arXiv subjects

Yan Liang

Publications and source records attributed to Yan Liang.

At least 19 recordsLinked to original sources

arXiv2026

Benchmarking Machine Learning Emulators of Stellar Evolution for Precision Asteroseismology

Fast and accurate stellar evolution emulators---surrogate models that approximate expensive simulation outputs with machine learning (ML)---are powerful tools for modern stellar characterization, hierarchical inference, and population synthesis. We analyze the grid density required for reliable emulation by training ML algorithms on main-sequence models with masses M=[0.7,1.2] solar masses. This range is challenging to emulate due to rapidly varying evolutionary behavior caused by the radiative-to-convective core transition, as well as the requirement to match the part-per-thousand seismic precision that has been delivered for such stars from the NASA Kepler mission. Generating grids from analytical models, as well as MESA, YREC, MIST, and ASTEC, we compare linear interpolation, k-nearest neighbors, random forests, and neural networks (NNs) in interpolating the stellar observables: T_eff, L, Delta nu, and nu_max. While NNs outperform other methods, sparse grids induce localized failures in the core-transition region, resulting in unstable derivatives, ensemble disagreement, and fragmented posterior distributions during inference. Performance gains from denser grids are non-uniform, suggesting that adaptive grid generation should be favored over uniform refinement. Finally, we show that NN ensembles allow for localized uncertainty propagation, more accurately reflecting emulator reliability across parameter space than global uncertainty estimates. As we consider only the two-dimensional case of varying only stellar mass and age along the main sequence, these results represent a lower bound on the challenge in emulating stellar evolution simulations for precision asteroseismology.

astro-ph.SR↗

arXiv2026

AESTRA II: Generative Spectral Modeling of the Sun as a Star for Precise Radial Velocities

The detection of Earth analogs with extreme-precision radial velocities (EPRVs) is limited by spectral variability from stellar activity, telluric absorption, and instrumental systematics. We apply AESTRA, a generative spectrum modeling framework, to NEID Sun-as-a-star observations. AESTRA empirically decomposes the spectra into stellar line-shape variability, micro-telluric absorption, and continuum variability without external atmospheric or stellar templates. After removing the learned telluric and continuum components, we train a low-dimensional representation of the spectrum to infer activity-driven apparent RVs jointly with candidate Doppler signals. We evaluate the method with 500 single-planet injection-recovery tests spanning periods of 2.5 to 400 days and semi-amplitudes of K = 0.1 to 0.7 m s^-1, calibrating the detection criterion to yield zero spurious detections. At this matched confidence level, AESTRA recovers 238 injected planets, including 13 with K < 0.3 m s^-1, whereas traditional CCF-based activity-indicator detrending recovers 9 planets and none below K = 0.5 m s^-1.

astro-ph.EP↗

arXiv2026

Coincidence-pumping upconversion detector based on passively synchronized fiber laser system

We experimentally demonstrated a high-performance frequency upconversion detector for telecom-band photons based on a passively synchronized fiber laser system. The involved coincidence pumping technique enabled to spectrally convert the pulsed infrared photons into the visible regime with a conversion efficiency of 72\%. The overall detection efficiency of the upconversion detector reached to 30\% with a low noise equivalent power of $3\times10^{-17}\ \text{W/Hz}^{1/2}$. In contrast to previous demonstrations, the whole upconversion detection system was constructed in an all-polarization-maintaining fiber structure, thus favoring substantial improvement of compactness and robustness. Moreover, the long-term stability was manifested by at least ten-hour operation with a relative fluctuation of count rates as small as 0.26\%. The achieved features here would be desirable in many practical applications requiring efficient and robust coherent manipulation of pulsed optical fields by nonlinear frequency conversion.

physics.optics↗

arXiv2026

Mid-Infrared Single-Photon Edge Enhanced Imaging based on Nonlinear Vortex Filtering

Edge enhanced imaging via the spiral phase contrast enables to reveal the phase or amplitude gradients of a target, which has been proved useful in feature recognition, machine vision, and object identification. A long quest is to extend the operation wavelength into the mid-infrared (MIR) region, as highly demanded in various fields including infrared sensing, astronomic observation, and biomedical diagnosis. Here, we demonstrated ultra-sensitive MIR imaging at the single-photon level based on nonlinear frequency upconversion, where the spectrally converted replica of the MIR object image at 3070 nm was captured by a silicon electron multiplying charged coupled device. The imaging sensitivity was significantly improved by the coincidence pulsed pumping with a spectro-temporal optimization. Furthermore, the edge enhancement has been realized by imprinting the spiral phase pattern of the pump onto the upconverted field at the Fourier plane within the nonlinear crystal. Such a nonlinear spatial filter not only provided an effective way to implement the required high-fidelity vortex screening in the edge enhanced detection, but also rendered the MIR illumination into a visible image in an efficient and low-noise fashion. The presented system for MIR edge enhanced imaging might facilitate immediate applications in label-free histopathological diagnosis and non-destructive defect inspection.

physics.optics↗

arXiv2026

Single-photon time-stretch infrared spectroscopy

Sensitive mid-infrared (MIR) spectroscopy is highly demanded in various fields ranging from industrial inspection, biomedical diagnosis to astronomical observation. However, the detection sensitivity of conventional MIR spectrometers has been severely limited by excessive noises for existing infrared sensors, which hinders widespread use in photon-scarce scenarios. Here, we devise and implement a broadband MIR single-photon time-stretch spectrometer based on high-fidelity spectral upconversion and time-correlated coincidence counting. Specifically, a nanophotonic supercontinuum illumination covering 2.4-4.2 $\mu$m is nonlinearly converted to the near-infrared band, where low-loss single-mode fiber and high-performance silicon detector can be leveraged to facilitate dispersive operation and sensitive detection, respectively. The arrival time for the dispersed upconversion photons is precisely registered with a low-timing-jitter photon counter, which enables us to obtain a high spectral resolution about 0.5 cm$^{-1}$ under a low-light-level illumination down to 0.14 photons/nm/pulse. In comparison to previous MIR upconversion spectrometers, the presented time-stretch architecture favors single-pixel simplicity and high-throughput acquisition for the single-photon spectral measurement. The achieved MIR spectroscopic features of broadband spectral coverage, sub-wavenumber resolution, single-photon sensitivity, and room-temperature operation would stimulate immediate applications in material and life sciences.

physics.optics↗

arXiv2026

Wide-field mid-infrared hyperspectral imaging beyond video rate

Mid-infrared hyperspectral imaging has become an indispensable tool to spatially resolve chemical information in a wide variety of samples. However, acquiring three-dimensional data cubes is typically time-consuming due to the limited speed of raster scanning or wavelength tuning, which impedes real-time visualization with high spatial definition across broad spectral bands. Here, we devise and implement a high-speed, wide-field mid-infrared hyperspectral imaging system relying on broadband parametric upconversion of high-brightness supercontinuum illumination at the Fourier plane. The upconverted replica is spectrally decomposed by a rapid acousto-optic tunable filter, which records high-definition monochromatic images at a frame rate of 10 kHz based on a megapixel silicon camera. Consequently, the hyperspectral imager allows us to acquire 100 spectral bands over 2600-4085 cm$^{-1}$ in 10 ms, corresponding to a refreshing rate of 100 Hz. Moreover, the angular dependence of phase matching in the image upconversion is leveraged to realize snapshot operation with spatial multiplexing for multiple spectral channels, which may further boost the spectral imaging rate. The high acquisition rate, wide-field operation, and broadband spectral coverage could open new possibilities for high-throughput characterization of transient processes in material and life sciences.

physics.optics↗

arXiv2026

Highly-sensitive mid-infrared upconversion detection based on external-cavity pump enhancement

Sensitive mid-infrared (MIR) detection is highly demanded in various applications, ranging from remote sensing, infrared surveillance, environmental monitoring to industrial inspection. Among others, upconversion infrared detectors have recently attracted increasing attention due to the advantageous features of high sensitivity, fast response, and room-temperature operation. However, it remains challenging to realize high-performance passive MIR sensing due to the stringent requirement of high-power continuous-wave pumping. Here, we propose and implement a high-efficiency and low-noise MIR upconversion detection system based on pumping enhancement via a low-loss optical cavity. Specifically, a single-longitudinal-mode pump at 1064 nm is significantly enhanced by a factor of 36, thus allowing for a peak conversion efficiency of up to 22\% at an intra-cavity average power of 55 W. The corresponding noise equivalent power is achieved as low as 0.3 fW/Hz$^{1/2}$, which indicates at least a ten-fold improvement over previous results. Notably, the involved single-frequency pumping would facilitate high-fidelity spectral mapping, which is particularly attractive for high-precision MIR upconversion spectroscopy in photon-starved scenarios.

physics.optics↗

arXiv2026

Mid-Infrared Single-Photon Compressive Spectroscopy

Sensitive mid-infrared (MIR) spectroscopy plays an indispensable role in various photon-starved conditions. However, the detection sensitivity of conventional MIR spectrometers is severely limited by excessive noises of the involved infrared sensors, especially for multi-pixel arrays in parallel spectral acquisition. Here, we devise and implement an ultra-sensitive MIR single-pixel spectrometer, which relies on high-fidelity spectral upconversion and wavelength-encoding compressive measurement. Specifically, a MIR nanophotonic supercontinuum from 3.1 to 3.9 $\mu$m is nonlinearly converted to the near-infrared band via synchronous chirped-pulse pumping, which facilitates both the precise spectral mapping and sensitive upconversion detection. The upconverted signal is then spatially dispersed onto a programmable digital micromirror device, before being registered by a single-element silicon detector. Consequently, the spectral information can be deciphered from the correlation between encoded patterns and recorded measurements, which results in a spectral resolution of 0.5 cm$^{-1}$ under an illumination flux down to 0.01 photons/nm/pulse. Moreover, we demonstrate faithful reconstructions at sub-Nyquist sampling rates by using the compressive sensing algorithm, which leads to a 95\% reduction in data acquisition time. The presented single-pixel computational spectrometer features wavelength multiplexing, high throughput, and efficient sampling, which thus paves a new way for sensitive and fast spectroscopic analysis at the single-photon level.

physics.optics↗

arXiv2026

High-resolution mid-infrared single-photon upconversion ranging

Single-photon laser ranging has widespread applications in remote sensing and target recognition. However, highly-sensitive light detection and ranging (LiDAR) has long been restricted in visible or near-infrared bands. An appealing quest is to extend the operation wavelength into the mid-infrared (MIR) region, which calls for an infrared photon counting system at high detection sensitivity and precise temporal resolution. Here, we devise and demonstrate a MIR upconversion LiDAR based on nonlinear asynchronous optical sampling. Specifically, the infrared probe is interrogated in a nonlinear crystal by a train of pump pulses at a slightly different repetition rate, which favors for a temporal optical scanning at a picosecond timing resolution and a kilohertz refreshing rate over $\sim$50 ns. Moreover, the cross-correlation upconversion trace is temporally stretched by a factor of 2$\times$10$^4$, which can thus be recorded by a low-bandwidth silicon detector. In combination with time-correlated photon-counting technique, the achieved effective resolution is about two orders of magnitude better than the timing jitter of the detector itself, which facilitates a ranging precision of 4 $\mu$m under a low detected flux of 8$\times$10$^{-5}$ photons per pulse. The presented MIR time-of-flight range finder is featured with single-photon sensitivity and high positioning resolution, which would be particularly useful in infrared sensing and imaging in photon-starved scenarios.

physics.optics↗

arXiv2026

Mid-infrared single-pixel imaging via two-photon optical encoding

Mid-infrared (MIR) imaging offers powerful capabilities for label-free chemical analysis, yet its practical deployment remains hindered by the high cost and cryogenic complexity of conventional cameras. Two-photon absorption (TPA) provides a promising route to room-temperature MIR detection, but existing TPA imagers based on raster scanning or array detectors are constrained by slow acquisition speed or limited detection sensitivity. Here we present a scanning-free MIR single-pixel imaging approach based on non-degenerate TPA in a silicon detector. The involved spatial encoding is realized by a near-infrared structured pump with a resolution of 7 $\mu$m, thus allowing high-fidelity MIR optical modulation through the phase-matching-free nonlinear interaction. Consequently, the spatially modulated TPA response is intrinsically integrated in the single-element photodetector, which favors computational reconstruction of the impinging MIR image by correlating measured intensities and predetermined patterns. Notably, the use of advanced algorithms of compressed sensing and deep learning facilitate image recovery under sub-Nyquist sampling with a compression ratio of 10\% and photon-starved illumination with an incident light flux of 0.5 pJ/pulse. Furthermore, a multispectral imaging over 2.5-3.8 $\mu$m is manifested for chemical discrimination of plastic films. The presented architecture would offer a broadband and sensitive alternative for MIR imaging in various fields ranging from biomedical diagnostics to material inspection.

physics.optics↗

arXiv2026

From Compression to Accountability: Harmless Copyright Protection for Dataset Distillation

Large-scale datasets have been a key driving force behind the rapid progress of deep learning, but their storage, computational, and energy costs have become increasingly prohibitive. Dataset distillation (DD) mitigates this problem by synthesizing compact yet informative datasets, thereby enabling efficient model training and storage. However, the ease of copying and distributing distilled datasets introduces serious risks of copyright infringement and data leakage. Existing protection methods are primarily designed for raw datasets rather than distilled datasets, and typically rely on backdoor-triggered malicious behaviors, which may raise security concerns. In this paper, we observe that deep neural networks tend to memorize subpopulation distributions during training, resulting in a systematic prediction bias, where models perform better on samples aligned with memorized subpopulations. Motivated by this observation, we propose SubPopMark, a harmless subpopulation-driven protection framework for distilled datasets. SubPopMark consists of two stages. First, the Copyright Verification Marker(CVM) optimization stage injects a class-consistent subpopulation bias while preserving the original optimization trajectory. Second, the User-Specific Tracing Marker (USTM) optimization stage further introduces user-distinguishable perturbations into the CVM-augmented data. To enable black-box verification and tracing, we construct a reference behavior bank by collecting model outputs over carefully designed test sets that cover both standard and subpopulation-shifted data distributions. The provenance of a suspicious model is then inferred by comparing its output behavior signature with the bank and identifying the most consistent reference behavior pattern.

cs.CR↗

arXiv2026

EMA: Efficient Model Adaptation for Learning-based Systems

Machine learning (ML) is increasingly applied to optimize system performance in tasks such as resource management and network simulation. Unlike traditional ML tasks (e.g., image classification), networked systems often operate in heterogeneous, long-running, and dynamic environment states, where input conditions (e.g., network loads) and operational objectives can shift over time and across settings. Existing learning-based systems offer little support for adaptation, resulting in costly model training, extensive data collection, degraded system performance, and slow responsiveness. This paper presents EMA, the first model adaptation system supporting learning-based systems to adapt to evolving environments with minimal operational overhead. EMA takes a system-driven, data-centric approach that accommodates diverse system and model designs while addressing two key deployment challenges. First, it reduces expensive model training by introducing state transformers that align the input state of a new environment with previously similar states, allowing models to warm-start adaptation. Second, it addresses the often-overlooked yet costly process of data labeling--collecting ground truth for exploring and training on various system decisions--by prioritizing labeling high-utility data while balancing the tradeoff between training and labeling cost. Evaluations on eight representative learning-based systems show that EMA reduces adaptation costs (e.g., GPU training time) by 14.9-42.4% while improving system performance (e.g., network throughput) by 6.9-31.3%.

cs.LG↗

arXiv2026

HexiSeq: Accommodating Long Context Training of LLMs over Heterogeneous Hardware

Long-context training of large language models (LLMs) is commonly distributed with Context Parallelism (CP) and Head Parallelism (HP), but existing training systems largely assume homogeneous GPU meshes. This paper extends CP and HP to heterogeneous GPU clusters with mixed GPU models and non-uniform network bandwidths, a common setting in production training. We introduce HexiSeq, a system that supports fully asymmetric CP--HP partitioning by assigning sequence shards and attention heads according to device compute, memory, and communication capabilities. We formalize heterogeneous CP--HP allocation as a constrained optimization problem and develop an efficient hierarchical scheduler for finding optimal schedules. We evaluate HexiSeq against state-of-the-art CP and HP baselines on both real and simulated heterogeneous clusters. Across models from 3B to 70B parameters and context lengths up to one million tokens, HexiSeq improves throughput by $1.11\times$ on average and up to $1.19\times$ on mixed H100--A100 testbeds, and by $1.36\times$ on average and up to $1.72\times$ in simulations with 32--128 GPUs spanning up to four GPU models. On FLOP-comparable pairs against homogeneous clusters, HexiSeq reaches throughput close to the strongest homogeneous baseline, showing that heterogeneous clusters can be used efficiently for long-context LLM training.

cs.DC↗

arXiv2026

Trust the Unreliability: Inward Backward Dynamic Unreliability Driven Coreset Selection for Medical Image Classification

Efficiently managing and utilizing large-scale medical imaging datasets with limited resources presents significant challenges. While coreset selection helps reduce computational costs, its effectiveness in medical data remains limited due to inherent complexity, such as large intra-class variation and high inter-class similarity. To address this, we revisit the training process and observe that neural networks consistently produce stable confidence predictions and better remember samples near class centers in training. However, concentrating on these samples may complicate the modeling of decision boundaries. Hence, we argue that the more unreliable samples are, in fact, the more informative in helping build the decision boundary. Based on this, we propose the Dynamic Unreliability-Driven Coreset Selection(DUCS) strategy. Specifically, we introduce an inward-backward unreliability assessment perspective: 1) Inward Self-Awareness: The model introspects its behavior by analyzing the evolution of confidence during training, thereby quantifying uncertainty of each sample. 2) Backward Memory Tracking: The model reflects on its training tracking by tracking the frequency of forgetting samples, thus evaluating its retention ability for each sample. Next, we select unreliable samples that exhibit substantial confidence fluctuations and are repeatedly forgotten during training. This selection process ensures that the chosen samples are near the decision boundary, thereby aiding the model in refining the boundary. Extensive experiments on public medical datasets demonstrate our superior performance compared to state-of-the-art(SOTA) methods, particularly at high compression rates.

cs.CV↗

arXiv2025

Competing Lattice and Defect Dynamics Govern Terahertz-Induced Ferroelectricity in Quantum Paraelectric SrTiO$_3$

Intense terahertz (THz) pulses induce transient inversion-symmetry breaking in quantum paraelectric SrTiO$_3$, yet the underlying mechanism remains controversial. Using fields up to $\sim$1.1 MV/cm, we reveal spatially inhomogeneous THz-field-induced second harmonic generation (TFISH) governed by competing lattice and defect dynamics. Short-lived coherent antiferrodistortive (AFD) modes suppress dipole correlations within $\sim$5 ps, while heavily damped soft/AFD modes and a defect-induced low-frequency mode ($\sim$0.1-0.3 THz) jointly prevent long-range ferroelectric coherence in oxygen-vacancy-rich regions. Collective modes manifested by oscillatory TFISH components exhibit softening followed by hardening below a critical temperature $T^*\simeq$28 K, confirming transient ferroelectric order where defects are sparse. These results reconcile conflicting interpretations, establish defect-mediated competition as a central regulator of light-induced ferroelectricity, and open routes to ultrafast control of quantum materials.

cond-mat.mtrl-sci↗

arXiv2025

Reconstructive comb spectroscopy: A single-pixel detection paradigm beyond dual-comb limitations

Frequency comb spectroscopy has revolutionized broadband molecular fingerprinting with mode-defined resolution. While dual-comb spectroscopy stands as a dominant paradigm for high-resolution measurements, it relies on mutually coherent dual combs, and its applicability to non-cooperative sensing is limited by the requirement for phase-sensitive detection and controlled optical returns. Here, we introduce reconstructive comb spectroscopy, a fundamentally different paradigm that eliminates these constraints. By integrating a mode-programmable optical comb with a computational sensing scheme based on single-pixel detection, our method achieves picometer-level spectral resolution over a 10-nm (1.27-THz) instantaneous bandwidth, with single-photon sensitivity down to 10^-4 photons per pulse, and compressed spectral acquisition at 2.5% sampling while maintaining reconstruction errors below 10%. We demonstrate robust performance through scattering media and from non-cooperative targets. These capabilities establish reconstructive comb spectroscopy as a new platform for gas sensing, with broad applicability in remote atmospheric monitoring, industrial leak detection, and standoff chemical-threat identification.

physics.optics↗

arXiv2025

The Impact of Orbital Anisotropy Assumptions in Lensing-Dynamics Modeling

We investigate potential systematic biases introduced by assumptions regarding stellar orbital anisotropy in joint lensing-dynamics modeling. Our study employs the massive early-type galaxies from the TNG100 simulation at redshifts z = 0.2, 0.5, and 0.7. Based on the simulated galaxies, we generate a self-consistent mock dataset containing both lensing and stellar kinematic observables. This is achieved through taking the potential composed of both dark matter and baryons of the simulated galaxies, plus the radial variation of the stellar orbit anisotropy depicted by a logistic function. By integrating constraints from both lensing and stellar kinematics, we separate the contributions of stars and dark matter inside the galaxies. Under three commonly adopted stellar anisotropy assumptions (isotropic orbits, constant anisotropy, and the Osipkov-Merritt profile), the model inferences suggest that the systematic biases in the total stellar mass and central dark matter fraction are not significant. Specifically, the total stellar mass on average is underestimated by less than $0.03\pm0.10$ $\rm dex$ while the dark matter fraction experiences only a statistically insignificant increase of less than $2\%\pm10\%$ at the population level. The dark matter inner density slope in our tests is over-predicted by $0.15\pm0.2$. Additionally, these lacks of significant biases are insensitive to the discrepancies between the assumed anisotropy in modeling and the ground truth orbital anisotropy of mock sample. Our results suggest that conventional assumptions regarding orbital anisotropy, such as an isotropic profile or the Osipkov-Merritt model, would not introduce a significant systematic bias when inferring galaxy mass density distribution at the population level.

astro-ph.GA↗

arXiv2025

LexInstructEval: Lexical Instruction Following Evaluation for Large Language Models

The ability of Large Language Models (LLMs) to precisely follow complex and fine-grained lexical instructions is a cornerstone of their utility and controllability. However, evaluating this capability remains a significant challenge. Current methods either rely on subjective and costly human evaluation or on automated LLM-as-a-judge systems, which suffer from inherent biases and unreliability. Existing programmatic benchmarks, while objective, often lack the expressiveness to test intricate, compositional constraints at a granular level. To address these limitations, we introduce LexInstructEval, a new benchmark and evaluation framework for fine-grained lexical instruction following. Our framework is built upon a formal, rule-based grammar that deconstructs complex instructions into a canonical triplet. This grammar enables the systematic generation of a diverse dataset through a multi-stage, human-in-the-loop pipeline and facilitates objective verification via a transparent, programmatic engine. We release our dataset and open-source evaluation tools to facilitate further research into the controllability and reliability of LLMs.

cs.CL↗