arXiv subjects

Ao Liu

Publications and source records attributed to Ao Liu.

At least 19 recordsLinked to original sources

arXiv2026

Learning Boson Star Solution Families with Physics-Informed Neural Networks

Computing boson star families traditionally requires repeated solution of nonlinear eigenvalue boundary-value problems and careful numerical continuation through turning points. We develop a physics-informed neural network (PINN) that learns the map from the physical parameters and radial coordinate directly to the scalar and metric fields over an equilibrium solution manifold. Regularity and asymptotic boundary conditions are incorporated into the network output, while the training objective combines pointwise supervision, Einstein-Klein-Gordon residuals, and curve-level constraints on the Arnowitt-Deser-Misner mass and Noether charge. A trained model generates a complete configuration in a single forward pass. Across representative one-, two-, and three-branch families, the method reconstructs the mass-frequency spirals and conserved quantities, including configurations on inner branches that require delicate continuation in conventional solvers. These results establish physics-informed surrogate learning as a practical route to amortized exploration of nonlinear self-gravitating solution families.

gr-qc↗

arXiv2026

ParasGB: A Graph Benchmark Suite for Parasitic Estimation on AMS Circuits

As chip manufacturing processes advance to deep submicron nodes, parasitic interconnect effects increasingly dominate the performance of analog and mixed-signal (AMS) circuits and often lead to costly layout iterations. This makes early-stage estimation of parasitic capacitance and resistance important for parasitic-aware design exploration before full physical implementation. However, progress on GNN-based parasitic modeling has been hindered by the lack of public, high-fidelity RC benchmarks that support reproducible evaluation. To address this gap, we introduce ParasGB, the first open-source benchmark suite for pre-layout parasitic parameter prediction on circuit graphs. ParasGB provides large-scale, heterogeneous RC networks extracted with commercial EDA tools from tape-out-proven designs, together with a unified evaluation protocol covering node-level ground capacitance, edge-level resistance, and edge-level coupling capacitance. Within this framework, we benchmark diverse GNN architectures using a standardized training pipeline and expose challenges such as extreme label imbalance, long-tailed parasitic distributions, and strong structural heterogeneity. By establishing a physically grounded and standardized benchmark for early-stage parasitic prediction, ParasGB provides an open platform for reproducible research on circuit graph learning and parasitic-aware model development. All datasets, preprocessing scripts, and configurations are publicly available in our code repository https://github.com/ShenShan123/ParasGB.git.

cs.LG↗

arXiv2026

The Cost of Perfect English: Pragmatic Flattening and the Erasure of Authorial Voice in L2 Writing Supported by GenAI

The integration of Generative AI (GenAI) into language learning offers second language (L2) writers powerful tools for text optimization. However, pursuing native-like fluency often sacrifices sociopragmatic diversity. Investigating "pragmatic flattening" - the systematic erasure of culturally preferred politeness and authorial stance - this study conducts a comparative analysis of argumentative essays by Chinese B2-level university students from the ICNALE corpus. The original texts were polished via the APIs of four leading Large Language Models at a zero-temperature setting for reproducibility. Findings reveal a nuanced "dimensional divergence" within the Semantic Preservation Paradox. While models corrected lexicogrammatical errors and retained propositional meaning, sociopragmatic interventions were bifurcated. In the interactive dimension, all models showed a drastic collapse of dialogic engagement markers, turning negotiated discourse into monologic assertions. Conversely, in the epistemic stance dimension, models showed architecture-based variability: some aggressively scrubbed epistemic markers, while others reinforced tentative hedging as decontextualized algorithmic caution. This confirms that while GenAI enhances accuracy, it systematically overwrites L2 writers' unique rhetorical identities into a homogenized Anglo-American paradigm. We argue that future instruction must move beyond error correction, advocating for Critical AI Literacy to empower multilingual writers to use GenAI for linguistic enhancement while safeguarding sociopragmatic diversity and rhetorical agency.

cs.CL↗

arXiv2026

R2G: A Multi-View Circuit Graph Benchmark Suite from RTL to GDSII

Graph neural networks (GNNs) are increasingly applied to physical design tasks such as congestion prediction and wirelength estimation, yet progress is hindered by inconsistent circuit representations and the absence of controlled evaluation protocols. We present R2G (RTL-to-GDSII), a multi-view circuit-graph benchmark suite that standardizes five stage-aware views with information parity (every view encodes the same attribute set, differing only in where features attach) over 30 open-source IP cores (up to $10^6$ nodes/edges). R2G provides an end-to-end DEF-to-graph pipeline spanning synthesis, placement, and routing stages, together with loaders, unified splits, domain metrics, and reproducible baselines. By decoupling representation choice from model choice, R2G isolates a confound that prior EDA and graph-ML benchmarks leave uncontrolled. In systematic studies with GINE, GAT, and ResGatedGCN, we find: (i) view choice dominates model choice, with Test R$^2$ varying by more than 0.3 across representations for a fixed GNN; (ii) node-centric views generalize best across both placement and routing; and (iii) decoder-head depth (3--4 layers) is the primary accuracy driver, turning divergent training into near-perfect predictions (R$^2$$>$0.99). Code and datasets are available at https://github.com/ShenShan123/R2G.

cs.CV↗

arXiv2026

Accelerating Black Hole Image Generation via Latent Space Diffusion Models

Interpreting horizon-scale black hole images currently relies on computationally intensive General Relativistic Ray Tracing (GRRT) simulations, which pose a significant bottleneck for rapid parameter exploration and high-precision tests of strong-field gravity. We demonstrate that physically accurate black hole images, synthesized from magnetized accretion flows, inherently reside on a low-dimensional manifold-encoding the essential features of spacetime geometry, plasma distribution, and relativistic emission. Leveraging this structure, we introduce a physics-conditioned diffusion model that operates in a compact latent space to generate high-fidelity black hole imagery directly from physical parameters. The model accurately reproduces critical observational signatures from full GRRT simulations-such as shadow diameter, photon-ring structure, and relativistic brightness asymmetry-while achieving over a fourfold reduction in computational expense. Compared with the previous generation of denoising diffusion models, the proposed approach achieves significant improvements in image quality, reconstruction fidelity, and parameter estimation accuracy, while reducing the average inference time per black hole image from 5.25 seconds to 1.15 seconds. Our work establishes diffusion-based latent models as efficient and scalable substitutes for traditional radiative transfer solvers, offering a practical framework toward real-time modeling and inference for next-generation black hole imaging.

gr-qc↗

arXiv2025

Lattice design of a storage-ring-based light source for generating high-power fully coherent EUV radiation

We present the physical design and systematic optimization of a high-performance storage ring tailored for the generation of high-power coherent radiation, with particular emphasis on the extreme ultraviolet (EUV) regime. The proposed ring adopts a Double Bend Achromat (DBA) lattice configuration and integrates 12 superconducting wigglers to significantly enhance radiation damping and minimize the natural emittance. And a bypass line is adopted to generate high power coherent radiation. Comprehensive linear and nonlinear beam dynamics analyses have been conducted to ensure beam stability and robustness across the operational parameter space. The optimized design achieves a natural emittance of approximately 0.8 nm and a longitudinal damping time of around 1.4 ms, enabling the efficient buildup of coherent radiation. Three-dimensional numerical simulations, incorporating the previously proposed angular dispersion-induced microbunching (ADM) mechanism, further confirm the system's capability to generate high-power EUV coherent radiation, with output powers reaching the order of several hundred watts. These results underscore the strong potential of the proposed design for applications in coherent photon science and EUV lithography.

physics.acc-ph↗

arXiv2025

PolyJailbreak: Cross-Modal Jailbreaking Attacks on Black-Box Multimodal LLMs

Multimodal large language models (MLLMs) have become integral to a wide range of real-world applications by jointly reasoning over text and visual inputs. However, despite recent advances in safety alignment, MLLMs remain vulnerable to jailbreak attacks, where carefully crafted inputs can bypass safety mechanisms and elicit harmful responses. In this work, we investigate the security vulnerabilities of MLLMs in text-vision scenarios and propose a novel black-box jailbreak framework, named PolyJailbreak. We first identify a phenomenon, termed multimodal safety asymmetry, where visual alignment introduces uneven safety constraints across modalities and weakens overall robustness. We analyze attention dynamics and latent representations in MLLMs, revealing that visual inputs can disrupt cross-modal information flow and reduce the model's ability to separate benign and malicious intents. Motivated by these findings, we propose PolyJailbreak, which organizes the discovered vulnerabilities into a structured library of reusable Atomic Strategy Primitives to enable step-wise transformations from harmful intents to effective jailbreak inputs. Guided by these primitives, a reinforcement learning-based multi-agent optimization process automatically adapts attacks to the target model without access to internal parameters. Extensive experiments on a wide range of MLLMs demonstrate that PolyJailbreak consistently outperforms state-of-the-art jailbreak baselines, with an average improvement of 18.15% in attack success rate and a success rate exceeding 95% on commercial black-box models, including GPT-4o and Gemini.

cs.CR↗

arXiv2025

ReLook: Vision-Grounded RL with a Multimodal LLM Critic for Agentic Web Coding

While Large Language Models (LLMs) excel at algorithmic code generation, they struggle with front-end development, where correctness is judged on rendered pixels and interaction. We present ReLook, an agentic, vision-grounded reinforcement learning framework that empowers an agent to close a robust generate--diagnose--refine loop by invoking a multimodal LLM (MLLM) as a tool. During training, the agent uses the MLLM-in-the-loop both as a visual critic--scoring code with screenshots--and as a source of actionable, vision-grounded feedback; a strict zero-reward rule for invalid renders anchors renderability and prevents reward hacking. To prevent behavioral collapse, we introduce Forced Optimization, a strict acceptance rule that admits only improving revisions, yielding monotonically better trajectories. At inference, we decouple the critic and run a lightweight, critic-free self-edit cycle, keeping latency comparable to base decoding while retaining most of the gains. Across three widely used benchmarks, ReLook consistently outperforms strong baselines in vision-grounded front-end code generation, highlighting the benefits of agentic perception, visual rewards, and training-inference decoupling.

cs.LG↗

arXiv2025

Identifying Microlensing by Compact Dark Matter through Diffraction Patterns in Gravitational Waves with Machine Learning

Gravitational wave lensing, particularly microlensing by compact dark matter (DM), offers a unique avenue to probe the nature of dark matter. However, conventional detection methods are often computationally expensive, inefficient, and sensitive to waveform systematics. In this work, we introduce the Wavelet Convolution Detector (WCD), a deep learning framework specifically designed to identify wave-optics diffraction patterns imprinted in gravitationally lensed signals. The WCD integrates multi-scale wavelet analysis within residual convolutional blocks to efficiently extract time-frequency interference structures, and is trained on a realistically generated dataset incorporating compact DM mass functions and astrophysical lensing probabilities. This work is the first machine learning-based approach capable of identifying such wave-optics signatures in lensed gravitational waves. Tested on simulated binary black hole events, the model achieves 92.2\% accuracy (AUC=0.965), with performance rising to AUC$\sim$0.99 at high SNR. Crucially, it maintains high discriminative power across a wide range of lens masses without retraining, demonstrating particular strength in the low-impact-parameter and high-lens-mass regimes where wave-optics effects are most pronounced. Compared to Bayesian inference, the WCD provides orders-of-magnitude faster inference, making it a scalable and efficient tool for discovering compact DM through lensed gravitational waves in the era of third-generation detectors.

astro-ph.IM↗

arXiv2025

Identification of Strongly Lensed Gravitational Wave Events Using Squeeze-and-Excitation Multilayer Perceptron Data-efficient Image Transformer

With the advancement of third-generation gravitational wave detectors, the identification of strongly lensed gravitational wave (GW) events is expected to play an increasingly vital role in cosmology and fundamental physics. However, traditional Bayesian inference methods suffer from combinatorial computational overhead as the number of events grows, making real-time analysis infeasible. To address this, we propose a deep learning model named Squeeze-and-Excitation Multilayer Perceptron Data-efficient Image Transformer (SEMD), based on Vision Transformers, which classifies strongly lensed GW events by modeling morphological similarity between time-frequency spectrogram pairs. By integrating Squeeze-and-Excitation attention mechanisms and multilayer perceptrons , SEMD achieves strong feature extraction and discrimination. Trained and evaluated on simulated datasets using Advanced LIGO and Einstein Telescope noise, the model demonstrates robustness and generalization across different detector sensitivities and physical conditions, highlighting the promise of deep learning for rapid identification of strongly lensed GW signals.

astro-ph.IM↗

arXiv2025

AutoCodeBench: Large Language Models are Automatic Code Benchmark Generators

Large Language Models (LLMs) have demonstrated remarkable capabilities across various domains, with code generation emerging as a key area of focus. While numerous benchmarks have been proposed to evaluate their code generation abilities, these benchmarks face several critical limitations. First, they often rely on manual annotations, which are time-consuming and difficult to scale across different programming languages and problem complexities. Second, most existing benchmarks focus primarily on Python, while the few multilingual benchmarks suffer from limited difficulty and uneven language distribution. To address these challenges, we propose AutoCodeGen, an automated method for generating high-difficulty multilingual code generation datasets without manual annotations. AutoCodeGen ensures the correctness and completeness of test cases by generating test inputs with LLMs and obtaining test outputs through a multilingual sandbox, while achieving high data quality through reverse-order problem generation and multiple filtering steps. Using this novel method, we introduce AutoCodeBench, a large-scale code generation benchmark comprising 3,920 problems evenly distributed across 20 programming languages. It is specifically designed to evaluate LLMs on challenging, diverse, and practical multilingual tasks. We evaluate over 30 leading open-source and proprietary LLMs on AutoCodeBench and its simplified version AutoCodeBench-Lite. The results show that even the most advanced LLMs struggle with the complexity, diversity, and multilingual nature of these tasks. Besides, we introduce AutoCodeBench-Complete, specifically designed for base models to assess their few-shot code generation capabilities. We hope the AutoCodeBench series will serve as a valuable resource and inspire the community to focus on more challenging and practical multilingual code generation scenarios.

cs.CL↗

arXiv2025

Managing, Analyzing and Sharing Research Data with Gen3 Data Commons

Gen3 is an open-source data platform for building data commons. A data commons is a cloud-based data platform for managing, analyzing, and sharing data with a research community. Gen3 has been used to build over a dozen data commons that in aggregate contain over 28 PB of data and 64 million FAIR data objects. To set up a Gen3 data commons, you first define a data model. Gen3 then autogenerates 1) a data portal for searching and exploring data in the commons; 2) a data portal for submitting data to the commons; and 3) FAIR APIs for accessing the data programmatically. Gen3 is built over a small number of standards-based software services, which are designed to support current and future Gen3 components so that Gen3 can interoperate with other data platforms and data ecosystems.

cs.DC↗

arXiv2025

Physical parameter regression from black hole images via a multiscale adaptive neural network

High-precision regression of physical parameters from black hole images generated by General Relativistic Ray Tracing (GRRT) is essential for investigating spacetime curvature and advancing black hole astrophysics. However, due to limitations in observational resolution, high observational costs, and imbalanced distributions of positive and negative samples, black hole images often suffer from data scarcity, sparse parameter spaces, and complex structural characteristics. These factors pose significant challenges to conventional regression methods based on simplified physical models. To overcome these challenges, this study introduces Multiscale Adaptive Network (MANet) , a novel regression framework grounded in deep learning. MANet integrates an Adaptive Channel Attention (ACA) module to selectively enhance features in physically informative regions. Meanwhile, a Multiscale Enhancement Feature Pyramid (MEFP) is employed to capture fine-grained spatial structures such as photon rings and accretion disks, while alleviating information loss due to downsampling. Experimental evaluations on GRRT-simulated datasets demonstrate that MANet substantially improves parameter estimation accuracy and generalization capability in high-dimensional parameter spaces, outperforming existing baseline approaches. This framework presents a promising avenue for high-precision parameter regression in Event Horizon Telescope (EHT) data analysis and broader astrophysical imaging applications characterized by sparse and noisy data.

astro-ph.IM↗

arXiv2025

ArtifactsBench: Bridging the Visual-Interactive Gap in LLM Code Generation Evaluation

The generative capabilities of Large Language Models (LLMs) are rapidly expanding from static code to dynamic, interactive visual artifacts. This progress is bottlenecked by a critical evaluation gap: established benchmarks focus on algorithmic correctness and are blind to the visual fidelity and interactive integrity that define modern user experiences. To bridge this gap, we introduce ArtifactsBench, a new benchmark and paradigm for the automated, multimodal evaluation of visual code generation. Our framework programmatically renders each generated artifact and captures its dynamic behavior through temporal screenshots. This visual evidence, alongside the source code, is then assessed by a Multimodal LLM (MLLM)-as-Judge, which is rigorously guided by a fine-grained, per-task checklist to ensure holistic and reproducible scoring. We construct a new benchmark of 1,825 diverse tasks and evaluate over 30 leading LLMs. Our automated evaluation achieves a striking 94.4% ranking consistency with WebDev Arena, the gold-standard for human preference in web development, and over 90% pairwise agreement with human experts. This establishes ArtifactsBench as the first framework to reliably automate the assessment of human-perceived quality at scale. Our analysis provides a high-resolution map of the current SOTA, revealing that generalist models often outperform domain-specific ones. We open-source ArtifactsBench, including the benchmark, evaluation harness, and baseline results at https://artifactsbenchmark.github.io/, to provide the community with a scalable and accurate tool to accelerate the development of user-centric generative models.

cs.CL↗

arXiv2025

Hunyuan-TurboS: Advancing Large Language Models through Mamba-Transformer Synergy and Adaptive Chain-of-Thought

As Large Language Models (LLMs) rapidly advance, we introduce Hunyuan-TurboS, a novel large hybrid Transformer-Mamba Mixture of Experts (MoE) model. It synergistically combines Mamba's long-sequence processing efficiency with Transformer's superior contextual understanding. Hunyuan-TurboS features an adaptive long-short chain-of-thought (CoT) mechanism, dynamically switching between rapid responses for simple queries and deep "thinking" modes for complex problems, optimizing computational resources. Architecturally, this 56B activated (560B total) parameter model employs 128 layers (Mamba2, Attention, FFN) with an innovative AMF/MF block pattern. Faster Mamba2 ensures linear complexity, Grouped-Query Attention minimizes KV cache, and FFNs use an MoE structure. Pre-trained on 16T high-quality tokens, it supports a 256K context length and is the first industry-deployed large-scale Mamba model. Our comprehensive post-training strategy enhances capabilities via Supervised Fine-Tuning (3M instructions), a novel Adaptive Long-short CoT Fusion method, Multi-round Deliberation Learning for iterative improvement, and a two-stage Large-scale Reinforcement Learning process targeting STEM and general instruction-following. Evaluations show strong performance: overall top 7 rank on LMSYS Chatbot Arena with a score of 1356, outperforming leading models like Gemini-2.0-Flash-001 (1352) and o4-mini-2025-04-16 (1345). TurboS also achieves an average of 77.9% across 23 automated benchmarks. Hunyuan-TurboS balances high performance and efficiency, offering substantial capabilities at lower inference costs than many reasoning models, establishing a new paradigm for efficient large-scale pre-trained models.

cs.CL↗

arXiv2025

Advancing Cosmological Parameter Estimation and Hubble Parameter Reconstruction with Long Short-Term Memory and Efficient-Kolmogorov-Arnold Networks

In this work, we propose a novel approach for cosmological parameter estimation and Hubble parameter reconstruction using Long Short-Term Memory (LSTM) networks and Efficient-Kolmogorov-Arnold Networks (Ef-KAN). LSTM networks are employed to extract features from observational data, enabling accurate parameter inference and posterior distribution estimation without relying on solvable likelihood functions. This method achieves performance comparable to traditional Markov Chain Monte Carlo (MCMC) techniques, offering a computationally efficient alternative for high-dimensional parameter spaces. By sampling from the reconstructed data and comparing it with mock data, our designed LSTM constraint procedure demonstrates the superior performance of this method in terms of constraint accuracy, and effectively captures the degeneracies and correlations between the cosmological parameters. Additionally, the Ef-KAN model is introduced to reconstruct the Hubble parameter H(z) from both observational and mock data. Ef-KAN is entirely data-driven approach, free from prior assumptions, and demonstrates superior capability in modeling complex, non-linear data distributions. We validate the Ef-KAN method by reconstructing the Hubble parameter, demonstrating that H(z) can be reconstructed with high accuracy. By combining LSTM and Ef-KAN, we provide a robust framework for cosmological parameter inference and Hubble parameter reconstruction, paving the way for future research in cosmology, especially when dealing with complex datasets and high-dimensional parameter spaces.

astro-ph.CO↗

arXiv2025

A Brain-Inspired Perception-Decision Driving Model Based on Neural Pathway Anatomical Alignment

In the realm of autonomous driving, conventional approaches for vehicle perception and decision-making primarily rely on sensor input and rule-based algorithms. However, these methodologies often suffer from lack of interpretability and robustness, particularly in intricate traffic scenarios. To tackle this challenge, we propose a novel brain-inspired driving (BID) framework. Diverging from traditional methods, our approach harnesses brain-inspired perception technology to achieve more efficient and robust environmental perception. Additionally, it employs brain-inspired decision-making techniques to facilitate intelligent decision-making. The experimental results show that the performance has been significantly improved across various autonomous driving tasks and achieved the end-to-end autopilot successfully. This contribution not only advances interpretability and robustness but also offers fancy insights and methodologies for further advancing autonomous driving technology.

cs.RO↗

arXiv2025

BAN: Neuroanatomical Aligning in Auditory Recognition between Artificial Neural Network and Human Cortex

Drawing inspiration from neurosciences, artificial neural networks (ANNs) have evolved from shallow architectures to highly complex, deep structures, yielding exceptional performance in auditory recognition tasks. However, traditional ANNs often struggle to align with brain regions due to their excessive depth and lack of biologically realistic features, like recurrent connection. To address this, a brain-like auditory network (BAN) is introduced, which incorporates four neuroanatomically mapped areas and recurrent connection, guided by a novel metric called the brain-like auditory score (BAS). BAS serves as a benchmark for evaluating the similarity between BAN and human auditory recognition pathway. We further propose that specific areas in the cerebral cortex, mainly the middle and medial superior temporal (T2/T3) areas, correspond to the designed network structure, drawing parallels with the brain's auditory perception pathway. Our findings suggest that the neuroanatomical similarity in the cortex and auditory classification abilities of the ANN are well-aligned. In addition to delivering excellent performance on a music genre classification task, the BAN demonstrates a high BAS score. In conclusion, this study presents BAN as a recurrent, brain-inspired ANN, representing the first model that mirrors the cortical pathway of auditory recognition.

q-bio.NC↗