies are approximately an order of magnitude lower than achieved via CLASS pair-differencing orthogonal detector pairs without modulation.

0 citations0 reviewsNo code linkedOpen access
preprint2022arXiv

Neural Annotation Refinement: Development of a New 3D Dataset for Adrenal Gland Analysis

The human annotations are imperfect, especially when produced by junior practitioners. Multi-expert consensus is usually regarded as golden standard, while this annotation protocol is too expensive to implement in many real-world projects. In this study, we propose a method to refine human annotation, named Neural Annotation Refinement (NeAR). It is based on a learnable implicit function, which decodes a latent vector into represented shape. By integrating the appearance as an input of implicit functions, the appearance-aware NeAR fixes the annotation artefacts. Our method is demonstrated on the application of adrenal gland analysis. We first show that the NeAR can repair distorted golden standards on a public adrenal gland segmentation dataset. Besides, we develop a new Adrenal gLand ANalysis (ALAN) dataset with the proposed NeAR, where each case consists of a 3D shape of adrenal gland and its diagnosis label (normal vs. abnormal) assigned by experts. We show that models trained on the shapes repaired by the NeAR can diagnose adrenal glands better than the original ones. The ALAN dataset will be open-source, with 1,584 shapes for adrenal gland diagnosis, which serves as a new benchmark for medical shape analysis. Code and dataset are available at https://github.com/M3DV/NeAR.

0 citations0 reviewsNo code linkedOpen access
preprint2022arXiv

The structural order of protein hydration water

The ability of water to dissolve biomolecules is crucial for our life. It has been shown that protein has a profound effect on the behavior of water in its hydration shell, which in turn affects the structure and function of the protein. However, there is still no consensus on whether protein promotes or destroys the structural order of water in its hydration shell until today, because of the lack of proper structural descriptor incorporating hydrogen-bond (H-bond) information for water at the protein/water interface. Here we performed all-atom molecular dynamics simulations of lysozyme protein in water and analyzed the H-bond structure of protein hydration water by using a newly developed structural descriptor. We find that the protein promotes local structural ordering of the hydration water while has a negligible effect on the strength of individual H-bond. These findings are fundamental to the structure and function of biomolecules and provide new insights into the hydration of protein in water.

0 citations0 reviewsNo code linkedOpen access
preprint2022arXiv

What to expect from dynamical modelling of cluster haloes II. Investigating dynamical state indicators with Random Forest

We investigate the importances of various dynamical features in predicting the dynamical state (DS) of galaxy clusters, based on the Random Forest (RF) machine learning approach. We use a large sample of galaxy clusters from the Three Hundred Project of hydrodynamical zoomed-in simulations, and construct dynamical features from the raw data as well as from the corresponding mock maps in the optical, X-ray, and Sunyaev-Zel'dovich (SZ) channels. Instead of relying on the impurity based feature importance of the RF algorithm, we directly use the out-of-bag (OOB) scores to evaluate the importances of individual features and different feature combinations. Among all the features studied, we find the virial ratio, $η$, to be the most important single feature. The features calculated directly from the simulations and in 3-dimensions carry more information on the DS than those constructed from the mock maps. Compared with the features based on X-ray or SZ maps, features related to the centroid positions are more important. Despite the large number of investigated features, a combination of up to three features of different types can already saturate the score of the prediction. Lastly, we show that the most sensitive feature $η$ is strongly correlated with the well-known half-mass bias in dynamical modelling. Without a selection in DS, cluster halos have an asymmetric distribution in $η$, corresponding to an overall positive half-mass bias. Our work provides a quantitative reference for selecting the best features to discriminate the DS of galaxy clusters in both simulations and observations.

0 citations0 reviewsNo code linkedOpen access
preprint2021arXiv

Research on Fitness Function of Two Evolution Algorithms Used for Neutron Spectrum Unfolding

When evolution algorithms are used to unfold the neutron energy spectrum, fitness function design is an important fundamental work for evaluating the quality of the solution, but it has not attracted much attention. In this work, we investigated the performance of eight fitness functions attached to the genetic algorithm (GA) and the differential evolution algorithm (DEA) used for unfolding four neutron spectra selected from the IAEA 403 report. Experiments show that the fitness functions with a maximum in the GA can limit the ability of the population to percept the fitness change, but the ability can be made up in the DEA. The fitness function with a feature penalty term helps to improve the performance of solutions, and the fitness function using the standard deviation and the Chi-squared result shows the balance between the algorithm and the spectra. The results also show that the DEA has good potential for neutron energy spectrum unfolding. The purposes of this work are to provide evidence for structuring and modifying the fitness functions and to suggest some genetic operations that should receive attention when using the fitness function to unfold neutron spectra.

0 citations0 reviewsNo code linkedOpen access
preprint2021arXiv

Unfolding the Neutron Spectra from a Water-Pumping-Injection Multi-layered Concentric Sphere Neutron Spectrometer Using a Self-Adaptive Differential Evolution Algorithm

A self-adaptive differential evolution neutron spectrum unfolding algorithm (SDENUA) was established in this paper to unfold the neutron spectra obtained from a Water-pumping-injection Multi-layered concentric sphere Neutron Spectrometer (WMNS). Specifically, the neutron fluence bounds were estimated to accelerate the algorithm convergence, the minimum error between the optimal solution and the input neutron counts with relative uncertainties was limited to 10-6 to avoid useless calculation. Furthermore, the crossover probability and scaling factor were controlled self-adaptively. FLUKA Monte Carlo was used to simulate the readings of the WMNS under (1) a spectrum of Cf-252 and (2) its spectrum after being moderated, (3) a spectrum used for BNCT, and (4) a reactor spectrum, and the measured neutron counts unfolded by using the SDENUA. The uncertainties of the measured neutron count and the response matrix are considered in the SDENUA, which does not require complex parameter tuning and the priori default spectrum. Results indicate that the solutions of the SDENUA are more in agreement with the IAEA spectra than that of the MAXED and GRAVEL in UMG 3.1, and the errors of the final results calculated by SDENUA are under 12%. The established SDENUA has potential applications for unfolding spectra from the WMNS.

0 citations0 reviewsNo code linkedOpen access
preprint2020arXiv

Normed ideal perturbation of irreducible operators in semifinite von Neumann factors

In [10], Halmos proved an interesting result that the set of irreducible operators is dense in $\mathcal B(\mathcal H)$ in the sense of Hilbert-Schmidt approximation. In a von Neumann algebra $\mathcal M$ with separable predual, an operator $a\in \mathcal M$ is said to be {irreducible in} $\mathcal M$ if $W^*(a)$ is an irreducible subfactor of $\mathcal M$, i.e., $W^*(a)'\cap \mathcal M={\mathbb C} \cdot I$. In this paper, let $Φ(\cdot)$ be a $\Vert\cdot\Vert$-dominating, unitarily invariant norm (see Definition 2.1), where by $\Vert\cdot\Vert$ we denote the operator norm. We prove that in every semifinite von Neumann factor $\mathcal M$ with separable predual, if the norm $Φ(\cdot)$ satisfies a natural restriction introduced in (1.1), then irreducible operators are $Φ(\cdot)$-norm dense in $\mathcal M$. In particular, the operator norm $\Vert\cdot\Vert$ and the $\max\{\Vert\cdot\Vert, \Vert\cdot\Vert_p\}$-norm (for each $p>1$) naturally satisfy the condition in (1.1), where $τ$ is a faithful, normal, semifinite, tracial weight and $\Vert x\Vert_p=τ(|x|^p)^{1/p}$ for all $x\in \mathcal M \cap L^{p}(\mathcal M,τ)$ (see [18, Preliminaries]). This can be viewed as a (stronger) analogue of a theorem of Halmos in [10], proved with different techniques developed in semifinite, properly infinite von Neumann factors. Meanwhile, for every $\Vert\cdot\Vert$-dominating, unitarily invariant norm $Φ(\cdot)$, we develop another method to prove that each normal operator in $\mathcal M$ is a sum of an irreducible operator in $\mathcal M$ and an arbitrarily small $Φ(\cdot)$-norm perturbation, where the $Φ(\cdot)$-norm isn't restricted by (1.1). Particularly, the $Φ(\cdot)$-norm can be the $\max\{\Vert\cdot\Vert, \Vert\cdot\Vert_1\}$-norm.

0 citations0 reviewsNo code linkedOpen access
preprint2020arXiv

Updated constraints on superconducting cosmic strings from the astronomy of fast radio bursts

In this article we update constraints on superconducting cosmic strings (SCSs) in light of the recent observational developments of fast radio bursts (FRBs) astronomy. Assuming strings follow an exponential distribution characterized by current, we show that two parameters in our context, which are the characteristic tension ($Gμ$) and a parameter which describes the aforementioned exponential distribution ($I_c$), can be constrained by FRB experiments. Particularly, we investigate data sets from Parkes and ASKAP. We looked through a parameter space where $Gμ\sim [10^{-17}, 10^{-12}]$ and $I_c\sim [10^{-1}, 10^{2}]$ GeV, and found our results show that Parkes jointly with ASKAP can constrain the parameter space for SCSs.

0 citations0 reviewsNo code linkedOpen access
preprint2019arXiv

Distinct signature of two local structural motifs of liquid water in the scattering function

Liquids generally become more ordered upon cooling. However, it has been a long-standing debate on whether such structural ordering in liquid water takes place continuously or discontinuosly: continuum vs. mixture models. Here, by computer simulations of three popular water models and analysis of recent scattering experiment data, we show that, in the structure factor of water, there are two overlapped peaks hidden in the apparent "first diffraction peak", one of which corresponds to the neighboring O-O distance as in ordinary liquids and the other to the longest periodicity of density waves in a tetrahedral structure. This unambiguously proves the coexistence of two local structural motifs. Our findings not only provide key clues to settle long-standing controversy on the water structure but also allow experimental access to the degree and range of structural ordering in liquid water.

0 citations0 reviewsNo code linkedOpen access