arXiv subjects

Hongjun An

Publications and source records attributed to Hongjun An.

At least 19 recordsLinked to original sources

arXiv2026

Sequence Search: Automated Sequence Design using Neural Architecture Search

Developing an MR sequence is challenging and remains largely constrained by human intuition. Recently, AI-driven approaches have been proposed; however, most require an initial sequence for parameter optimization or extensive training datasets, limiting their general applicability. In this study, we propose "Sequence Search," an automated sequence design framework based on neural architecture search. The method takes tissue properties, imaging parameters, and design objectives as inputs and generates pulse sequences satisfying the design objectives, without requiring prior knowledge of conventional sequence structures. Sequence Search iteratively generates candidate sequences through neural architecture search and optimizes them via a differentiable Bloch simulator and objective-specific loss functions using gradient-based learning. The framework successfully replicated conventional spin-echo, T2-weighted spin-echo, and inversion recovery sequences. Less intuitive solutions were also discovered, such as three-RF spin-echo-like sequences with reduced RF energy and refocusing phases deviating from the conventional Hahn-echo. This work establishes a generalizable framework for automated MR sequence design, highlighting the potential to explore configurations beyond conventional designs based on human intuition.

cs.AI↗

arXiv2026

Silicon Bureaucracy and AI Test-Oriented Education: Contamination Sensitivity and Score Confidence in LLM Benchmarks

Public benchmarks increasingly govern how large language models (LLMs) are ranked, selected, and deployed. We frame this benchmark-centered regime as Silicon Bureaucracy and AI Test-Oriented Education, and argue that it rests on a fragile assumption: that benchmark scores directly reflect genuine generalization. In practice, however, such scores may conflate exam-oriented competence with principled capability, especially when contamination and semantic leakage are difficult to exclude from modern training pipelines. We therefore propose an audit framework for analyzing contamination sensitivity and score confidence in LLM benchmarks. Using a router-worker setup, we compare a clean-control condition with noisy conditions in which benchmark problems are systematically deleted, rewritten, and perturbed before being passed downstream. For a genuinely clean benchmark, noisy conditions should not consistently outperform the clean-control baseline. Yet across multiple models, we find widespread but heterogeneous above-baseline gains under noisy conditions, indicating that benchmark-related cues may be reassembled and can reactivate contamination-related memory. These results suggest that similar benchmark scores may carry substantially different levels of confidence. Rather than rejecting benchmarks altogether, we argue that benchmark-based evaluation should be supplemented with explicit audits of contamination sensitivity and score confidence.

cs.AI↗

arXiv2026

Ruyi2 Technical Report

Large Language Models (LLMs) face significant challenges regarding deployment costs and latency, necessitating adaptive computing strategies. Building upon the AI Flow framework, we introduce Ruyi2 as an evolution of our adaptive model series designed for efficient variable-depth computation. While early-exit architectures offer a viable efficiency-performance balance, the Ruyi model and existing methods often struggle with optimization complexity and compatibility with large-scale distributed training. To bridge this gap, Ruyi2 introduces a stable "Familial Model" based on Megatron-LM. By using 3D parallel training, it achieves a 2-3 times speedup over Ruyi, while performing comparably to same-sized Qwen3 models. These results confirm that family-based parameter sharing is a highly effective strategy, establishing a new "Train Once, Deploy Many" paradigm and providing a key reference for balancing architectural efficiency with high-performance capabilities.

cs.CL↗

arXiv2026

CreditAudit: 2$^\text{nd}$ Dimension for LLM Evaluation and Selection

Leaderboard scores on public benchmarks have been steadily rising and converging, with many frontier language models now separated by only marginal differences. However, these scores often fail to match users' day to day experience, because system prompts, output protocols, and interaction modes evolve under routine iteration, and in agentic multi step pipelines small protocol shifts can trigger disproportionate failures, leaving practitioners uncertain about which model to deploy. We propose CreditAudit, a deployment oriented credit audit framework that evaluates models under a family of semantically aligned and non adversarial system prompt templates across multiple benchmarks, reporting mean ability as average performance across scenarios and scenario induced fluctuation sigma as a stability risk signal, and further mapping volatility into interpretable credit grades from AAA to BBB via cross model quantiles with diagnostics that mitigate template difficulty drift. Controlled experiments on GPQA, TruthfulQA, and MMLU Pro show that models with similar mean ability can exhibit substantially different fluctuation, and stability risk can overturn prioritization decisions in agentic or high failure cost regimes. By providing a 2D and grade based language for regime specific selection, CreditAudit supports tiered deployment and more disciplined allocation of testing and monitoring effort, enabling more objective and trustworthy model evaluation for real world use.

cs.AI↗

arXiv2026

Single-Pixel Vision-Language Model for Intrinsic Privacy-Preserving Behavioral Intelligence

Adverse social interactions, such as bullying, harassment, and other illicit activities, pose significant threats to individual well-being and public safety, leaving profound impacts on physical and mental health. However, these critical events frequently occur in privacy-sensitive environments like restrooms, and changing rooms, where conventional surveillance is prohibited or severely restricted by stringent privacy regulations and ethical concerns. Here, we propose the Single-Pixel Vision-Language Model (SP-VLM), a novel framework that reimagines secure environmental monitoring. It achieves intrinsic privacy-by-design by capturing human dynamics through inherently low-dimensional single-pixel modalities and inferring complex behavioral patterns via seamless vision-language integration. Building on this framework, we demonstrate that single-pixel sensing intrinsically suppresses identity recoverability, rendering state-of-the-art face recognition systems ineffective below a critical sampling rate. We further show that SP-VLM can nonetheless extract meaningful behavioral semantics, enabling robust anomaly detection, people counting, and activity understanding from severely degraded single-pixel observations. Combining these findings, we identify a practical sampling-rate regime in which behavioral intelligence emerges while personal identity remains strongly protected. Together, these results point to a human-rights-aligned pathway for safety monitoring that can support timely intervention without normalizing intrusive surveillance in privacy-sensitive spaces.

cs.CV↗

arXiv2026

New Hard X-Ray and Multiwavelength Study of the PeVatron Candidate PWN G0.9+0.1 in the Galactic Center Region

We present a new X-ray study and multiwavelength spectral energy distribution (SED) modeling of the young pulsar wind nebula (PWN) powered by the energetic pulsar PSR J1747-2809, inside the composite supernova remnant (SNR) G0.9+0.1, located in the Galactic Center region. The source is detected by NuSTAR up to 30 keV with evidence for the synchrotron burnoff effect in the changing spatial morphology with increasing energy. The broadband 2-30 keV spectrum of PWN G0.9+0.1 is modeled by a single power law with photon index $\Gamma=2.11\pm0.07$. We combined the new X-ray data with the multiwavelength observations in radio, GeV, and TeV gamma rays and modeled the SED, applying a one-zone and a multi-zone leptonic model. The comparison of the models is successful, as we obtained physically compatible results in the two cases. Through the one-zone model, we constrain the age of the system to $\sim2.2$ kyr, as well as reproduce the observed PWN and SNR radio sizes. In both the one-zone and multi-zone leptonic models, the electron injection spectrum is well-described by a single power law with spectral index $p \sim 2.6$ and a maximum electron energy of $\sim2$ PeV, suggesting the source could be a leptonic PeVatron candidate. We estimate the average magnetic field to be $B_{\rm PWN} \sim 20\ \mu$G. We also report the serendipitous NuSTAR detection of renewed X-ray activity from the very faint X-ray transient XMMU J174716.1-281048 and characterize its spectrum.

astro-ph.HE↗

arXiv2026

Are LLMs Vulnerable to Preference-Undermining Attacks (PUA)? A Factorial Analysis Methodology for Diagnosing the Trade-off between Preference Alignment and Real-World Validity

Large Language Model (LLM) training often optimizes for preference alignment, rewarding outputs that are perceived as helpful and interaction-friendly. However, this preference-oriented objective can be exploited: manipulative prompts can steer responses toward user-appeasing agreement and away from truth-oriented correction. In this work, we investigate whether aligned models are vulnerable to Preference-Undermining Attacks (PUA), a class of manipulative prompting strategies designed to exploit the model's desire to please user preferences at the expense of truthfulness. We propose a diagnostic methodology that provides a finer-grained and more directive analysis than aggregate benchmark scores, using a factorial evaluation framework to decompose prompt-induced shifts into interpretable effects of system objectives (truth- vs. preference-oriented) and PUA-style dialogue factors (directive control, personal derogation, conditional approval, reality denial) within a controlled $2 \times 2^4$ design. Surprisingly, more advanced models are sometimes more susceptible to manipulative prompts. Beyond the dominant reality-denial factor, we observe model-specific sign reversals and interactions with PUA-style factors, suggesting tailored defenses rather than uniform robustness. These findings offer a novel, reproducible factorial evaluation methodology that provides finer-grained diagnostics for post-training processes like RLHF, enabling better trade-offs in the product iteration of LLMs by offering a more nuanced understanding of preference alignment risks and the impact of manipulative prompts.

cs.CR↗

arXiv2025

Theoretical Foundations of Scaling Law in Familial Models

Neural scaling laws have become foundational for optimizing large language model (LLM) training, yet they typically assume a single dense model output. This limitation effectively overlooks "Familial models, a transformative paradigm essential for realizing ubiquitous intelligence across heterogeneous device-edge-cloud hierarchies. Transcending static architectures, familial models integrate early exits with relay-style inference to spawn G deployable sub-models from a single shared backbone. In this work, we theoretically and empirically extend the scaling law to capture this "one-run, many-models" paradigm by introducing Granularity (G) as a fundamental scaling variable alongside model size (N) and training tokens (D). To rigorously quantify this relationship, we propose a unified functional form L(N, D, G) and parameterize it using large-scale empirical runs. Specifically, we employ a rigorous IsoFLOP experimental design to strictly isolate architectural impact from computational scale. Across fixed budgets, we systematically sweep model sizes (N) and granularities (G) while dynamically adjusting tokens (D). This approach effectively decouples the marginal cost of granularity from the benefits of scale, ensuring high-fidelity parameterization of our unified scaling law. Our results reveal that the granularity penalty follows a multiplicative power law with an extremely small exponent. Theoretically, this bridges fixed-compute training with dynamic architectures. Practically, it validates the "train once, deploy many" paradigm, demonstrating that deployment flexibility is achievable without compromising the compute-optimality of dense baselines.

cs.LG↗

arXiv2025

The Advanced X-ray Imaging Satellite (AXIS) Community Science Book

The AXIS Community Science Book represents the collective effort of 592 scientists worldwide to define the transformative science enabled by the Advanced X-ray Imaging Satellite (AXIS), a next-generation X-ray mission selected by NASA's Astrophysics Probe Program for Phase A study. AXIS will advance the legacy of high-angular-resolution X-ray astronomy with ~1.5'' imaging over a wide 24' field of view and an order of magnitude greater collecting area than Chandra in the 0.3-12 keV band. Combining sharp imaging, high throughput, and rapid response capabilities, AXIS will open new windows on virtually every aspect of modern astrophysics, exploring the birth and growth of supermassive black holes, the feedback processes that shape galaxies, the life cycles of stars and exoplanet environments, and the nature of compact stellar remnants, supernova remnants, and explosive transients. This book compiles 138 community-contributed science cases developed by five Science Working Groups focused on AGN and supermassive black holes, galaxy evolution and feedback, compact objects and supernova remnants, stellar physics and exoplanets, and time-domain and multi-messenger astrophysics. Together, these studies establish the scientific foundation for next-generation X-ray exploration in the 2030s and highlight strong synergies with facilities of the 2030s, such as JWST, Roman, Rubin/LSST, SKA, ALMA, ngVLA, and next-generation gravitational-wave and neutrino networks.

astro-ph.HE↗

arXiv2025

X-ray studies of PSR J1838$-$0655 and its wind nebula associated with HESS J1837$-$069 and 1LHAASO J1837$-$0654u

We analyzed X-ray data from Chandra, XMM-Newton, NICER, and NuSTAR to characterize the properties of the pulsar PSR J1838$-$0655 and its pulsar wind nebula (PWN) associated with HESS J1837$-$069. Based on 5.5 years of NICER monitoring, we detected a glitch around MJD 59300, characterized by a fractional frequency jump of approximately $2\times 10^{-6}$. We constructed semi-phase-coherent timing solutions for pre- and post-glitch epochs, allowing for phase alignment of multi-instrument data and a subsequent measurement of the pulsed spectrum of the pulsar. This analysis confirmed previously-reported spectral curvature and revealed a peak energy of $73^{+85}_{-26}$ keV in the pulsar's spectral energy distribution (SED), based on a logpar model fit of the pulsed spectrum. We discuss these findings within the framework of pulsar magnetospheric emission scenarios. The PWN's X-ray spectrum is well-described by a power law with a photon index of $2.1\pm0.3$, softer than previously-reported measurements. We also characterized the X-ray emission from another extended X-ray source AX J1837.3$-$0652 within the extent of HESS J1837$-$069. Based on the spatial and spectral properties of these X-ray sources, we propose a leptonic emission scenario for HESS J1837$-$069 and demonstrate its feasibility through SED modeling. Finally, we discuss the implications of our model results and alternative scenarios for the gamma-ray emission.

astro-ph.HE↗

arXiv2025

Multi-wavelength Study of HESS J0632+057: New Insights into Pulsar-Disk Interaction

We present an analysis of new multi-wavelength observations of the TeV gamma-ray binary HESS J0632+057, conducted using SALT, Swift, NuSTAR, and VERITAS in 2023--2024. By combining these new data with archival observations, we confirm previous suggestions of orbital variability in the source's X-ray spectrum, including increased X-ray absorption at the orbital phase interval of $\phi\approx0.3\textrm{--}0.4$. The source's X-ray flux within this phase interval seems to have exhibited a significant change on an orbital timescale. Additionally, occasional short-term variations in the X-ray band on a timescale of less than 3 days have been observed. The measured duration of the increased absorbing column density and the flux variability timescales can provide clues about the interaction between the putative pulsar and the Be companion's disk if, as previously suggested, the pulsar crosses the disk at this phase interval. Moreover, the new contemporaneous X-ray and TeV observations around the pulsar-crossing phases revealed independent variability in the X-ray and TeV fluxes, contrary to a previous observation of concurrent flux increases. While these observations alone cannot provide definitive conclusions, we discuss our results in the context of pulsar-disk interaction and intrabinary shock emission scenarios.

astro-ph.HE↗

arXiv2025

AI Flow: Perspectives, Scenarios, and Approaches

Pioneered by the foundational information theory by Claude Shannon and the visionary framework of machine intelligence by Alan Turing, the convergent evolution of information and communication technologies (IT/CT) has created an unbroken wave of connectivity and computation. This synergy has sparked a technological revolution, now reaching its peak with large artificial intelligence (AI) models that are reshaping industries and redefining human-machine collaboration. However, the realization of ubiquitous intelligence faces considerable challenges due to substantial resource consumption in large models and high communication bandwidth demands. To address these challenges, AI Flow has been introduced as a multidisciplinary framework that integrates cutting-edge IT and CT advancements, with a particular emphasis on the following three key points. First, device-edge-cloud framework serves as the foundation, which integrates end devices, edge servers, and cloud clusters to optimize scalability and efficiency for low-latency model inference. Second, we introduce the concept of familial models, which refers to a series of different-sized models with aligned hidden features, enabling effective collaboration and the flexibility to adapt to varying resource constraints and dynamic scenarios. Third, connectivity- and interaction-based intelligence emergence is a novel paradigm of AI Flow. By leveraging communication networks to enhance connectivity, the collaboration among AI models across heterogeneous nodes achieves emergent intelligence that surpasses the capability of any single model. The innovations of AI Flow provide enhanced intelligence, timely responsiveness, and ubiquitous accessibility to AI services, paving the way for the tighter fusion of AI techniques and communication systems.

cs.AI↗

arXiv2025

Large model enhanced computational ghost imaging

Ghost imaging (GI) achieves 2D image reconstruction through high-order correlation of 1D bucket signals and 2D light field information, particularly demonstrating enhanced detection sensitivity and high-quality image reconstruction via efficient photon collection in scattering media. Recent investigations have established that deep learning (DL) can substantially enhance the ghost imaging reconstruction quality. Furthermore, with the emergence of large models like SDXL, GPT-4, etc., the constraints of conventional DL in parameters and architecture have been transcended, enabling models to comprehensively explore relationships among all distinct positions within feature sequences. This paradigm shift has significantly advanced the capability of DL in restoring severely degraded and low-resolution imagery, making it particularly advantageous for noise-robust image reconstruction in GI applications. In this paper, we propose the first large imaging model with 1.4 billion parameters that incorporates the physical principles of GI (GILM). The proposed GILM implements a skip connection mechanism to mitigate gradient explosion challenges inherent in deep architectures, ensuring sufficient parametric capacity to capture intricate correlations among object single-pixel measurements. Moreover, GILM leverages multi-head attention mechanism to learn spatial dependencies across pixel points during image reconstruction, facilitating the extraction of comprehensive object information for subsequent reconstruction. We validated the effectiveness of GILM through a series of experiments, including simulated object imaging, imaging objects in free space, and imaging object located 52 meters away in underwater environment. The experimental results show that GILM effectively analyzes the fluctuation trends of the collected signals, thereby optimizing the recovery of the object's image from the acquired data.

eess.IV↗

arXiv2024

Investigation of the non-thermal X-ray emission from the supernova remnant CTB 37B hosting the magnetar CXOU J171405.7$-$381031

We present a detailed X-ray investigation of a region (S1) exhibiting non-thermal X-ray emission within the supernova remnant (SNR) CTB 37B hosting the magnetar CXOU J171405.7$-$381031. Previous analyses modeled this emission with a power law (PL), inferring various values for the photon index ($\Gamma$) and absorbing column density ($N_{\rm H}$). Based on these, S1 was suggested to be the SNR shell, a background pulsar wind nebula (PWN), or an interaction region between the SNR and a molecular cloud. Our analysis of a larger dataset favors a steepening (broken or curved PL) spectrum over a straight PL, with the best-fit broken power-law (BPL) parameters of $\Gamma=1.23\pm0.23$ and $2.24\pm0.16$ below and above a break at $5.57\pm0.52$ keV, respectively. However, a simple PL or srcut model cannot be definitively ruled out. For the BPL model, the inferred $N_{\rm H}=(4.08\pm0.72)\times 10^{22}\rm \ cm^{-2}$ towards S1 is consistent with that of the SNR, suggesting a physical association. The BPL-inferred spectral break $\Delta \Gamma \approx 1$ and hard $\Gamma$ can be naturally explained by a non-thermal bremsstrahlung (NTB) model. We present an evolutionary NTB model that reproduces the observed spectrum, which indicates the presence of sub-relativistic electrons within S1. However, alternate explanations for S1, an unrelated PWN or the SNR shock with unusually efficient acceleration, cannot be ruled out. We discuss these explanations and their implications for gamma-ray emission from CTB 37B, and describe future observations that could settle the origin of S1.

astro-ph.HE↗

arXiv2024

Revisiting the Intrabinary Shock Model for Millisecond Pulsar Binaries: Radiative Losses and Long-Term Variability

Spectrally hard X-ray emission with double-peak light curves (LCs) and orbitally modulated gamma rays have been observed in some millisecond pulsar binaries, phenomena attributed to intrabinary shocks (IBSs). While the existing IBS model by Sim, An, and Wadiasingh (2024) successfully explains these high-energy features observed in three pulsar binaries, it neglects particle energy loss within the shock region. We refine this IBS model to incorporate radiative losses of X-ray emitting electrons and positrons, and verify that the losses have insignificant impact on the observed LCs and spectra of the three binaries. Applying our refined IBS model to the X-ray bright pulsar binary PSR J1723-2837, we predict that it can be detected by the Cherenkov Telescope Array. Additionally, we propose that the long-term X-ray variability observed in XSS J12270-4859 and PSR J1723-2837 is due to changes in the shape of their IBSs. Our modeling of the X-ray variability suggests that these IBS shape changes may alter the extinction of the companion's optical emission, potentially explaining the simultaneous optical and X-ray variability observed in XSS J12270-4859. We present the model results and discuss their implications.

astro-ph.HE↗

arXiv2024

Physics in Next-token Prediction

We discovered the underlying physics in Next-token Prediction (NTP). We identified the law of information conservation within NTP and proposed the First Law of Information Capacity (IC-1), demonstrating that the essence of intelligence emergence in auto-regressive models is fundamentally a process of information transfer. We also introduced Landauer's Principle into NTP, formulating the Second Law of Information Capacity (IC-2), which establishes the relationship between auto-regressive model training and energy consumption. Additionally, we presented several corollaries, which hold practical significance for production practices. Finally, we demonstrate the consistency between the Law of Information Capacity and the Scaling Law for Neural Language Models, the Knowledge Capacity Scaling Laws, and the Scaling Laws for Precision.

cs.LG↗

arXiv2024

SentenceVAE: Enable Next-sentence Prediction for Large Language Models with Faster Speed, Higher Accuracy and Longer Context

Current large language models (LLMs) primarily utilize next-token prediction method for inference, which significantly impedes their processing speed. In this paper, we introduce a novel inference methodology termed next-sentence prediction, aiming at enhancing the inference efficiency of LLMs. We present Sentence Variational Autoencoder (SentenceVAE), which includes a Sentence Encoder to compress multiple tokens in a sentence into a single token, and a Sentence Decoder to reconstruct it. By integrating SentenceVAE into the input and output layers of LLMs, we develop Sentence-level LLMs (SLLMs) that employ a sentence-by-sentence inference method. In addition, the SentenceVAE module of SLLMs can maintain the integrity of the original semantic content by segmenting the context into sentences, thereby improving accuracy while boosting inference speed. Moreover, compared to previous LLMs, SLLMs process fewer tokens over equivalent context length, significantly reducing memory demands for self-attention computation and facilitating the handling of longer context. Extensive experiments on Wanjuan dataset have revealed that the proposed method can accelerate inference speed by 204~365%, reduce perplexity (PPL) to 46~75% of its original metric, and decrease memory overhead by 86~91% for the equivalent context length, compared to previous token-by-token methods.

cs.AI↗

arXiv2024

StreakNet-Arch: An Anti-scattering Network-based Architecture for Underwater Carrier LiDAR-Radar Imaging

In this paper, we introduce StreakNet-Arch, a real-time, end-to-end binary-classification framework based on our self-developed Underwater Carrier LiDAR-Radar (UCLR) that embeds Self-Attention and our novel Double Branch Cross Attention (DBC-Attention) to enhance scatter suppression. Under controlled water tank validation conditions, StreakNet-Arch with Self-Attention or DBC-Attention outperforms traditional bandpass filtering and achieves higher $F_1$ scores than learning-based MP networks and CNNs at comparable model size and complexity. Real-time benchmarks on an NVIDIA RTX 3060 show a constant Average Imaging Time (54 to 84 ms) regardless of frame count, versus a linear increase (58 to 1,257 ms) for conventional methods. To facilitate further research, we contribute a publicly available streak-tube camera image dataset contains 2,695,168 real-world underwater 3D point cloud data. More importantly, we validate our UCLR system in a South China Sea trial, reaching an error of 46mm for 3D target at 1,000 m depth and 20 m range. Source code and data are available at https://github.com/BestAnHongjun/StreakNet .

cs.CV↗