arXiv subjects

Yu Luo

Publications and source records attributed to Yu Luo.

At least 91 records · Page 5Linked to original sources

arXiv2025

ViT-based Local Volume dwarf galaxy Identificationin (VIDA) in the CSST survey

Identifying dwarf galaxies within the Local Volume is crucial for constraining the luminosity function of satellite galaxies in the nearby universe. We report the detection capabilities of dwarf galaxies within the Local Volume using the Chinese Space Station Telescope (CSST). Based on the simulated imaging data of CSST, we develop a detection and classification pipeline that combines traditional image-based search techniques with advanced machine learning classification models. The simulated Local Volume dwarf galaxies can be identified using a pre-processing method for "extended source detection", followed by classification with a pretrained ViT-Base model. This pipeline achieves a true positive rate (TPR) exceeding 85% with a false positive rate (FPR) of only 0.1%. We quantify the detection completeness of Local Volume dwarf galaxies across a three-dimensional parameter space defined by absolute magnitude ($M_V$), half-light radius ($R_h$), and heliocentric distance, based on simulated single-exposure CSST wide-field imaging survey data. For unresolved or semi-resolved dwarf galaxies, our method achieves a significantly deeper absolute magnitude detection limit compared to catalog-based approaches, reaching $M_V = -7$ within 10 \Mpc. By combining this image-based approach with traditional stellar catalog-based "matched filter" techniques, our automated framework established in this work can identify dwarf galaxies within 20 \Mpc for the CSST mission.

astro-ph.GA↗

arXiv2025

The Jiutian simulations for the CSST extra-galactic surveys

We provide an overview of the Jiutian simulations, a hybrid simulation suite for the China Space Survey Telescope (CSST) extragalactic surveys. It consists of four complementary modules: the primary runs with high resolutions with the fiducial concordance cosmology, the emulator runs exploring the parameter uncertainties around the fiducial cosmology, the reconstruction runs intended for recovering the observed Universe position by position, and the extension runs employing extended cosmologies beyond the standard model. For the primary runs, two independent pipelines are adopted to construct subhaloes and merger trees. On top of them, four sets of mock galaxy light-cone catalogs are produced from semi-analytical models and subhalo abundance matching, providing a variety of observational properties including galaxy SED, emission lines, lensing distortions, and mock images. The 129 emulator runs are used to train the CSST emulator, achieving one percent accuracy in predicting the matter power spectrum over $k\leq 10h{\rm Mpc}^{-1}$ and $z\leq 2$. The reconstruction runs employ a number of subgrid baryonic models to predict the evolution and galaxy population resembling certain regions in the real Universe with constrained initial conditions, enabling controlled investigation of galaxy formation on top of structure formation. The extension runs cover models with warm dark matter, $f(R)$ gravity, interacting dark energy, and nonzero neutrino masses, revealing differences in the cosmic structure under alternative cosmological models. We introduce the specifications for each run, the data products derived from them, the corresponding pipeline developments, and present some main tests. Using the primary runs, we also show that the subhalo peak mass functions of different levels are approximately universal. These simulations form a comprehensive and open library for CSST surveys and beyond.

astro-ph.CO↗

arXiv2025

Pangu Embedded: An Efficient Dual-system LLM Reasoner with Metacognition

This work presents Pangu Embedded, an efficient Large Language Model (LLM) reasoner developed on Ascend Neural Processing Units (NPUs), featuring flexible fast and slow thinking capabilities. Pangu Embedded addresses the significant computational costs and inference latency challenges prevalent in existing reasoning-optimized LLMs. We propose a two-stage training framework for its construction. In Stage 1, the model is finetuned via an iterative distillation process, incorporating inter-iteration model merging to effectively aggregate complementary knowledge. This is followed by reinforcement learning on Ascend clusters, optimized by a latency-tolerant scheduler that combines stale synchronous parallelism with prioritized data queues. The RL process is guided by a Multi-source Adaptive Reward System (MARS), which generates dynamic, task-specific reward signals using deterministic metrics and lightweight LLM evaluators for mathematics, coding, and general problem-solving tasks. Stage 2 introduces a dual-system framework, endowing Pangu Embedded with a "fast" mode for routine queries and a deeper "slow" mode for complex inference. This framework offers both manual mode switching for user control and an automatic, complexity-aware mode selection mechanism that dynamically allocates computational resources to balance latency and reasoning depth. Experimental results on benchmarks including AIME 2024, GPQA, and LiveCodeBench demonstrate that Pangu Embedded with 7B parameters, outperforms similar-size models like Qwen3-8B and GLM4-9B. It delivers rapid responses and state-of-the-art reasoning quality within a single, unified model architecture, highlighting a promising direction for developing powerful yet practically deployable LLM reasoners.

cs.CL↗

arXiv2025

One-shot manipulation of coherence in dynamic quantum resource theory

A fundamental problem in quantum information is to understand the operational significance of quantum resources. Quantum resource theories (QRTs) provide a powerful theoretical framework that aids in analyzing and comprehending the operational meaning of these resources. Early resource theories primarily focused on analyzing static quantum resources. Recently, there has been growing interest in the study of dynamic quantum resources. In this paper, we utilize superchannel theory to describe the dynamic resource theory of quantum coherence. In this dynamic resource theory, we treat classical channels as free channels and consider two classes of free superchannels that preserve channel incoherence (maximally incoherent superchannels (MISC) and dephasing-covariant incoherent superchannels (DISC)) as free resources. We regard the quantum Fourier transform as the golden unit of dynamic coherence resources. We first establish the one-shot theory of dynamic coherence cost and dynamic coherence distillation, which involves converting the quantum Fourier transform into an arbitrary quantum channel using MISC and DISC. Next, we introduce a class of free superchannels known as $δ$-MISC, which asymptotically generate negligible dynamic coherence. Finally, we provide upper and lower bounds for the one-shot catalytic dynamic coherence cost of quantum channels under the action of these $δ$-MISC superchannels.

quant-ph↗

arXiv2025

On the moduli description of ramified unitary local models of signature $(n-1,1)$

We provide a moduli description of the ramified unitary local model of signature $(n-1,1)$ with arbitrary parahoric level structure, assuming the residue field has characteristic not equal to $2$, thereby confirming a conjecture of Smithling. Our approach involves writing down explicit equations for the special fiber and proving that they define a normal, Cohen-Macaulay scheme, which is also of independent interest. As applications, we obtain moduli descriptions for: (1) ramified unitary Pappas-Zhu local models with arbitrary parahoric level; (2) the irreducible components of their special fiber in the maximal parahoric case; (3) integral models of ramified unitary Shimura varieties with arbitrary (quasi-)parahoric level.

math.AG↗

arXiv2025

Pangu Ultra MoE: How to Train Your Big MoE on Ascend NPUs

Sparse large language models (LLMs) with Mixture of Experts (MoE) and close to a trillion parameters are dominating the realm of most capable language models. However, the massive model scale poses significant challenges for the underlying software and hardware systems. In this paper, we aim to uncover a recipe to harness such scale on Ascend NPUs. The key goals are better usage of the computing resources under the dynamic sparse model structures and materializing the expected performance gain on the actual hardware. To select model configurations suitable for Ascend NPUs without repeatedly running the expensive experiments, we leverage simulation to compare the trade-off of various model hyperparameters. This study led to Pangu Ultra MoE, a sparse LLM with 718 billion parameters, and we conducted experiments on the model to verify the simulation results. On the system side, we dig into Expert Parallelism to optimize the communication between NPU devices to reduce the synchronization overhead. We also optimize the memory efficiency within the devices to further reduce the parameter and activation management overhead. In the end, we achieve an MFU of 30.0% when training Pangu Ultra MoE, with performance comparable to that of DeepSeek R1, on 6K Ascend NPUs, and demonstrate that the Ascend system is capable of harnessing all the training stages of the state-of-the-art language models. Extensive experiments indicate that our recipe can lead to efficient training of large-scale sparse language models with MoE. We also study the behaviors of such models for future reference.

cs.CL↗

arXiv2025

Multi-segment Soft Robot Control via Deep Koopman-based Model Predictive Control

Soft robots, compared to regular rigid robots, as their multiple segments with soft materials bring flexibility and compliance, have the advantages of safe interaction and dexterous operation in the environment. However, due to its characteristics of high dimensional, nonlinearity, time-varying nature, and infinite degree of freedom, it has been challenges in achieving precise and dynamic control such as trajectory tracking and position reaching. To address these challenges, we propose a framework of Deep Koopman-based Model Predictive Control (DK-MPC) for handling multi-segment soft robots. We first employ a deep learning approach with sampling data to approximate the Koopman operator, which therefore linearizes the high-dimensional nonlinear dynamics of the soft robots into a finite-dimensional linear representation. Secondly, this linearized model is utilized within a model predictive control framework to compute optimal control inputs that minimize the tracking error between the desired and actual state trajectories. The real-world experiments on the soft robot "Chordata" demonstrate that DK-MPC could achieve high-precision control, showing the potential of DK-MPC for future applications to soft robots.

cs.RO↗

arXiv2025

Kudla-Rapoport conjecture for unramified maximal parahoric level

We prove the Kudla-Rapoport conjecture for unramified unitary groups with maximal parahoric level structure. Our approach differs from the local proof given in Li-W.Zhang. We reduce the conjecture to a global intersection problem using local-global compatibility. Then we apply an inductive procedure based on the modularity of generating series of global special divisors. This strategy follows the framework developed in the proof of the arithmetic fundamental lemma from W.Zhang and Mihatsch-W.Zhang and arithmetic transfer identities from Z.Zhang and Luo-Mihatsch-Z.Zhang.

math.NT↗

arXiv2025

Roadmap on Nonlocality in Photonic Materials and Metamaterials

Photonic technologies continue to drive the quest for new optical materials with unprecedented responses. A major frontier in this field is the exploration of nonlocal (spatially dispersive) materials, going beyond the local, wavevector-independent assumption traditionally made in optical material modeling. On one end, the growing interest in plasmonic, polaritonic and quantum materials has revealed naturally occurring nonlocalities, emphasizing the need for more accurate models to predict and design their optical responses. This has major implications also for topological, nonreciprocal, and time-varying systems based on these material platforms. Beyond natural materials, artificially structured materials--metamaterials and metasurfaces--can provide even stronger and engineered nonlocal effects, emerging from long-range interactions or multipolar effects. This is a rapidly expanding area in the field of photonic metamaterials, with open frontiers yet to be explored. In the case of metasurfaces, in particular, nonlocality engineering has become a powerful tool for designing strongly wavevector-dependent responses, enabling enhanced wavefront control, spatial compression, multifunctional devices, and wave-based computing. Furthermore, nonlocality and related concepts play a critical role in defining the ultimate limits of what is possible in optics, photonics, and wave physics. This Roadmap aims to survey the most exciting developments in nonlocal photonic materials, highlight new opportunities and open challenges, and chart new pathways that will drive this emerging field forward--toward new scientific discoveries and technological advancements.

cond-mat.mes-hall↗

arXiv2025

A Deep Learning Model for Coronary Artery Segmentation and Quantitative Stenosis Detection in Angiographic Images

Coronary artery disease (CAD) is a leading cause of cardiovascular-related mortality, and accurate stenosis detection is crucial for effective clinical decision-making. Coronary angiography remains the gold standard for diagnosing CAD, but manual analysis of angiograms is prone to errors and subjectivity. This study aims to develop a deep learning-based approach for the automatic segmentation of coronary arteries from angiographic images and the quantitative detection of stenosis, thereby improving the accuracy and efficiency of CAD diagnosis. We propose a novel deep learning-based method for the automatic segmentation of coronary arteries in angiographic images, coupled with a dynamic cohort method for stenosis detection. The segmentation model combines the MedSAM and VM-UNet architectures to achieve high-performance results. After segmentation, the vascular centerline is extracted, vessel diameter is computed, and the degree of stenosis is measured with high precision, enabling accurate identification of arterial stenosis. On the mixed dataset (including the ARCADE, DCA1, and GH datasets), the model achieved an average IoU of 0.6308, with sensitivity and specificity of 0.9772 and 0.9903, respectively. On the ARCADE dataset, the average IoU was 0.6303, with sensitivity of 0.9832 and specificity of 0.9933. Additionally, the stenosis detection algorithm achieved a true positive rate (TPR) of 0.5867 and a positive predictive value (PPV) of 0.5911, demonstrating the effectiveness of our model in analyzing coronary angiography images. SAM-VMNet offers a promising tool for the automated segmentation and detection of coronary artery stenosis. The model's high accuracy and robustness provide significant clinical value for the early diagnosis and treatment planning of CAD. The code and examples are available at https://github.com/qimingfan10/SAM-VMNet.

eess.IV↗

arXiv2025

HALURust: Exploiting Hallucinations of Large Language Models to Detect Vulnerabilities in Rust

As an emerging programming language, Rust has rapidly gained popularity and recognition among developers due to its strong emphasis on safety. It employs a unique ownership system and safe concurrency practices to ensure robust safety. Despite these safeguards, security in Rust still presents challenges. Since 2018, 442 Rust-related vulnerabilities have been reported in real-world applications. The limited availability of data has resulted in existing vulnerability detection tools performing poorly in real-world scenarios, often failing to adapt to new and complex vulnerabilities. This paper introduces HALURust, a novel framework that leverages hallucinations of large language models (LLMs) to detect vulnerabilities in real-world Rust scenarios. HALURust leverages LLMs' strength in natural language generation by transforming code into detailed vulnerability analysis reports. The key innovation lies in prompting the LLM to always assume the presence of a vulnerability. If the code sample is vulnerable, the LLM provides an accurate analysis; if not, it generates a hallucinated report. By fine-tuning LLMs on these hallucinations, HALURust can effectively distinguish between vulnerable and non-vulnerable code samples. HALURust was evaluated on a dataset of 81 real-world vulnerabilities, covering 447 functions and 18,691 lines of code across 54 applications. It outperformed existing methods, achieving an F1 score of 77.3%, with over 10% improvement. The hallucinated report-based fine-tuning improved detection by 20\% compared to traditional code-based fine-tuning. Additionally, HALURust effectively adapted to unseen vulnerabilities and other programming languages, demonstrating strong generalization capabilities.

cs.CR↗

arXiv2025

Step-Video-T2V Technical Report: The Practice, Challenges, and Future of Video Foundation Model

We present Step-Video-T2V, a state-of-the-art text-to-video pre-trained model with 30B parameters and the ability to generate videos up to 204 frames in length. A deep compression Variational Autoencoder, Video-VAE, is designed for video generation tasks, achieving 16x16 spatial and 8x temporal compression ratios, while maintaining exceptional video reconstruction quality. User prompts are encoded using two bilingual text encoders to handle both English and Chinese. A DiT with 3D full attention is trained using Flow Matching and is employed to denoise input noise into latent frames. A video-based DPO approach, Video-DPO, is applied to reduce artifacts and improve the visual quality of the generated videos. We also detail our training strategies and share key observations and insights. Step-Video-T2V's performance is evaluated on a novel video generation benchmark, Step-Video-T2V-Eval, demonstrating its state-of-the-art text-to-video quality when compared with both open-source and commercial engines. Additionally, we discuss the limitations of current diffusion-based model paradigm and outline future directions for video foundation models. We make both Step-Video-T2V and Step-Video-T2V-Eval available at https://github.com/stepfun-ai/Step-Video-T2V. The online version can be accessed from https://yuewen.cn/videos as well. Our goal is to accelerate the innovation of video foundation models and empower video content creators.

cs.CV↗

arXiv2025

Detecting Code Vulnerabilities with Heterogeneous GNN Training

Detecting vulnerabilities in source code is a critical task for software security assurance. Graph Neural Network (GNN) machine learning can be a promising approach by modeling source code as graphs. Early approaches treated code elements uniformly, limiting their capacity to model diverse relationships that contribute to various vulnerabilities. Recent research addresses this limitation by considering the heterogeneity of node types and using Gated Graph Neural Networks (GGNN) to aggregate node information through different edge types. However, these edges primarily function as conduits for passing node information and may not capture detailed characteristics of distinct edge types. This paper presents Inter-Procedural Abstract Graphs (IPAGs) as an efficient, language-agnostic representation of source code, complemented by heterogeneous GNN training for vulnerability prediction. IPAGs capture the structural and contextual properties of code elements and their relationships. We also propose a Heterogeneous Attention GNN (HAGNN) model that incorporates multiple subgraphs capturing different features of source code. These subgraphs are learned separately and combined using a global attention mechanism, followed by a fully connected neural network for final classification. The proposed approach has achieved up to 96.6% accuracy on a large C dataset of 108 vulnerability types and 97.8% on a large Java dataset of 114 vulnerability types, outperforming state-of-the-art methods. Its applications to various real-world software projects have also demonstrated low false positive rates.

cs.CR↗

arXiv2025

Photometric Objects Around Cosmic Webs (PAC) Delineated in a Spectroscopic Survey. VIII. Revisiting the Lensing is Low Effect

The issue of over-predicting the galaxy-galaxy lensing (GGL) signal using conventional galaxy-halo connection models has become well-known as the ``Lensing is Low'' problem, which has been extensively investigated using the Baryon Oscillation Spectroscopic Survey (BOSS) galaxy samples. This issue is also tightly related to the so-called $S_8$ tension. By applying our Photometric objects Around Cosmic webs (PAC) method to the BOSS survey and the DESI deep photometric survey, we obtained hundreds of cross-correlation measurements to establish an accurate galaxy-halo connection for BOSS galaxies through the halo abundance matching technique (Paper IV). With this galaxy-halo connection, we show in this work that the predicted GGL signals for BOSS galaxies both in the Planck and WMAP Universes actually agree very well with the GGL measurements. We find the best-fitting value $S_8 = 0.8294 \pm 0.0110$, $0.8073 \pm 0.0372$ and $0.8189 \pm 0.0440$ for the CMASS samples with the source galaxies from HSC, DES and KiDS image surveys, respectively. Our work indicates that accurate modeling of the lens population is so critical to interpret the GGL observation. For the scale of $r_p < 0.6\,h^{-1}\rm{Mpc}$, our GGL prediction for LOWZ samples are also in good agreement with the observations of HSC and DES. However, the GGL observation of KiDS is much lower on the small scale. Our results indicate that no significant baryon feedback is needed to suppress the small scale clustering unless the the GGL observation of KiDS on the small scale will be confirmed.

astro-ph.CO↗

arXiv2025

Searching for Nearby Diffuse Dwarf Galaxies in the COSMOS Field

It remains challenging to systematically survey nearby diffuse dwarf galaxies and address the formation mechanism of this population distinguishing from regular ones. We carry out a pilot search for these galaxies in the COSMOS field using the deep \textit{HST}/F814W imaging data. We report three diffuse dwarf galaxies satisfying the criteria: (1) redshift $z<0.2$, (2) effective radius $r_{\rm e}>1.0''$, and (3) central surface brightness $μ_{\rm 0}>24$ mag arcsec$^{-2}$. Two of the three galaxies, COSMOS-UDG1 and COSMOS-UDG2, are recognized as ultra-diffuse galaxies (UDGs) with redshift $z=0.130$ and $0.049$, respectively. The third galaxy, COSMOS-dw1, is spectroscopically confirmed as a dwarf galaxy at $z=0.004$. We derive the physical properties through fitting their spectral energy distributions (SEDs) extracted from deep multiwavelength observations. COSMOS-dw1 has a stellar mass of $5.6_{-2.7}^{+2.5}\times10^{6}$ M$_{\odot}$, harboring neutral hydrogen gas of mass $4.90\pm0.90\times10^{6}$ M$_{\odot}$, hinting that this galaxy may be in the nascent stages of quenching. The estimated dynamical mass of $3.4\times10^{7}\,M_{\odot}$ further suggests that COSMOS-dw1 is predominantly of dark matter. COSMOS-UDG1 and COSMOS-UDG2 exhibit comparable stellar masses of $\sim 2\times10^{8}$ M$_{\odot}$. Notably, COSMOS-UDG1 is younger and more metal-rich than COSMOS-UDG2 and COSMOS-dw1. Conversely, COSMOS-UDG2 and COSMOS-dw1 have similar stellar metallicities, yet COSMOS-UDG2 is older than COSMOS-dw1. All three galaxies adhere to the stellar mass-metallicity relation (MZR) for dwarf galaxies in the local Universe, implying they belong to the dwarf galaxy population.

astro-ph.GA↗

arXiv2025

Simulation as Reality? The Effectiveness of LLM-Generated Data in Open-ended Question Assessment

The advancement of Artificial Intelligence (AI) has created opportunities for e-learning, particularly in automated assessment systems that reduce educators' workload and provide timely feedback to students. However, developing effective AI-based assessment tools remains challenging due to the substantial resources required for collecting and annotating real student data. This study investigates the potential and gap of simulative data to address this limitation. Through a two-phase experimental study, we examined the effectiveness and gap of Large Language Model generated synthetic data in training educational assessment systems. Our findings reveal that while simulative data demonstrates promising results in training automated assessment models, outperforming state-of-the-art GPT-4o in most question types, its effectiveness has notable limitations. Specifically, models trained on synthetic data show excellent performance in simulated environment but need progress when applied to real-world scenarios. This performance gap highlights the limitations of only using synthetic data in controlled experimental settings for AI training. The absence of real-world noise and biases, which are also present in over-processed real-world data, contributes to this limitation. We recommend that future development of automated assessment agents and other AI tools should incorporate a mixture of synthetic and real-world data, or introduce more realistic noise and biases patterns, rather than relying solely on synthetic or over-processed data.

cs.CY↗

arXiv2025

Conditional Diffusion-based Parameter Generation for Quantum Approximate Optimization Algorithm

The Quantum Approximate Optimization Algorithm (QAOA) is a hybrid quantum-classical algorithm that shows promise in efficiently solving the MaxCut problem, a representative example of combinatorial optimization. However, its effectiveness heavily depends on the parameter optimization pipeline, where the parameter initialization strategy is nontrivial due to the non-convex and complex optimization landscapes characterized by issues with low-quality local minima. Recent inspiration comes from the diffusion of classical neural network parameters, which has demonstrated that neural network training can benefit from generating good initial parameters through diffusion models. Therefore, in this work, we formulate the problem of finding good initial parameters as a generative task and propose the initial parameter generation scheme through dataset-conditioned pre-trained parameter sampling. Concretely, the generative machine learning model, specifically the denoising diffusion probabilistic model (DDPM), is trained to learn the distribution of pretrained parameters conditioned on the graph dataset. Intuitively, our proposed framework aims at effectively distilling knowledge from pre-trained parameters to generate well-performing initial parameters for QAOA. Compared to random parameter initialization, experiments on various-sized Max-Cut problem instances consistently show that our conditional DDPM is capable of improving the approximation ratio by as much as 14.4%, 11.0%, 11.4% and 7.49%, 8.31%, 6.08% on average for random, regular, and Watts-Strogatz graphs, respectively. Additionally, the experimental results also indicate that the conditional DDPM trained on small problem instances can be extrapolated to larger ones, improving the approximation ratio by up to 28.4% and 12.1% on average.

quant-ph↗

arXiv2025

Topologically protected edge states in time photonic crystals with chiral symmetry

Time photonic crystals are media in which their electromagnetic parameters are modulated periodically in time, showing promising applications in non-resonant lasers and particle accelerators, among others. Traditionally utilized to study space photonic crystals, topological band theory has also been translated recently to analyze time photonic crystals with time inversion symmetry, enabling the construction of the temporal version of topological edge states. However, temporal disorder can readily break time inversion symmetry in practice, hence likely destroying the edge states associated with this type of time photonic crystals. To overcome this limitation, here we propose a new class of time photonic crystals presenting chiral symmetry instead, whose edge states exhibit superior robustness over the time-reversal-symmetry-protected counterparts. Our time photonic crystal is equivalent to a temporal version of the Su-Schrieffer-Heeger model, and the chiral symmetry of this type of time photonic crystals quantizes the winding number defined in the Bloch frequency band. Remarkably, random temporal disorders do not impact the eigenfrequencies of these chiral-symmetry-protected edge states, while instead enhancing their temporal localizations. Our findings thus provide a promising paradigm to control field amplification with exceptional robustness as well as being a feasible platform to investigate various topological phases in time-varying media.

physics.optics↗