arXiv subjects

Cheng Cheng

Publications and source records attributed to Cheng Cheng.

At least 73 records · Page 4Linked to original sources

arXiv2025

JWST's PEARLS: A z=6 quasar in a train-wreck galaxy merger system

We present JWST NIRSpec integral field spectroscopy observations of the z=5.89 quasar NDWFS J1425+3254 from 0.6-5.3 microns, covering the rest-frame ultraviolet and optical at a spectral resolution of R~100. The quasar has a black hole mass of $M_{\rm{BH}}=(1.4\substack{+3.1\\-1.0})\times10^9 M_\odot$ and an Eddington ratio of $L_{\rm{Bol}}/L_{\rm{Edd}}=0.3\substack{+0.6\\-0.2}$, as implied from the broad Balmer H$α$ and H$β$ lines. The quasar host has significant ongoing obscured star formation, as well as a quasar-driven outflow with velocity $6050\substack{+460\\-630}$ km/s and ionised outflow rate of $1650\substack{+130\\-1230}M_\odot$yr$^{-1}$. This is possibly one of the most extreme outflows in the early Universe. The data also reveal that two companion galaxies are merging with the quasar host. The north-eastern companion galaxy is relatively old and very massive, with a luminosity-weighted stellar age of $65\substack{+9\\-4}$ Myr, stellar mass of $(3.6\substack{+0.6\\-0.3})\times10^{11} M_\odot$, and star-formation rate (SFR) of ~15-30 $M_\odot$yr$^{-1}$. A bridge of gas connects this companion galaxy and the host, confirming their ongoing interaction. A second merger is occurring between the quasar host and a much younger companion galaxy to the south, with a stellar age of $6.7\pm1.8$ Myr, stellar mass of $(1.9\pm0.4)\times10^{10} M_\odot$, and SFR of ~40-65 $M_\odot$yr$^{-1}$. There is also another galaxy in the field, likely in the foreground at z=1.135, which could be gravitationally lensing the quasar with magnification $1<μ<2$, and, thus, <0.75 mag. Overall, the system is a 'train-wreck' merger of three galaxies, with star formation and extreme quasar activity that were likely triggered by these ongoing interactions.

astro-ph.GA↗

arXiv2025

Multi-Modal Machine Learning Framework for Predicting Early Recurrence of Brain Tumors Using MRI and Clinical Biomarkers

Accurately predicting early recurrence in brain tumor patients following surgical resection remains a clinical challenge. This study proposes a multi-modal machine learning framework that integrates structural MRI features with clinical biomarkers to improve postoperative recurrence prediction. We employ four machine learning algorithms -- Gradient Boosting Machine (GBM), Random Survival Forest (RSF), CoxBoost, and XGBoost -- and validate model performance using concordance index (C-index), time-dependent AUC, calibration curves, and decision curve analysis. Our model demonstrates promising performance, offering a potential tool for risk stratification and personalized follow-up planning.

cs.LG↗

arXiv2025

A Multimodal Deep Learning Framework for Early Diagnosis of Liver Cancer via Optimized BiLSTM-AM-VMD Architecture

This paper proposes a novel multimodal deep learning framework integrating bidirectional LSTM, multi-head attention mechanism, and variational mode decomposition (BiLSTM-AM-VMD) for early liver cancer diagnosis. Using heterogeneous data that include clinical characteristics, biochemical markers, and imaging-derived variables, our approach improves both prediction accuracy and interpretability. Experimental results on real-world datasets demonstrate superior performance over traditional machine learning and baseline deep learning models.

cs.LG↗

arXiv2025

Lonely Little Red Dots: Challenges to the AGN-nature of little red dots through their clustering and spectral energy distributions

Observations with the James Webb Space Telescope (JWST) reveal a previously unseen population of compact red objects, known as ``little red dots`` (LRDs). We study a new photometrically selected sample of 124 LRDs in the redshift range $z$ $\sim$ 3 - 10 selected from NIRCam coverage of the CEERS, NEP-TDF, JADES and JEMS surveys. For JADES, the NEP-TDF and CEERS, we compare SED models with and without AGN components and analyse the impact of an AGN component on the goodness of fit using the Bayesian information criterion (BIC). We find that whilst the $χ^{2}$ of the majority of models containing AGN components is improved compared to models without AGN components, we show that the BIC suggests models without AGN are a more appropriate fit to LRD SEDs, especially when MIRI data is available. We also measure LRD clustering in the CEERS field, JADES field, and NEP-TDF, where we compare the spatial distribution of LRDs and galaxies with Kolmogorov-Smirnov tests of equality of distribution. We find that the neighbourhood of LRDs tends to be less dense compared to galaxies at all selections and masses and at similar redshifts. We further measure upper limit estimates for the halo masses of LRDs using abundance matching. Whilst the population of LRDs could be a mixture of several different inherent populations, as a whole it does appear that these systems are mostly hosting compact galaxies or star clusters in formation.

astro-ph.GA↗

arXiv2025

Conjugate phase retrieval on graphs and with applications in shift-invariant spaces

In this paper, we study the conjugate phase retrieval for complex-valued \mbox{signals} residing on graphs, and explore its applications to shift-invariant spaces. Given a complex-valued graph signal $\bf f$ residing on the graph $\mathcal G$, we introduce a graph ${\mathcal G}_{\bf f}$ and show that its connectivity is sufficient to determine $\bf f$ up to a global unimodular constant and conjugation. We then construct two explicit graph models and show that graph signals residing on them can be recovered, up to a unimodular constant and conjugation, from its absolute values on the vertices and the relative magnitudes between neighboring vertices. Building on this graph-based framework, we apply our results to shift-invariant spaces generated by real-valued functions. For signals in the Paley-Wiener space, we show that any complex-valued function can be recovered, up to a unimodular constant and conjugation, from structured phaseless samples taken at three times the Nyquist rate. For more general shift invariant spaces, we establish the conjugate phase retrievability of signals from phaseless samples collected on a discrete sampling set, in conjunction with relative magnitude measurements between neighboring sample points. Two numerical reconstruction algorithms are introduced to recover the signals in the Paley-Wiener space and general shift-invariant spaces, up to a unimodular constant and conjugation, from the given phaseless measurements.

math.FA↗

arXiv2025

JWST/MIRI reveals the true number density of massive galaxies in the early Universe

Early JWST studies reporting an unexpected abundance of massive galaxies at $z \sim 5$--$8$ challenge galaxy formation models in the $Λ$CDM framework. Previous stellar mass ($M_\star$) estimates suffered from large uncertainties due to the lack of rest-frame near-infrared data. Using deep JWST/NIRCam and MIRI photometry from PRIMER, we systematically analyze massive galaxies at $z \sim 3$--$8$, leveraging rest-frame $\gtrsim 1\,μ$m constraints. We find MIRI is critical for robust $M_\star$ measurements for massive galaxies at $z > 5$: excluding MIRI overestimates $M_\star$ by $\sim 0.4$ dex on average for $M_\star > 10^{10}\,M_\odot$ galaxies, with no significant effects at lower masses. This reduces number densities of $M_\star > 10^{10}\,M_\odot$ ($10^{10.3}\,M_\odot$) galaxies by $\sim 36\%$ ($55\%$). MIRI inclusion also reduces ``Little Red Dot'' (LRD) contamination in massive galaxy samples, lowering the LRD fraction from $\sim 32\%$ to $\sim 13\%$ at $M_\star > 10^{10.3}\,M_\odot$. Assuming pure stellar origins, LRDs exhibit $M_\star \sim 10^{9\text{--}10.5}\,M_\odot$ with MIRI constraints, rarely exceeding $10^{10.5}\,M_\odot$. Within standard $Λ$CDM, our results indicate a moderate increase in the baryon-to-star conversion efficiency ($ε$) toward higher redshifts and masses at $z > 3$. For the most massive $z \sim 8$ galaxies, $ε\sim 0.3$, compared to $ε\lesssim 0.2$ for typical galaxies at $z < 3$. This result is consistent with models where high gas densities and short free-fall times suppress stellar feedback in massive high-$z$ halos.

astro-ph.GA↗

arXiv2025

Detecting Galactic Rings in the DESI Legacy Imaging Surveys with Semi-Supervised Deep Learning

The ring structures of disk galaxies are vital for understanding galaxy evolution and dynamics. However, due to the scarcity of ringed galaxies and challenges in their identification, traditional methods often struggle to efficiently obtain statistically significant samples. To address this, this study employs a novel semi-supervised deep learning model, GC-SWGAN, aimed at identifying galaxy rings from high-resolution images of the DESI Legacy Imaging Surveys. We selected over 5,000 confirmed ringed galaxies from the Catalog of Southern Ringed Galaxies (CSRG) and the Northern Ringed Galaxies from the GZ2 catalog (GZ2-CNRG), both verified by morphology expert R. J. Buta, to create an annotated training set. Additionally, we incorporated strictly selected non-ringed galaxy samples from the Galaxy Zoo 2 dataset and utilized unlabelled data from DESI Legacy Surveys to train our model. Through semi-supervised learning, the model significantly reduced reliance on extensive annotated data while enhancing robustness and generalization. On the test set, it demonstrated exceptional performance in identifying ringed galaxies. With a probability threshold of 0.5, the classification accuracy reached 97\%, with precision and recall for ringed galaxies at 94\% and 93\%, respectively. Building on these results, we predicted 750,000 galaxy images from the DESI Legacy Imaging Surveys with r-band apparent magnitudes less than 17.0 and redshifts in the range 0.0005 < z < 0.25, compiling the largest catalog of ringed galaxies to date, containing 62,962 galaxies with ring structures. This catalog provides essential data for subsequent research on the formation mechanisms and evolutionary history of galaxy rings.

astro-ph.GA↗

arXiv2025

Probing Obscured Star Formation in Galaxy Clusters Using JWST Medium Band Images: 3.3$μ\rm m$ PAH Emitter Sample in Abell 2744

Star-forming galaxies in galaxy clusters play a crucial role in understanding the advanced stages of galaxy evolution within dense environments. We present a sample of 3.3$μ$m PAH-bright galaxies in the Abell 2744 (A2744) galaxy cluster. Using F430M medium band images, we select PAH emitters in the galaxy cluster, which capture the 3.3$μ$m PAH emission at the redshift of A2744. Our multi-wavelength study demonstrates consistent star formation rates (SFRs) derived from PAH emission and SED fitting, indicating the 3.3 $μ$m PAH flux estimated from medium band image alone can reveal the entirety of star formation, immune to dust obscuration. We find that the PAH emitters are located in relatively low mass surface density regions of A2744, with SFRs aligning with the field star-forming main sequence at $z=0.3$. The PAH emission morphologies show more asymmetry than that of the F444W image when asymmetry index $> 0.4$. With these results, we suggest that these star-forming galaxies in A2744 are in the stage of falling into the cluster from the field, and have not been quenched yet. We further explore a potential link between these galaxies and cosmic filaments being accreted onto the cluster, which may channel gas inflows to fuel star formation. JWST medium-band imaging provides a powerful new tool for identifying heavily dust-obscured star-forming populations. Future HI and low-J CO observations should be prioritized to resolve the cold gas kinematics and star formation processes in these systems, which would directly test the role of environmental stripping versus filamentary gas supply.

astro-ph.GA↗

arXiv2025

An H$α$ Cloud in the HI Tail: Recent Star Formation in the Outskirts of NGC 4258 Revealed by Nanshan 1-m Telescope

We present first-light deep H$α$ imaging taken with the Nanshan 1-meter wide-field telescope on the local galaxy NGC 4258, alongside archival data from Hubble Space telescope (HST), Westerbork Synthesis Radio Telescope, and The Dark Energy Camera Legacy Survey. The H$α$ image shows ongoing star formation not only inside the galaxy but also in an HI cloud in the eastern HI tail, which is roughly 16 kpc away from the main galaxy. The HST images reveal several ultra-blue compact objects ($\rm F555W - F814W <-0.5 mag,\, FWHM\sim 0.2''$) in the H$α$ bright region, aligned with the HI tail, suggesting the presence of young open cluster candidate in the HI tail. Our results suggest that wide field H$α$ imaging is a valuable tool for investigating recent star formation in the extended regions of NGC 4258. Furthermore, the star formation in diffuse HI tails could highlight an potential aspect of galaxy stellar halo formation, warranting further investigation of the impact of star formation in halos on galaxy evolution.

astro-ph.GA↗

arXiv2025

IRT-Router: Effective and Interpretable Multi-LLM Routing via Item Response Theory

Large language models (LLMs) have demonstrated exceptional performance across a wide range of natural language tasks. However, selecting the optimal LLM to respond to a user query often necessitates a delicate balance between performance and cost. While powerful models deliver better results, they come at a high cost, whereas smaller models are more cost-effective but less capable. To address this trade-off, we propose IRT-Router, a multi-LLM routing framework that efficiently routes user queries to the most suitable LLM. Inspired by Item Response Theory (IRT), a psychological measurement methodology, IRT-Router explicitly models the relationship between LLM capabilities and user query attributes. This not only enables accurate prediction of response performance but also provides interpretable insights, such as LLM abilities and query difficulty. Additionally, we design an online query warm-up technique based on semantic similarity, further enhancing the online generalization capability of IRT-Router. Extensive experiments on 20 LLMs and 12 datasets demonstrate that IRT-Router outperforms most baseline methods in terms of effectiveness and interpretability. Its superior performance in cold-start scenarios further confirms the reliability and practicality of IRT-Router in real-world applications. Code is available at https://github.com/Mercidaiha/IRT-Router.

cs.AI↗

arXiv2025

LoRA-Gen: Specializing Large Language Model via Online LoRA Generation

Recent advances have highlighted the benefits of scaling language models to enhance performance across a wide range of NLP tasks. However, these approaches still face limitations in effectiveness and efficiency when applied to domain-specific tasks, particularly for small edge-side models. We propose the LoRA-Gen framework, which utilizes a large cloud-side model to generate LoRA parameters for edge-side models based on task descriptions. By employing the reparameterization technique, we merge the LoRA parameters into the edge-side model to achieve flexible specialization. Our method facilitates knowledge transfer between models while significantly improving the inference efficiency of the specialized model by reducing the input context length. Without specialized training, LoRA-Gen outperforms conventional LoRA fine-tuning, which achieves competitive accuracy and a 2.1x speedup with TinyLLaMA-1.1B in reasoning tasks. Besides, our method delivers a compression ratio of 10.1x with Gemma-2B on intelligent agent tasks.

cs.CL↗

arXiv2025

The Internal Kinematics, Stellar Population, and Gas-phase Properties of The Pseudobulge in An Ultra-diffuse Galaxy: AGC721966

Leveraging spectroscopic data from the Sloan Digital Sky Survey, we conduct a comprehensive analysis of the central stellar velocity dispersion, stellar population properties, star formation history, and gas-phase chemical abundances in AGC721966, a unique ultra-diffuse galaxy (UDG) harboring a pseudobulge. Our findings reveal that the pseudobulge formed in the early universe but underwent a recent episode of rejuvenated star formation. The system exhibits a mass-weighted (light-weighted) stellar population age of $τ_{\star}\sim 7.4\pm2.5$ ($2.9\pm1.5$)~Gyr, a stellar metallicity of [M/H]$\sim -0.62\pm0.26$ ($-0.55\pm0.20$), an $α$-element enhancement of [$α$/Fe]$\sim 0.36\pm0.09$ ($0.37\pm0.07$), and a gas-phase oxygen abundance of \Oabund$\sim 8.15\pm0.03$. The central stellar velocity dispersion is measured as $σ_{\rm c}\sim 57.9\pm15.7$~km/s. These results provide robust evidence supporting the early halo-halo merging formation scenario proposed by \cite{Rong25}, while unequivocally ruling out the ``failed'' $L^{\star}$ formation model, at least for AGC721966. Furthermore, through systematic application of the baryonic Tully-Fisher relation, we establish that these pseudobulge-hosting UDGs are neither misidentified nuclear star cluster-bearing dwarf galaxies nor bulge-dominated massive galaxies, thereby affirming their distinct evolutionary pathway.

astro-ph.GA↗

arXiv2025

CogMath: Assessing LLMs' Authentic Mathematical Ability from a Human Cognitive Perspective

Although large language models (LLMs) show promise in solving complex mathematical tasks, existing evaluation paradigms rely solely on a coarse measure of overall answer accuracy, which are insufficient for assessing their authentic capabilities. In this paper, we propose \textbf{CogMath}, which comprehensively assesses LLMs' mathematical abilities through the lens of human cognition. Specifically, inspired by psychological theories, CogMath formalizes human reasoning process into 3 stages: \emph{problem comprehension}, \emph{problem solving}, and \emph{solution summarization}. Within these stages, we investigate perspectives such as numerical calculation, knowledge, and counterfactuals, and design a total of 9 fine-grained evaluation dimensions. In each dimension, we develop an ``\emph{Inquiry}-\emph{Judge}-\emph{Reference}'' multi-agent system to generate inquiries that assess LLMs' mastery from this dimension. An LLM is considered to truly master a problem only when excelling in all inquiries from the 9 dimensions. By applying CogMath on three benchmarks, we reveal that the mathematical capabilities of 7 mainstream LLMs are overestimated by 30\%-40\%. Moreover, we locate their strengths and weaknesses across specific stages/dimensions, offering in-depth insights to further enhance their reasoning abilities.

cs.AI↗

arXiv2025

Towards Foundation Model on Temporal Knowledge Graph Reasoning

Temporal Knowledge Graphs (TKGs) store temporal facts with quadruple formats (s, p, o, t). Existing Temporal Knowledge Graph Embedding (TKGE) models perform link prediction tasks in transductive or semi-inductive settings, which means the entities, relations, and temporal information in the test graph are fully or partially observed during training. Such reliance on seen elements during inference limits the models' ability to transfer to new domains and generalize to real-world scenarios. A central limitation is the difficulty in learning representations for entities, relations, and timestamps that are transferable and not tied to dataset-specific vocabularies. To overcome these limitations, we introduce the first fully-inductive approach to temporal knowledge graph link prediction. Our model employs sinusoidal positional encodings to capture fine-grained temporal patterns and generates adaptive entity and relation representations using message passing conditioned on both local and global temporal contexts. Our model design is agnostic to temporal granularity and time span, effectively addressing temporal discrepancies across TKGs and facilitating time-aware structural information transfer. As a pretrained, scalable, and transferable model, POSTRA demonstrates strong zero-shot performance on unseen temporal knowledge graphs, effectively generalizing to novel entities, relations, and timestamps. Extensive theoretical analysis and empirical results show that a single pretrained model can improve zero-shot performance on various inductive temporal reasoning scenarios, marking a significant step toward a foundation model for temporal KGs.

cs.AI↗

arXiv2025

HaploOmni: Unified Single Transformer for Multimodal Video Understanding and Generation

With the advancement of language models, unified multimodal understanding and generation have made significant strides, with model architectures evolving from separated components to unified single-model frameworks. This paper explores an efficient training paradigm to build a single transformer for unified multimodal understanding and generation. Specifically, we propose a multimodal warmup strategy utilizing prior knowledge to extend capabilities. To address cross-modal compatibility challenges, we introduce feature pre-scaling and multimodal AdaLN techniques. Integrating the proposed technologies, we present the HaploOmni, a new single multimodal transformer. With limited training costs, HaploOmni achieves competitive performance across multiple image and video understanding and generation benchmarks over advanced unified models. All codes will be made public at https://github.com/Tencent/HaploVLM.

cs.CV↗

arXiv2025

From Objectives to Questions: A Planning-based Framework for Educational Mathematical Question Generation

Automatically generating high-quality mathematical problems that align with educational objectives is a crucial task in NLP-based educational technology. Traditional generation methods focus primarily on textual quality, but they often overlook educational objectives. Moreover, these methods address only single-dimensional, simple question generation, failing to meet complex, multifaceted educational requirements. To address these challenges, we constructed and annotated EduMath, a dataset of 16k mathematical questions with multi-dimensional educational objectives. Based on this dataset, we developed EQGEVAL, which incorporates three evaluation dimensions and is designed to assess the ability of models to generate educational questions. Drawing inspiration from teachers' problem design processes, we propose the Educational Question Planning with self-Reflection (EQPR) method for educational mathematical question generation, following a "plan-evaluate-optimize" approach. Specifically, by combining planning algorithm based on Monte Carlo Tree Search with the generative capabilities of Large Language Models, we continuously optimize questions through iterative feedback. This self-optimization mechanism ensures that the generated questions both fit the educational context and strategically achieve specific basic educational objectives. Through extensive experiments based on EQGEVAL, we have demonstrated that EQPR achieves significant improvements in generating questions that meet multi-dimensional educational objectives.

cs.CL↗

arXiv2025

Skywork Open Reasoner 1 Technical Report

The success of DeepSeek-R1 underscores the significant role of reinforcement learning (RL) in enhancing the reasoning capabilities of large language models (LLMs). In this work, we present Skywork-OR1, an effective and scalable RL implementation for long Chain-of-Thought (CoT) models. Building on the DeepSeek-R1-Distill model series, our RL approach achieves notable performance gains, increasing average accuracy across AIME24, AIME25, and LiveCodeBench from 57.8% to 72.8% (+15.0%) for the 32B model and from 43.6% to 57.5% (+13.9%) for the 7B model. Our Skywork-OR1-32B model surpasses both DeepSeek-R1 and Qwen3-32B on the AIME24 and AIME25 benchmarks, while achieving comparable results on LiveCodeBench. The Skywork-OR1-7B and Skywork-OR1-Math-7B models demonstrate competitive reasoning capabilities among models of similar size. We perform comprehensive ablation studies on the core components of our training pipeline to validate their effectiveness. Additionally, we thoroughly investigate the phenomenon of entropy collapse, identify key factors affecting entropy dynamics, and demonstrate that mitigating premature entropy collapse is critical for improved test performance. To support community research, we fully open-source our model weights, training code, and training datasets.

cs.LG↗

arXiv2025

AD-AGENT: A Multi-agent Framework for End-to-end Anomaly Detection

Anomaly detection (AD) is essential in areas such as fraud detection, network monitoring, and scientific research. However, the diversity of data modalities and the increasing number of specialized AD libraries pose challenges for non-expert users who lack in-depth library-specific knowledge and advanced programming skills. To tackle this, we present AD-AGENT, an LLM-driven multi-agent framework that turns natural-language instructions into fully executable AD pipelines. AD-AGENT coordinates specialized agents for intent parsing, data preparation, library and model selection, documentation mining, and iterative code generation and debugging. Using a shared short-term workspace and a long-term cache, the agents integrate popular AD libraries like PyOD, PyGOD, and TSLib into a unified workflow. Experiments demonstrate that AD-AGENT produces reliable scripts and recommends competitive models across libraries. The system is open-sourced to support further research and practical applications in AD.

cs.CL↗