arXiv subjects

Yi Zhang

Publications and source records attributed to Yi Zhang.

At least 469 records · Page 26Linked to original sources

arXiv2025

Trustworthy Text-to-Image Diffusion Models: A Timely and Focused Survey

Text-to-Image (T2I) Diffusion Models (DMs) have garnered widespread attention for their impressive advancements in image generation. However, their growing popularity has raised ethical and social concerns related to key non-functional properties of trustworthiness, such as robustness, fairness, security, privacy, factuality, and explainability, similar to those in traditional deep learning (DL) tasks. Conventional approaches for studying trustworthiness in DL tasks often fall short due to the unique characteristics of T2I DMs, e.g., the multi-modal nature. Given the challenge, recent efforts have been made to develop new methods for investigating trustworthiness in T2I DMs via various means, including falsification, enhancement, verification \& validation and assessment. However, there is a notable lack of in-depth analysis concerning those non-functional properties and means. In this survey, we provide a timely and focused review of the literature on trustworthy T2I DMs, covering a concise-structured taxonomy from the perspectives of property, means, benchmarks and applications. Our review begins with an introduction to essential preliminaries of T2I DMs, and then we summarise key definitions/metrics specific to T2I tasks and analyses the means proposed in recent literature based on these definitions/metrics. Additionally, we review benchmarks and domain applications of T2I DMs. Finally, we highlight the gaps in current research, discuss the limitations of existing methods, and propose future research directions to advance the development of trustworthy T2I DMs. Furthermore, we keep up-to-date updates in this field to track the latest developments and maintain our GitHub repository at: https://github.com/wellzline/Trustworthy_T2I_DMs

cs.LG↗

arXiv2025

AlphaAlign: Incentivizing Safety Alignment with Extremely Simplified Reinforcement Learning

Large language models (LLMs), despite possessing latent safety understanding from their vast pretraining data, remain vulnerable to generating harmful content and exhibit issues such as over-refusal and utility degradation after safety alignment. Current safety alignment methods often result in superficial refusal shortcuts or rely on intensive supervision for reasoning-based approaches, failing to fully leverage the model's intrinsic safety self-awareness. We propose \textbf{AlphaAlign}, a simple yet effective pure reinforcement learning (RL) framework with verifiable safety reward designed to incentivize this latent safety awareness through proactive safety reasoning.} AlphaAlign employs a dual-reward system: a verifiable safety reward encourages correctly formatted and explicitly justified refusals for harmful queries while penalizing over-refusals, and a normalized helpfulness reward guides high-quality responses to benign inputs. This allows the model to develop proactive safety reasoning capabilities without depending on supervised safety-specific reasoning data. AlphaAlign demonstrates three key advantages: (1) Simplicity and efficiency, requiring only binary prompt safety labels and minimal RL steps for substantial improvements. (2) Breaking the safety-utility trade-off, by enhancing refusal of harmful content and reducing over-refusals, while simultaneously maintaining or even improving general task performance and robustness to unseen jailbreaks. (3) Deep alignment, fostering proactive safety reasoning that generates explicit safety rationales rather than relying on shallow refusal patterns.

cs.AI↗

arXiv2025

Survey of GenAI for Automotive Software Development: From Requirements to Executable Code

Adoption of state-of-art Generative Artificial Intelligence (GenAI) aims to revolutionize many industrial areas by reducing the amount of human intervention needed and effort for handling complex underlying processes. Automotive software development is considered to be a significant area for GenAI adoption, taking into account lengthy and expensive procedures, resulting from the amount of requirements and strict standardization. In this paper, we explore the adoption of GenAI for various steps of automotive software development, mainly focusing on requirements handling, compliance aspects and code generation. Three GenAI-related technologies are covered within the state-of-art: Large Language Models (LLMs), Retrieval Augmented Generation (RAG), Vision Language Models (VLMs), as well as overview of adopted prompting techniques in case of code generation. Additionally, we also derive a generalized GenAI-aided automotive software development workflow based on our findings from this literature review. Finally, we include a summary of a survey outcome, which was conducted among our automotive industry partners regarding the type of GenAI tools used for their daily work activities.

cs.SE↗

arXiv2025

Neural Architecture Search generated Phase Retrieval Net for Real-time Off-axis Quantitative Phase Imaging

In off-axis Quantitative Phase Imaging (QPI), artificial neural networks have been recently applied for phase retrieval with aberration compensation and phase unwrapping. However, the involved neural network architectures are largely unoptimized and inefficient with low inference speed, which hinders the realization of real-time imaging. Here, we propose a Neural Architecture Search (NAS) generated Phase Retrieval Net (NAS-PRNet) for accurate and fast phase retrieval. NAS-PRNet is an encoder-decoder style neural network, automatically found from a large neural network architecture search space through NAS. By modifying the differentiable NAS scheme from SparseMask, we learn the optimized skip connections through gradient descent. Specifically, we implement MobileNet-v2 as the encoder and define a synthesized loss that incorporates phase reconstruction loss and network sparsity loss. NAS-PRNet has achieved high-fidelity phase retrieval by achieving a peak Signal-to-Noise Ratio (PSNR) of 36.7 dB and a Structural SIMilarity (SSIM) of 86.6% as tested on interferograms of biological cells. Notably, NAS-PRNet achieves phase retrieval in only 31 ms, representing 15x speedup over the most recent Mamba-UNet with only a slightly lower phase retrieval accuracy.

eess.IV↗

arXiv2025

CASCADE Your Datasets for Cross-Mode Knowledge Retrieval of Language Models

Language models often struggle with cross-mode knowledge retrieval -- the ability to access knowledge learned in one format (mode) when queried in another. We demonstrate that models trained on multiple data sources (e.g., Wikipedia and TinyStories) exhibit significantly reduced accuracy when retrieving knowledge in a format different from its original training mode. This paper quantitatively investigates this phenomenon through a controlled study of random token sequence memorization across different modes. We first explore dataset rewriting as a solution, revealing that effective cross-mode retrieval requires prohibitively extensive rewriting efforts that follow a sigmoid-like relationship. As an alternative, we propose CASCADE, a novel pretraining algorithm that uses cascading datasets with varying sequence lengths and computing losses on only the second half of each training sequence to capture knowledge at different scales. Our experiments demonstrate that CASCADE outperforms dataset rewriting approaches, even when compressed into a single model with a unified loss function. This work provides both qualitative evidence of cross-mode retrieval limitations and a practical solution to enhance language models' ability to access knowledge independently of its presentational format.

cs.LG↗

arXiv2025

Compress Any Segment Anything Model (SAM)

Due to the excellent performance in yielding high-quality, zero-shot segmentation, Segment Anything Model (SAM) and its variants have been widely applied in diverse scenarios such as healthcare and intelligent manufacturing. Therefore, effectively compressing SAMs has become an increasingly pressing practical need. In this study, we propose Birkhoff, a novel data-free compression algorithm for SAM and its variants. Unlike quantization, pruning, distillation, and other compression methods, Birkhoff embodies versatility across model types, agility in deployment, faithfulness to the original model, and compactness in model size. Specifically, Birkhoff introduces a novel compression algorithm: Hyper-Compression, whose core principle is to find a dense trajectory to turn a high-dimensional parameter vector into a low-dimensional scalar. Furthermore, Birkhoff designs a dedicated linear layer operator, HyperLinear, to fuse decompression and matrix multiplication to significantly accelerate inference of the compressed SAMs. Extensive experiments on 18 SAMs in the COCO, LVIS, and SA-1B datasets show that Birkhoff performs consistently and competitively in compression time, compression ratio, post-compression performance, and inference speed. For example, Birkhoff can achieve a compression ratio of 5.17x on SAM2-B, with less than 1% performance drop without using any fine-tuning data. Moreover, the compression is finished within 60 seconds for all models.

cs.CV↗

arXiv2025

Electric-Field Reconstruction for Radio Detection of Inclined Air Showers in Three Polarizations

Accurate reconstruction of the electric field produced by extensive air showers is essential for the radio-detection technique, as the key parameters of interest of the primary particles that generated the showers are the amplitude, polarization, frequency spectrum, and energy fluence carried by the electric field at each receiving radio antenna. Conventional electric-field reconstruction methods primarily focus on antennas with two horizontal polarizations. In this work, we introduce an analytic $χ^2$ minimization method that is applicable to both two and three polarizations. This solution has been verified for simple and realistic antenna responses, with a particular focus on inclined air showers. Our method achieves standard deviations better than 4\% and 6\% for the estimation of the Hilbert peak envelope amplitude of the electric field and the energy fluence, respectively, with an antenna-response-dependent bias. Additionally, we have studied the dependence of the method with arrival direction showing that it has a good performance in the zenith range from 63$^\circ$ up to 80$^\circ$. This work also demonstrates that incorporating vertically polarized antennas enhances the precision of the reconstruction, leading to a more accurate and reliable electric-field estimation for inclined air showers.

astro-ph.IM↗

arXiv2025

Bridging Classical and Learning-based Iterative Registration through Deep Equilibrium Models

Deformable medical image registration is traditionally formulated as an optimization problem. While classical methods solve this problem iteratively, recent learning-based approaches use recurrent neural networks (RNNs) to mimic this process by unrolling the prediction of deformation fields in a fixed number of steps. However, classical methods typically converge after sufficient iterations, but learning-based unrolling methods lack a theoretical convergence guarantee and show instability empirically. In addition, unrolling methods have a practical bottleneck at training time: GPU memory usage grows linearly with the unrolling steps due to backpropagation through time (BPTT). To address both theoretical and practical challenges, we propose DEQReg, a novel registration framework based on Deep Equilibrium Models (DEQ), which formulates registration as an equilibrium-seeking problem, establishing a natural connection between classical optimization and learning-based unrolling methods. DEQReg maintains constant memory usage, enabling theoretically unlimited iteration steps. Through extensive evaluation on the public brain MRI and lung CT datasets, we show that DEQReg can achieve competitive registration performance, while substantially reducing memory consumption compared to state-of-the-art unrolling methods. We also reveal an intriguing phenomenon: the performance of existing unrolling methods first increases slightly then degrades irreversibly when the inference steps go beyond the training configuration. In contrast, DEQReg achieves stable convergence with its inbuilt equilibrium-seeking mechanism, bridging the gap between classical optimization-based and modern learning-based registration methods.

eess.IV↗

arXiv2025

Progress of the GRANDProto300 Project

GRANDProto300 (hereafter referred to as GP300) is a pioneering prototype array of the GRAND experiment. It consists of 300 radio antennas and will cover an area of 200 km$^2$ in a radio-quiet region of western China. Serving as a test bench for the GRAND experiment, GRANDProto300 aims to achieve autonomous radio detection and reconstruction of highly inclined air showers. It is designed to detect ultra-high-energy cosmic rays in the energy range of $10^{16.5}$-$10^{18}$ eV at a rate comparable to that of the Pierre Auger Observatory. Over the past two years, significant improvements have been made to both the hardware and firmware of GP300. Currently, 65 antenna units have been deployed at the site by June 2025. We present the current status of detector commissioning, including updates on hardware, calibration results such as GPS timing and antenna positioning. Additionally, we discuss the solar radio bursts associated with solar flares, the galactic radio emissions detected, and preliminary cosmic ray surveys.

astro-ph.HE↗

arXiv2025

Non-Hermitian Delocalization Induced by Residue Imaginary Velocity

The dichotomy of localization versus delocalization is a historic topic central to quantum and condensed matter physics. We discover a new delocalization mechanism attributed to a residue imaginary (part of) velocity $\operatorname{Im}(v)$, feasible for ground states or low-temperature states of non-Hermitian quantum systems under periodic boundary conditions. In sharp contrast to conventional formalisms through extended wavefunctions, these target systems exhibit delocalization in collective physical properties such as correlation and entanglement (of the Fermi Seas) despite sometimes localized left and right single-particle eigenstates, as we demonstrate numerically and generalize to scenarios with finite temperatures and interaction. Interestingly, disorder contributing to $\operatorname{Im}(v)$ may also allow strong-disorder delocalization. Thus, the nontrivial physics of $\operatorname{Im}(v)$ significantly enriches our understanding of delocalization and harbors interesting experiments and practical applications.

quant-ph↗

arXiv2025

Physics-Informed Neural ODEs for Temporal Dynamics Modeling in Cardiac T1 Mapping

Spin-lattice relaxation time ($T_1$) is an important biomarker in cardiac parametric mapping for characterizing myocardial tissue and diagnosing cardiomyopathies. Conventional Modified Look-Locker Inversion Recovery (MOLLI) acquires 11 breath-hold baseline images with interleaved rest periods to ensure mapping accuracy. However, prolonged scanning can be challenging for patients with poor breathholds, often leading to motion artifacts that degrade image quality. In addition, $T_1$ mapping requires voxel-wise nonlinear fitting to a signal recovery model involving an iterative estimation process. Recent studies have proposed deep-learning approaches for rapid $T_1$ mapping using shortened sequences to reduce acquisition time for patient comfort. Nevertheless, existing methods overlook important physics constraints, limiting interpretability and generalization. In this work, we present an accelerated, end-to-end $T_1$ mapping framework leveraging Physics-Informed Neural Ordinary Differential Equations (ODEs) to model temporal dynamics and address these challenges. Our method achieves high-accuracy $T_1$ estimation from a sparse subset of baseline images and ensures efficient null index estimation at test time. Specifically, we develop a continuous-time LSTM-ODE model to enable selective Look-Locker (LL) data acquisition with arbitrary time lags. Experimental results show superior performance in $T_1$ estimation for both native and post-contrast sequences and demonstrate the strong benefit of our physics-based formulation over direct data-driven $T_1$ priors.

eess.IV↗

arXiv2025

Test mass charge management in the detection of gravitational waves in space based on UV micro-LED

As an alternative to the ultraviolet light emitting diode(UV LED), the feasibility of utilizing UV micro-LED in the charge management in the detection of gravitational waves in space is experimentally studied. Compared with UV LED, micro-LED is more compact in size, has better current spreading, faster response time and longer operating life. Performance characteristics of micro-LEDs were measured, with peak wavelength of 254 nm, 262 nm, 274 nm, and 282 nm for each respective micro-LED, and the photoelectric effect was demonstrated. The effectiveness of micro-LED based charge management experiments were demonstrated using above micro-LEDs mounted on a cubical test mass, and different discharge rates were achieved by varying the drive current and duty cycle using pulse width modulation(PWM). Laboratory data was also shown to demonstrate the space qualification of the micro-LED device, the key electrical and optical characteristics of the micro-LEDs showed less than 5% variation. The results of the qualification bring the micro-LED device Technology Readiness Level(TRL) to TRL-5. TRL-6 will be reached provided additional radiation and thermal tests are conducted and in a position ready to be flown and further tested in space.

physics.ins-det↗

arXiv2025

Knowing You Don't Know: Learning When to Continue Search in Multi-round RAG through Self-Practicing

Retrieval Augmented Generation (RAG) has shown strong capability in enhancing language models' knowledge and reducing AI generative hallucinations, driving its widespread use. However, complex tasks requiring multi-round retrieval remain challenging, and early attempts tend to be overly optimistic without a good sense of self-skepticism. Current multi-round RAG systems may continue searching even when enough information has already been retrieved, or they may provide incorrect answers without having sufficient information or knowledge. Existing solutions either require large amounts of expensive human-labeled process supervision data or lead to subpar performance. This paper aims to address these limitations by introducing a new framework, SIM-RAG, to explicitly enhance RAG systems' self-awareness and multi-round retrieval capabilities. To train SIM-RAG, we first let a RAG system self-practice multi-round retrieval, augmenting existing question-answer pairs with intermediate inner monologue reasoning steps to generate synthetic training data. For each pair, the system may explore multiple retrieval paths, which are labeled as successful if they reach the correct answer and unsuccessful otherwise. Using this data, we train a lightweight information sufficiency Critic. At inference time, the Critic evaluates whether the RAG system has retrieved sufficient information at each round, guiding retrieval decisions and improving system-level self-awareness through in-context reinforcement learning. Experiments across multiple prominent RAG benchmarks show that SIM-RAG is an effective multi-round RAG solution. Furthermore, this framework is system-efficient, adding a lightweight component to RAG without requiring modifications to existing LLMs or search engines, and data-efficient, eliminating the need for costly human-annotated mid-step retrieval process supervision data.

cs.AI↗

arXiv2025

Phantom-Data : Towards a General Subject-Consistent Video Generation Dataset

Subject-to-video generation has witnessed substantial progress in recent years. However, existing models still face significant challenges in faithfully following textual instructions. This limitation, commonly known as the copy-paste problem, arises from the widely used in-pair training paradigm. This approach inherently entangles subject identity with background and contextual attributes by sampling reference images from the same scene as the target video. To address this issue, we introduce \textbf{Phantom-Data, the first general-purpose cross-pair subject-to-video consistency dataset}, containing approximately one million identity-consistent pairs across diverse categories. Our dataset is constructed via a three-stage pipeline: (1) a general and input-aligned subject detection module, (2) large-scale cross-context subject retrieval from more than 53 million videos and 3 billion images, and (3) prior-guided identity verification to ensure visual consistency under contextual variation. Comprehensive experiments show that training with Phantom-Data significantly improves prompt alignment and visual quality while preserving identity consistency on par with in-pair baselines.

cs.CV↗

arXiv2025

Learnable-Differentiable Finite Volume Solver for Accelerated Simulation of Flows

Simulation of fluid flows is crucial for modeling physical phenomena like meteorology, aerodynamics, and biomedicine. Classical numerical solvers often require fine spatiotemporal grids to satisfy stability, consistency, and convergence conditions, leading to substantial computational costs. Although machine learning has demonstrated better efficiency, they typically suffer from issues of interpretability, generalizability, and data dependency. Hence, we propose a learnable and differentiable finite volume solver, called LDSolver, designed for efficient and accurate simulation of fluid flows on spatiotemporal coarse grids. LDSolver comprises two key components: (1) a differentiable finite volume solver, and (2) an learnable module providing equivalent approximation for fluxes (derivatives and interpolations), and temporal error correction on coarse grids. Even with limited training data (e.g., only a few trajectories), our model could accelerate the simulation while maintaining a high accuracy with superior generalizability. Experiments on different flow systems (e.g., Burgers, decaying, forced and shear flows) show that LDSolver achieves state-of-the-art performance, surpassing baseline models with notable margins.

cs.LG↗

arXiv2025

From Collapse to Stability: A Knowledge-Driven Ensemble Framework for Scaling Up Click-Through Rate Prediction Models

Click-through rate (CTR) prediction plays a crucial role in modern recommender systems. While many existing methods utilize ensemble networks to improve CTR model performance, they typically restrict the ensemble to only two or three sub-networks. Whether increasing the number of sub-networks consistently enhances CTR model performance to align with scaling laws remains unclear. In this paper, we investigate larger ensemble networks and find three inherent limitations in commonly used ensemble methods: (1) performance degradation as the number of sub-networks increases; (2) sharp declines and high variance in sub-network performance; and (3) significant discrepancies between sub-network and ensemble predictions. Meanwhile, we analyze the underlying causes of these limitations from the perspective of dimensional collapse: the collapse within sub-networks becomes increasingly severe as the number of sub-networks grows, leading to a lower knowledge abundance. In this paper, we employ knowledge transfer methods, such as Knowledge Distillation (KD) and Deep Mutual Learning (DML), to address the aforementioned limitations. We find that KD enables CTR models to better follow scaling laws, while DML reduces variance among sub-networks and minimizes discrepancies with ensemble predictions. Furthermore, by combining KD and DML, we propose a model-agnostic and hyperparameter-free Knowledge-Driven Ensemble Framework (KDEF) for CTR Prediction.

cs.IR↗

arXiv2025

The Average Soft X-ray Spectra of eROSITA Active Galactic Nuclei

Context. AGNs are strong X-ray emitters shaped by disk-corona interactions. The soft excess (0.5-2.0 keV) reveals key information about the "warm corona" bridging the disk and hot corona. Yet, how this feature evolves with accretion properties remains poorly constrained, especially in large samples using spectral stacking. Aims. The eROSITA All-Sky Survey (eRASS:5) provides an unprecedented sample. We investigate how the average AGN X-ray spectra evolve with accretion parameters, and explore disk-corona connection by further combining stacked UV data. Methods. We developed Xstack, a novel tool that stacks rest-frame X-ray spectra and responses while preserving spectral shape through optimized weighting. We stack 17929 AGNs ("spec-z" sample, 23 Ms) with similar X-ray loudness alpha_ox, UV luminosity L_UV, and 4159 AGNs ("BH-mass" sample, 3 Ms) with similar Eddington ratio lambda_Edd and black hole mass M_BH. The resulting stacked X-ray spectra are analyzed with a phenomenological model. We further fit the stacked optical-UV-Xray SED with AGNSED model. Results. Soft excess strengthens strongly with alpha_ox and lambda_Edd (~5), while the hard X-ray spectral shape remains largely unchanged, supporting that soft excess is dominated by warm corona rather than reflection. AGNSED modeling reveals that warm corona radius (R_g units) generally increases with lambda_Edd and decreases with M_BH, or equivalently the disk-to-warm-corona transition consistently occurs near 1e4 K. The hot corona contracts with lambda_Edd and is unaffected by M_BH, aligning with disk evaporation predictions. Conclusions. The soft excess likely originates from a warm corona, with the disk to warm corona transition tied to hydrogen ionization near 1e4 K - supporting earlier eFEDS-HSC stacking results (Hagen et al. 2024). This study shows the strength of spectral stacking in probing AGN disk-corona physics.

astro-ph.HE↗

arXiv2025

Impact of the large-scale cosmic web on the X-ray emitting circumgalactic medium

The hot circumgalactic medium (CGM), probed by X-ray observations, plays a central role in understanding gas flows that drive a galaxy's evolution. While CGM properties have been widely studied, the influence of a galaxy's large-scale cosmic environment on the hot gas content remains less explored. We investigate how the large-scale cosmic web affects the X-ray surface brightness (XSB) profiles of galaxies in the context of cosmological simulations. We use our novel IllustrisTNG-based lightcone, spanning $0.03 \leq z \leq 0.3$, first developed in our previous work, and generate self-consistent mock X-ray observations, using intrinsic gas cell information. We apply the filament finder DisPerSE on the galaxy distributions to identify the cosmic filaments within the lightcone. We classify central galaxies into five distinct large-scale environment (LSE) categories: clusters and massive groups, cluster outskirts, filaments, filament-void transition regions, and voids/walls. We find that the X-ray surface brightness profiles (XSB) of central galaxies of dark matter halos in filaments with $M_{\rm 200m} >10^{12}\ M_\odot$ are X-ray brighter than those in voids and walls, with $20-45%$ deviations in the radial range of $\sim (0.3-0.5) R_{\rm 200m}$. We investigate the source of this enhancement and find that the filament galaxies show higher average gas densities, temperatures, and metallicities compared to voids/walls galaxies. Our results demonstrate that the impact of the large-scale cosmic environment is imprinted on the hot CGM's X-ray emission. Future theoretical work on studying the effect of assembly history, connectivity, and gas accretion on galaxies in filaments and voids would help to further our understanding of the impact of the environment on X-ray observations.

astro-ph.GA↗