arXiv subjects

Yi Zhong

Publications and source records attributed to Yi Zhong.

At least 37 records · Page 2Linked to original sources

arXiv2026

Quasi-Simultaneous Broadband Spectral Energy Distributions of a Sample of Fermi Blazars -- I. Correlation Results

Blazars' non-thermal emission shows rapid variability across all wavelengths, so spectral energy distributions (SEDs) built from quasi-simultaneous data are crucial for revealing the jets physical properties. In this work, we construct quasi-simultaneous broadband SEDs for 93 Fermi blazars (56 FSRQs, 35 BL Lacs, and 2 blazar candidates of uncertain type), fit both peaks with cubic functions to allow for potential asymmetries, and examine correlations among key parameters. Our main results are summarized as follows: (1) We find that synchrotron peak frequency and curvature are only weakly related, suggesting that charged particles are accelerated by mixed acceleration mechanism. (2) The blazar sequence is confirmed in the observer's frame through negative correlations of both the bolometic luminosity $\log L_{\rm bol}$ and the Compton dominance $\log Y$ with the synchrotron peak frequency $\log ν_{\rm syn}^{\rm peak}$. After correcting for Doppler boosting, a weak positive correlation emerges between $\log L_{\rm bol}$ and $\log ν_{\rm syn}^{\rm peak}$. FSRQs and BL Lacs exhibit distinct correlation patterns within the blazar sequence, indicating differences in cooling mechanisms. (3) Using variability time lags between 0.1-1 GeV and 1-300 GeV light curves, we estimate lower limits of Doppler factors for 4 sources, providing a jet-speed diagnostic anchored directly to the $γ$-ray emission zone.

astro-ph.HE↗

arXiv2026

Breaking Modality Heterogeneity in Low-Bit Quantization for Large Vision-Language Models

Low-bit post-training quantization (PTQ) is a pivotal technique for deploying Vision-Language Models (VLMs) on resource-constrained devices. However, existing PTQ methods often degrade VLMs' accuracy due to the heterogeneous activation distributions of text and vision modalities during quantization. We find that this cross-modal heterogeneity is distributed unevenly across channels: a small subset of channels contains most modality-specific outliers, and these outliers typically reside in different channels for each modality. Motivated by this, we propose SplitQ, a channel-Splitting-driven post-training Quantization framework. At its core, SplitQ introduces a novel Modality-specific Outlier Channel Decoupling (MOCD) module that effectively isolates salient modality-specific outlier channels with minimal overhead. To further address the remaining cross-modal distribution discrepancies, we design an Adaptive Cross-Modal Calibration (ACC) module that employs dual lightweight learnable branches to dynamically mitigate modality-induced quantization errors. Extensive experiments on popular VLMs demonstrate that SplitQ significantly outperforms existing approaches across 6 popular multi-modal datasets under all evaluated quantization settings, including W4A8, W4A4, W3A3, and W3A2. Notably, SplitQ preserves 93.5% of FP16 performance under the challenging W3A3 setting (69.5 vs. 74.3), pushing the efficiency frontier for deploying advanced VLMs. Our code is available at https://github.com/EMVision-NK/SplitQ

cs.CV↗

arXiv2026

Safety Alignment as Continual Learning: Mitigating the Alignment Tax via Orthogonal Gradient Projection

Safety post-training can improve the harmfulness and policy compliance of Large Language Models (LLMs), but it may also reduce general utility, a phenomenon often described as the \emph{alignment tax}. We study this trade-off through the lens of continual learning: sequential alignment stages expose the model to shifted data distributions and objectives, and their gradients may interfere with directions that support previously acquired general capabilities. This view does not claim that all alignment degradation has a single cause; rather, it provides a useful first-order mechanism for mitigating one important source of capability regression. We propose \textbf{O}rthogonal \textbf{G}radient \textbf{P}rojection for \textbf{S}afety \textbf{A}lignment (\textbf{OGPSA}), a lightweight update rule that estimates a low-rank reference subspace from gradients on a small set of general-capability data and removes from each safety gradient the component lying in this subspace. The resulting update is the steepest local safety-descent direction subject to first-order preservation constraints on the reference objectives. OGPSA is compatible with standard post-training pipelines and avoids large-scale replay, although it introduces periodic reference-gradient computation. Across Supervised Fine-Tuning (SFT), Direct Preference Optimization (DPO), and sequential SFT$\rightarrow$DPO settings, OGPSA improves the observed safety--utility trade-off over standard baselines. Under the sequential SFT$\rightarrow$DPO pipeline, the average performance gain increases from 33.98\% to 42.74\% on Qwen2.5-7B-Instruct and from 19.74\% to 32.98\% on Llama3.1-8B-Instruct. We have open sourced our code at https://github.com/SunGL001/OGPSA.

cs.LG↗

arXiv2026

MePo: Meta Post-Refinement for Rehearsal-Free General Continual Learning

To cope with uncertain changes of the external world, intelligent systems must continually learn from complex, evolving environments and respond in real time. This ability, collectively known as general continual learning (GCL), encapsulates practical challenges such as online datastreams and blurry task boundaries. Although leveraging pretrained models (PTMs) has greatly advanced conventional continual learning (CL), these methods remain limited in reconciling the diverse and temporally mixed information along a single pass, resulting in sub-optimal GCL performance. Inspired by meta-plasticity and reconstructive memory in neuroscience, we introduce here an innovative approach named Meta Post-Refinement (MePo) for PTMs-based GCL. This approach constructs pseudo task sequences from pretraining data and develops a bi-level meta-learning paradigm to refine the pretrained backbone, which serves as a prolonged pretraining phase but greatly facilitates rapid adaptation of representation learning to downstream GCL tasks. MePo further initializes a meta covariance matrix as the reference geometry of pretrained representation space, enabling GCL to exploit second-order statistics for robust output alignment. MePo serves as a plug-in strategy that achieves significant performance gains across a variety of GCL benchmarks and pretrained checkpoints in a rehearsal-free manner (e.g., 15.10\%, 13.36\%, and 12.56\% on CIFAR-100, ImageNet-R, and CUB-200 under Sup-21/1K). Our source code is available at \href{https://github.com/SunGL001/MePo}{MePo}

cs.AI↗

arXiv2026

Entanglement probes of gravitational Kaluza-Klein spectra: signal hierarchy and model discrimination

Quantum-gravity-induced entanglement of masses (QGEM) provides a phase-sensitive probe of extra-dimensional corrections to the Newtonian potential at submillimeter separations. We compare three representative Kaluza-Klein spectral scenarios: the Randall-Sundrum II (RSII) and Arkani-Hamed-Dimopoulos-Dvali (ADD) models, and the case of a gapped continuum modeled by a Pöschl-Teller potential. We evaluate the entangling phase, concurrence, and normalized phase-response profiles over $d=40$-$80\,μ\mathrm{m}$ using representative benchmark parameters guided by current short-range gravity tests. In this range, the signal exhibits a stable hierarchy: ADD $>$ gapped $>$ RSII. For conservative experimental parameters, the ADD signal surpasses the nominal entanglement threshold at smaller separations, whereas the gapped benchmark is resolvable only at the lower end of the window, and RSII remains below resolution. In a more optimistic near-term scenario, all three spectral signatures comfortably exceed the threshold. We further show that normalized distance scans of the phase response clearly separate the RSII benchmark from the ADD and gapped cases, whereas ADD and the gapped continuum remain nearly indistinguishable in normalized profile. QGEM phase observables therefore provide a complementary discriminator of Kaluza-Klein spectral structure at submillimeter scales.

gr-qc↗

arXiv2026

FlyPrompt: Brain-Inspired Random-Expanded Routing with Temporal-Ensemble Experts for General Continual Learning

General continual learning (GCL) challenges intelligent systems to learn from single-pass, non-stationary data streams without clear task boundaries. While recent advances in continual parameter-efficient tuning (PET) of pretrained models show promise, they typically rely on multiple training epochs and explicit task cues, limiting their effectiveness in GCL scenarios. Moreover, existing methods often lack targeted design and fail to address two fundamental challenges in continual PET: how to allocate expert parameters to evolving data distributions, and how to improve their representational capacity under limited supervision. Inspired by the fruit fly's hierarchical memory system characterized by sparse expansion and modular ensembles, we propose FlyPrompt, a brain-inspired framework that decomposes GCL into two subproblems: expert routing and expert competence improvement. FlyPrompt introduces a randomly expanded analytic router for instance-level expert activation and a temporal ensemble of output heads to dynamically adapt decision boundaries over time. Extensive theoretical and empirical evaluations demonstrate FlyPrompt's superior performance, achieving up to 11.23%, 12.43%, and 7.62% gains over state-of-the-art baselines on CIFAR-100, ImageNet-R, and CUB-200, respectively. Our source code is available at https://github.com/AnAppleCore/FlyGCL.

cs.LG↗

arXiv2026

SpikCommander: A High-performance Spiking Transformer with Multi-view Learning for Efficient Speech Command Recognition

Spiking neural networks (SNNs) offer a promising path toward energy-efficient speech command recognition (SCR) by leveraging their event-driven processing paradigm. However, existing SNN-based SCR methods often struggle to capture rich temporal dependencies and contextual information from speech due to limited temporal modeling and binary spike-based representations. To address these challenges, we first introduce the multi-view spiking temporal-aware self-attention (MSTASA) module, which combines effective spiking temporal-aware attention with a multi-view learning framework to model complementary temporal dependencies in speech commands. Building on MSTASA, we further propose SpikCommander, a fully spike-driven transformer architecture that integrates MSTASA with a spiking contextual refinement channel MLP (SCR-MLP) to jointly enhance temporal context modeling and channel-wise feature integration. We evaluate our method on three benchmark datasets: the Spiking Heidelberg Dataset (SHD), the Spiking Speech Commands (SSC), and the Google Speech Commands V2 (GSC). Extensive experiments demonstrate that SpikCommander consistently outperforms state-of-the-art (SOTA) SNN approaches with fewer parameters under comparable time steps, highlighting its effectiveness and efficiency for robust speech command recognition.

cs.SD↗

arXiv2026

Energy-Based Cell Association in Nonuniform Renewable Energy-Powered Cellular Networks: Analysis and Optimization of Carbon Efficiency

The increasing global push for carbon reduction highlights the importance of integrating renewable energy into the supply chain of cellular networks. However, due to the stochastic nature of renewable energy generation and the uneven load distribution across base stations, the utilization rate of renewable energy remains low. To address these challenges, this paper investigates the trade-off between carbon emissions and downlink throughput in cellular networks, offering insights into optimizing both network performance and sustainability. The renewable energy state of base station batteries and the number of occupied channels are modeled as a quasi-birth-death process. We construct models for the probability of channel blocking, average successful transmission probability for users, downlink throughput, carbon emissions, and carbon efficiency based on stochastic geometry. Based on these analyses, an energy-based cell association scheme is proposed to optimize the carbon efficiency of cellular networks. The results show that, compared to the closest cell association scheme, the energy-based cell association scheme is capable of reducing the carbon emissions of the network by 13.0% and improving the carbon efficiency by 11.3%.

eess.SP↗

arXiv2025

PROMISE: Prompt-Attentive Hierarchical Contrastive Learning for Robust Cross-Modal Representation with Missing Modalities

Multimodal models integrating natural language and visual information have substantially improved generalization of representation models. However, their effectiveness significantly declines in real-world situations where certain modalities are missing or unavailable. This degradation primarily stems from inconsistent representation learning between complete multimodal data and incomplete modality scenarios. Existing approaches typically address missing modalities through relatively simplistic generation methods, yet these approaches fail to adequately preserve cross-modal consistency, leading to suboptimal performance. To overcome this limitation, we propose a novel multimodal framework named PROMISE, a PROMpting-Attentive HIerarchical ContraStive LEarning approach designed explicitly for robust cross-modal representation under conditions of missing modalities. Specifically, PROMISE innovatively incorporates multimodal prompt learning into a hierarchical contrastive learning framework, equipped with a specially designed prompt-attention mechanism. This mechanism dynamically generates robust and consistent representations for scenarios where particular modalities are absent, thereby effectively bridging the representational gap between complete and incomplete data. Extensive experiments conducted on benchmark datasets, along with comprehensive ablation studies, clearly demonstrate the superior performance of PROMISE compared to current state-of-the-art multimodal methods.

cs.CV↗

arXiv2025

Why Do Open-Source LLMs Struggle with Data Analysis? A Systematic Empirical Study

Large Language Models (LLMs) hold promise in automating data analysis tasks, yet open-source models face significant limitations in these kinds of reasoning-intensive scenarios. In this work, we investigate strategies to enhance the data analysis capabilities of open-source LLMs. By curating a seed dataset of diverse, realistic scenarios, we evaluate model behavior across three core dimensions: data understanding, code generation, and strategic planning. Our analysis reveals three key findings: (1) Strategic planning quality serves as the primary determinant of model performance; (2) Interaction design and task complexity significantly influence reasoning capabilities; (3) Data quality demonstrates a greater impact than diversity in achieving optimal performance. We leverage these insights to develop a data synthesis methodology, demonstrating significant improvements in open-source LLMs' analytical reasoning capabilities. Code is available at https://github.com/zjunlp/DataMind.

cs.CL↗

arXiv2025

Cross-link RTS/CTS for MLO mm-Wave WLANs

The directional RTS/CTS mechanism of mm-wave Wi-Fi hardly resolves the hidden terminal problem perfectly. This paper proposes cross-link RTS/CTS under multi-link operation (MLO) to address this problem and introduces a novel point process, named the generalized RTS/CTS hard-core process (G-HCP), to model the spatial transceiver relationships under the RTS/CTS mechanism, including the directional case and the omnidirectional case. Analytical expressions are derived for the intensity, the mean interference, an approximation of the success probability, and the expected number of hidden nodes for the directional RTS/CTS mechanism. Theoretical and numerical results demonstrate the performance difference between two RTS/CTS mechanisms. The cross-link RTS/CTS mechanism ensures higher link quality at the cost of reduced network throughput. In contrast, the directional RTS/CTS sacrifices the link quality for higher throughput. Our study reveals a fundamental trade-off between link reliability and network throughput, providing critical insights into the selection and optimization of RTS/CTS mechanisms in next-generation WLAN standards.

cs.NI↗

arXiv2025

S$^2$M-Former: Spiking Symmetric Mixing Branchformer for Brain Auditory Attention Detection

Auditory attention detection (AAD) aims to decode listeners' focus in complex auditory environments from electroencephalography (EEG) recordings, which is crucial for developing neuro-steered hearing devices. Despite recent advancements, EEG-based AAD remains hindered by the absence of synergistic frameworks that can fully leverage complementary EEG features under energy-efficiency constraints. We propose S$^2$M-Former, a novel spiking symmetric mixing framework to address this limitation through two key innovations: i) Presenting a spike-driven symmetric architecture composed of parallel spatial and frequency branches with mirrored modular design, leveraging biologically plausible token-channel mixers to enhance complementary learning across branches; ii) Introducing lightweight 1D token sequences to replace conventional 3D operations, reducing parameters by 14.7$\times$. The brain-inspired spiking architecture further reduces power consumption, achieving a 5.8$\times$ energy reduction compared to recent ANN methods, while also surpassing existing SNN baselines in terms of parameter efficiency and performance. Comprehensive experiments on three AAD benchmarks (KUL, DTU and AV-GC-AAD) across three settings (within-trial, cross-trial and cross-subject) demonstrate that S$^2$M-Former achieves comparable state-of-the-art (SOTA) decoding accuracy, making it a promising low-power, high-performance solution for AAD tasks. Code is available at https://github.com/JackieWang9811/S2M-Former.

cs.LG↗

arXiv2025

Analysis of SINR Coverage in LEO Satellite Networks through Spatial Network Calculus

We introduce a new analytical framework, developed based on the spatial network calculus, for performance assessment of Low Earth Orbit (LEO) satellite networks. Specifically, we model the satellites' spatial positions as a strong ball-regulated point process on the sphere. Under this model, proximal points in space exhibit a locally repulsive property, reflecting the fact that intersatellite links are protected by a safety distance and would not be arbitrarily close. Subsequently, we derive analytical lower bounds on the conditional coverage probabilities under Nakagami-$m$ and Rayleigh fading, respectively. These expressions have a low computational complexity, enabling efficient numerical evaluations. We validate the effectiveness of our theoretical model by contrasting the coverage probability obtained from our analysis with that estimated from a Starlink constellation. The results show that our analysis provides a tight lower bound on the actual value and, surprisingly, matches the empirical simulations almost perfectly with a 1 dB shift. This demonstrates our framework as an appropriate theoretical model for LEO satellite networks.

cs.IT↗

arXiv2025

Pattern-Aware Diffusion Synthesis of fMRI/dMRI with Tissue and Microstructural Refinement

Magnetic resonance imaging (MRI), especially functional MRI (fMRI) and diffusion MRI (dMRI), is essential for studying neurodegenerative diseases. However, missing modalities pose a major barrier to their clinical use. Although GAN- and diffusion model-based approaches have shown some promise in modality completion, they remain limited in fMRI-dMRI synthesis due to (1) significant BOLD vs. diffusion-weighted signal differences between fMRI and dMRI in time/gradient axis, and (2) inadequate integration of disease-related neuroanatomical patterns during generation. To address these challenges, we propose PDS, introducing two key innovations: (1) a pattern-aware dual-modal 3D diffusion framework for cross-modality learning, and (2) a tissue refinement network integrated with a efficient microstructure refinement to maintain structural fidelity and fine details. Evaluated on OASIS-3, ADNI, and in-house datasets, our method achieves state-of-the-art results, with PSNR/SSIM scores of 29.83 dB/90.84\% for fMRI synthesis (+1.54 dB/+4.12\% over baselines) and 30.00 dB/77.55\% for dMRI synthesis (+1.02 dB/+2.2\%). In clinical validation, the synthesized data show strong diagnostic performance, achieving 67.92\%/66.02\%/64.15\% accuracy (NC vs. MCI vs. AD) in hybrid real-synthetic experiments. Code is available in \href{https://github.com/SXR3015/PDS}{PDS GitHub Repository}

cs.CV↗

arXiv2025

BrainCSD: A Hierarchical Consistency-Driven MoE Foundation Model for Unified Connectome Synthesis and Multitask Brain Trait Prediction

Functional and structural connectivity (FC/SC) are key multimodal biomarkers for brain analysis, yet their clinical utility is hindered by costly acquisition, complex preprocessing, and frequent missing modalities. Existing foundation models either process single modalities or lack explicit mechanisms for cross-modal and cross-scale consistency. We propose BrainCSD, a hierarchical mixture-of-experts (MoE) foundation model that jointly synthesizes FC/SC biomarkers and supports downstream decoding tasks (diagnosis and prediction). BrainCSD features three neuroanatomically grounded components: (1) a ROI-specific MoE that aligns regional activations from canonical networks (e.g., DMN, FPN) with a global atlas via contrastive consistency; (2) a Encoding-Activation MOE that models dynamic cross-time/gradient dependencies in fMRI/dMRI; and (3) a network-aware refinement MoE that enforces structural priors and symmetry at individual and population levels. Evaluated on the datasets under complete and missing-modality settings, BrainCSD achieves SOTA results: 95.6\% accuracy for MCI vs. CN classification without FC, low synthesis error (FC RMSE: 0.038; SC RMSE: 0.006), brain age prediction (MAE: 4.04 years), and MMSE score estimation (MAE: 1.72 points). Code is available in \href{https://github.com/SXR3015/BrainCSD}{BrainCSD}

q-bio.NC↗

arXiv2025

Domain Generalizable Continual Learning

To adapt effectively to dynamic real-world environments, intelligent systems must continually acquire new skills while generalizing them to diverse, unseen scenarios. Here, we introduce a novel and realistic setting named domain generalizable continual learning (DGCL): a model learns sequential tasks with each involving a single domain, aiming to perform well across all encountered tasks and domains. This setting poses unique challenges in acquiring, retaining, and leveraging both semantic- and domain-relevant information for robust generalization. Although state-of-the-art continual learning (CL) methods have employed pre-trained models (PTMs) to enhance task-specific generalization, they typically assume identical training and testing domains for each task and therefore perform poorly in DGCL. To this end, we propose adaptive Domain Transformation (DoT), an innovative PTMs-based approach tailored to DGCL. Inspired by the distributed-plus-hub theory of the human brain, DoT disentangles semantic- and domain-relevant information in representation learning, and adaptively transforms task representations across various domains for output alignment, ensuring balanced and generalized predictions. DoT serves as a plug-in strategy that greatly facilitates state-of-the-art CL baselines under both full parameter tuning and parameter-efficient tuning paradigms in DGCL, validated by extensive experiments. Also, DoT is shown to accumulate domain-generalizable knowledge from DGCL, and ensure resource efficiency with a lightweight implementation.

cs.LG↗

arXiv2025

AutoAssert 1: A LoRA Fine-Tuned LLM Model for Efficient Automated Assertion Generation

As the complexity of software systems continues to increase, the demand for automated testing and maintenance tools is growing exponentially. To meet this urgent need, we propose a new assertion generation method based on Hardware Description Language (HDL). This method combines a lightweight, parameter-adjustable large language model (LLM) with the Unsloth platform to automatically generate test cases, thereby significantly reducing training costs without sacrificing accuracy or generalization performance. Empirical evaluation shows that our method can efficiently generate assertions that strictly conform to the hardware logic. This framework provides a robust and flexible solution to modern software testing and maintenance challenges. https://github.com/liusu-orange/AutoAssert-1 and https://gitee.com/OpenBPU/auto-assert1 are the locations of the source code.

cs.SE↗

arXiv2025

Efficient Preparation of Fermionic Superfluids in an Optical Dipole Trap through Reinforcement Learning

We demonstrate a reinforcement learning (RL) based control framework for optimizing evaporative cooling in the preparation of strongly interacting degenerate Fermi gases of Li6. Using a Soft Actor-Critic (SAC) algorithm, the system autonomously explores a high-dimensional parameter space to learn optimal cooling trajectories. Compared to conventional exponential ramps, our method achieves up to 130% improvement in atomic density within a 0.5 second, revealing non-trivial control strategies that balance fast evaporation and thermalization. While our current optimization focuses on the evaporation stage, future integration of other cooling stages, such as grey molasses cooling, could further extend RL to the full preparation pipeline. Our result highlights the promise of RL as a general tool for closed-loop quantum control and automated calibration in complex atomic physics experiments.

cond-mat.quant-gas↗