arXiv subjects

Cheng Li

Publications and source records attributed to Cheng Li.

At least 37 records · Page 2Linked to original sources

arXiv2026

SPHEREx 3.3 Micron Aromatic Emission in the Magellanic Clouds: Separating Dust Column, Excitation, and Environmental Suppression

Bright emission in the 3.3um unidentified infrared band, often described as an aromatic infrared band, can trace dust column, radiative excitation, or the survival and emissivity of its carriers. We combine SPHEREx maps of the Magellanic Clouds with HERITAGE dust columns, a TIR heating proxy, Brα, Hα, H II-region catalogs, and SAGE 8um imaging. We measure continuum-subtracted 3.3um feature excess, X_{3.3}, and Brαexcess, X_{Brα}, in matched cells. In each Cloud, a quiescent diffuse baseline depending on dust surface density, Σ_d, and TIR/Σ_d predicts X_{3.3} with R^2=0.628 in the LMC and 0.288 in the SMC. Relative to this expectation, LMC H II interiors are under-luminous by -0.223^{+0.035}_{-0.040} dex; local-background measurements give -0.153^{+0.014}_{-0.043} dex despite a total-X_{3.3} enhancement of +0.749 dex. SMC H II-region residuals are consistent with zero over its lower-X_{Brα} regime. A matched comparison with stellar-subtracted 8um dust intensity, I_{8,dust}, shows that column and heating predict I_{8,dust} more tightly than X_{3.3}, and that the high-X_{Brα} LMC deficit is stronger for X_{3.3} than for I_{8,dust}. The deficit is therefore selective to the isolated 3.3um feature, not uniform in all aromatic-band-dominated emission. This result suggests that the small aromatic/aliphatic carbonaceous carriers traced by the isolated 3.3um feature are more susceptible to processing in active regions than the larger or differently emitting carbonaceous-grain population dominating the 8um aromatic complexes.

astro-ph.GA↗

arXiv2026

PointQ-Bench: Benchmarking Diagnostic and Interpretable Point Cloud Quality Assessment

Point cloud quality plays a critical role in 3D acquisition, reconstruction, rendering, and perception, yet existing point cloud quality assessment (PCQA) research remains largely centered on scalar score prediction. In practical inspection scenarios, quality assessment often involves identifying defects, characterizing dominant issue types, assessing downstream usability, and providing evidence-supported descriptions, which are not explicitly evaluated by current benchmarks. We introduce PointQ-Bench, a benchmark designed to extend PCQA from scalar scoring toward comprehensive quality understanding. PointQ-Bench consists of 3,083 point clouds spanning authentic scans, simulated distortions, and AI-generated content, covering eight major issue types. Each sample is annotated with mean opinion scores (MOS), quality levels, issue tags, expert-grounded descriptions, and 12,332 question-answer pairs. The benchmark supports three perception-oriented tasks: anomaly sensing, defect diagnosis, and usability grading, as well as a cognition-oriented task of open-ended quality reporting. To evaluate free-form quality descriptions, we further propose SSFRQ-5D, a five-dimensional evaluation protocol validated through human-AI agreement analysis. Extensive experiments on 14 vision-language models and traditional PCQA baselines reveal a consistent perception-diagnosis gap: while current models exhibit emerging abilities in coarse defect perception, they struggle with grounded diagnosis and quality calibration. Strong 2D MLLMs generally outperform existing 3D VLMs, and the benefit of additional views or point-level inputs is non-uniform, varying across tasks, data sources, and models, particularly under boundary-ambiguous conditions. Overall, PointQ-Bench provides a diagnostic testbed for advancing reliable and interpretable point cloud quality understanding.

cs.CV↗

arXiv2026

Train What You Deploy:Token-Faithful Post-Training of a Production Coding

Existing post-training pipelines for coding and terminal agents suffer severe token and control fidelity errors: simplified training environments mismatch production deployments, and offline token reconstruction from agent logs distorts original prompts and conflates policy calls with background model operations. We present a fidelity-aware training coupling framework that retains trainer-side sampling over original prompts, eliminates spurious model calls via a negotiated training protocol, and restricts loss computation to verifiable token spans with closed-failure guarantees. We further propose Certified Divergence Proximal Policy Optimization (C-DPPO), which establishes tight two-sided TV certification bounds, adaptive-K rules, budget-aware sequence guarantees, and error-robust policy masking atop standard DPPO. Evaluated on matched Baize5B and Baize10B models with identical training and test protocols on TMax-100, C-DPPO yields a consistent +3.0-point performance gain over standard DPPO across model scales. Certificate audits validate the reliability and full operational coverage of our certified training pipeline.

cs.AI↗

arXiv2026

Beyond Appearance: Can Multimodal Large Language Models Exploit Vertical Structure for Remote Sensing Natural Scene Understanding?

Multimodal large language models (MLLMs) have advanced rapidly in remote-sensing analysis, yet existing evaluations remain predominantly 2D-centric. Because spectrally confused regions can appear nearly identical yet differ substantially in vertical structure, appearance alone is often insufficient for reliable semantic interpretation in natural scenes. Vertical structure therefore provides decision-critical physical evidence, yet whether current MLLMs can effectively perceive, ground, and utilize such geometric evidence remains underexplored. To bridge this gap, we introduce VertiCue-Bench, the first diagnostic benchmark that uses controlled interventions to probe whether vertical height evidence is actually perceived, grounded, and utilized, and we establish a three-stage evidence-utilization framework of Perception--Grounding--Utilization. By constructing a Representation Intervention Spectrum spanning multiple presentation and interaction modalities, including Raw Visual, Tool-assisted, and Oracle Text conditions, together with controlled counterfactual tests, we conduct an in-depth disentangled diagnosis across 10 state-of-the-art models. Our experiments reveal and formally characterize the Vertical Structure Utilization Gap. Although current models exhibit emerging geometric perception capabilities, they still struggle to accurately ground vertical evidence to relevant spatial entities and integrate it into high-level semantic decisions. This finding identifies a critical bottleneck in developing physically grounded and geometry-aware remote-sensing MLLMs.

cs.CV↗

arXiv2026

Production probability of super-heavy nuclei in fusion

The synthesis of super-heavy nuclei (SHN) through fusion reactions is a critical area of nuclear physics, offering insights into nuclear stability and the limits of the periodic table. However, theoretical predictions of evaporation residue cross sections $ σ_{\rm {ER} }$ remain challenging due to large uncertainties arising from complex reaction mechanisms and sensitive model parameters. In this work, the average production probabilities of SHN with atomic number $Z\ge 110$ are systematically analyzed, based on barrier tunneling concept. Together with the empirical barrier distribution method for describing capture, a phenomenological formula, EBD3, is proposed, which reproduces 58 measured $ σ_{\rm {ER} }$ within one order of magnitude, with a root-mean-square deviation of 0.369. The formula successfully captures key quantities in fission-like process, including fission barrier height, mass asymmetry, depth of capture pocket and the effective fusion barrier height. Predictions for the synthesis of element 119 are presented, identifying promising projectile-target combinations such as $^{50}$Ti + $^{249}$Bk with a maximum cross section of $54.9_{-29.0}^{+61.3}$ fb at excitation energy of 36.5 MeV. The maximum cross section falls to $3.2_{-1.7}^{+3.6}$ fb for $^{54}$Cr + $^{243}$Am at the optimal incident energy of 244 MeV.

nucl-th↗

arXiv2026

MortarBench: Evaluating Mortgage Loan Origination Agents

Loan origination is the process by which a lender creates a new loan, from application and underwriting through approval and funding. This process serves a critical role in evaluating the eligibility and level of risk posed by an applicant. Recently, firms have begun using mortgage loan agents to augment human loan officers, despite a lack of any public benchmark. To fill this gap, we present MortarBench, a loan origination agent benchmark. MortarBench uses a financial data synthesis and mutation pipeline to generate examples with broad edge case coverage that match real-world distributions and questions. We find that state-of-the-art large language models (LLMs) perform poorly, with closed-source models achieving at most 77.1\% exact match accuracy. We also discover systematic biases in LLM perception of foreignness related to non-English names. Noting these weaknesses, we introduce CRIT, a confidence calibration framework. Our method increases accuracy to 80.5\% while improving risk management steering and reducing bias.

cs.LG↗

arXiv2026

SDSS-IV MaNGA: Star Formation Cessation in Low-redshift Galaxies. III. Dependence on Quenching Criteria

This paper is the third in a series of studies investigating star formation cessation in nearby galaxies on kiloparsec scales. Using the final SDSS-IV MaNGA data release, we ask how the inferred importance of global, local, and environmental properties depends on the operational definition of quenched regions. We classify spaxels as star-forming, reliably quenched, or potentially quenched by accounting for measurement uncertainties, and train random forest classifiers with a parameter set chosen for direct comparison with previous work. For reliably quenched regions, the local stellar mass surface density $Σ_\ast$ consistently has the highest feature importance, independent of quenching definition. By contrast, the high importance of central velocity dispersion $σ_c$, previously interpreted as evidence for galaxy-wide AGN feedback, is recovered mainly when potentially quenched regions are included. The leading parameter also varies with stellar mass: $Σ_{\rm 1kpc}$ is most important below $\sim10^{10.2}\,\textrm{M}_{\odot}$, whereas local quantities such as $Σ_\ast$ and $σ_\ast$ become more prominent at high masses. These results show that quenching criteria and uncertainty treatment can reconcile apparently discrepant feature-importance studies. AGN-related processes may contribute to ambiguous regions, but the reliably quenched population is most tightly linked to high local stellar density.

astro-ph.GA↗

arXiv2026

Full-field mapping of spatially varying polarization entanglement generated from spontaneous parametric down-conversion

Two-photon states generated from spontaneous parametric down-conversion (SPDC) can display entanglement in all degrees of freedom (DoFs) of light, including spatial, temporal, and polarization. The coupling between different DoFs of a two-photon state has been shown to display rich structures that enable novel and robust information processing schemes. While existing literature has studied these couplings by post-selecting the SPDC field, a comprehensive understanding of the inherent spatial-polarization coupling produced in the SPDC process is still lacking. This work produces a full spatial map of the polarization entanglement generated across the entire SPDC field, which contains an entire class of near-maximally polarization-entangled states with an average concurrence of 0.8303 +/- 0.0004. The spatial-polarization coupling manifests as radially or linearly varying polarization-entangled states, whose wavefunctions are dependent on the transverse momenta of the down-converted photons and the pump beam, respectively. Our study lays important groundwork for further exploiting the coupling between entanglement in different DoFs for future quantum technologies

quant-ph↗

arXiv2026

Evidence for Azimuthally Anisotropic MgII Absorption around DESI Luminous Red Galaxies

We use luminous red galaxies (LRGs) and background quasar spectra from DESI DR1 to measure the mean MgII equivalent width around massive quiescent galaxies as a function of projected radius and azimuth relative to the projected major axis. Our forced-measurement approach assigns a MgII doublet-window EW to every LRG-quasar pair, including spectra without individually detected absorbers, and subtracts a redshift-matched random-control signal measured from the same normalized quasar spectra. The all-angle profile declines smoothly over $r_p=0.01$-$1.0$ proper Mpc and shows stronger inner absorption at $0.4<z_{\rm LRG}<0.75$ than at $0.75<z_{\rm LRG}<1.1$. Superposed on this radial and redshift dependence, sightlines near the LRG major axis show enhanced absorption relative to the minor axis at $r_p\simeq30$-$80$ kpc. For the full $0.4<z_{\rm LRG}<1.1$ sample, the integrated fiducial major-minus-minor difference over $r_p=0.03$-$0.077$ Mpc is $0.184\pm0.075$ Angstrom, with a position-angle randomization probability $p=0.013$. The redshift split shows no significant evolution in the anisotropy amplitude, even though the inner EW profile itself evolves. Our results provide evidence for a localized major-axis enhancement of MgII-bearing cool/warm gas around LRGs, showing that azimuthal CGM structure is measurable in massive quiescent halos as well as in star-forming systems.

astro-ph.GA↗

arXiv2026

Enhancement of alpha-decay by positive hexadecapole deformation

Whether hexadecapole deformation ($β_4$) influences $α$ decay remains controversial: machine-learning analyses suggest a strong link to cluster preformation, while empirical formulas find only marginal effects. We show that this discrepancy originates in the treatment of shell effects---without an explicit shell correction, residuals near magic numbers are absorbed into the deformation coefficients, obscuring the genuine $β_4$ dependence. Adding the inverse Casten factor $C_{pn}$, which encodes valence proton--neutron correlations relative to the nearest closed shells, and the parent-nucleus deformation $β_4^{(p)}$ to the Royer formula reduces the root-mean-square deviation from 0.309 to 0.184 for 192 even--even nuclei. The fitted negative $β_4^{(p)}$ coefficient shows that positive hexadecapole deformation systematically shortens half-lives, consistent with enhanced $α$-cluster preformation at locally convex surface regions. For the $Z=94$ isotopic chain (Pu), where pronounced $β_4^{(p)}>0$ occurs, the original Royer formula overestimates half-lives by up to $\sim\!0.5$~dex, providing a clear, testable signature of this surface-preformation effect. The same correction also improves the UDL and yields predictions for 1060 even--even nuclei.

nucl-th↗

arXiv2026

DeaMoE: Efficient MoE Structure for Fast Small-Batch Decoding

Mixture-of-Experts (MoE) models have been widely adopted in real-time interactive applications such as coding assistants, real-time audio-video interaction systems. To meet the extremely low response latency requirements of these scenarios, practitioners commonly employ small-batch decoding, under which MoE inference becomes memory-bound and is severely bottlenecked by expert weight loading. However, this bottleneck has received limited attention, and existing solutions such as post-training weight compression or fine-grained expert design during pre-training either degrade model accuracy or introduce additional computation and communication overhead. To tackle this issue, we propose DeaMoE, a decoding-efficient MoE architecture, in which the experts are grouped into several departments, and the experts belonging to the same department share most parameters since they come from the same professional field, and additionally each expert contains a few private parameters to reflect its uniqueness. Moreover, we design customized two-stage routing strategy for DeaMoE to avoid redundant loading, under which DeaMoE greatly improves the efficiency during LLM decoding. Compared with vanilla MoE, DeaMoE reduces per-step loaded weights by up to 50.9% and achieves up to 1.33 end-to-end TPOT speedup for the pre-trained 7B model on A40, and up to 2.00x and 1.97x peak speedup for DeepSeek-V3 on A40 and H100 in microbenchmarks.

cs.LG↗

arXiv2026

Kimi K2.5: Visual Agentic Intelligence

We introduce Kimi K2.5, an open-source multimodal agentic model designed to advance general agentic intelligence. K2.5 emphasizes the joint optimization of text and vision so that two modalities enhance each other. This includes a series of techniques such as joint text-vision pre-training, zero-vision SFT, and joint text-vision reinforcement learning. Building on this multimodal foundation, K2.5 introduces Agent Swarm, a self-directed parallel agent orchestration framework that dynamically decomposes complex tasks into heterogeneous sub-problems and executes them concurrently. Extensive evaluations show that Kimi K2.5 achieves state-of-the-art results across various domains including coding, vision, reasoning, and agentic tasks. Agent Swarm also reduces latency by up to $4.5\times$ over single-agent baselines. We release the post-trained Kimi K2.5 model checkpoint to facilitate future research and real-world applications of agentic intelligence.

cs.CL↗

arXiv2026

Juno Microwave Observations Reveal Jupiter's Deep Alkali-Chlorine Relation

The longest-wavelength channel of the Juno Microwave Radiometer (MWR) probes Jupiter's kilobar atmosphere through free electrons produced by sodium and potassium ionization. Under equilibrium chemistry the electron abundance is the small residual of the charge balance between alkali cations and the anions Cl- and HS-. Chlorine is not directly measurable in Jupiter's deep atmosphere because gaseous HCl is removed from the observable atmosphere by NH4Cl condensation, whereas sulfur has been measured by the Galileo probe. The MWR-derived electron measurement therefore constrains the alkali-to-chlorine ratio rather than the alkali abundance alone. We combine the MWR observations with equilibrium chemistry and microwave radiative transfer in a Bayesian framework, finding that the deep gas-phase elemental alkali-to-chlorine abundance ratio is (Na+K)/Cl = 0.05 over 0.3-5 times solar in chlorine, about 180 times below the protosolar ratio of 8.7. At 3 times solar chlorine, the inferred alkali metallicity is 1.6 x 10^-2 times solar (1 sigma: 1.2 x 10^-2 - 2.7 x 10^-2 times solar), while at low chlorine abundance HS- sets an alkali floor near 10^-3 times solar. The inferred gas-phase alkali abundance exceeds the ~10^-5 times solar threshold by more than two orders of magnitude and rules out the long-proposed global kilobar radiative zone. Because sodium and potassium are refractory whereas chlorine is volatile, the inferred ratio provides a new diagnostic of the rock-to-ice balance in the solids accreted by Jupiter. This compositional interpretation assumes equilibrium chemistry; if lofted mineral clouds instead control the electron abundance under disequilibrium conditions, the inferred alkali-chlorine relationship need not hold.

astro-ph.EP↗

arXiv2026

Alkali Metallicity, Mineral Clouds, and Deep Atmospheric Variability on Jupiter

The bulk elemental abundances of Jupiter provide critical insights into its formation history and interior structure. Recent observations by the Juno Microwave Radiometer (MWR) reveal a deep Jovian atmosphere significantly depleted in electrons, implying an alkali metal (Na, K) abundance of 10^-1 - 10^-5 times solar. This depletion stands in sharp contrast to the supersolar volatile enrichments measured by the Galileo probe. We propose that this apparent depletion arises from mineral cloud-induced processes deep in the atmosphere. We explore two physical mechanisms using thermochemical and microphysical modeling. In the "chemical sequestration" scenario, vigorous vertical mixing lofts deep refractory condensates (e.g., spinel) into the 1000-2000 bar region, where they react to form alkali feldspars (albite) and feldspathoids (leucite), efficiently sequestering gaseous Na and K. In the "dust-catalyzed recombination" scenario, the bulk alkali inventory remains gaseous, but the free electron density is suppressed by dust-plasma interactions. Thermally emitted alkali ions from the surfaces of micron-sized iron and silicate grains significantly increase the cation density, driving rapid recombination of free electrons. Both mechanisms allow for a bulk solar or even supersolar alkali inventory while suppressing the electron density to match Juno observations. Analyzing an extended dataset of MWR observations with 61 perijoves, we detect spatial variability in the deep atmosphere that suggests modulation by mineral clouds. Our findings challenge the traditional rainout framework, unveiling a deep "mineralogical zone" in Jupiter shaped by dynamics and heterogeneous chemistry, resembling the photospheres of hot exoplanets and brown dwarfs.

astro-ph.EP↗

arXiv2026

AuroraRL: Fast, Fault-Tolerant, and Cost-Efficient Reinforcement Learning over Decentralized Network

LLM reinforcement learning (RL) requires frequent synchronization of large model parameters between the trainer and distributed rollout actors. High-throughput RL post-training therefore relies on dedicated RDMA HPC/cloud clusters, an infrastructure cost most organizations cannot absorb. A natural alternative is to aggregate loosely-coupled GPUs over standard Ethernet and WAN links, but this commodity connectivity cannot sustain full-weight broadcasts: synchronizing an 8B model can take over 100~seconds on bandwidth-limited links, while rollout generation typically takes tens of seconds. Toward making RL practical in this regime, we observe that RL fine-tuning yields highly sparse per-step updates, with only around 1\% of parameter elements changing. On top of this insight, we present AuroraRL, a novel high-performance RL training system that preserves bit-exact updates without dropping or quantizing information, designed for commodity-networked, loosely-coupled GPU resources. AuroraRL represents each step as a sparse delta checkpoint, pipelines delta extraction with multi-stream transmission, overlaps transfer with rollout generation, and coordinates heterogeneous workers with throughput- and bandwidth-aware scheduling plus lease-based fault tolerance. Across Qwen3 4B--14B models deployed in up to four geographic regions, AuroraRL shrinks per-step weight transfer by 79$\times$ on Qwen3-8B, delivers 1.3--9.5$\times$ higher throughput than dense-broadcast baselines (PrimeRL-Full, async-tolerant, multi-stream variants), and brings end-to-end training within 8.91\% of an ideal RDMA single-datacenter baseline, while transparently tolerating common failures and preserving training accuracy. By leveraging on-demand, cross-cloud GPUs over commodity links, AuroraRL delivers 1.21--1.59$\times$ higher tokens per dollar than reserved RDMA clusters at comparable throughput.

cs.DC↗

arXiv2026

RSC-GestureNet: Reliability-Aware Selective Causal Recognition of Chinese Traffic Police Gestures

Traffic police gestures are safety-critical perception cues for autonomous driving. A deployable recognizer must infer commands causally from continuous full-frame video, remain stable around transitional arm motion, and avoid over-trusting corrupted pose measurements. This study presents RSC-GestureNet, a reliability-aware selective causal recognizer, for Chinese traffic police gestures. The model treats pose confidence as a first-class signal: unreliable joints are down weighted during graph reasoning, temporal evidence is aggregated causally, and calibrated predictions are selectively emitted through a reliability-aware inference rule. We further introduce CTPGesture-C, a reproducible feature-level corruption benchmark with seven pose/RGB degradation families, and an RGB-level diagnostic in which corrupted frames are reprocessed by MediaPipe before recognition. On the complete official CTPGesture v1 split (134,424 labeled frames and 33,451 causal windows), RSC-GestureNet achieves 93.33+-0.24% accuracy, 91.71+-0.27% macro-F1, 91.69+-0.29% online macro-F1, 98.80+-0.07% Early@10, 0.153+-0.013 s TTC, and the best robust macro-F1 among evaluated methods. Under the same split and causal protocol, it exceeds reproduced traffic-specific MD-GCN and HLP-GCN baselines by 3.23-4.11 macro-F1 points and 2.15-3.07 online-F1 points. These results, together with calibration, selective-risk, statistical, adaptive-branching, and image-level re-extraction analyses, indicate that explicit pose-reliability modeling improves early, stable, and robust traffic-command recognition.

cs.CV↗

arXiv2026

Correlations of ALMA CO(2-1) with JWST mid-infrared fluxes down to scale of $\lesssim$100 parsec in nearby star-forming galaxies from PHANGS

We investigate the correlations of CO (2-1) emission (${I_{\rm CO}}$) with PAH (${I_{\rm F770W, PAH}}$ and ${I_{\rm F1130W}}$) and dust (${I_{\rm F2100W}}$) emission down to scales of $\lesssim$ 100 pc, by applying ${\tt raddest}$, a novel regression technique recently developed by T. Jing & C. Li (2025) that effectively handles uncertainties and outliers in datasets, to 19 nearby star-forming galaxies in the PHANGS sample. We find that for the majority of the data points in all galaxies, the scaling of ${I_{\rm CO}}$ with ${I_{\rm F770W, PAH}}$, ${I_{\rm F1130W}}$, and ${I_{\rm F2100W}}$ can be well described by log-log linear relations, though with substantial dependence on ionization conditions (i.e., HII-like, composite-like, and AGN-like). Under given ionization conditions, significant galaxy-to-galaxy variations are identified, and are primarily attributed to variations of intercept $b$, which exhibits clear bimodality. This bimodality is related to the normalized overall host galaxy star formation rate, such as specific star formation and star formation efficiency. The differences in slope $k$ and intrinsic scatter $σ$ across different MIR bands (${I_{\rm F770W, PAH}}$, ${I_{\rm F1130W}}$, and ${I_{\rm F2100W}}$) are minor compared to their galaxy-to-galaxy variations. All parameters ($k$, $b$, and $σ$) depend on the spatial scale of measurement, suggesting that the coupling among CO, PAH, and dust is regulated by different mechanisms at varying scales. We identify deviations from the log-log linear relation in the brightest regions, primarily characterized by a flattening of the slope. No significant (3$σ$) correlations are found between global properties and the best-fit parameters. We discuss the comparison to previous studies and plausible physics behind the statistical results obtained in this work.

astro-ph.GA↗

arXiv2026

Galactic HII regions in LAMOST Medium-Resolution Spectroscopic Survey of Nebulae

Based on LAMOST Medium-Resolution Spectroscopic Survey of Nebulae (MRS-N) data and WISE Galactic HII region catalog, we construct a sample of 280 Galactic HII regions and candidates in the Outer Galaxy (80$^{\circ}$ $\lesssim$ l $\lesssim$ 220$^{\circ}$). Using MRS-N optical spectra, we measure four emission lines (H$α$, [NII]$λ$6584, [SII]$λλ$6717,6731) and use line-ratios to spectroscopically confirm 255 HII regions, including 90 previously "Known" HII regions and 165 newly classified ones. We measure their $T_{\rm e}$, $n_{\rm e}$ and oxygen abundance, and determine distances via associated OB stars and the kinematic method. The sample spans $R_{\rm gal}$ from 8.16 to 15.36 kpc, enabling investigation of radial gradients in physical properties. We find [NII]/H$α$ and [SII]/H$α$ decrease with increasing $R_{\rm gal}$, while [SII]/[NII] remains nearly flat; these trends are quite different from diffuse ionized gas (DIG). We derive the $T_{\rm e}$ gradient of 344.530 $\pm$ 78.083 K kpc$^{-1}$, and the $\log n_{\rm e}$ gradient of -0.143 $\pm$ 0.041 cm$^{-3}$ kpc$^{-1}$. Oxygen abundance shows a steep slope of -0.044 $\pm$ 0.010 dex kpc$^{-1}$ in the inner disk and a shallow slope of -0.016 $\pm$ 0.005 dex kpc$^{-1}$ in the outer disk, with a global slope of -0.014 $\pm$ 0.005 dex kpc$^{-1}$. We also examine the two-dimensional distributions of $T_{\rm e}$, $n_{\rm e}$, and oxygen abundance, and find the gradients vary with azimuth. There is no obvious difference between spiral arm and interarm regions, and no trend appears along individual arms. From [NII]/H$α$-[SII]$λ$6717/H$α$ diagram, HII regions have a S$^+$/S ratio (0.32), lower than DIG (0.43); however, heavy overlap prevents clear separation from this diagram alone.

astro-ph.GA↗