Searcharxiv⌕ Search

arXiv subjects

Jun Dong

Publications and source records attributed to Jun Dong.

At least 19 recordsLinked to original sources

DuoFlow: Towards Synchronized Speech and Motion Generation with Fine-Grained Control

An embodied speaker's expressiveness depends not only on what is said, but also on how the voice sounds and how the body moves. However, existing joint speech-motion generation models are largely driven implicitly by transcript semantics or intermediate speech representations, offering little explicit control over motion hierarchy, body parts, or vocal style. To bridge this gap, we construct MotionSpeech, a structured benchmark for controllable joint speech-motion generation that includes transcripts, voice-style descriptions, and motion conditions at two hierarchical levels. Building on this benchmark, we propose DuoFlow, a dual-stream latent flow transformer that jointly denoises speech and motion under a shared noise level and a unified second-based temporal coordinate. Modality-specific residual connectors inject structured conditions into the corresponding streams, layer-wise bidirectional cross-modal attention coordinates their generation, and a monotonic phoneme pathway combines normalized progress anchors with CTC supervision to strengthen speech temporal modeling. DuoFlow provides a unified interface for fine-grained motion and vocal-style control and demonstrates the ability to synthesize high-quality, synchronized, and instruction-consistent speech-motion sequences. Project page, https://jundong-dev.github.io/DuoFlow-project-page.

cs.CV↗

OptiSAR-Net++: A Large-Scale Benchmark and Transformer-Free Framework for Cross-Domain Remote Sensing Visual Grounding

Remote sensing visual grounding (RSVG) aims to localize specific targets in remote sensing images using natural language expressions. However, existing methods are restricted to single-sensor domains, i.e., either optical or synthetic aperture radar (SAR), limiting their real-world applicability. In this paper, we introduce the Cross-Domain RSVG (CD-RSVG) task and construct OptSAR-RSVG, the first large-scale benchmark dataset for this setting. To tackle the challenges of cross-domain feature modeling, computational inefficiency, and fine-grained semantic discrimination, we propose OptiSAR-Net++. Our framework features a patch-level Low-Rank Adaptation Mixture of Experts (PL-MoE) for efficient cross-domain feature decoupling. To mitigate the substantial computational overhead of Transformer decoding frameworks, we adopt a CLIP-based contrastive paradigm and further incorporate dynamic adversarial negative sampling, thereby transforming generative regression into an efficient cross-modal matching process. Additionally, a text-guided dual-gate fusion module (TGDF-SSA) and a region-aware auxiliary head are introduced to enhance semantic-visual alignment and spatial modeling. Extensive experiments demonstrate that OptiSAR-Net++ achieves SOTA performance on both OptSAR-RSVG and DIOR-RSVG benchmarks, offering significant advantages in localization accuracy and efficiency. The model and dataset have been made publicly available at https://github.com/JunDong-dev/OptiSAR-Net-PlusPlus.

cs.CV↗

SignGPT: Toward LLM-Mediated Sign Language Interaction through Gloss-Free Translation and Generation

Large language models (LLMs) provide limited support for sign language interaction. Unifying sign language translation (SLT) and generation (SLG) to enable sign language as both input and output can reduce switching between separate models during sign-text interaction. We present SignGPT, a unified, pose-based framework for gloss-free SLT and SLG. SignGPT integrates part-aware hierarchical representations of body, hand, and facial motion into a shared language model and employs asymmetric multi-token prediction and progressive training for bidirectional modeling. We evaluate SignGPT on How2Sign (ASL) and Phoenix-2014T (DGS) through benchmark comparisons, qualitative analyses, and component ablations. An exploratory study with 12 Deaf ASL signers assesses an LLM-mediated sign-to-sign response pipeline, highlighting the potential of unified modeling to support sign language conversation (SLC).

cs.CV↗

Accelerated Convergence of a Second-Order Dynamical System and its Application to Splitting Algorithms for Comonotone Inclusions

This paper introduces a novel second-order dynamical system driven by a forward-backward splitting operator for solving the structured inclusion $0\in(A+B)(x)$ in a real Hilbert space, where $A$ is a maximal $ρ$-comonotone operator and $B$ is a $ν$-cocoercive operator. The well-posedness of the system is established, and Lyapunov analysis yields accelerated convergence rates of order $o(\frac{1}{t})$ for the velocity and $o(\frac{1}{t^2})$ for the forward-backward residual, together with weak convergence of the trajectories to $\operatorname{zer}(A+B)$. Temporal discretization further leads to a class of double inertial Halpern forward-backward splitting algorithms that encompasses the classical forward-backward splitting method and its inertial variants as special cases. Numerical experiments on split feasibility, sparse signal recovery, and image deblurring illustrate the effectiveness of the proposed algorithm.

math.OC↗

MM-BrowseComp: A Comprehensive Benchmark for Multimodal Browsing Agents

AI agents with advanced reasoning and tool-use capabilities have demonstrated impressive performance in web browsing for deep search. However, existing benchmarks such as BrowseComp primarily focus on textual content, overlooking the prevalence of multimodal content. To bridge this gap, we introduce MM-BrowseComp, a novel benchmark comprising 400 challenging, hand-crafted questions designed to evaluate multimodal retrieval and reasoning capabilities. Unlike prior work, MM-BrowseComp incorporates visual prompts and necessitates the extraction of key evidence from web images and videos to complete questions, rendering text-only approaches insufficient. Additionally, we provide a verified checklist for each question, enabling fine-grained analysis of multimodal dependencies and reasoning paths. Our comprehensive evaluation of 27 state-of-the-art models reveals that even leading models like GPT-5-High with tools achieve only 24.25\% accuracy, highlighting the suboptimal multimodal browsing capabilities, establishing MM-BrowseComp as a rigorous new standard for the field.

cs.CL↗

Residual group-like symmetries in selection rules without group actions

We analyze loop-induced group-like symmetries in theories where fields are labeled by basis elements of a fusion algebra constructed from the conjugacy classes of finite groups. Although the fusion rules for conjugacy classes are in general violated at loop level, residual group-like symmetries, including both Abelian and non-Abelian ones, remain exact through a procedure referred to as ``groupification''. By examining various conjugacy classes of finite groups realized in heterotic string theory on non-Abelian orbifolds, we identify an approximate discrete symmetry that controls the magnitude of loop-induced couplings. As a result, most parameters appearing in non-invertible selection rules are natural in the sense of 't Hooft. Furthermore, we discuss anomalies of the groupification symmetry, which can impose additional constraints on models with non-invertible fusion rules.

hep-th↗

Yukawa Textures with Enhanced Symmetries in Heterotic Calabi-Yau Compactifications

We clarify the structure of Yukawa couplings and mass matrices for matter fields in heterotic string theory on smooth Calabi-Yau threefolds with standard embedding. The topological structure of Calabi-Yau threefolds leads to interesting Yukawa textures that cannot be derived from group-theoretical symmetries, e.g., the so-called Weinberg texture in the case of two generations of matter fields. Furthermore, we find that a $U(2)$ flavor symmetry, which plays an important role in controlling higher-dimensional operators in the Standard Model effective field theory, emerges at specific loci in the moduli space of multi-Higgs fields. Small perturbations around these loci generate semi-realistic patterns of quark masses and mixings.

hep-th↗

From Code Foundation Models to Agents and Applications: A Comprehensive Survey and Practical Guide to Code Intelligence

Large language models (LLMs) have fundamentally transformed automated software development by enabling direct translation of natural language descriptions into functional code, driving commercial adoption through tools like Github Copilot (Microsoft), Cursor (Anysphere), Trae (ByteDance), and Claude Code (Anthropic). While the field has evolved dramatically from rule-based systems to Transformer-based architectures, achieving performance improvements from single-digit to over 95\% success rates on benchmarks like HumanEval. In this work, we provide a comprehensive synthesis and practical guide (a series of analytic and probing experiments) about code LLMs, systematically examining the complete model life cycle from data curation to post-training through advanced prompting paradigms, code pre-training, supervised fine-tuning, reinforcement learning, and autonomous coding agents. We analyze the code capability of the general LLMs (GPT-4, Claude, LLaMA) and code-specialized LLMs (StarCoder, Code LLaMA, DeepSeek-Coder, and QwenCoder), critically examining the techniques, design decisions, and trade-offs. Further, we articulate the research-practice gap between academic research (e.g., benchmarks and tasks) and real-world deployment (e.g., software-related code tasks), including code correctness, security, contextual awareness of large codebases, and integration with development workflows, and map promising research directions to practical needs. Last, we conduct a series of experiments to provide a comprehensive analysis of code pre-training, supervised fine-tuning, and reinforcement learning, covering scaling law, framework selection, hyperparameter sensitivity, model architectures, and dataset comparisons.

cs.SE↗

Coupling Selection Rules in Heterotic Calabi-Yau Compactifications

We study coupling selection rules of chiral matter fields in heterotic string theory with standard embedding. These selection rules are determined by topological properties of Calabi-Yau threefolds. We classify coupling selection rules on complete intersection Calabi-Yau threefolds for $h^{1,1}\leq 5$. It is found that all of these selection rules for $h^{1,1}\leq 5$ are understood by combinations of only five types of fusion rules.

hep-th↗

Iterative Distortion Cancellation Algorithms for Single-Sideband Systems

We propose an iterative distortion cancellation algorithm to digitally mitigate the impact of double-sideband dither signal amplitude from the automatic bias control module on Kramers-Kronig receivers without modifying physical layer structures. The algorithm utilizes the KK relation for initial signal decisions and reconstructs the distortion caused by dither signals. Experimental tests in back-to-back showed it improved tolerance to dither amplitudes up to 10% Vπ. For 80-km fiber transmission, the algorithm increased the receiver sensitivity by more than 1 dB, confirming the effectiveness of the proposed distortion cancellation method.

eess.SP↗

On discrete gauging and non-invertible selection rules

We clarify selection rules of conjugacy classes of several finite discrete groups where we deal with both gauged and ungauged cases. We find that the selection rules enjoy finite Abelian or non-Abelian discrete symmetries originating from the inner and/or outer automorphism of underlying discrete groups. Since the selection rules of conjugacy classes do not obey conventional group-like selection rules, they open up new coupling selection rules of fields which are labeled by the conjugacy classes.

hep-th↗

Real-Time AIoT for AAV Antenna Interference Detection via Edge-Cloud Collaboration

In the fifth-generation (5G) era, eliminating communication interference sources is crucial for maintaining network performance. Interference often originates from unauthorized or malfunctioning antennas, and radio monitoring agencies must address numerous sources of such antennas annually. Unmanned aerial vehicles (UAVs) can improve inspection efficiency. However, the data transmission delay in the existing cloud-only (CO) artificial intelligence (AI) mode fails to meet the low latency requirements for real-time performance. Therefore, we propose a computer vision-based AI of Things (AIoT) system to detect antenna interference sources for UAVs. The system adopts an optimized edge-cloud collaboration (ECC+) mode, combining a keyframe selection algorithm (KSA), focusing on reducing end-to-end latency (E2EL) and ensuring reliable data transmission, which aligns with the core principles of ultra-reliable low-latency communication (URLLC). At the core of our approach is an end-to-end antenna localization scheme based on the tracking-by-detection (TBD) paradigm, including a detector (EdgeAnt) and a tracker (AntSort). EdgeAnt achieves state-of-the-art (SOTA) performance with a mean average precision (mAP) of 42.1% on our custom antenna interference source dataset, requiring only 3 million parameters and 14.7 GFLOPs. On the COCO dataset, EdgeAnt achieves 38.9% mAP with 5.4 GFLOPs. We deployed EdgeAnt on Jetson Xavier NX (TRT) and Raspberry Pi 4B (NCNN), achieving real-time inference speeds of 21.1 (1088) and 4.8 (640) frames per second (FPS), respectively. Compared with CO mode, the ECC+ mode reduces E2EL by 88.9%, increases accuracy by 28.2%. Additionally, the system offers excellent scalability for coordinated multiple UAVs inspections. The detector code is publicly available at https://github.com/SCNU-RISLAB/EdgeAnt.

eess.SP↗

You Sense Only Once Beneath: Ultra-Light Real-Time Underwater Object Detection

Despite the remarkable achievements in object detection, the model's accuracy and efficiency still require further improvement under challenging underwater conditions, such as low image quality and limited computational resources. To address this, we propose an Ultra-Light Real-Time Underwater Object Detection framework, You Sense Only Once Beneath (YSOOB). Specifically, we utilize a Multi-Spectrum Wavelet Encoder (MSWE) to perform frequency-domain encoding on the input image, minimizing the semantic loss caused by underwater optical color distortion. Furthermore, we revisit the unique characteristics of even-sized and transposed convolutions, allowing the model to dynamically select and enhance key information during the resampling process, thereby improving its generalization ability. Finally, we eliminate model redundancy through a simple yet effective channel compression and reconstructed large kernel convolution (RLKC) to achieve model lightweight. As a result, forms a high-performance underwater object detector YSOOB with only 1.2 million parameters. Extensive experimental results demonstrate that, with the fewest parameters, YSOOB achieves mAP50 of 83.1% and 82.9% on the URPC2020 and DUO datasets, respectively, comparable to the current SOTA detectors. The inference speed reaches 781.3 FPS and 57.8 FPS on the T4 GPU (TensorRT FP16) and the edge computing device Jetson Xavier NX (TensorRT FP16), surpassing YOLOv12-N by 28.1% and 22.5%, respectively.

cs.CV↗

Ultrasound Nodule Segmentation Using Asymmetric Learning with Simple Clinical Annotation

Recent advances in deep learning have greatly facilitated the automated segmentation of ultrasound images, which is essential for nodule morphological analysis. Nevertheless, most existing methods depend on extensive and precise annotations by domain experts, which are labor-intensive and time-consuming. In this study, we suggest using simple aspect ratio annotations directly from ultrasound clinical diagnoses for automated nodule segmentation. Especially, an asymmetric learning framework is developed by extending the aspect ratio annotations with two types of pseudo labels, i.e., conservative labels and radical labels, to train two asymmetric segmentation networks simultaneously. Subsequently, a conservative-radical-balance strategy (CRBS) strategy is proposed to complementally combine radical and conservative labels. An inconsistency-aware dynamically mixed pseudo-labels supervision (IDMPS) module is introduced to address the challenges of over-segmentation and under-segmentation caused by the two types of labels. To further leverage the spatial prior knowledge provided by clinical annotations, we also present a novel loss function namely the clinical anatomy prior loss. Extensive experiments on two clinically collected ultrasound datasets (thyroid and breast) demonstrate the superior performance of our proposed method, which can achieve comparable and even better performance than fully supervised methods using ground truth annotations.

cs.CV↗

Integration of a new Cryogenic Liquefier into the IB-1 Cryogenic Test Facility

The increase over the last years of the testing activities related to quantum systems, SRF cavities for the PIP-II and the LCLS-II projects, as well as superconducting magnets for the HL-LHC project and Fusion research activities, has required the addition of a new Helium cryogenic plant into the existing IB-1 Industrial Cryogenic Test Facility. The new cryogenic plant is composed of a cryogenic liquefier (Cold Box) able to provide up to 340 L/h, a 4 kL Dewar and two Mycom compressors providing up to 120 g/s. AL-AT (Air Liquide Advanced Technologies) has taken part of this project by designing and manufacturing the cryogenic liquefier. This new cryogenic plant is connected through a cryogenic distribution system to a 10 kL Dewar, which is part of the existing cryogenic test facility, itself composed of another Cold Box and a Sullair compressor. The new cryogenic plant has two main operating modes: one allows to transfer liquid helium at 1.7 bar between the two Dewars, the other allows to transfer supercritical Helium at 2 bar or more between the new Cold Box and the 10 kL Dewar. The entire industrial cryogenic facility is handled by a common Inventory Control System, composed of three control valves, and 9 tanks giving a total buffer volume of more than 1000 m3. This paper presents the technical features of the new Helium cryogenic plant, as well as the main results of the liquefier commissioning phase and details of the helium transfer between the two Dewars, making the connection between the cryogenic plants at the IB-1 Industrial Cryogenic Test Facility.

physics.acc-ph↗

Direct Imaging of Contacts and Forces in Colloidal Gels

Colloidal dispersions are prized as model systems to understand basic properties of materials, and are central to a wide range of industries from cosmetics to foods to agrichemicals. Among the key developments in using colloids to address challenges in condensed matter is to resolve the particle coordinates in 3D, allowing a level of analysis usually only possible in computer simulation. However in amorphous materials, relating mechanical properties to microscopic structure remains problematic. This makes it rather hard to understand, for example, mechanical failure. Here we address this challenge by studying the contacts and the forces between particles, as well as their positions. To do so, we use a colloidal model system (an emulsion) in which the interparticle forces and local stress can be linked to the microscopic structure. We demonstrate the potential of our method to reveal insights into the failure mechanisms of soft amorphous solids by determining local stress in a colloidal gel. In particular, we identify `force chains' of load-bearing droplets, and local stress anisotropy, and investigate their connection with locally rigid packings of the droplets.

cond-mat.soft↗

Observation of orbital ordering and origin of the nematic order in FeSe

To elucidate the origin of nematic order in FeSe, we performed field-dependent 77Se-NMR measurements on single crystals of FeSe. We observed orbital ordering from the splitting of the NMR spectra and Knight shift and a suppression of it with magnetic field B0 up to 16 T applied parallel to the Fe-planes. There is a significant change in the distribution and magnitude of the internal magnetic field across the orbital ordering temperature Torb while stripe-type antiferromagnetism is absent. Giant antiferromagnetic (AFM) spin fluctuations measured by the NMR spin-lattice relaxation are gradually developed starting at ~ 40 K, which is far below the nematic ordering temperature Tnem. These results demonstrate that orbital ordering is the origin of the nematic order, and the AFM spin fluctuation is the driving mechanism of superconductivity in FeSe under the presence of the nematic order.

cond-mat.supr-con↗

Opposed flow focusing: evidence of a second order jetting transition

We propose a novel microfluidic "opposed-flow" geometry in which the continuous fluid phase is fed into a junction in a direction opposite the dispersed phase. This pulls out the dispersed phase into a micron-sized jet, which decays into micron-sized droplets. As the driving pressure is tuned to a critical value, the jet radius vanishes as a power law down to sizes below 1 $μ$m. By contrast, the conventional "coflowing" junction leads to a first order jetting transition, in which the jet disappears at a finite radius of several $μ$m, to give way to a "dripping" state, resulting in much larger droplets. We demonstrate the effectiveness of our method by producing the first microfluidic silicone oil emulsions with a sub micron particle radius, and utilize these droplets to produce colloidal clusters.

cond-mat.soft↗