arXiv subjects

Huan Chen

Publications and source records attributed to Huan Chen.

At least 19 recordsLinked to original sources

arXiv2026

Toward an Organizational Science of Multi-Agent LLM Systems: Decoupling Who, How, and Which Algorithm

Multi-agent frameworks built on large language models (LLMs) routinely entangle three logically distinct concerns: who is on the team (organization), how members align (coordination), and which algorithm fuses their work (collaboration protocol). IMACS (Intelligent Multi-Agent Collaboration System) separates the three into orthogonal, independently swappable layers. Classic organizational theory (Belbin roles, Mintzberg coordination, RACI accountability) becomes executable, validated configuration, and the framework places six published collaboration algorithms behind a common interface while exposing roles, coordination, and accountability as independently configurable factors. We use this separation to conduct controlled comparisons in which organizational assignments vary while the collaboration protocol is held fixed. It also turns protocol choice into a variable that can be learned: Adaptive Org Routing, a contextual-bandit meta-protocol, selects a protocol per task under an explicit quality-cost tradeoff, outperforms every fixed protocol in a controlled study, and trains online on real benchmark and LLM-judge rewards. The ablations expose a mechanism. Accountability placement changes outcomes exactly when the protocol routes the deliverable through the accountable agent, and the winning placement flips across model families, so organizational design cannot be hard-coded; it must be revalidated, or learned, for each model binding.

cs.AI↗

arXiv2026

OwlPath: Lossless Knowledge Compression for LLM Bug Repair

LLM-based software engineering agents are constrained by limited context windows: roughly 100K tokens must store structurally relevant code subsets to resolve bugs. Standard retrieval models treat code as plain text, forcing agents to resolve multi-hop dependencies including subclass chains, transitive callers and interface implementations through slow trial and error. We tackle this limitation with lossless knowledge compression, encoding source code into an OWL2 ontology to answer structural queries using minimal relevant code fragments. We present OwlPath, an OWL2 reasoning layer atop CodeGraph, a widely used code intelligence platform with 500K+ GitHub stars, offering a unified CLI for structural code retrieval. Powered by tree-sitter parsing, OwlPath supports multi-language repositories (Python, JavaScript, TypeScript, Go, etc.) and encodes language-specific semantics into a unified OWL2 ontology. It adopts two complementary modules. First, a transitive-closure engine fetches all structurally linked symbols via single SPARQL property-path queries, capturing multi-hop relations missed by string matching. Second, the OWL Software Knowledge Map (OWL-SKM) precomputes a compact 3KB summary with module trees, core APIs and issue-related symbols, directing agents to target modules in the first query. Evaluated on 18 SWE-bench Pro instances, OwlPath obtains a 68.4% strict-apply rate versus 66.7% for the CodeGraph baseline, cutting token usage by 28.8% and runtime by 39.5%. In offline retrieval tests over 67 instances, OwlPath improves recall 2.06 times (0.464 vs 0.226) and reaches 88.1% hit rate compared to CodeGraph's 59.7%. On a 37-question structural retrieval benchmark, recall rises from 4.4% to 28.8%, with 69-80% accuracy on transitive caller and interface tasks.

cs.SE↗

arXiv2026

Two-fluid $f$-mode oscillations of dark-matter-admixed neutron stars

We study quadrupolar $f$-mode oscillations of dark-matter-admixed neutron stars (DANSs) in full general relativity (GR). The ordinary component is described by microscopic Brueckner-Hartree-Fock matter matched to the Shen2020 crust, while the dark matter (DM) component is treated as a cold self-interacting fermion fluid coupled to ordinary matter only by gravity. For fixed-DM-fraction sequences we solve the polar two-fluid perturbation equations with an outgoing gravitational-wave (GW) boundary condition, obtaining complex eigenfrequencies rather than only real mode frequencies. The spectrum contains two principal $f$-like sequences. Their local character can be ordinary-matter-led, DM-led, or mixed, and is diagnosed using the component kinetic energies, the displacement overlap, and the cancellation of the matter quadrupole. A main result is that, for intermediate DM fractions, one of the two-fluid branches can become weakly radiating, with damping times enhanced by several orders of magnitude. The same calculation gives the outgoing Zerilli amplitude and the GW damping time, which we use to estimate the GW energy required to reach a prescribed detector threshold. Thus the analysis extends previous two-fluid Cowling studies by retaining metric perturbations and the radiative boundary condition.

gr-qc↗

arXiv2026

Cooling of Hybrid Stars with a 2SC+$ $ Phase

Recently, Fujimoto, Fukushima & Weise (2019) have proposed a new colour-superconductive state, 2SC+$ $ phase, which can be smoothly connected to the low-density baryon superfluidity in contrast to the 2SC phase. In this scenario, the neutron ${}^3P_2$ superfluidity on the low-density side of the phase transition is inherited by unpaired $d$-quarks in the 2SC phase on the high-density side. Since this could be realized in hybrid stars (neutron stars containing hadronic and quark matter), the 2SC+$ $ phase may change the properties of neutron stars compared to the traditional 2SC phase. In this work, we study the thermal evolution of hybrid stars with the 2SC+$ $ phase for the first time. We find that NSs with the 2SC+$ $ phase become hotter than those with the 2SC phase, and are close to the CFL phase. The ${}^{3}P_2$ superfluidity plays an important role in cooling curves with not the 2SC but 2SC+$ $ phases due to the suppression of quark $\beta$ decay. We therefore point out that, if the scenario of 2SC+$ $ phase is true, it could be specified through low-temperature observations such as Vela, 3C58, Vela Jr., and Vela-like pulsar.

nucl-th↗

arXiv2026

Partial Fairness Awareness: Belief-Guided Strategic Mechanism for Strategic Agents

Strategic machine learning investigates scenarios where agents manipulate their features to receive favorable decisions from predictive models. To address fairness concerns intrinsic to strategic classification, recent work has introduced group-specific fairness constraints. However, current fairness-aware approaches face a fundamental dilemma in the issue of fairness exposure: making these constraints public enables strategic manipulation and can lead to fairness reversal, while keeping them hidden may reduce social welfare and discourage genuine improvement. To fill this gap, we subsequently propose the problem of partial fairness awareness (PFA), as our theoretical analysis informs that such a dilemma can be mitigated by releasing the candidate set of fairness constraints and concealing the grounding constraint. To be specific, we introduce a belief-guided strategic mechanism, wherein agents iteratively interact with the decision system and maintain a belief distribution over the candidate set of fairness constraints. This belief-guided process enables agents, through iterative interaction and feedback, to update their belief distribution over the candidate set, thereby gradually aligning their belief with the grounding fairness constraint employed by the system. Extensive experiments on real-world and synthetic datasets demonstrate that PFA achieves lower group fairness gaps, higher acceptance of truly qualified individuals, and more stable outcomes compared to fully public or private fairness regimes.

cs.LG↗

arXiv2026

Reinforcement Learning from Denoising Feedback

Policy loss estimation remains a fundamental and long-standing challenge in reinforcement learning (RL) for diffusion language models (DLMs). We introduce Reinforcement Learning from Denoising Feedback (RLDF), a novel training paradigm that leverages feedback obtained from rollout and training processes to facilitate accurate and efficient policy loss estimation. To balance the trade-off between computational efficiency and estimation effectiveness, RLDF optimizes the model toward the clipped clean state from intermediate noisy states, combined with weighted timestep sampling over denoising timesteps. Extensive experiments demonstrate that RLDF achieves consistent and substantial improvements in both performance and generalizability across two representative DLM architectures, LLaDA and Dream, on multiple reasoning benchmarks. Our work lays a principled foundation for scalable reinforcement learning in diffusion language models. We build Drift, a training framework for DLMs, available at https://github.com/ant-research/Drift.

cs.CL↗

arXiv2026

Quasiradial oscillations of rotating hybrid neutron stars

We investigate fundamental quasiradial oscillations in slow-rotation approximation of pure and hybrid neutron stars, employing equations of state of nuclear matter from Brueckner-Hartree-Fock theory or the relativistic mean field model, and of quark matter from the Dyson-Schwinger quark model, performing a Gibbs construction for the mixed phase in hybrid stars. Characteristic differences between neutron-star and hybrid-star fundamental quasiradial oscillation frequencies during spin-down are pointed out.

nucl-th↗

arXiv2026

Hyperspectral Image Classification via Efficient Global Spectral Supertoken Clustering

Hyperspectral image classification demands spatially coherent predictions and precise boundary delineation. Yet prevailing superpixel-based methods face an inherent contradiction: clustering aggregates similar pixels into regions, but the subsequent classifier operates pixel-wise, undermining regional consistency. Consequently, existing approaches do not guarantee region-level, boundary-aligned classification. To address this limitation, we propose the Dual-stage Spectrum-Constrained Clustering-based Classifier (DSCC), an end-to-end framework that explicitly decouples clustering from classification by first grouping spectral similar and spatially proximate pixels into spectral supertokens and then performing token-level prediction. At its core, DSCC computes an image-level multi-criteria feature distance between pixels and centers, followed by a locality-aware assignment regularization, enabling the generation of boundary-preserving spectral supertokens. A density-isolation based center selection further yields representative, well-separated centers, reducing redundancy and improving robustness to scale variation. To accommodate mixed land-cover compositions within each token, we introduce a soft-label scheme that encodes class proportions and improves robustness for mixed-class tokens. DSCC attains a CF1 of 0.728 at 197.75 FPS on the WHU-OHS dataset, offering a superior accuracy-efficiency trade-off compared with state-of-the-art methods. Extensive experiments further validate the effectiveness and generality of the proposed dual-stage paradigm for hyperspectral image classification. The source code is available at https://github.com/laprf/DSCC.

cs.CV↗

arXiv2026

Properties of Galactic Outflows Driven by Starburst at Cosmic Noon: Insights from Hydrodynamical Simulations

We investigate starburst-driven galactic outflows in low-mass galaxies ($9.0 < \log(M_*/M_\odot) < 10.0$) at cosmic noon using high-resolution 3D hydrodynamical simulations based on a framework that can reproduce the multiphase outflows in M82. The simulations produce starbursts lasting 20-30 Myr, with peak star formation rates of 2-68 M$_\odot \,\rm{yr}^{-1}$. Outflow properties vary strongly with time, radial distance to galaxy center, stellar mass, and gas fraction, exhibiting velocities of 50-1000 $\,\rm{km\,s}^{-1}$, mass outflow rates of 0.3-20 M$_\odot \,\rm{yr}^{-1}$, and mass loading factors, $\eta_\mathrm{M}$, of 0.24-6.26. The cool phase ($8000 < T \le 2 \times 10^4$ K) dominates the outflow, and properties of the cool and warm phases are broadly consistent with observations. At $M_*= 10^{9.5}\,M_\odot$, average $\eta_\mathrm{M}$ for the total, cool, and warm phases are $\sim$1.2, 0.75, and 0.25, respectively. The mass loading factor decreases with increasing galaxy stellar mass, but increases with star formation rate. Given strong temporal and spatial evolution, scaling slopes from limited samples should be treated with caution. Our total $\eta_\mathrm{M}$ values are higher than FIRE-2 by 0.06 dex but lower than EAGLE and TNG50 by 0.50 and 0.84 dex. Accounting for methodological differences in outflow measurement reduces these gaps to 0.2-0.4 dex, suggesting that part of the discrepancy between observations and simulations reported in the literature may arise from inconsistent definitions and measurement methods, though differences in individual phases persist. Larger observational and simulation samples, together with consistent methods for measuring outflow properties, are required to draw robust conclusions about the scaling relations of galactic outflows.

astro-ph.GA↗

arXiv2026

Up to 36x Speedup: Mask-based Parallel Inference Paradigm for Key Information Extraction in MLLMs

Key Information Extraction (KIE) from visually-rich documents (VrDs) is a critical task, for which recent Large Language Models (LLMs) and Multi-Modal Large Language Models (MLLMs) have demonstrated strong potential. However, their reliance on autoregressive inference, which generates outputs sequentially, creates a significant efficiency bottleneck, especially as KIE tasks often involve extracting multiple, semantically independent fields. To overcome this limitation, we introduce PIP: a Parallel Inference Paradigm for KIE. Our approach reformulates the problem by using "[mask]" tokens as placeholders for all target values, enabling their simultaneous generation in a single forward pass. To facilitate this paradigm, we develop a tailored mask pre-training strategy and construct large-scale supervised datasets. Experimental results show that our PIP-models achieve a 5-36x inference speedup with negligible performance degradation compared to traditional autoregressive base models. By substantially improving efficiency while maintaining high accuracy, PIP paves the way for scalable and practical real-world KIE solutions.

cs.CL↗

arXiv2026

HyperCOD: The First Challenging Benchmark and Baseline for Hyperspectral Camouflaged Object Detection

RGB-based camouflaged object detection struggles in real-world scenarios where color and texture cues are ambiguous. While hyperspectral image offers a powerful alternative by capturing fine-grained spectral signatures, progress in hyperspectral camouflaged object detection (HCOD) has been critically hampered by the absence of a dedicated, large-scale benchmark. To spur innovation, we introduce HyperCOD, the first challenging benchmark for HCOD. Comprising 350 high-resolution hyperspectral images, It features complex real-world scenarios with minimal objects, intricate shapes, severe occlusions, and dynamic lighting to challenge current models. The advent of foundation models like the Segment Anything Model (SAM) presents a compelling opportunity. To adapt the Segment Anything Model (SAM) for HCOD, we propose HyperSpectral Camouflage-aware SAM (HSC-SAM). HSC-SAM ingeniously reformulates the hyperspectral image by decoupling it into a spatial map fed to SAM's image encoder and a spectral saliency map that serves as an adaptive prompt. This translation effectively bridges the modality gap. Extensive experiments show that HSC-SAM sets a new state-of-the-art on HyperCOD and generalizes robustly to other public HSI datasets. The HyperCOD dataset and our HSC-SAM baseline provide a robust foundation to foster future research in this emerging area.

cs.CV↗

arXiv2025

SatSAM2: Motion-Constrained Video Object Tracking in Satellite Imagery using Promptable SAM2 and Kalman Priors

Existing satellite video tracking methods often struggle with generalization, requiring scenario-specific training to achieve satisfactory performance, and are prone to track loss in the presence of occlusion. To address these challenges, we propose SatSAM2, a zero-shot satellite video tracker built on SAM2, designed to adapt foundation models to the remote sensing domain. SatSAM2 introduces two core modules: a Kalman Filter-based Constrained Motion Module (KFCMM) to exploit temporal motion cues and suppress drift, and a Motion-Constrained State Machine (MCSM) to regulate tracking states based on motion dynamics and reliability. To support large-scale evaluation, we propose MatrixCity Video Object Tracking (MVOT), a synthetic benchmark containing 1,500+ sequences and 157K annotated frames with diverse viewpoints, illumination, and occlusion conditions. Extensive experiments on two satellite tracking benchmarks and MVOT show that SatSAM2 outperforms both traditional and foundation model-based trackers, including SAM2 and its variants. Notably, on the OOTB dataset, SatSAM2 achieves a 5.84% AUC improvement over state-of-the-art methods. Our code and dataset will be publicly released to encourage further research.

cs.CV↗

arXiv2025

Multi-Aspect Cross-modal Quantization for Generative Recommendation

Generative Recommendation (GR) has emerged as a new paradigm in recommender systems. This approach relies on quantized representations to discretize item features, modeling users' historical interactions as sequences of discrete tokens. Based on these tokenized sequences, GR predicts the next item by employing next-token prediction methods. The challenges of GR lie in constructing high-quality semantic identifiers (IDs) that are hierarchically organized, minimally conflicting, and conducive to effective generative model training. However, current approaches remain limited in their ability to harness multimodal information and to capture the deep and intricate interactions among diverse modalities, both of which are essential for learning high-quality semantic IDs and for effectively training GR models. To address this, we propose Multi-Aspect Cross-modal quantization for generative Recommendation (MACRec), which introduces multimodal information and incorporates it into both semantic ID learning and generative model training from different aspects. Specifically, we first introduce cross-modal quantization during the ID learning process, which effectively reduces conflict rates and thus improves codebook usability through the complementary integration of multimodal information. In addition, to further enhance the generative ability of our GR model, we incorporate multi-aspect cross-modal alignments, including the implicit and explicit alignments. Finally, we conduct extensive experiments on three well-known recommendation datasets to demonstrate the effectiveness of our proposed method.

cs.IR↗

arXiv2025

Breaking the Gradient Barrier: Unveiling Large Language Models for Strategic Classification

Strategic classification~(SC) explores how individuals or entities modify their features strategically to achieve favorable classification outcomes. However, existing SC methods, which are largely based on linear models or shallow neural networks, face significant limitations in terms of scalability and capacity when applied to real-world datasets with significantly increasing scale, especially in financial services and the internet sector. In this paper, we investigate how to leverage large language models to design a more scalable and efficient SC framework, especially in the case of growing individuals engaged with decision-making processes. Specifically, we introduce GLIM, a gradient-free SC method grounded in in-context learning. During the feed-forward process of self-attention, GLIM implicitly simulates the typical bi-level optimization process of SC, including both the feature manipulation and decision rule optimization. Without fine-tuning the LLMs, our proposed GLIM enjoys the advantage of cost-effective adaptation in dynamic strategic environments. Theoretically, we prove GLIM can support pre-trained LLMs to adapt to a broad range of strategic manipulations. We validate our approach through experiments with a collection of pre-trained LLMs on real-world and synthetic datasets in financial and internet domains, demonstrating that our GLIM exhibits both robustness and efficiency, and offering an effective solution for large-scale SC tasks.

cs.LG↗

arXiv2025

Semantic4Safety: Causal Insights from Zero-shot Street View Imagery Segmentation for Urban Road Safety

Street-view imagery (SVI) offers a fine-grained lens on traffic risk, yet two fundamental challenges persist: (1) how to construct street-level indicators that capture accident-related features, and (2) how to quantify their causal impacts across different accident types. To address these challenges, we propose Semantic4Safety, a framework that applies zero-shot semantic segmentation to SVIs to derive 11 interpretable streetscape indicators, and integrates road type as contextual information to analyze approximately 30,000 accident records in Austin. Specifically, we train an eXtreme Gradient Boosting (XGBoost) multi-class classifier and use Shapley Additive Explanations (SHAP) to interpret both global and local feature contributions, and then apply Generalized Propensity Score (GPS) weighting and Average Treatment Effect (ATE) estimation to control confounding and quantify causal effects. Results uncover heterogeneous, accident-type-specific causal patterns: features capturing scene complexity, exposure, and roadway geometry dominate predictive power; larger drivable area and emergency space reduce risk, whereas excessive visual openness can increase it. By bridging predictive modeling with causal inference, Semantic4Safety supports targeted interventions and high-risk corridor diagnosis, offering a scalable, data-informed tool for urban road safety planning.

cs.CV↗

arXiv2025

GBA-UBF : A Large-Scale and Fine-Grained Building Function Classification Dataset in the Greater Bay Area

Rapid urbanization in the Guangdong-Hong Kong-Macao Greater Bay Area (GBA) has created urgent demand for high-resolution, building-level functional data to support sustainable spatial planning. Existing land use datasets suffer from coarse granularity and difficulty in capturing intra-block heterogeneity. To this end, we present the Greater Bay Area Urban Building Function Dataset (GBA-UBF), a large-scale, fine-grained dataset that assigns one of five functional categories to nearly four million buildings across six core GBA cities. We proposed a Multi-level Building Function Optimization (ML-BFO) method by integrating Points of Interest (POI) records and building footprints through a three-stage pipeline: (1) candidate label generation using spatial overlay with proximity weighting, (2) iterative refinement based on neighborhood label autocorrelation, and (3) function-related correction informed by High-level POI buffers. To quantitatively validate results, we design the Building Function Matching Index (BFMI), which jointly measures categorical consistency and distributional similarity against POI-derived probability heatmaps. Comparative experiments demonstrate that GBA-UBF achieves significantly higher accuracy, with a BMFI of 0.58. This value markedly exceeds that of the baseline dataset and exhibits superior alignment with urban activity patterns. Field validation further confirms the dataset's semantic reliability and practical interpretability. The GBA-UBF dataset establishes a reproducible framework for building-level functional classification, bridging the gap between coarse land use maps and fine-grained urban analytics. The dataset is accessible at https://github.com/chenchs0629/GBA-UBF, and the data will undergo continuous improvement and updates based on feedback from the community.

cs.CY↗

arXiv2025

$f$-mode oscillations of protoneutron stars

We investigate nonradial $f$-mode oscillations of protoneutron stars in full general relativity, employing equations of state described by the Brueckner-Hartree-Fock theory or the relativistic mean field model, while assuming isentropy and fixed lepton fractions for the internal structure. The validity of various universal relations for cold neutron stars involving $f$-mode characteristics and macroscopic properties of the star is confirmed for those isentropic protoneutron stars. Prospects of observations are also discussed. According to simulation results, we then model details of the thermal and trapping profiles in a PNS with the canonical mass. The corresponding $f$-mode frequencies and gravitational-wave strain amplitudes are presented. The validity of the universal relations during the evolution to the formation of a cold neutron star is confirmed.

astro-ph.HE↗

arXiv2025

Hyperspectral Remote Sensing Images Salient Object Detection: The First Benchmark Dataset and Baseline

The objective of hyperspectral remote sensing image salient object detection (HRSI-SOD) is to identify objects or regions that exhibit distinct spectrum contrasts with the background. This area holds significant promise for practical applications; however, progress has been limited by a notable scarcity of dedicated datasets and methodologies. To bridge this gap and stimulate further research, we introduce the first HRSI-SOD dataset, termed HRSSD, which includes 704 hyperspectral images and 5327 pixel-level annotated salient objects. The HRSSD dataset poses substantial challenges for salient object detection algorithms due to large scale variation, diverse foreground-background relations, and multi-salient objects. Additionally, we propose an innovative and efficient baseline model for HRSI-SOD, termed the Deep Spectral Saliency Network (DSSN). The core of DSSN is the Cross-level Saliency Assessment Block, which performs pixel-wise attention and evaluates the contributions of multi-scale similarity maps at each spatial location, effectively reducing erroneous responses in cluttered regions and emphasizes salient regions across scales. Additionally, the High-resolution Fusion Module combines bottom-up fusion strategy and learned spatial upsampling to leverage the strengths of multi-scale saliency maps, ensuring accurate localization of small objects. Experiments on the HRSSD dataset robustly validate the superiority of DSSN, underscoring the critical need for specialized datasets and methodologies in this domain. Further evaluations on the HSOD-BIT and HS-SOD datasets demonstrate the generalizability of the proposed method. The dataset and source code are publicly available at https://github.com/laprf/HRSSD.

cs.CV↗