Searcharxiv⌕ Search

arXiv subjects

Bin Chen

Publications and source records attributed to Bin Chen.

At least 145 records · Page 8Linked to original sources
arXiv2026

A Density-Delay Law for Stable Event-Driven State Progression in Open Distributed Systems

Distributed systems in which concurrent proposals are mutually exclusive face a fundamental stability constraint under network delay. In open systems where global state progression is event-driven rather than round-driven, propagation delay creates a conflict window within which overlapping proposals may generate competing branches. This paper derives a density-delay law for such exclusive state progression processes. Under independent proposal arrivals and bounded propagation delay, overlap is approximated by a Poisson model and fork depth is represented by a birth-death process. The analysis shows that maintaining bounded fork depth as the number of participants grows requires the density-delay product $λΔ$ to remain $O(1)$, implying that aggregate proposal intensity must stay bounded and yielding an inverse-scaling law $g(N)=O(1/N)$ at the unit level. Simulation experiments across varying network sizes and propagation delays align with a common density-delay curve, supporting the predicted scaling behavior. The result provides a compact law for stable event-driven state progression in open distributed systems and offers a scaling-based interpretation of Bitcoin-style difficulty adjustment as a decentralized way to regulate effective event density.

cs.DC↗
arXiv2026

CoPRS: Learning Positional Prior from Chain-of-Thought for Reasoning Segmentation

Existing works on reasoning segmentation either connect hidden features from a language model directly to a mask decoder or represent positions in text, which limits interpretability and semantic detail. To solve this, we present CoPRS, a Multi-modal Chain-of-Thought (MCoT)-based positional perception model that bridges language reasoning to segmentation through a differentiable and interpretable positional prior instantiated as a heatmap. By making the reasoning process clear via MCoT and expressing it as a dense, differentiable heatmap, this interface enhances interpretability and diagnostic analysis and yields more concentrated evidence on the target. A learnable concentration token aggregates features of the image and reasoning text to generate this positional prior, which is decoded to precise masks through a lightweight decoder, providing a direct connection between reasoning and segmentation. Across the RefCOCO series and ReasonSeg, CoPRS matches or surpasses the best reported metrics on each standard split under comparable protocols, with performance at or above the prior state of the art across both validation and test partitions. Extensive experiments demonstrate a strong positive correlation among the CoT trajectory, the generated heatmap, and the decoded mask, supporting an interpretable alignment between the reasoning output and downstream mask generation. Collectively, these findings support the utility of this paradigm in bridging reasoning and segmentation and show advantages in concentration driven by reasoning and in more precise mask prediction. Code has been released at https://github.com/ZhenyuLU-Heliodore/CoPRS.

cs.CV↗
arXiv2026

A weighted cycle-localization inequality

In 1959, Erdős and Gallai showed that every $2$-connected graph $G$ contains a cycle of length at least $\frac{2|E(G)|}{|V(G)|-1}$. This result was subsequently extended to weighted graphs by Bondy and Fan in 1991. A natural local variant of this problem arises by considering, for each edge $e\in E(G)$, the quantity $c(e)$, defined as the length of the longest cycle in $G$ containing $e$ (with $c(e)=2$ if $e$ is a bridge). Zhao and Zhang recently proved that for every graph $G$ on $n$ vertices satisfies $\sum_{e\in E(G)}\frac{1}{c(e)}\le \frac{n-1}{2}.$ In this note, we establish a weighted generalization of this inequality. For a weighted graph $(G,w)$ with positive edge weights, let $C_w(e)$ denote the maximum weight of a cycle containing $e$ (setting $C_w(e)=2w(e)$ if $e$ is a bridge). We prove that $$ \sum_{e\in E(G)}\frac{w(e)}{C_w(e)}\le \frac{n-1}{2}. $$ Our result can be viewed as a weighted local analogue of the Bondy-Fan theorem, thereby establishing a correspondence between the global and local perspectives. Furthermore, we present a broad class of graphs attaining equality and derive necessary conditions for equality.

math.CO↗
arXiv2026

PromptHub: Enhancing Multi-Prompt Visual In-Context Learning with Locality-Aware Fusion, Concentration and Alignment

Visual In-Context Learning (VICL) aims to complete vision tasks by imitating pixel demonstrations. Recent work pioneered prompt fusion that combines the advantages of various demonstrations, which shows a promising way to extend VICL. Unfortunately, the patch-wise fusion framework and model-agnostic supervision hinder the exploitation of informative cues, thereby limiting performance gains. To overcome this deficiency, we introduce PromptHub, a framework that holistically strengthens multi-prompting through locality-aware fusion, concentration and alignment. PromptHub exploits spatial priors to capture richer contextual information, employs complementary concentration, alignment, and prediction objectives to mutually guide training, and incorporates data augmentation to further reinforce supervision. Extensive experiments on three fundamental vision tasks demonstrate the superiority of PromptHub. Moreover, we validate its universality, transferability, and robustness across out-of-distribution settings, and various retrieval scenarios. This work establishes a reliable locality-aware paradigm for prompt fusion, moving beyond prior patch-wise approaches. Code is available at https://github.com/luotc-why/ICLR26-PromptHub.

cs.CV↗
arXiv2026

High Quality Underwater Image Compression with Adaptive Color Correction

With the increasing exploration and exploitation of the underwater world, underwater images have become a critical medium for human interaction with marine environments, driving extensive research into their efficient transmission and storage. However, contemporary underwater image compression algorithms fail to adequately address the impact of water refraction and scattering on light waves, which not only elevate training complexity but also result in suboptimal compression performance. To tackle this limitation, we propose High Quality Underwater Image Compression (HQUIC), a novel framework designed to handle the unique illumination conditions and color shifts inherent in underwater images, thereby achieving superior compression performance. HQUIC first incorporates an Adaptive Lighting and Tone Correction (ALTC) module to adaptively predict the attenuation coefficients and global light information of images, effectively alleviating issues stemming from variations in illumination and tone across underwater images. Secondly, it dynamically weights multi-scale frequency components, prioritizing information critical to distortion quality while discarding redundant details. Furthermore, we introduce a tone adjustment loss to enable the model to better balance discrepancies among different color channels. Comprehensive evaluations on diverse underwater datasets validate that HQUIC outperforms state-of-the-art compression methods, demonstrating its effectiveness.

cs.CV↗
arXiv2026

GLM-OCR Technical Report

GLM-OCR is an efficient 0.9B-parameter compact multimodal model designed for real-world document understanding. It combines a 0.4B-parameter CogViT visual encoder with a 0.5B-parameter GLM language decoder, achieving a strong balance between computational efficiency and recognition performance. To address the inefficiency of standard autoregressive decoding in deterministic OCR tasks, GLM-OCR introduces a Multi-Token Prediction (MTP) mechanism that predicts multiple tokens per step, significantly improving decoding throughput while keeping memory overhead low through shared parameters. At the system level, a two-stage pipeline is adopted: PP-DocLayout-V3 first performs layout analysis, followed by parallel region-level recognition. Extensive evaluations on public benchmarks and industrial scenarios show that GLM-OCR achieves competitive or state-of-the-art performance in document parsing, text and formula transcription, table structure recovery, and key information extraction. Its compact architecture and structured generation make it suitable for both resource-constrained edge deployment and large-scale production systems.

cs.CL↗
arXiv2026

On the Limiting Behavior of $L^2$-Critical Pseudo-Relativistic Fermi Systems

We consider ground states of a pseudo-relativistic Fermi system in the $L^2$-critical case. We prove that the system admits ground states, if and only if the attractive strength $a$ satisfies $0<a<D_{4/3,2}$, where $D_{4/3,2}\in(0, \infty)$ is the optimal constant of a dual fractional Lieb--Thirring inequality. The limiting behavior of ground states for the system is further analyzed as $a\nearrow D_{4/3,2}$. As a byproduct, the qualitative properties of optimizers for the dual fractional Lieb-Thirring inequality are also investigated.

math.AP↗
arXiv2026

Modeling Inverse Ellipsometry Problem via Flow Matching with a Large-Scale Dataset

Inverse ellipsometry, i.e., reconstructing optical constants and film thickness from the measured phase difference $Δ$ and amplitude ratio $Ψ$, is a fundamentally ill-posed problem. Traditional solutions rely on slow, expert-driven iterative fitting, while the development of machine learning approaches has been severely limited by the lack of large-scale, physically consistent datasets. To address this gap, we introduce \textbf{EllipBench}, a comprehensive benchmark comprising over 8 million high-precision samples spanning 98 thin-film materials and 5 substrates. Building upon this benchmark, we conduct a systematic evaluation of a broad spectrum of methods, including traditional machine learning models, deep neural networks, and Physics-Informed Neural Networks, and show that existing paradigms consistently struggle to fully resolve the inverse ellipsometry task. To better capture its inherent ambiguity, we further propose a novel \textbf{Decoupled Conditional Flow Matching (DCFM)} framework. Rather than formulating the problem as deterministic point-to-point regression, DCFM explicitly decouples geometric film thickness and incorporates it as a robust physical condition to guide a continuous vector field for modeling the inverse probability distribution of wavelength-dependent optical constants. Combined with a gradient detachment strategy and physics-based constraints, our joint architecture effectively mitigates intrinsic physical ambiguities and delivers a robust and accurate solution for inverse ellipsometry.

cs.LG↗
arXiv2026

An improvement bound on a problem of Picasarri-Arrieta and Rambaud

Let $k$ and $\ell$ be positive integers. A cycle with two blocks $C(k,\ell)$ is a digraph consisting of two internally vertex disjoint directed paths of lengths $k$ and $\ell$ with the same initial vertex and terminal vertex. Picasarri-Arrieta and Rambaud (European J. Combin., 2024) proved that for any $k\geq 2$, every digraph of minimum out-degree at least two and girth at least $8k-6$ contains a subdivision of $C(k,k)$. They also construct a family of digraphs showing that the girth cannot be reduced to $k-1$, and posed the problem of determining the minimum girth such that every digraph of minimum out-degree at least two contains a subdivision of $C(k,k)$. In this paper, we improve the lower bound on the girth from $8k-6$ to $4k+2$, and construct a family of digraphs in which every member has minimum out-degree two and girth $k$ but contains no subdivision of $C(k,k)$. Thus our results show that the girth in question lies between $k+1$ and $4k+2$.

math.CO↗
arXiv2026

Hybrid Integration of Quantum Dot Single Photon Sources with Lithium Tantalate Photonics for On Chip Routing

A promising pathway towards scalable quantum photonic processors involves the simultaneous integration of deterministic single-photon sources, low-loss photonic circuitry, and fast reconfigurability. Thin-film lithium tantalate on insulator (LTOI) offers an exceptional electro-optic response and low optical loss at 900 nm wavelength band, yet its lack of efficient quantum emitters has hindered progress toward fully integrated quantum technologies. Here, we demonstrate heterogeneous integration of indium arsenide quantum dots (QDs) with low-loss reconfigurable LTOI waveguides using micro-transfer printing. By directly butt-coupling tapered gallium arsenide waveguides with inversely tapered LTOI waveguides, we achieve robust and alignment-tolerant inter-waveguide coupling. The hybrid chip operates at cryogenic temperatures, enabling deterministic routing of successively emitted single photons from the QDs with a halfwave voltage-length product, confirming the cryogenic stability of LTOI's electro-optic coefficient. These results establish the first demonstration of high-speed on-chip routing of single photons with hybrid QD-LTOI circuits, providing a scalable pathway toward integrated quantum photonic processors.

physics.optics↗
arXiv2026

OARS: Process-Aware Online Alignment for Generative Real-World Image Super-Resolution

Aligning generative real-world image super-resolution models with human visual preference is challenging due to the perception--fidelity trade-off and diverse, unknown degradations. Prior approaches rely on offline preference optimization and static metric aggregation, which are often non-interpretable and prone to pseudo-diversity under strong conditioning. We propose OARS, a process-aware online alignment framework built on COMPASS, a MLLM-based reward that evaluates the LR to SR transition by jointly modeling fidelity preservation and perceptual gain with an input-quality-adaptive trade-off. To train COMPASS, we curate COMPASS-20K spanning synthetic and real degradations, and introduce a three-stage perceptual annotation pipeline that yields calibrated, fine-grained training labels. Guided by COMPASS, OARS performs progressive online alignment from cold-start flow matching to full-reference and finally reference-free RL via shallow LoRA optimization for on-policy exploration. Extensive experiments and user studies demonstrate consistent perceptual improvements while maintaining fidelity, achieving state-of-the-art performance on Real-ISR benchmarks.

cs.CV↗
arXiv2026

A Commensal Radio-Only Cosmic Ray Detector at the Owens Valley Radio Observatory Long Wavelength Array

The brief (10 nanoseconds) transient radio emission from cosmic ray air showers carries key information about the energy and mass composition of high energy cosmic rays, but anthropogenic radio frequency interference has historically prevented radio-based cosmic ray studies from being carried out independently from other types of detectors. We describe a cosmic ray detection system for the Owens Valley Radio Observatory Long Wavelength Array that searches for radio emission from cosmic ray air showers without relying on an external trigger, and runs alongside the other observing modes of the array. The OVRO-LWA, located in Eastern California, recently completed an expansion to 352 dual-polarization antennas and new signal processing infrastructure. In order to detect cosmic rays in the presence of radio frequency interference (RFI), initial event classification and RFI rejection is performed on Field Programmable Gate Array boards, which each process a sampled voltage timeseries from both polarizations of a subarray of 32 antennas. Each board uses dedicated RFI veto antennas outside the air shower radio footprint to reject RFI events. We present the trigger design, RFI flagging strategy, and candidate cosmic rays.

astro-ph.IM↗
arXiv2026

DeAR: Fine-Grained VLM Adaptation by Decomposing Attention Head Roles

Prompt learning is a dominant paradigm for adapting pre-trained Vision-Language Models (VLMs) to downstream tasks. However, existing methods often rely on a simplistic, layer-centric view, assuming shallow layers capture general features while deep layers handle task-specific knowledge. This assumption results in uncontrolled interactions between learnable tokens and original tokens. Task-specific knowledge could degrades the model's core generalization and creates a trade-off between task adaptation and the preservation of zero-shot generalization. To address this, we challenge the layer-centric view and propose \textbf{DeAR}, a framework that achieves fine-grained VLM adaptation by \textbf{De}composing \textbf{A}ttention head \textbf{R}oles. We posit that the functional specialization within VLMs occurs not between layers, but at the finer-grained level of individual attention heads in the deeper layers. Based on this insight, we introduce a novel metric, Concept Entropy, to systematically classify attention heads into distinct functional roles: \textit{Attribute}, \textit{Generalization}, and \textit{Mixed}. Guided by these roles, we introduce specialized attribute tokens and a Role-Based Attention Mask mechanism to precisely control information flow, ensuring generalization heads remain isolated from task-specific knowledge. We further incorporate a Task-Adaptive Fusion Strategy for inference. Extensive experiments on fifteen datasets show that DeAR achieves a strong balance between task adaptation and generalization, outperforming previous methods across various tasks.

cs.CV↗
arXiv2026

Looking Back and Forth: Cross-Image Attention Calibration and Attentive Preference Learning for Multi-Image Hallucination Mitigation

Although large vision-language models (LVLMs) have demonstrated remarkable capabilities, they are prone to hallucinations in multi-image tasks. We attribute this issue to limitations in existing attention mechanisms and insufficient cross-image modeling. Inspired by this, we propose a structured hallucination mitigation framework involving Cross-Image Attention calibration and Preference Learning (CAPL). CAPL explicitly enhances inter-image interactions at the architectural level while reinforcing reliance on genuine cross-image evidence during training, thereby improving the model's perception and modeling of cross-image associations. Specifically, we (i) introduce a selectable image token interaction attention mechanism to establish fine-grained cross-image entity alignment and information flow; (ii) design a cross-image modeling-based preference optimization strategy that contrasts reasoning outcomes under full inter-image interaction and those obtained when images are mutually invisible, encouraging the model to ground its predictions in authentic visual evidence and mitigating erroneous inferences driven by textual priors. Experimental results demonstrate that CAPL consistently improves performance across multiple model architectures, achieving stable gains on both multi-image hallucination and general benchmarks. Notably, performance on single-image visual tasks remains stable or slightly improves, indicating strong generalization capability.

cs.CV↗
arXiv2026

RealOSR: Latent Guidance Boosts Diffusion-based Real-world Omnidirectional Image Super-Resolutions

Omnidirectional image super-resolution (ODISR) aims to upscale low-resolution (LR) omnidirectional images (ODIs) to high-resolution (HR), catering to the growing demand for detailed visual content across a $ 180^{\circ}\times360^{\circ}$ viewport. Existing ODISR methods are limited by simplified degradation assumptions (e.g., bicubic downsampling), failing to model and exploit the real-world degradation information. Recent latent-based diffusion approaches using condition guidance suffer from slow inference due to their hundreds of updating steps and frequent use of VAE. To tackle these challenges, we propose \textbf{RealOSR}, a diffusion-based framework tailored for real-world ODISR, featuring efficient latent-based condition guidance within a one-step denoising paradigm. Central to efficient latent-based condition guidance is the proposed \textbf{Latent Gradient Alignment Routing (LaGAR)}, a lightweight module that enables effective pixel-latent space interactions and simulates gradient descent directly in the latent space, thereby leveraging the semantic richness and multi-scale features captured by the denoising UNet. Compared to the recent diffusion-based ODISR method, OmniSSR, RealOSR achieves significant improvements in visual quality and over \textbf{200$\times$} inference acceleration. Our code and models will be released upon acceptance.

eess.IV↗
arXiv2026

Improved Adversarial Diffusion Compression for Real-World Video Super-Resolution

While many diffusion models have achieved impressive results in real-world video super-resolution (Real-VSR) by generating rich and realistic details, their reliance on multi-step sampling leads to slow inference. One-step networks like SeedVR2, DOVE, and DLoRAL alleviate this through condensing generation into one single step, yet they remain heavy, with billions of parameters and multi-second latency. Recent adversarial diffusion compression (ADC) offers a promising path via pruning and distilling these models into a compact AdcSR network, but directly applying it to Real-VSR fails to balance spatial details and temporal consistency due to its lack of temporal awareness and the limitations of standard adversarial learning. To address these challenges, we propose an improved ADC method for Real-VSR. Our approach distills a large diffusion Transformer (DiT) teacher DOVE equipped with 3D spatio-temporal attentions, into a pruned 2D Stable Diffusion (SD)-based AdcSR backbone, augmented with lightweight 1D temporal convolutions, achieving significantly higher efficiency. In addition, we introduce a dual-head adversarial distillation scheme, in which discriminators in both pixel and feature domains explicitly disentangle the discrimination of details and consistency into two heads, enabling both objectives to be effectively optimized without sacrificing one for the other. Experiments demonstrate that the resulting compressed AdcVSR model reduces complexity by 95% in parameters and achieves an 8$\times$ acceleration over its DiT teacher DOVE, while maintaining competitive video quality and efficiency.

cs.CV↗
arXiv2026

Few Made It Out: A Multi-Messenger Study of an In Situ Solar Energetic Electron Event Driven by a Solar Jet

When in situ solar energetic electron (SEE) events are closely associated with nonthermal flares, the escaping electron population is frequently observed to be much smaller than the nonthermal-radiation-emitting population near the solar surface. If a single accelerated population drives both signatures, the physical mechanism causing this severe deficit of upward-propagating electrons remains poorly understood.Focusing on one of the 2022 November 10--12 SEE events associated with recurrent solar jets and interplanetary type III radio bursts, we present a new, combined microwave--X-ray analysis using the Expanded Owens Valley Solar Array (EOVSA) and the Spectrometer/Telescope for Imaging X-rays (STIX) aboard Solar Orbiter. This synergy enables, for the first time for such an event, spatially resolved diagnostics over a broad energy spectrum of the near-Sun energetic electrons, complemented by in situ measurements made by spacecraft at multiple heliocentric longitudes and distances. Consistent with earlier results based on in situ and X-ray data, our results show that only 0.1--1\% of energetic electrons escape into interplanetary space. Crucially, the new microwave spectral imaging analysis suggests that energetic electrons are strongly concentrated in a compact region just above a mini-flare arcade at the base of the jet spire, and that their number density decreases by at least two orders of magnitude in the direction of the jet spire away from this region. This steep gradient, revealed by the microwave diagnostics, points to efficient local acceleration and trapping in the region analogous to the above-the-looptop ``magnetic bottle'' region in major eruptive flares, allowing only a small fraction of electrons to access open magnetic field lines and enter interplanetary space.

astro-ph.SR↗
arXiv2026

GLM-5: from Vibe Coding to Agentic Engineering

We present GLM-5, a next-generation foundation model designed to transition the paradigm of vibe coding to agentic engineering. Building upon the agentic, reasoning, and coding (ARC) capabilities of its predecessor, GLM-5 adopts DSA to significantly reduce training and inference costs while maintaining long-context fidelity. To advance model alignment and autonomy, we implement a new asynchronous reinforcement learning infrastructure that drastically improves post-training efficiency by decoupling generation from training. Furthermore, we propose novel asynchronous agent RL algorithms that further improve RL quality, enabling the model to learn from complex, long-horizon interactions more effectively. Through these innovations, GLM-5 achieves state-of-the-art performance on major open benchmarks. Most critically, GLM-5 demonstrates unprecedented capability in real-world coding tasks, surpassing previous baselines in handling end-to-end software engineering challenges. Code, models, and more information are available at https://github.com/zai-org/GLM-5.

cs.LG↗