arXiv subjects

Hao Liang

Publications and source records attributed to Hao Liang.

At least 19 recordsLinked to original sources

arXiv2026

Hidden Perturbative Structure in Attosecond Streaking

Attosecond streaking is inherently multiphoton, yet we uncover a hidden perturbative structure revealing the underlying continuum--continuum angular-momentum propensity. By filtering the inversion-odd response in helium streaking data at the IR fundamental, we isolate an angular distribution dominated by the $P_1$ and $P_3$ Legendre components, whose complex coefficients yield an effective $s/d$ two-photon amplitude ratio. Multichannel, two-electron time-dependent Schrödinger equation calculations demonstrate that the characteristic Fano propensity survives electron correlation. Incorporating the experimental XUV attochirp reproduces the measured emission-edge trend, while near-threshold calculations predict resonant amplitude and phase signatures associated with the bound He $np$ Rydberg series. This symmetry- and frequency-filtering approach offers a general spectroscopic framework with potential applications to atomic, molecular, and solid-state systems where attosecond streaking is established.

physics.atom-ph↗

arXiv2026

SpatialScout: Can VLM Agents Actively Explore and Map 3D Scenes?

Spatial intelligence requires agents to actively acquire visual evidence and integrate observations into a coherent representation of the three-dimensional world. Yet spatial question answering offers limited evidence of whether models can construct accurate metric maps through autonomous exploration. We introduce SpatialScout, a benchmark for evaluating vision-language model (VLM) agents on active object-level 3D mapping. SpatialScout comprises 80 diverse indoor scenes containing 3,140 target objects. Given an object inventory, RGB observations, camera calibration, and camera poses, agents control a camera to explore each scene and estimate object centers in a shared coordinate frame. The task requires coordinating viewpoint selection, cross-view association, geometric estimation, and map revision under bounded interaction budgets. Evaluation jointly measures localization accuracy and target coverage through object-scale overlap scores that penalize missing predictions. Experiments with frontier VLMs reveal substantial limitations: the strongest evaluated model, Claude Opus 5.5, achieves a mean overlap score of 41.71%, with 47.23% of targets exceeding an intersection-over-union (IoU) threshold of 0.5 and only 3.25% exceeding 0.9. These results reveal a substantial gap between recovering coarse object layouts and constructing complete, precise metric maps. SpatialScout provides a quantitative testbed for developing agents that build reliable 3D spatial representations through active perception. Code is available at https://github.com/JesseZZZZZ/SpatialScout.

cs.CV↗

arXiv2026

Highly integrated quantum key distribution transmitter enabled by silicon photonics

Quantum key distribution (QKD) provides information-theoretic security independent of computational assumptions, yet the bulk and cost of current systems hinder large-scale deployment. Although integrated photonic technologies have enabled highly integrated QKD chips, practical QKD transmitters are still predominantly implemented as rack-mounted systems. Here, we demonstrate a highly integrated standalone QKD transmitter that integrates all essential functionalities required for practical QKD operation within a compact platform built around a silicon photonic encoding chip. The transmitter occupies only $167 \times 56 \times 21~\mathrm{mm}^3$ ($\sim 0.2~\mathrm{dm}^3$), comparable in size to a half-height, half-width PCIe card and more than 30 times smaller in volume than a conventional 1U rack-mounted system. Paired with a conventional discrete-component receiver, it achieves a secure key rate of 219.1 kbps over 51.3 km of standard single-mode fiber, with performance comparable to that of conventional discrete-component implementations. This work bridges the gap between photonic chip integration and deployable QKD hardware, marking an important step toward transitioning QKD transmitters from conventional rack-mounted equipment to compact board-level platforms.

quant-ph↗

arXiv2026

Differential Privacy as a Perk: Federated Learning over Multiple-Access Fading Channels with a Multi-Antenna Base Station

Federated Learning (FL) is a distributed learning paradigm that preserves privacy by eliminating the need to exchange raw data during training. In its prototypical edge instantiation with underlying wireless transmissions enabled by analog over-the-air computing (AirComp), referred to as \emph{over-the-air FL (AirFL)}, the inherent channel noise plays a unique role of \emph{frenemy} in the sense that it degrades training due to noisy global aggregation while providing a natural source of randomness for privacy-preserving mechanisms, formally quantified by \emph{differential privacy (DP)}. It remains, nevertheless, challenging to effectively harness such channel impairments, as prior arts, under assumptions of either simple channel models or restricted types of loss functions, mostly considering (local) DP enhancement with a single-round or non-convergent bound on privacy loss. In this paper, we study AirFL over multiple-access fading channels with a multi-antenna base station (BS) subject to user-level DP requirements. Despite a recent study, which claimed in similar settings that artificial noise (AN) must be injected to ensure DP in general, we demonstrate, on the contrary, that DP can be gained as a \emph{perk} even \emph{without} employing any AN. Specifically, we derive a novel bound on DP that converges under general bounded-domain assumptions on model parameters, along with a convergence bound with general smooth and non-convex loss functions. Next, we optimize over receive beamforming and power allocations to characterize the optimal convergence-privacy trade-offs, which also reveal explicit conditions in which DP is achievable without compromising training. Finally, our theoretical findings are validated by extensive numerical results.

cs.LG↗

arXiv2026

Time performance of USTC-IME LGAD under synchrotron light source focused X-ray

The time performance of Low Gain Avalanche Diodes (LGADs), designed by the University of Science and Technology of China (USTC) and fabricated by the Institute of Microelectronics of Chinese Academy of Sciences (IME), was characterized at the Shanghai Synchrotron Radiation Facility (SSRF). The experiment was conducted at the BL16B1 beamline, which delivers a focused X-ray beam with a diameter of 500 \textmu\text{m}, a repetition period of 2 ns, and a photon energy of 10 keV. Using a fast oscilloscope, waveforms containing continuous signal pulses were recorded within a 50 ns time window. The LGAD under test successfully resolved the 2 ns period of the SSRF. To mitigate pile-up effects and extract pulse-by-pulse information from the acquired waveforms, a waveform-level global template fitting method was employed. The time resolution was then estimated using a combined profile likelihood approach, with a restricted maximum-likelihood (REML) correction accounting for the per-waveform degrees of freedom, yielding a value of 136.8 $\pm$ 0.15 (stat.) $\pm$ 11.9 (syst.) ps. The complete reconstruction and fitting chain was validated end-to-end using simulated waveforms in which the timing deviations of individual large pulses were drawn from Gaussian distributions with fixed standard deviations. The effect of random photon absorption depth on the time resolution of LGADs was studied through dedicated simulations.

physics.ins-det↗

arXiv2026

K12-KGraph: A Curriculum-Aligned Knowledge Graph for Benchmarking and Training Educational LLMs

Large language models are increasingly used in K-12 education, but existing benchmarks mainly test exam question answering rather than understanding how curriculum knowledge is structured and visually presented. We call this capability curriculum cognition. It covers prerequisite chains, concept taxonomies, experiment-concept links, pedagogical sequencing, and visual grounding. We introduce K12-KGraph, a curriculum-aligned knowledge graph extracted from official People's Education Press textbooks in mathematics, physics, chemistry, and biology across primary, middle, and high school. It contains nine node types and fourteen relation types covering curriculum structure and visual grounding. From this graph, we derive K12-Bench, a 23,640-question multi-select benchmark with five task families: Ground, Prereq, Neighbor, Evidence, and Locate. We also build K12-Train, a graph-guided supervised fine-tuning corpus of 7,335 samples, including 2,267 text-only QA pairs and 5,068 multimodal VQA pairs. On K12-Bench, Gemini-3-Flash achieves only 57 percent exact match and Gemma-4-31B-IT reaches 46 percent, with Prereq and Neighbor being the hardest tasks. Our training experiments show that domain-specific supervision can reduce this gap. Under a matched 2,300-sample budget, K12-Train-Text consistently outperforms equally sized subsets of eight mainstream instruction-tuning corpora on GaokaoBench and EduEval. For vision-language models, K12-Train-Full achieves the best overall results on Gaokao-MM, MDK12-medium, and K12Vista among all compared training configurations, despite using fewer samples than the full DataFlow and WizardLM baselines. It also surpasses both text-only and multimodal-only variants, showing that textual and visual supervision are complementary. We release the graph, benchmark, training data, and complete construction pipeline.

cs.CL↗

arXiv2026

Why GRPO Needs Normalization: A Local-Curvature Perspective on Adaptive Gradients

Reinforcement learning (RL) has become a key driver of language model reasoning. Among RL algorithms, Group Relative Policy Optimization (GRPO) is the de facto standard, avoiding the need for a critic by using per-prompt baselines and variance normalization. Yet why and when this normalization helps remains unclear. In this work, we answer both questions through the lens of local curvature of the sequence-level policy gradient: standard deviation normalization implements an adaptive gradient. Theoretically, we prove that under mild conditions GRPO attains an improved convergence rate over unnormalized REINFORCE, with gains characterized by the average within-prompt reward standard deviation across prompts and iterations. We further introduce IS-GRPO, an importance-sampling variant of GRPO whose expected update remains aligned with the full gradient, and prove a convergence guarantee of the same form as REINFORCE's, which is tighter in practice during real training runs by a factor we measure. Empirically, on GSM8K and MATH we validate the curvature--variance link and measure the quantities that govern our bounds along real training runs. At 1.5B scale, per-prompt normalization outperforms both unnormalized and globally normalized baselines, with gains that emerge in a middle phase of training where per-prompt variances become heterogeneous. At 7B scale, the normalization schemes become statistically indistinguishable while IS-GRPO retains a small lead, in line with the theory's prediction that the gains require heterogeneity.

cs.LG↗

arXiv2026

DataFlex: A Unified Framework for Data-Centric Dynamic Training of Large Language Models

Data-centric training has emerged as a promising direction for improving large language models (LLMs) by optimizing not only model parameters but also the selection, composition, and weighting of training data during optimization. However, existing approaches to data selection, data mixture optimization, and data reweighting are often developed in isolated codebases with inconsistent interfaces, hindering reproducibility, fair comparison, and practical integration. In this paper, we present DataFlex, a unified data-centric dynamic training framework built upon LLaMA-Factory. DataFlex supports three major paradigms of dynamic data optimization: sample selection, domain mixture adjustment, and sample reweighting, while remaining fully compatible with the original training workflow. It provides extensible trainer abstractions and modular components, enabling a drop-in replacement for standard LLM training, and unifies key model-dependent operations such as embedding extraction, inference, and gradient computation, with support for large-scale settings including DeepSpeed ZeRO-3. We conduct comprehensive experiments across multiple data-centric methods. Dynamic data selection consistently outperforms static full-data training on MMLU across both Mistral-7B and Llama-3.2-3B. For data mixture, DoReMi and ODM improve both MMLU accuracy and corpus-level perplexity over default proportions when pretraining Qwen2.5-1.5B on SlimPajama at 6B and 30B token scales. DataFlex also achieves consistent runtime improvements over original implementations. These results demonstrate that DataFlex provides an effective, efficient, and reproducible infrastructure for data-centric dynamic training of LLMs.

cs.LG↗

arXiv2026

An upper-symbol approach to $Φ^4_d$ limits of Bose gases

We study the high-temperature derivation of $Φ^4_d$ measures from grand-canonical Bose gases on two- and three-dimensional tori. Building on the upper-symbol variational approach of the author's earlier work, we use relative entropy to give a simple, unified proof of relative free-energy convergence and Hilbert--Schmidt convergence of all fixed-order rescaled reduced density matrices to their Hartree counterparts. For interaction ranges $\varepsilon=λ^η$, where $λ$ is the inverse temperature, these results hold for every fixed $0<η<1/2$ in two dimensions and $0<η<1/26$ in three. In two dimensions, we improve the result of Jougla and Rougerie (2026) and the local $Φ^4_2$ derivation of Fröhlich, Knowles, Schlein and Sohinger (2025). Combined with classical approximation, the two-dimensional result gives the first local $Φ^4_2$ derivation valid for every fixed exponent $0<η<1/2$, allowing polynomial exponents arbitrarily close to the diluteness threshold from below. In three dimensions, we use the identification of the local measure and the stationary SPDE estimates of Nam, R.~Zhu and X.~Zhu (2025). From these estimates we strengthen their classical correlation convergence to the Hilbert--Schmidt norm. This yields the local limit at every fixed order in the same polynomial range.

math-ph↗

arXiv2026

Nonlocal Cubic Density Gibbs Measures from Bosonic Gibbs States with Three-Body Interactions

We study the high-temperature mean-field limit of grand-canonical bosonic Gibbs states on the torus with renormalized nonlocal three-body interactions. In dimensions two and three, we construct the limiting nonlinear classical Gibbs measure and prove convergence of the relative free energy and of the reduced density matrices of every fixed order; in three dimensions, a smallness condition on the interaction is imposed. The proof combines the density-channel representation of the interaction with a coherent-state variational method based on the upper-symbol representation of the free Gibbs state. The same framework also contains, as a special case, the homogeneous positive-type model studied by Lewin, Nam, and Rougerie (2021).

math-ph↗

arXiv2026

Cut pairs and Morse splitting of finitely generated groups

Bowditch's theorem for hyperbolic groups establishes a fundamental correspondence between splittings over two-ended subgroups and the existence of local cut points in the Gromov boundary. While analogous results have been obtained for CAT(0) and relatively hyperbolic groups, no general theorem of this type exists for arbitrary finitely generated groups. The Morse boundary, introduced by Charney-Sultan and extended by Cordes, provides a quasi-isometry invariant boundary for any finitely generated group that naturally generalizes the Gromov boundary. In this paper, we prove that a splitting of a finitely generated group with connected Morse boundary over a two-ended Morse subgroup gives rise to a separating pair of points in the Morse boundary.

math.GT↗

arXiv2026

OmniEdu: Open Foundation Models for Learning and Teaching

Educational foundation models must solve problems, understand curriculum structure, diagnose learner difficulties, and provide appropriate instructional support. Existing educational language models often focus on either problem solving or tutoring, with training mixtures organized by source or task rather than capability. We present OmniEdu, an open family of foundation models for K-12 learning and teaching. Its instruction-tuning corpus combines over 100 educational resources and general instruction sources, organized around four capabilities: subject competence, curriculum grounding, diagnostic reasoning, and pedagogical action and scaffolding. Our pipeline integrates deterministic cleaning, semantic auditing and rewriting, task-specific quality scoring, token-budgeted diversity selection, and pedagogical instruction assignment. It yields 69,999 examples and 15.96M supervised response tokens, including 60,951 education-specific examples. We fine-tune 4B, 9B, and 27B models and evaluate curriculum grounding, K-12 problem solving, and pedagogical tutoring, alongside general capability. Education-oriented tuning consistently improves all three educational benchmark groups across model scales. OmniEdu-27B achieves 63.12% EM and 76.69% F1 on K12-Bench, 85.89% on MathFish, 86.95% on EDUMATH, and 78.74% in MathTutorBench's Scaffold setting. It also achieves the highest Teaching average on LongTutor among the evaluated models, at 3.02. These results demonstrate the value of curated, capability-balanced supervision for adapting general language models to educational tasks spanning problem solving, curriculum understanding, and instructional support.

cs.CL↗

arXiv2026

A High Gain Preamplifier Board for Low Charge Semiconductor Detectors

We present a high-gain, low-noise preamplifier board designed for reading out low-charge semiconductor detectors--specifically Low-Gain Avalanche Detectors (LGADs) and three-dimensional (3D) silicon sensors--in high-energy physics applications. The circuit employs a three-stage architecture featuring a discrete SiGe:C bipolar junction transistor (BJT)-based transimpedance amplifier (TIA) front-end followed by two resistive feedback amplification stages using LTC6431 chips. This configuration achieves a charge gain of 115.27 mV ns/fC with excellent linearity over an input range of 0.5 fC to 25 fC and a wide bandwidth spanning from 34.0 MHz to 594.3 MHz. Experimental evaluations coupled with LGAD detectors demonstrate a time resolution of 36.41 ps and an equivalent noise charge (ENC) of 0.18 fC at 20 °C. Furthermore, tests with conventional PIN sensors without a gain layer yield a timing resolution of 76.10 ps with an ENC of 0.10 fC, and tests with the 3D silicon detector yield 39.40 ps with an ENC of 0.09 fC, all measured under the same thermal condition. All configurations confirm the board's capability in low-signal regimes, significantly outperforming previous reference designs. A six-channel variant of the board has also been developed to support position-sensitive measurements. These findings demonstrate the board's suitability for laboratory-based 4D tracking detector characterisation, while simultaneously providing the groundwork for dedicated ASIC development.

physics.ins-det↗

arXiv2026

Measurement of $\mathrm{^{75}As}(\mathrm{n},γ)\mathrm{^{76}As}$ reaction relevant to 0$νββ$ decay searches of $\mathrm{^{76}Ge}$ and astrophysical $s$-process temperatures

The cross sections and resonance structures of $\rm^{75}As$(n,$γ$)$\rm^{76}As$ reaction are critical to the neutrinoless double-$β$ (0$νββ$) decay searches of $\rm^{76}Ge$, the $s$-process nucleosynthesis of nuclear astrophysics, and Neutron Resonance Capture Analysis for determining the elemental and isotopic composition of archaeological and cultural heritage. We report a high-precision measurement of the $\rm^{75}As$ neutron capture cross sections from 1~eV to 1~MeV, performed at the Back-n facility of the China Spallation Neutron Source using the Time-of-Flight method. In the resolved resonance region, nineteen resonance structures of $\rm^{75}As$(n,$γ$)$\rm^{76}As$ reaction have been discovered for the first time, and inconsistencies between evaluated libraries has been resolved. Resonance parameters for the newly observed structures were extracted with the $R$-matrix code SAMMY. These findings will help refine the theoretical predictions of half-lives and decay constants for the double-$β$ decay searches of $\rm^{76}Ge$. Astrophysical Maxwellian-averaged cross sections were calculated based on the averaged cross sections in the unresolved resonance region. And the $\rm^{75}As$(n,$γ$)$\rm^{76}As$ reaction rates were derived over the astrophysically relevant temperature range in both the main and weak $s$-processes nucleosynthesis. The present reaction rates deviate significantly from the recent theoretical predictions, and the uncertainties are significantly reduced.

nucl-ex↗

arXiv2026

RayOrch: Programming and Executing Lineage-Controlled Multi-Grain Dataflows for Foundation-Model Data Preparation

Preparing high quality training data for foundation models requires scalable pipelines that transform heterogeneous documents and videos into structured records. Such pipelines expand each parent item into an ordered and input dependent sequence of children, whose counts may be long tailed. GPUs should batch children across parents while preserving parent relationships, child order, completion status, and result routing. Existing systems either hide parallelism behind coarse grained jobs or expose flat records that force applications to manage lineage and regrouping. We present RayOrch, a programming model and distributed execution engine that preserves parent child relations throughout execution. Programs declare ordered variable cardinality expansions and matching gathers. The compiler validates each pair, while the runtime records child membership, immediate parents, immutable ordinals, and terminal states. Per Call FIFO Ready Queues batch ready children across parents. Gathers reconstruct results from declared membership and ordinals rather than batch boundaries or completion order. Parents can advance as soon as all required children become terminal. Typed parent scoped failures suppress undispatched siblings of the failed parent while allowing unrelated parents to continue. On NVIDIA H20 GPUs, RayOrch achieves 15.14 times speedup when scaling MinerU from 4 to 64 GPUs and 7.82 times speedup when scaling a video pipeline from 8 to 64 GPUs. It reduces end to end time by 13.1 percent versus Ray Data and 29.0 percent versus Daft on MinerU, and by 16.0 percent versus Ray Data on Docling. Code available at https://github.com/OpenDCAI/RayOrch .

cs.DC↗

arXiv2026

CapGeo-Bench: Decoupling Visual Perception from Reasoning and Evaluating Geometric Understanding

While Multimodal Large Language Models (MLLMs) have achieved remarkable success in difficult purely textual mathematical reasoning tasks, even advanced closed-source models such as GPT-o3 still struggle with geometric problems. This discrepancy motivates us to investigate the root cause: is the bottleneck of multimodal geometric reasoning rooted in reasoning itself, or in the perception of geometric information from diagrams? To answer this question, we conduct an exploratory experiment and find that providing high-quality captions consistently and substantially boosts performance of MLLMs, empirically validating the visual perception bottleneck in geometric reasoning. However, MLLMs' capabilities in visual geometric perception remain insufficiently evaluated. To fill the gap, we propose CapGeo-Bench, a benchmark of 4,641 high-quality figure-caption pairs equipped with a fine-grained keypoint-based evaluation metric. Specifically, CapGeo-Bench encompasses three geometric classes and is divided into four difficulty levels according to the complexity of geometric primitives. Captions are scored at a fine-grained level by extracting keypoints along three dimensions-geometric elements, spatial relationships, and numerical attributes. CapGeo-Bench provides a reliable instrument and guideline for evaluating and advancing MLLMs' geometric perception capabilities. Our code and data are publicly available at: https://github.com/YuYingLi0/CapGeo.

cs.CV↗

arXiv2026

DataFlex-RL: An Evaluation Platform for RLVR Data Policies

Data policies for reinforcement learning with verifiable rewards (RLVR) determine which rollouts are used, how strongly they are weighted, and which domains contribute to subsequent training batches. We introduce DataFlex-RL, an evaluation platform for comparing these choices under a common GRPO recipe. Our primary experiment evaluates 13 configurations across 12 matched seeds using Qwen2.5-7B-Base and 12 mathematics, logic, and science benchmarks. Uniform GRPO improves the domain-balanced average accuracy by 7.76 percentage points over the untrained checkpoint. None of the eight rollout-selection or reweighting methods achieves a paired 95% confidence interval that excludes zero relative to uniform sampling, and none of the three adaptive mixtures outperforms a fixed equal mixture at the same level of precision. A corrected 12-seed extension on Llama-3.1-8B-Base places the additional methods on the same score scale as the original controls, but does not reveal a consistent winner in terms of observed mean performance. We also quantify evaluation sensitivity by rescoring nine Qwen2.5-7B-Instruct runs using a math-heavy six-benchmark summary, consisting of five mathematics benchmarks and GPQA-Diamond but no logic benchmark, and comparing it with the domain-balanced 12-benchmark summary. The resulting rankings are negatively correlated, with a correlation coefficient of -0.33, whereas summaries that retain all 12 benchmarks largely agree. Across the controlled settings studied here, changing the data policy measurably changes the training process but does not produce a reproducible improvement over uniform training.

cs.LG↗

arXiv2026

TraceAV-Bench: Benchmarking Multi-Hop Trajectory Reasoning over Long Audio-Visual Videos

Real-world audio-visual understanding requires chaining evidence that is sparse, temporally dispersed, and split across the visual and auditory streams, whereas existing benchmarks largely fail to evaluate this capability. They restrict videos to short clips, isolate modalities, or reduce questions to one-hop perception. We introduce TraceAV-Bench, the first benchmark to jointly evaluate multi-hop reasoning over long audio-visual trajectories and multimodal hallucination robustness. TraceAV-Bench comprises 2,200 rigorously validated multiple-choice questions over 578 long videos, totaling 339.5 hours, spanning 4 evaluation dimensions and 15 sub-tasks. Each question is grounded in an explicit multi-hop evidence trajectory that averages 3.68 hops across a 15.1-minute temporal span. The dataset is built by a three-step pipeline followed by a strict quality assurance process. Evaluation of multiple representative OmniLLMs on TraceAV-Bench reveals that the benchmark poses a persistent challenge across all models, with the strongest closed-source model (Gemini 3.1 Pro) reaching only 68.29% on general tasks, and the best open-source model (Ming-Flash-Omni-2.0) reaching 51.70%, leaving substantial headroom. Further analysis shows that current models struggle to gather and combine evidence across long audio-visual videos, with performance varying across modalities and reasoning tasks. We hope TraceAV-Bench will facilitate progress toward reliable long-form audio-visual reasoning in OmniLLMs.

cs.CV↗