arXiv subjects

Chao Liu

Publications and source records attributed to Chao Liu.

At least 55 records · Page 3Linked to original sources

arXiv2026

SemGeoNav:A Safety-Guided Visual Navigation Approach with Semantic Reasoning and Geometric Planning

Learning-based visual navigation has enhanced semantic goal-reaching capabilities. However, due to their black-box nature, purely end-to-end models often lack explicit geometric constraints, leading to unpredictable and unreliable obstacle avoidance in open environments. Conversely, traditional geometric planners ensure safety but struggle with high-dimensional visual targets. To address these limitations, we propose SemGeoNav, a novel hierarchical visual navigation framework.It tightly integrates the high-level semantic reasoning of end-to-end models with the reliable local planning ability of geometry-based methods, achieving robust image-based navigation while significantly improving obstacle avoidance. Furthermore, we introduce a temporal trajectory smoothing mechanism to ensure continuous and stable robot motion. We evaluated SemGeoNav on a Unitree Go2 quadruped robot in real-world environments. The results demonstrate that SemGeoNav outperforms existing representative methods, including ViNT and NoMaD, achieving higher success rates and shorter navigation times.

cs.RO↗

arXiv2026

Hamiltonian-Aware ADAPT Variational Quantum Eigensolver for Molecular Ground-State Simulation

Designing compact ansätze in Variational Quantum Eigensolver (VQE) is crucial for solving energetic problems of practical molecules on near-term quantum devices. However, existing Adaptive Derivative-Assembled Pseudo-Trotter (ADAPT) ansätze face two challenges: improper operator selection and accumulation of degraded operators. In this paper, we propose the Hamiltonian-Aware (HA) ADAPT-VQE algorithm to address these issues. First, we establish a novel excitation operator selection criterion. It breaks the local constraint of existing criteria by incorporating Hamiltonian information, prioritizes physically meaningful excitation operators, and incurs no extra classical or quantum computational overhead. Furthermore, we develop a problem-adaptive method for discriminating and pruning redundant excitation operators stemming from improper selection and inevitable degradation. This method balances redundant operator pruning and convergence guarantee, and is applicable to ansätze with arbitrary scales. Systematic numerical experiments on typical strongly correlated molecular systems demonstrate that our HA-ADAPT-VQE avoids energy plateaus and outperforms baseline algorithms in terms of energy error, ansatz size, and measurement cost. This work offers an efficient, robust ansatz construction paradigm, facilitating the development and practical deployment of large-scale VQE in quantum chemistry.

quant-ph↗

arXiv2026

Proofs on singularity-free solutions and scalarization in nonlinear Einstein-scalar-Gauss-Bonnet cosmology

The search for singularity-free cosmological solutions has become a highly active topic in the physics community in recent years, yet existing results are largely numerical or based on asymptotic analysis. To place these developments on a firm mathematical footing, we rigorously establish the global existence and estimates of a class of singularity-free cosmological solutions to the fully nonlinear Einstein--scalar system in Einstein-scalar-Gauss-Bonnet gravity with quadratic coupling, providing proofs of previous numerical results in mathematical perspective. We further prove nonlinear scalarization triggered by a Gauss-Bonnet-induced tachyonic instability. Our analysis relies on a novel structural identity, the power identity, which yields decoupled differential inequalities for the Hubble parameter. This framework provides a new method for converting numerical evidence into a mathmatical proof for nonlinear systems.

math.AP↗

arXiv2026

High-power Test and System Integration of Direct RF Sampling Based LLRF Control and Monitoring System for S-Band Accelerating Structures

High precision Low-level RF (LLRF) control and monitoring systems for future particle accelerators will be a significant technical challenge as the requirements in performance, flexibility and affordability become increasingly stringent. We have developed an RF system-on-chip (RFSoC) based next-generation LLRF (NG-LLRF) for S-band accelerating structures, which samples and synthesizes the RF pulses directly without the analog mixers used for traditional LLRF systems. The platform delivered considerably better performance than the requirements of the targeted applications, such as the upgrades for Next Linear Collider Test Accelerator (NLCTA) and test facilities at SLAC. As part of the upgrade program, we also developed a custom solid-state amplifier (SSA) to deliver RF pulses at the desired power level of the klystron. Integration of the LLRF with the SSA and the high-power test facility could be challenging. The power levels and RF pulse stability at each stage of the high-power RF drive system must be optimized to deliver the desired RF performance. In this paper, the integration procedure and the test and characterization results at each stage of integration will be summarized, analyzed and discussed. This integration is an essential step for the full deployment of the NG-LLRF system to test facilities and accelerators in different frequency bands.

physics.acc-ph↗

arXiv2026

Stellar Density Classification and Regression for CSST Multi-color Imaging Using Deep Learning

The Chinese Space Station Survey Telescope (CSST) aims to map the universe across an unprecedented dynamic range of stellar densities, spanning from extragalactic voids to the crowded Galactic center (e.g. a few stars and galaxies in the voids and $>10^5$ stars per detector in Galactic center). However, processing such heterogeneous data with a general source extraction pipeline introduces significant systematic uncertainties, standard algorithms exhibit poor accuracy in crowded fields and suffer from increased astrometric uncertainty in void regions. To mitigate these systematics, we propose a hierarchical, two-stage deep learning model for adaptive data reduction. The first stage ('classification') employs a ResNet-34 model to classify images into six discrete density categories, achieving $98.83\%$ in global accuracy. This classification acts as a critical decision gate, ensuring high calibration accuracy in the crowded fields. In the second stage ('regression'), a ResNet-50 regression model predicts the bright stars ($<23.5$ mag) in the field, which is essential for astrometric calibration, achieving a mean absolute error (MAE) of 0.0824 dex. By decoupling density characterization from source extraction, our model ensures that photometric and astrometric algorithms are optimally matched to the stellar density environment, thereby enhancing the fidelity and homogeneity of CSST as well as future large sky survey data products.

astro-ph.IM↗

arXiv2026

Comparative analysis of missing data imputation methods for CSST survey: Impact on photometric redshift estimation performance

Improving the accuracy of photometric redshifts (photo-$z$) is essential for reliable statistical studies of cosmology and galaxy evolution. However, missing photometric bands are a common observational challenge that can significantly degrade photo-$z$ estimation accuracy. In this work, we present a systematic evaluation of data imputation methods aimed at improving photo-$z$ performance. We benchmark a range of representative machine learning (ML) and deep learning (DL) architectures, identifying k-nearest neighbors (KNN) and the attention-based SAITS model as the leading performers. These models are then applied to China Space Station Survey Telescope (CSST) mock data to assess their performance under realistic observational conditions. Our results show that KNN yields the highest accuracy under idealized missing completely at random (MCAR) conditions with complete training sets, whereas robustness tests reveal that SAITS significantly outperforms KNN when training data is incomplete or when applied to realistic mixed-mechanism scenarios. We find that domain consistency between training and testing missingness patterns is a prerequisite for optimal performance, highlighting the risks of domain shift in supervised regression tasks. Furthermore, our analysis demonstrates that while general imputation models are highly effective for MCAR and missing at random (MAR) data, they are detrimental when applied to missing not at random (MNAR) data arising from flux limits, as statistical models fail to capture the physical information inherent in these non-detections. Consequently, we advocate for more sophisticated architectures capable of disentangling stochastic missingness from physical non-detections to address these distinct mechanisms individually.

astro-ph.GA↗

arXiv2026

DynaFix: Iterative Automated Program Repair Driven by Execution-Level Dynamic Information

Automated Program Repair (APR) aims to automatically generate correct patches for buggy programs. Recent approaches leveraging large language models (LLMs) have shown promise but face limitations. Most rely solely on static analysis, ignoring runtime behaviors. Some attempt to incorporate dynamic signals, but these are often restricted to training or fine-tuning, or injected only once into the repair prompt, without iterative use. This fails to fully capture program execution. Current iterative repair frameworks typically rely on coarse-grained feedback, such as pass/fail results or exception types, and do not leverage fine-grained execution-level information effectively. As a result, models struggle to simulate human stepwise debugging, limiting their effectiveness in multi-step reasoning and complex bug repair. To address these challenges, we propose DynaFix, an execution-level dynamic information-driven APR method that iteratively leverages runtime information to refine the repair process. In each repair round, DynaFix captures execution-level dynamic information such as variable states, control-flow paths, and call stacks, transforming them into structured prompts to guide LLMs in generating candidate patches. If a patch fails validation, DynaFix re-executes the modified program to collect new execution information for the next attempt. This iterative loop incrementally improves patches based on updated feedback, similar to the stepwise debugging practices of human developers. We evaluate DynaFix on the Defects4J v1.2 and v2.0 benchmarks. DynaFix repairs 186 single-function bugs, a 10% improvement over state-of-the-art baselines, including 38 bugs previously unrepaired. It achieves correct patches within at most 35 attempts, reducing the patch search space by 70% compared with existing methods, thereby demonstrating both effectiveness and efficiency in repairing complex bugs.

cs.SE↗

arXiv2026

A Multi-modal Fusion Network for Star-Galaxy Classification from CSST Simulated Datasets

The distinction between stars and galaxies is a fundamental problem in the field of celestial classification. This issue has become challenging for these ongoing and upcoming digital surveys, which will produce terabytes and even petabytes of astronomical data. While deep learning offers a powerful solution for star-galaxy classification in large-scale datasets, most current approaches are limited by their reliance on catalog data alone, which consists primarily of multi-band magnitudes and imprecise morphological parameters. Therefore, we utilize China Space Station Telescope (CSST) simulation data to build a dataset with both image and photometric catalog, including 32,371 stars and 93,525 galaxies. A supervised deep learning network based on ResNet-50 and BiLSTM is proposed to improve the classification of two types of astronomical objects. The features of the catalog and image are integrated by the model, achieving 99.81% recall for galaxies and 99.66% recall for stars after training on GPU for 50 epochs. We evaluated the effects of data augmentation and multi-modal data fusion, which demonstrate that our model has commendable performance. Furthermore, our model also has a high accuracy rate for faint astronomical objects and high redshift galaxies, demonstrating its applicability to the upcoming CSST scientific data.

astro-ph.IM↗

arXiv2026

Toward Functional and Non-Functional Evaluation of Application-Level Code Generation

Large language models (LLMs) have achieved strong performance on code generation. However, most prior evaluations focus on snippet-level outputs, such as function generation or repository completion. These settings do not fully evaluate application-level code generation, where the goal is to produce a runnable repository with coherent multi-file structure, dependency support, and end-to-end executability. In addition, real-world software quality depends not only on functional correctness but also on non-functional quality attributes, such as maintainability and security. In this paper, we present RAL-Bench, a benchmark and evaluation framework for application-level code generation. For each task, RAL-Bench derives a concise natural-language requirement from a high-quality reference project, constructs black-box system tests for both functional correctness and non-functional quality attributes. It also retains only the candidate tests that pass on the reference repository. Under this unified evaluation protocol, functional correctness is measured by the system test pass rate, while non-functional quality is evaluated along five ISO/IEC 25010-inspired dimensions, with per-dimension diagnostics and reference-normalized scoring.We evaluate 16 frontier LLMs under a controlled zero-shot setting with greedy decoding. The results show that functional correctness remains the primary bottleneck in application-level code generation, while non-functional quality also remains challenging. Under our evaluation protocol, no model exceeds a 45\% functional score. These findings suggest that strong performance on existing code generation benchmarks does not yet translate to strong performance on application-level repository generation. This result highlights the need for evaluation settings that directly assess end-to-end repository generation rather than relying only on snippet-level success.

cs.SE↗

arXiv2026

SUAD: Solid-Channel Ultrasound Injection Attack and Defense to Voice Assistants

As a versatile AI application, voice assistants (VAs) have become increasingly popular, but are vulnerable to security threats. Attackers have proposed various inaudible attacks, but are limited by cost, distance, or LoS. Therefore, we propose \name~Attack, a long-range, cross-barrier, and interference-free inaudible voice attack via solid channels. We begin by thoroughly analyzing the dispersion effect in solid channels, revealing its unique impact on signal propagation. To avoid distortions in voice commands, we design a modular command generation model that parameterizes attack distance, victim audio, and medium dispersion features to adapt to variations in the solid-channel state. Additionally, we propose SUAD Defense, a universal defense that uses ultrasonic perturbation signals to block inaudible voice attacks (IVAs) without impacting normal speech. Since the attack can occur at arbitrary frequencies and times, we propose a training method that randomizes both time and frequency to generate perturbation signals that break ultrasonic commands. Notably, the perturbation signal is modulated to an inaudible frequency without affecting the functionality of voice commands for VAs. Experiments on six smartphones have shown that SUAD Attack achieves activation success rates above 89.8% and SUAD Defense blocks IVAs with success rates exceeding 98%.

cs.CR↗

arXiv2026

Stellar initial mass function in the 100-pc solar neighbourhood

The stellar initial mass function (IMF) is among the most fundamental distributions in astrophysics, defined as the mass spectrum of stars produced in a single star-formation event. Even in the solar neighbourhood, where measurements can be conducted via star counting, disentangling the IMF from observational effects remains challenging. In this work we introduce a new parametrisation of the stellar IMF in the 100-pc solar neighbourhood, leveraging the high-precision astrometric and photometric data from \textsl{Gaia} DR3: we model the colour-magnitude diagram of the field star population while accounting for observational uncertainties, Malmquist bias, Lutz-Kelker bias, variations in the mass-luminosity relation arising from metallicity differences, and the effects of unresolved binaries. In particular, we synthesise the binary population with a process imitating the dynamical evolution observed in star clusters to enforce that all components are drawn from the same IMF, while simultaneously recovering the observed present-day mass-ratio distribution. We determine an averaged stellar IMF over $0.25<m<1.0~M_{\odot}$ that aligns with canonical IMFs but achieves significantly tighter constraints: $α_1=0.75^{+0.06}_{-0.04}$, $α_2=2.07^{+0.04}_{-0.03}$, and a break point at $m_{\mathrm{break}}=0.40^{+0.01}_{-0.01}$ $\mathrm{M_{\odot}}$. Our inference also yields an averaged binary fraction over $0.25<m<1.0~M_{\odot}$ of approximately 26\%, and constrains the \textsl{Gaia} DR3 angular resolution to $1.11^{+0.11}_{-0.08}$ arcsec. We also provide the $ξ$-parameter for our IMF, which is $0.5070_{-0.0096}^{+0.0068}$, to facilitate direct comparison with other IMF determinations.

astro-ph.GA↗

arXiv2026

Evaluating the spatial intra-pixel sensitivity variations and influence based on space observation

Intra-pixel sensitivity variations (IPSVs) in charge-coupled devices (CCDs) and complementary metal-oxide-semiconductor (CMOS) detectors constitute a significant source of astrometric error for undersampled stellar observations. Since laboratory-based IPSV measurements suffer from limited applicability, we propose a computational method to directly infer IPSV from stellar images and validate it with simulated data. By minimizing the flux residuals between theoretical and observed stellar models through least-squares fitting, we can successfully recover the IPSV, which is treated as nearly identical across pixels. Simulations demonstrate that the reconstructed IPSV achieves high accuracy, and the instrumental point spread function (IPSF) restored using this IPSV improves stellar centroiding by nearly 30$\times$, effectively eliminating periodic pixel-phase errors. The method remains robust under different morphologies of IPSV and varying sampling conditions. Additionally, the framework can be extended to an iterative IPSF-IPSV closed-loop scheme that updates both components simultaneously, providing a practical pathway for continuous detector calibration in future space-based astronomical surveys.

astro-ph.IM↗

arXiv2026

CSST-PSFNet: A Point Spread Function Reconstruction Model for the CSST Based on Deep Learning

This paper presents CSST-PSFNet, a deep learning method for high-fidelity point spread function (PSF) reconstruction developed for the Chinese Space Station Survey Telescope (CSST). The model integrates a residual neural network, a lightweight Transformer architecture, and a variational latent representation to address key challenges in CSST imaging, including severe PSF undersampling, inter-band variability, and smooth spatial variation across the focal plane. Trained and validated on high-resolution star-PSF pairs generated by the CSST Main Survey Simulator, CSST-PSFNet achieves improved pixel-level accuracy and more precise recovery of shape parameters relevant to weak lensing compared to widely used PSFEx. On both the standard test dataset and a blurred dataset representing the upper bound of expected on-orbit PSF degradation, the model achieves a size residual precision below 0.005 and an ellipticity residual precision below 0.002. A weak-label adaptation experiment further shows that the model can recover PSFEx-level performance when the true PSF is unknown, demonstrating robustness in controlled degradation scenarios and weak-label adaptation experiments. These results indicate that CSST-PSFNet provides a flexible and extensible framework for future on-orbit PSF calibration in large-scale CSST surveys, with potential applications in weak-lensing cosmology and precision astrophysical measurements.

astro-ph.IM↗

arXiv2026

Galactic Stellar Halo Luminosity Function

We measure the luminosity function (LF) of the Milky Way's stellar halo, using a magnitude complete, distance limited sample of stars from $Gaia$ DR3. Stars with high transverse velocities are selected, to isolate a high purity sample of the local halo. We adopt a cutoff transverse velocity of 250$\,$km$\,$s$^{-1}$, yielding 24,471 stars, and compute the halo LF, taking into account the effects of sample selection criteria. The LF displays similar features as are found in the well-probed LF of nearby, metal-rich disk stars, showing a strong peak at an absolute magnitude of around $M_G=10$, and a flattening near $M_G\sim7$ (Wielen dip). The $Gaia$ sample yields the first measurement of the LF continuously from the dimmest main sequence halo stars (subdwarfs) at an absolute $M_G$ magnitude near 13 mag to bright giants at $M_G\sim-3$. We obtain a local stellar halo number density of $1.7\times10^{-4}$ stars$\,$pc$^{-3}$ and disk-to-halo ratio by stellar number density of 480:1. We convert the $Gaia$ $G$ band measurements for our sample stars to Johnson-Kron-Cousins $V$ band, compute the $V$-band halo LF, and compare it to previous studies published over many decades that cover a wide range of techniques used. We discuss applications of the LF to the measurement of the luminosity and stellar mass of the Milky Way halo.

astro-ph.GA↗

arXiv2026

A Robust Geometric Distortion Solution for Main Survey Camera of CSST

The advancement in sensitivity and field of view of next-generation wide-field survey telescopes requires astrometric measurements with high precision, even in the presence of significant geometric distortions. To address this challenge, we develop a Weighted Polynomial Distortion Correction in 2-Phase (WPDC-2P) method. This approach enhances stellar cross-matching, incorporates distance-based weighting into the traditional polynomial fitting, and employs a look-up table to absorb the remaining distortion residuals. Validated on simulated data from the Main Survey Camera of the \emph{Chinese Space Station Survey Telescope} (CSST), incorporating geometric distortions up to approximately $200$ pixels, the method achieves astrometric standard deviation ranging from 0.013 to 0.107 pixels (0.03 pixels for the $g$-1 detector) across all 18 detectors. Under extreme crowding conditions (e.g., globular cluster NGC 2298), the astrometric precision for the $g$-1 detector reaches 0.05-pixel level within the central region ($r_d < 4000$), despite a centroiding precision of $\sim$0.04 pixels. When applied to the Beijing-Arizona Sky Survey data, for which the standard pipeline delivers an astrometric uncertainty of $\sim$20 mas, our method reduces the positional scatter to $ σ_{Δα}=5.494$ mas (0.01 pixels) and $ σ_{Δδ}=9.981$ mas (0.02 pixels) using only a weighted 3rd-order polynomial correction. The method has been integrated into the CSST data processing pipeline and is prepared for further refinement using on-orbit calibration data.

astro-ph.IM↗

arXiv2026

The Average Age Map of the Galactic Bulge

The Galactic Bulge, as the center of the Galaxy, is the closest laboratory for studying galaxy formation and evolution. However, its study faces significant challenges due to heavy dust extinction. This paper is devoted to deriving the average age of the Galactic Bulge and investigating its spatial distribution. We utilize a high-precision PSF-fitting photometric catalogue in the $J$ and $K_{\mathrm{s}}$ bands observed by VISTA to study the average stellar ages within the Bulge. Red giant stars are employed as tracers, with their average distances determined using red clump stars as references. The average ages are fitted with stellar models. Our analysis reveals a systematic age gradient across the Galactic Bulge ($2^{\circ} < |b| < 8^{\circ}$). The mean stellar age increases significantly with galactic latitude, shifting from a younger population ($\sim 4.69^{+0.97}_{-0.81}$ Gyr) prevalent near the plane to a predominantly older population ($\sim 10.48^{+0.93}_{-0.85}$ Gyr) at higher latitudes. We hypothesize that the young stellar population at low latitudes is predominantly composed of a pseudo-bulge formed via disk/bar processes (incorporating contributions from recent star-forming activity in the Galactic center), whereas the older stellar population is associated with spheroidal bulges generated through early-stage collapse or accretion of debris from merged dwarf galaxies.

astro-ph.GA↗

arXiv2026

From Performers to Creators: Understanding Retired Women's Perceptions of Technology-Enhanced Dance Performance

Over 100 million retired women in China engage in dance, but their performances are constrained by limited resources and age-related decline. While interactive dance technologies can enhance artistic expression, existing systems are largely inaccessible to non-professional older dancers. This paper explores how interactive dance technologies can be designed with an age-sensitive approach to support retired women in enhancing their stage performance. We conducted two workshops with community-based retired women dancers, employing interactive dance and LLM-powered video generation probes in co-design activities. Findings indicate that age-sensitive adaptations, such as low-barrier keyword input, motion-aligned visual effects, and participatory scaffolds, lowered technical barriers and fostered a sense of authorship. These features enabled retired women to empower their stage, transitioning from passive recipients of stage design to empowered co-creators of performance. We outline design implications for incorporating interactive dance and artificial intelligence-generated content (AIGC) into the cultural practices of retired women, offering broader strategies for age-sensitive creative technologies.

cs.HC↗

arXiv2026

High Power RF Pulse Shaping Tests with NG-LLRF and Cool Copper Collider Prototype Structure

RF pulse modulation techniques are widely applied to shape RF pulses for various types of RF stations of particle accelerators. The amplitude and phase modulations are typically implemented with additional RF components that require drive or control electronics. For the RF system-on-chip (RFSoC) based next generation LLRF (NG-LLRF) platform, which we have developed in the last several years, RF modulation and demodulation are fully implemented in the digital domain. Therefore, arbitrary RF pulse shaping can be realized without any additional analogue components. We performed a range of high-power experiments with the NG-LLRF and a prototype Cool Copper Collider (C3) structure. In this paper, the RF field measured at different stages with different pulse shapes and peak power levels up to 5.4 MW and pulse width about 1 microsecond will be demonstrated and analyzed. The high precision pulse shaping schemes of the NG-LLRF can be applied to realize the phase modulation for a linear accelerator injector, the phase reversal for a pulse compressor, or the modulation required to compensate for the beam loading effect.

physics.acc-ph↗