arXiv subjects

Xinyu Zhang

Publications and source records attributed to Xinyu Zhang.

At least 217 records · Page 12Linked to original sources

arXiv2025

SimpleAR: Pushing the Frontier of Autoregressive Visual Generation through Pretraining, SFT, and RL

This work presents SimpleAR, a vanilla autoregressive visual generation framework without complex architecure modifications. Through careful exploration of training and inference optimization, we demonstrate that: 1) with only 0.5B parameters, our model can generate 1024x1024 resolution images with high fidelity, and achieve competitive results on challenging text-to-image benchmarks, e.g., 0.59 on GenEval and 79.66 on DPG; 2) both supervised fine-tuning (SFT) and Group Relative Policy Optimization (GRPO) training could lead to significant improvements on generation aesthectics and prompt alignment; and 3) when optimized with inference acceleraton techniques like vLLM, the time for SimpleAR to generate an 1024x1024 image could be reduced to around 14 seconds. By sharing these findings and open-sourcing the code, we hope to reveal the potential of autoregressive visual generation and encourage more participation in this research field. Code is available at https://github.com/wdrink/SimpleAR.

cs.CV↗

arXiv2025

Challenges in interpretability of additive models

We review generalized additive models as a type of ``transparent'' model that has recently seen renewed interest in the deep learning community as neural additive models. We highlight multiple types of nonidentifiability in this model class and discuss challenges in interpretability, arguing for restraint when claiming ``interpretability'' or ``suitability for safety-critical applications'' of such models.

cs.LG↗

arXiv2025

Thermodynamics of Chemical Reactions in Open Systems Reduced by PEA and QSSA

Previous studies have primarily focused on the nonequilibrium thermodynamics of chemical reaction networks (CRNs) occurring in closed systems. In contrast, CRNs in open systems exhibit much richer nonequilibrium phenomena due to sustained matter and energy exchange. Here, we bridge the quantitative relationships between essential thermodynamic quantities -- including the steady state, enthalpy, intrinsic Gibbs free energy, and entropy production rate -- in original mass-action equations and their PEA- or QSSA-reduced counterparts for open CRNs. Our analysis demonstrates that the thermodynamic structure, especially the second law of thermodynamics, of the full CRNs may not be preserved in reduced models when algebraic relations are imposed. Specifically, PEA-reduced models lose monotonicity in the intrinsic Gibbs free energy, whereas QSSA retains this property. These theoretical findings are further validated through analytical and numerical studies of two archetypal open systems: the Michaelis-Menten reactions and the phosphorylation-dephosphorylation cycle (PdPC). Our results provide a systematic framework for evaluating the fidelity of reduced models.

physics.chem-ph↗

arXiv2025

EveGuard: Defeating Vibration-based Side-Channel Eavesdropping with Audio Adversarial Perturbations

Vibrometry-based side channels pose a significant privacy risk, exploiting sensors like mmWave radars, light sensors, and accelerometers to detect vibrations from sound sources or proximate objects, enabling speech eavesdropping. Despite various proposed defenses, these involve costly hardware solutions with inherent physical limitations. This paper presents EveGuard, a software-driven defense framework that creates adversarial audio, protecting voice privacy from side channels without compromising human perception. We leverage the distinct sensing capabilities of side channels and traditional microphones, where side channels capture vibrations and microphones record changes in air pressure, resulting in different frequency responses. EveGuard first proposes a perturbation generator model (PGM) that effectively suppresses sensor-based eavesdropping while maintaining high audio quality. Second, to enable end-to-end training of PGM, we introduce a new domain translation task called Eve-GAN for inferring an eavesdropped signal from a given audio. We further apply few-shot learning to mitigate the data collection overhead for Eve-GAN training. Our extensive experiments show that EveGuard achieves a protection rate of more than 97 percent from audio classifiers and significantly hinders eavesdropped audio reconstruction. We further validate the performance of EveGuard across three adaptive attack mechanisms. We have conducted a user study to verify the perceptual quality of our perturbed audio.

cs.CR↗

arXiv2025

Active Sampling and Gaussian Reconstruction for Radio Frequency Radiance Field

Radio-frequency (RF) Radiance Field reconstruction is a challenging problem. The difficulty lies in the interactions between the propagating signal and objects, such as reflections and diffraction, which are hard to model precisely, especially when the shapes and materials of the objects are unknown. Previously, a neural network-based method was proposed to reconstruct the RF Radiance Field, showing promising results. However, this neural network-based method has some limitations: it requires a large number of samples for training and is computationally expensive. Additionally, the neural network only provides the predicted mean of the RF Radiance Field and does not offer an uncertainty model. In this work, we propose a training-free Gaussian reconstruction method for RF Radiance Field. Our method demonstrates that the required number of samples is significantly smaller compared to the neural network-based approach. Furthermore, we introduce an uncertainty model that provides confidence estimates for predictions at any selected position in the scene. We also combine the Gaussian reconstruction method with active sampling, which further reduces the number of samples needed to achieve the same performance. Finally, we explore the potential benefits of our method in a quasi-dynamic setting, showcasing its ability to adapt to changes in the scene without requiring the entire process to be repeated.

eess.SP↗

arXiv2025

Debiased Offline Representation Learning for Fast Online Adaptation in Non-stationary Dynamics

Developing policies that can adjust to non-stationary environments is essential for real-world reinforcement learning applications. However, learning such adaptable policies in offline settings, with only a limited set of pre-collected trajectories, presents significant challenges. A key difficulty arises because the limited offline data makes it hard for the context encoder to differentiate between changes in the environment dynamics and shifts in the behavior policy, often leading to context misassociations. To address this issue, we introduce a novel approach called Debiased Offline Representation for fast online Adaptation (DORA). DORA incorporates an information bottleneck principle that maximizes mutual information between the dynamics encoding and the environmental data, while minimizing mutual information between the dynamics encoding and the actions of the behavior policy. We present a practical implementation of DORA, leveraging tractable bounds of the information bottleneck principle. Our experimental evaluation across six benchmark MuJoCo tasks with variable parameters demonstrates that DORA not only achieves a more precise dynamics encoding but also significantly outperforms existing baselines in terms of performance.

cs.LG↗

arXiv2025

Autoregressive Action Sequence Learning for Robotic Manipulation

Designing a universal policy architecture that performs well across diverse robots and task configurations remains a key challenge. In this work, we address this by representing robot actions as sequential data and generating actions through autoregressive sequence modeling. Existing autoregressive architectures generate end-effector waypoints sequentially as word tokens in language modeling, which are limited to low-frequency control tasks. Unlike language, robot actions are heterogeneous and often include continuous values -- such as joint positions, 2D pixel coordinates, and end-effector poses -- which are not easily suited for language-based modeling. Based on this insight, we introduce a straightforward enhancement: we extend causal transformers' single-token prediction to support predicting a variable number of tokens in a single step through our Chunking Causal Transformer (CCT). This enhancement enables robust performance across diverse tasks of various control frequencies, greater efficiency by having fewer autoregression steps, and lead to a hybrid action sequence design by mixing different types of actions and using a different chunk size for each action type. Based on CCT, we propose the Autoregressive Policy (ARP) architecture, which solves manipulation tasks by generating hybrid action sequences. We evaluate ARP across diverse robotic manipulation environments, including Push-T, ALOHA, and RLBench, and show that ARP, as a universal architecture, matches or outperforms the environment-specific state-of-the-art in all tested benchmarks, while being more efficient in computation and parameter sizes. Videos of our real robot demonstrations, all source code and the pretrained models of ARP can be found at http://github.com/mlzxy/arp.

cs.RO↗

arXiv2025

Are Images Indistinguishable to Humans Also Indistinguishable to Classifiers?

The ultimate goal of generative models is to perfectly capture the data distribution. For image generation, common metrics of visual quality (e.g., FID) and the perceived truthfulness of generated images seem to suggest that we are nearing this goal. However, through distribution classification tasks, we reveal that, from the perspective of neural network-based classifiers, even advanced diffusion models are still far from this goal. Specifically, classifiers are able to consistently and effortlessly distinguish real images from generated ones across various settings. Moreover, we uncover an intriguing discrepancy: classifiers can easily differentiate between diffusion models with comparable performance (e.g., U-ViT-H vs. DiT-XL), but struggle to distinguish between models within the same family but of different scales (e.g., EDM2-XS vs. EDM2-XXL). Our methodology carries several important implications. First, it naturally serves as a diagnostic tool for diffusion models by analyzing specific features of generated data. Second, it sheds light on the model autophagy disorder and offers insights into the use of generated data: augmenting real data with generated data is more effective than replacing it. Third, classifier guidance can significantly enhance the realism of generated images.

cs.CV↗

arXiv2025

CP-AgentNet: Autonomous and Explainable Communication Protocol Design Using Generative Agents

Although DRL (deep reinforcement learning) has emerged as a powerful tool for making better decisions than existing hand-crafted communication protocols, it faces significant limitations: 1) Selecting the appropriate neural network architecture and setting hyperparameters are crucial for achieving desired performance levels, requiring domain expertise. 2) The decision-making process in DRL models is often opaque, commonly described as a 'black box.' 3) DRL models are data hungry. In response, we propose CP-AgentNet, the first framework designed to use generative agents for developing communication network protocols. This approach addresses these challenges by creating an autonomous system for protocol design, significantly reducing human effort. We developed LLMA (LLM-agents-based multiple access) and CPTCP (CP-Agent-based TCP) for heterogeneous environments. Our comprehensive simulations have demonstrated the efficient coexistence of LLMA and CPTCP with nodes using different types of protocols, as well as enhanced explainability.

cs.NI↗

arXiv2025

Non-equilibrium Capacitance Loss in EDLC Based on Ion Flux

Understanding the loss of non-equilibrium capacitance in the electric double layer capacitor (EDLC) during continuous charging compared to equilibrium capacitance is crucial for the design of high-rate energy storage devices. Due to three major challenges, including strong ion correlations induced by high concentrations at the interface, the difficulty in representing far from equilibrium, and numerical stability issues, current numerical simulations of ion distribution based on statistical mechanics struggle to achieve long-term accurate modeling under realistic parameters. Since non-equilibrium capacitance loss is a universal macroscopic phenomenon in EDLC, which exists independently of the microscopic model of the electrolyte, this study starts from macroscopic non-equilibrium phenomena, introduces a novel perspective centered on ion flux (IF) for the first time. By deriving the relationship among the charging current, IF, and displacement current, a novel capacitance model based on IF is first established. It is further demonstrated that the current relationship and voltage relationship of EDLC, when represented by conventional circuit models, exhibit inherent contradictions. Therefore, a new interpretative framework is proposed, wherein IF and displacement current are analogized to acceleration and deformation in an elastic medium, respectively. Inspired by this insight, this study reveals that the decrease in non-equilibrium capacitance with increasing charging current originates from the inherent resistance of IF to external charging. This tendency to restore equilibrium is the fundamental cause of non-equilibrium capacitance loss.

cond-mat.mtrl-sci↗

arXiv2025

Seedream 2.0: A Native Chinese-English Bilingual Image Generation Foundation Model

Rapid advancement of diffusion models has catalyzed remarkable progress in the field of image generation. However, prevalent models such as Flux, SD3.5 and Midjourney, still grapple with issues like model bias, limited text rendering capabilities, and insufficient understanding of Chinese cultural nuances. To address these limitations, we present Seedream 2.0, a native Chinese-English bilingual image generation foundation model that excels across diverse dimensions, which adeptly manages text prompt in both Chinese and English, supporting bilingual image generation and text rendering. We develop a powerful data system that facilitates knowledge integration, and a caption system that balances the accuracy and richness for image description. Particularly, Seedream is integrated with a self-developed bilingual large language model as a text encoder, allowing it to learn native knowledge directly from massive data. This enable it to generate high-fidelity images with accurate cultural nuances and aesthetic expressions described in either Chinese or English. Beside, Glyph-Aligned ByT5 is applied for flexible character-level text rendering, while a Scaled ROPE generalizes well to untrained resolutions. Multi-phase post-training optimizations, including SFT and RLHF iterations, further improve the overall capability. Through extensive experimentation, we demonstrate that Seedream 2.0 achieves state-of-the-art performance across multiple aspects, including prompt-following, aesthetics, text rendering, and structural correctness. Furthermore, Seedream 2.0 has been optimized through multiple RLHF iterations to closely align its output with human preferences, as revealed by its outstanding ELO score. In addition, it can be readily adapted to an instruction-based image editing model, such as SeedEdit, with strong editing capability that balances instruction-following and image consistency.

cs.CV↗

arXiv2025

Huge Stress-induced Adiabatic Temperature Change in a High-Toughness All-d-metal Heusler Alloy

The elastocaloric effect (eCE), referring to the thermal effect triggered by a uniaxial stress, provides a promising and versatile routine for green and high efficient thermal management. However, current eCE materials generally suffer from relatively low eCE and poor mechanical properties, hindering their practical applications. Here, we report a exceptionally huge eCE with a directly measured adiabatic temperature change of up to 57.2 K in a dual-phase all-d-metal Heusler Mn50Ni37.5Ti12.5 polycrystalline alloy, revealing an extra contribution to the latent heat during the stress-induced martensitic transformation from B2 to L10, and breaking the record of adiabatic temperature change for elastocaloric alloys. Moreover, thanks to the combined strengthening effect of d-d hybridization and well-dispersed secondary cubic γ phase, the alloy can endure a uniaxial stress up to 1760 MPa. Such an abnormal huge eCE is attributed to the combination of the enhanced entropy change associated with a stress-induced B2 to L10 martensitic transformation under higher stress, in contrast with the thermally induced B2 to 5-layer modulated structure one, and the high transformation fraction due to the multi-point nucleation facilitated by the γ phase dispersed in the main phase. This work provides insight into making full use of the transformation heat to enhance the caloric effect for high-efficient thermal management systems.

cond-mat.mtrl-sci↗

arXiv2025

PABBO: Preferential Amortized Black-Box Optimization

Preferential Bayesian Optimization (PBO) is a sample-efficient method to learn latent user utilities from preferential feedback over a pair of designs. It relies on a statistical surrogate model for the latent function, usually a Gaussian process, and an acquisition strategy to select the next candidate pair to get user feedback on. Due to the non-conjugacy of the associated likelihood, every PBO step requires a significant amount of computations with various approximate inference techniques. This computational overhead is incompatible with the way humans interact with computers, hindering the use of PBO in real-world cases. Building on the recent advances of amortized BO, we propose to circumvent this issue by fully amortizing PBO, meta-learning both the surrogate and the acquisition function. Our method comprises a novel transformer neural process architecture, trained using reinforcement learning and tailored auxiliary losses. On a benchmark composed of synthetic and real-world datasets, our method is several orders of magnitude faster than the usual Gaussian process-based strategies and often outperforms them in accuracy.

stat.ML↗

arXiv2025

A novel and efficient parameter estimation of the Lognormal-Rician turbulence model based on k-Nearest Neighbor and data generation method

In this paper, we propose a novel and efficient parameter estimator based on $k$-Nearest Neighbor ($k$NN) and data generation method for the Lognormal-Rician turbulence channel. The Kolmogorov-Smirnov (KS) goodness-of-fit statistical tools are employed to investigate the validity of $k$NN approximation under different channel conditions and it is shown that the choice of $k$ plays a significant role in the approximation accuracy. We present several numerical results to illustrate that solving the constructed objective function can provide a reasonable estimate for the actual values. The accuracy of the proposed estimator is investigated in terms of the mean square error. The simulation results show that increasing the number of generation samples by two orders of magnitude does not lead to a significant improvement in estimation performance when solving the optimization problem by the gradient descent algorithm. However, the estimation performance under the genetic algorithm (GA) approximates to that of the saddlepoint approximation and expectation-maximization estimators. Therefore, combined with the GA, we demonstrate that the proposed estimator achieves the best tradeoff between the computation complexity and the accuracy.

eess.SP↗

arXiv2025

Robust Watermarks Leak: Channel-Aware Feature Extraction Enables Adversarial Watermark Manipulation

Watermarking plays a key role in the provenance and detection of AI-generated content. While existing methods prioritize robustness against real-world distortions (e.g., JPEG compression and noise addition), we reveal a fundamental tradeoff: such robust watermarks inherently improve the redundancy of detectable patterns encoded into images, creating exploitable information leakage. To leverage this, we propose an attack framework that extracts leakage of watermark patterns through multi-channel feature learning using a pre-trained vision model. Unlike prior works requiring massive data or detector access, our method achieves both forgery and detection evasion with a single watermarked image. Extensive experiments demonstrate that our method achieves a 60\% success rate gain in detection evasion and 51\% improvement in forgery accuracy compared to state-of-the-art methods while maintaining visual fidelity. Our work exposes the robustness-stealthiness paradox: current "robust" watermarks sacrifice security for distortion resistance, providing insights for future watermark design.

cs.CV↗

arXiv2025

Effects of skull properties on long-pulsed transcranial focused ultrasound transmission

Transcranial low-intensity focused ultrasound can deliver energy to the brain in a minimally invasive manner for neuromodulation applications. However, continuous sonication through the skull introduces significant wave interactions, complicating precise energy delivery to the target. We present a comprehensive examination of intracranial acoustic fields generated by focused ultrasound transducers and assess the characteristics of cranial bone that affect acoustic transmission. Acoustic field maps were generated at 88 regions of interest across 10 historical and 2 Thiel-embalmed human skull specimens with sonication at frequencies of 220 kHz, 650 kHz, and 1000 kHz. The average peak pressure insertion losses for historical were 3.6$\pm$3.4 dB, 9.3$\pm$3.3 dB, and 14.8$\pm$5.8 dB, respectively, and for Thiel skulls, the respective losses were 2.9$\pm$1.8 dB, 9.4$\pm$2.6 dB, and 17.0$\pm$5.5 dB. The effect of skull thickness, skull density ratio, and skull curvature on intracranial peak pressure, power and focal area was investigated and linear fits produced. Several unfavorable focusing performances were observed in regions with excessive thickness variation. The effects of angulation and spacing between the transducer and the skull were also investigated. Preliminary findings indicate that wave superposition resulting from skull and transducer spacing could lead to a 30-40% uncertainty in peak recorded intracranial pressure.

physics.med-ph↗

arXiv2025

Classroom Simulacra: Building Contextual Student Generative Agents in Online Education for Learning Behavioral Simulation

Student simulation supports educators to improve teaching by interacting with virtual students. However, most existing approaches ignore the modulation effects of course materials because of two challenges: the lack of datasets with granularly annotated course materials, and the limitation of existing simulation models in processing extremely long textual data. To solve the challenges, we first run a 6-week education workshop from N = 60 students to collect fine-grained data using a custom built online education system, which logs students' learning behaviors as they interact with lecture materials over time. Second, we propose a transferable iterative reflection (TIR) module that augments both prompting-based and finetuning-based large language models (LLMs) for simulating learning behaviors. Our comprehensive experiments show that TIR enables the LLMs to perform more accurate student simulation than classical deep learning models, even with limited demonstration data. Our TIR approach better captures the granular dynamism of learning performance and inter-student correlations in classrooms, paving the way towards a ''digital twin'' for online education.

cs.HC↗

arXiv2025

Into the Void: Mapping the Unseen Gaps in High Dimensional Data

We present a comprehensive pipeline, augmented by a visual analytics system named ``GapMiner'', that is aimed at exploring and exploiting untapped opportunities within the empty areas of high-dimensional datasets. Our approach begins with an initial dataset and then uses a novel Empty Space Search Algorithm (ESA) to identify the center points of these uncharted voids, which are regarded as reservoirs containing potentially valuable novel configurations. Initially, this process is guided by user interactions facilitated by GapMiner. GapMiner visualizes the Empty Space Configurations (ESC) identified by the search within the context of the data, enabling domain experts to explore and adjust ESCs using a linked parallel-coordinate display. These interactions enhance the dataset and contribute to the iterative training of a connected deep neural network (DNN). As the DNN trains, it gradually assumes the task of identifying high-potential ESCs, diminishing the need for direct user involvement. Ultimately, once the DNN achieves adequate accuracy, it autonomously guides the exploration of optimal configurations by predicting performance and refining configurations, using a combination of gradient ascent and improved empty-space searches. Domain users were actively engaged throughout the development of our system. Our findings demonstrate that our methodology consistently produces substantially superior novel configurations compared to conventional randomization-based methods. We illustrate the effectiveness of our method through several case studies addressing various objectives, including parameter optimization, adversarial learning, and reinforcement learning.

cs.LG↗