arXiv subjects

Feng He

Publications and source records attributed to Feng He.

At least 19 recordsLinked to original sources

arXiv2026

From Research Goals to Scientific Closure: Evaluating Agents Through Evidence-Supported Physics Research

Can AI agents organize scientific investigations that lead to well-supported conclusions? To evaluate this capability, we introduce PhysClosureBench, a benchmark of 50 expert-verified, goal-specified research tasks spanning strong-field, attosecond, and neighboring physics areas. Each task specifies the research objective and essential scientific constraints, while leaving the agent to organize the investigation and determine how to generate and validate the evidence needed to answer the question. A 20-criterion Atomic Score measures local scientific accomplishment, while Research Closure (RC0--RC4) assesses how fully the investigation resolves the specified research question based on whether its indispensable evidence requirements and applicable scope conditions are satisfied. We evaluate seven frontier LLMs, where Claude-Opus-5 achieves the highest mean Atomic Score (6.29/10) but reaches RC4 in only 32.7% of attempts. Across the benchmark, 36.2% of attempts scoring at least 8/10 still fall short of RC4, while all seven models reach RC4 on substantially fewer tasks in all three runs than in at least one. Case-level analysis further illustrates how misinterpretation of generated results and unresolved physical applicability limits can lead to incomplete closure. Together, these results reveal a persistent gap between AI agents' local scientific accomplishments and their ability to achieve complete research closure.

cs.AI↗

arXiv2026

Cortex: Consolidating Context into Parametric Memory

Context can improve large language model performance, but these gains may be lost after context removal. Inspired by biological systems consolidation, we formulate Memory Consolidation (MC) as internalizing these gains while preserving non-target capabilities, connecting context-specified editing, skill learning, and behavioral steering under a common objective. We introduce a five-task-family testbed measuring consolidation and retention under single, batch, and sequential updates. We propose Cortex, which generates prospective queries to anticipate uses of context and distills context-conditioned outputs into mergeable low-rank or structural feed-forward network updates. Across 4B--120B dense and mixture-of-experts backbones, including both text-only and multimodal models, Cortex lies on the empirical consolidation--retention Pareto frontier in most evaluated settings. After 50 sequential factual edits on a 14B model, Cortex-Structural recovers 88.9% of the context-induced gain while retaining 99.8% of original non-target performance.

cs.CL↗

arXiv2026

PHRBench: A Behavioral Evaluation of Post-Hallucination Reasoning in LLMs

Hallucinated information can propagate through multi-stage LLM systems and become part of the context for subsequent reasoning. Existing studies of post-hallucination reasoning (PHR) mainly characterize changes in final outcomes and aggregate reasoning dynamics, leaving how models resolve hallucinated premises at the response level insufficiently understood. In this work, we introduce PHRBench, a controlled benchmark for behaviorally structured PHR across four domains and 18 large language models. PHRBench characterizes each reasoning trajectory independently of final-answer correctness through Hallucination Compliance, Hallucination Avoidance, and Heuristic Correction, and defines an insightful trajectory as successful correction that ultimately reaches the correct answer. Across 4820 controlled instances, we find that successful recovery remains relatively rare and is associated with more frequent belief updates along the reasoning trajectory. We further find that properties of the hallucinated prompt contain substantial predictive signal for successful recovery, with a lightweight predictor achieving an AUROC of 0.847. These findings provide a behavioral view of post-hallucination reasoning, characterizing how LLMs resolve erroneous context and when successful recovery is likely to occur.

cs.CL↗

arXiv2026

Verification Trap: Understanding Test-Time Selection Failures under False Premises in Code Generation

Test-time compute has become a central way to improve code generation: systems sample multiple candidate programs and use verifier-visible evidence to select the final output. This paradigm implicitly assumes that the verifier provides a corrective signal independent from the generator. We challenge this assumption under misleading task premises. When the generator and verifier share a false premise, they become coupled through a mistaken belief: the generator produces premise-consistent shortcuts, while the verifier supplies evidence that fails to expose them. Consequently, the selector may choose a hidden-test-wrong candidate even when a hidden-test-correct program exists in the pool. We call this failure mode Verification Trap. Across three code-generation benchmarks and five code models, false premises consistently degrade first-sample correctness, reduce selector-chosen correctness after 64-sample test-time selection, and amplify recoverable mis-selection. Mechanistically, verifier-written tests inherit the premise-level blind spot, reshaping verifier-visible candidate space away from hidden-test correctness. These traces make Verification Trap predictable before hidden execution: a lightweight gold-free predictor using verifier-visible features reaches 0.846 AUROC. Our results identify decoupled evidence as a key mitigation axis: coupled scaling provides limited recovery, whereas premise-agnostic robustness auditors recover substantial oracle headroom.

cs.CL↗

arXiv2026

GenMC: Real-Time Generative Monte Carlo Surrogate for Quantitative Photoacoustic Imaging

Photoacoustic (PA) imaging provides molecular and functional information about tissue, such as blood oxygen saturation (sO2), yet its clinical translation is hindered by inaccurate quantification. A major source of error is the spectral colouring effect, in which wavelength-dependent optical attenuation distorts the local optical fluence. Monte Carlo (MC) simulation is the gold standard for modelling light transport, but its computational demand precludes real-time use. Here, GenMC is presented, a deep generative framework based on a conditional generative adversarial network that estimates optical fluence distributions from tissue anatomy and literature-derived optical properties, with anatomical priors obtained from co-registered ultrasound images. Trained on MC-generated synthetic datasets, GenMC produces high-fidelity fluence maps in under 30 ms per frame, a four-orders-of-magnitude speed-up over conventional MC simulation, and reaches peak signal-to-noise ratios of up to 36.24 dB in vivo, outperforming UNet and Pix2Pix baselines. Validation in blood-mimicking phantoms and in 37 human volunteers spanning Fitzpatrick skin types III-V demonstrates improved accuracy, robustness, and physiological consistency of sO2 estimation. By enabling real-time, accurate, and reproducible quantification of tissue oxygenation, GenMC addresses a critical barrier to quantitative PA imaging and offers a general strategy for rapid, high-fidelity approximation of light transport in tissue.

physics.med-ph↗

arXiv2026

Quantum Stochastic Walks on the Permutation Group

How rapidly does order give way to randomness, and can quantum coherence accelerate this process? We address these questions through the paradigmatic problem of card shuffling, formulated as a random walk on the symmetric group $S_n$. We first recast the random-transposition walk studied by Diaconis and Shahshahani, as well as more general walks generated by conjugacy classes of $S_n$, in continuous time. We then identify the transition matrix of each classical walk with a permutation Hamiltonian generating a corresponding unitary quantum walk. Purely unitary evolution, however, does not generically converge to the uniform distribution in the classical sense of mixing: coherence preserves information rather than erasing it. We therefore embed the problem into a quantum stochastic walk, where coherent dynamics competes with the dissipative process responsible for classical mixing. In this setting, quantum coherence assists randomization. We prove that it can only decrease the distance from the uniform distribution in the computational basis and can therefore accelerate mixing. An analysis of the slowest mode yields a criterion for the coupling strength required to produce an appreciable speedup. Finally, numerical results reveal a scaling collapse of the ratio between quantum and classical mixing times onto a simple one-parameter form. Our results illustrate how coherence and dissipation can cooperate in the emergence of randomness in walks on permutation groups.

cond-mat.stat-mech↗

arXiv2026

Artificial Anisotropy Induced Bound States in the Continuum for Integrated Photonic Waveguide

Bound states in the continuum (BICs) enable counterintuitive light confinement without radiation loss, providing a powerful foundation for integrated photonic waveguides. However, existing BIC waveguides are predominantly realized through geometry-dependent designs, where the BIC condition is restricted to narrowly defined structural parameters, limiting design flexibility and practical applicability. Artificial optical anisotropy is introduced as a new design paradigm for BIC waveguides. Implemented using subwavelength-grating (SWG) metamaterials, continuously tailorable anisotropy provides an independent degree of freedom for deterministically reshaping the radiative continuum, enabling flexible formation and systematic control of BIC waveguides over a broad design space. Anisotropy-engineered symmetry breaking further enables controllable asymmetric radiation and precisely tailored field leakage. This paradigm transforms BIC waveguides from geometry-constrained structures into an anisotropy-engineered platform, establishing a general framework for programmable radiation engineering and next-generation integrated photonic devices.

physics.optics↗

arXiv2026

Attosecond Access to the Quantum Noise of Light

Bright squeezed light has entered strong-field atomic physics, creating an immediate need to characterize the quantum field delivered to the target on subcycle timescales. Here we show that attosecond streaking maps the coherent displacement and phase-sensitive covariance of a displaced squeezed field onto distinct harmonics: the mean photoelectron momentum follows an $ω$-periodic shift of the spectral center, whereas the momentum variance exhibits a $2ω$-periodic breathing. A Feynman--Vernon formulation represents Gaussian quantum light by stochastic vector-potential trajectories, enabling Coulomb-resolved TDSE simulations with a finite XUV gate and without explicit photon-state propagation. A coherent-state reference calibrates the streaking phase and variance background, enabling retrieval of the coherent amplitude and phase together with the squeezed-noise amplitude and squeezing phase. This establishes attosecond streaking as an in situ, gas-phase diagnostic of bright squeezed light.

quant-ph↗

arXiv2026

Anomalous radiation pressure in strong-field ionization driven by quantum light

We show that in strong-field ionization driven by bright squeezed vacuum, the mean longitudinal photoelectron momentum scales with the mean incident intensity $I$ as $I^{2/3}$, rather than linearly as under coherent-light driving. This anomalous scaling originates from a field-amplitude saddle-point structure: nonlinear tunneling selects two dominant field amplitudes of equal magnitude and opposite sign from the broad quantum fluctuations. Tracing over the final photon state erases their relative-phase information, leaving their common magnitude to determine the field-dependent longitudinal momentum shift and its $I^{2/3}$ scaling. Photon-number resolution instead preserves this coherence, producing parity-dependent modulations of the longitudinal momentum transfer through time-domain double-slit interference between two field pathways separated by half a cycle. These results establish longitudinal momentum transfer as a distinct observable in strong-field quantum optics that encodes both the photon statistics and the field coherence of intense quantum light.

quant-ph↗

arXiv2026

Catch the Patient, Not the AI: Collective Sensemaking in an Online Health Community

Patients and caregivers increasingly use artificial intelligence (AI) tools to interpret medical reports, weigh care decisions, and seek emotional support. Yet most research treats patient-facing AI as a private exchange between a user and a system. This study examines how AI-related content is taken up once users carry it back into the peer communities, using data from House086, China's largest online community for lymphoma patients and caregivers. We identified roughly 400 publicly accessible threads (2014-2026) through keyword searches and manual screening, extracted them into structured case profiles using a schema-prompted large language model, and conducted mixed-method analysis. After quality control, the verified analytic sample comprised 337 post-ChatGPT records. Members most often reported using AI for informational support, followed by second opinions and psychosocial support. Although members often introduced AI favorably, roughly one in six described feeling overwhelmed by AI output. When other members responded, they frequently engaged the poster's underlying medical or emotional intent while leaving the AI dimension unaddressed. This tendency persisted even in threads seeking triangulation between AI output and other information sources. When members did discuss the AI, they were more often cautious than endorsing. Rather than systematically auditing AI output, the community more often worked to deflate the false certainty it produced, placing a single AI answer back among multiple sources of judgment. This study argues that AI does not replace the interpretive work of online health communities, nor is it systematically audited by them. Instead, it shifts the locus of sensemaking downstream, so that the community continues to catch the person even when it does not catch the AI.

cs.HC↗

arXiv2026

LYRA: Label-Free Structural Synchronization and Resource Allocation for UAV Edge Networks

While deploying hierarchical vision models to process mission-critical tasks, UAV edge systems must adaptively update the models to sustain inference reliability under low-level environmental corruption. However, existing work has overlooked the optimal timing for model updates, the impracticality of relying on real-time expert labels, and the significant bandwidth and energy constraints of UAVs. This paper proposes a joint model update scheduling and resource allocation framework, aiming to maximize long-term semantic fidelity and resource efficiency of UAV edge intelligence systems. To address the challenge of label-free semantic evaluation, we formulate the Online Semantic Disagreement Rate (OSDR) as a proxy for timely update triggering, thereby enabling fine-grained Sensitivity-Aware Structural Synchronization (SASS). Furthermore, to overcome the curse of dimensionality in hybrid action spaces and effectively bound long-term energy budgets, we propose a Lyapunov-guided discrete reinforcement learning algorithm that performs action space dimensionality reduction and transforms constraints into virtual queue stability problems. The reported experimental results, based on real traffic traces, demonstrate that the proposed framework consistently outperforms representative baselines in semantic recovery efficiency and update triggering precision, by satisfying long-term energy budget and by reducing average risk backlog by up to 33.3\% in the dynamic environmental corruption scenario.

cs.NI↗

arXiv2026

DeepNC: A Fast GNN-based Pre-Verification Surrogate for TSN Configuration

Time-Sensitive Networking (TSN) is critical to deterministic communication in safety-critical domains, with formal verification such as Network Calculus (NC) serving as the cornerstone for schedulability guarantees. However, during automated configuration-space exploration, repeated schedulability analysis consumes over 90% of the total configuration time, becoming the primary bottleneck for large-scale TSN configurations. To address this challenge, we propose DeepNC, a novel pre-verification surrogate module that pioneers the structural fusion of NC principles into a Graph Neural Network (GNN) for TSN configuration-space exploration. Rather than replacing formal verification, DeepNC acts as a high-speed pre-verification filter, reserving computationally expensive formal verification only for promising candidates. Extensive evaluations demonstrate that DeepNC significantly improves worst-case delay prediction accuracy over state-of-the-art learning-based methods, increasing the average $R^2$ by 55.8% and reducing the average MAPE by 65.3%. More importantly, its high-fidelity regression substantially reduces the number of formal verification calls during configuration-space exploration by 93.25%, while accelerating NC-based verification by more than two orders of magnitude.

cs.NI↗

arXiv2026

Orbit-resolved spin holography: role of Coulomb focusing in target-dependent polarization

Strong-field photoelectron holography encodes ultrafast electron dynamics through momentum-space interference. However, the orbit-resolved origin of spider-like spin fringes and the mechanism underlying their target dependence remain unclear. Here, we resolve both issues by analyzing photoelectron spin textures generated during tunneling ionization. We use the Coulomb quantum-orbit strong-field approximation, benchmarked against time-dependent Schrödinger equation simulations for $\mathrm{He^+}$ and Xe, to separate orbital-channel and quantum-orbit contributions. Spider-like fringes arise from interference between $p$-orbital ionization channels with different magnetic quantum numbers within an individual orbit class and therefore do not require interorbit interference. The observable polarization along these fringes, however, depends on the balance among orbit-class contributions. The decomposition associates the opposite first-leg polarizations of $\mathrm{He^+}$ and Xe with different relative weights of laser-deflected and forward-scattered trajectories, consistent with target-dependent Coulomb focusing. Photoelectron spin textures thus complement momentum distributions as probes of Coulomb-driven strong-field dynamics.

physics.atom-ph↗

arXiv2026

HIVE: Understanding Post-Hallucination Reasoning in Vision Language Models

Hallucinations in vision language models (VLMs) are commonly treated as semantic errors, yet they often arise from partial or ambiguous visual evidence. Prior work mainly focuses on detecting or suppressing hallucinations at generation time, leaving the subsequent reasoning stage largely unexplored. In this work, we study Post Hallucination Reasoning (PHR), the stage in which hallucinated semantics enter the model's inference context and influence downstream predictions. To systematically investigate PHR, we introduce HIVE, Hallucination Inference and Verification Engine, an evaluation infrastructure that enables controlled comparisons between faithful and hallucinated captions. Across nine tasks and nine models, we observe structured modality dependent patterns: hallucinated captions often improve accuracy on vision language tasks, while text only tasks exhibit limited or unstable effects. Further analyses show that hallucinated cues broaden semantic coverage and reshape reasoning dynamics while preserving stable inference. These findings highlight that hallucinated semantics may influence downstream reasoning once they enter the model's inference context. Understanding this post hallucination stage is important for improving the reliability and interpretability of multimodal reasoning systems. Code is publicly available at https://github.com/hefengcs/HIVE.

cs.CV↗

arXiv2026

Spectral Decimation of Quantum Many-Body Hamiltonians

We develop a systematic theory of spectral decimation for quantum many-body Hamiltonians and show that it provides a quantitative probe of emergent symmetries in statistically mixed spectra. Building on an analytical description of statistical mixtures, we derive an explicit expression for the size of a characteristic symmetry sector (CSS), defined as the largest subsequence of levels exhibiting non-Poissonian correlations. The CSS dimension is shown to be the size-biased average of the underlying symmetry sectors, establishing a direct link between spectral statistics and Hilbert-space structure. We apply this framework to two paradigmatic settings: Hilbert-space fragmentation and disorder-induced many-body localization (MBL). In fragmented systems, the CSS reproduces the mixture prediction and isolates correlated subsectors even when the full spectrum appears nearly Poissonian. In the disordered Heisenberg chain, spectral decimation reveals the gradual emergence of integrability through a shrinking CSS, whose statistics exhibit signatures consistent with local integrals of motion. We introduce a characteristic symmetry entropy (CSE) as a finite-size scaling observable and extract, within accessible system sizes, the crossover exponents. Our results establish spectral decimation as a controlled, unbiased and computationally inexpensive diagnostic of hidden structure in many-body spectra, capable of distinguishing between chaotic dynamics, statistical mixtures, and emergent integrability.

cond-mat.stat-mech↗

arXiv2026

Neural Network Optimization Reimagined: Decoupled Techniques for Scratch and Fine-Tuning

With the accumulation of resources in the era of big data and the rise of pre-trained models in deep learning, optimizing neural networks for various tasks often involves different strategies for fine-tuning pre-trained models versus training from scratch. However, existing optimizers primarily focus on reducing the loss function by updating model parameters, without fully addressing the unique demands of these two major paradigms. In this paper, we propose DualOpt, a novel approach that decouples optimization techniques specifically tailored for these distinct training scenarios. For training from scratch, we introduce real-time layer-wise weight decay, designed to enhance both convergence and generalization by aligning with the characteristics of weight updates and network architecture. For more importantly fine-tuning, we integrate weight rollback with the optimizer, incorporating a rollback term into each weight update step. This ensures consistency in the weight distribution between upstream and downstream models, effectively mitigating knowledge forgetting and improving fine-tuning performance. Additionally, we extend the layer-wise weight decay to dynamically adjust the rollback levels across layers, adapting to the varying demands of different downstream tasks. Extensive experiments across diverse tasks, including image classification, object detection, semantic segmentation, and instance segmentation, demonstrate the broad applicability and state-of-the-art performance of DualOpt. Code is available at https://github.com/qklee-lz/OLOR-AAAI-2024.

cs.CV↗

arXiv2026

HG-Lane: High-Fidelity Generation of Lane Scenes under Adverse Weather and Lighting Conditions without Re-annotation

Lane detection is a crucial task in autonomous driving, as it helps ensure the safe operation of vehicles. However, existing datasets such as CULane and TuSimple contain relatively limited data under extreme weather conditions, including rain, snow, and fog. As a result, detection models trained on these datasets often become unreliable in such environments, which may lead to serious safety-critical failures on the road. To address this issue, we propose HG-Lane, a High-fidelity Generation framework for Lane Scenes under adverse weather and lighting conditions without requiring re-annotation. Based on this framework, we further construct a benchmark that includes adverse weather and lighting scenarios, containing 30,000 images. Experimental results demonstrate that our method consistently and significantly improves the performance of existing lane detection networks. For example, using the state-of-the-art CLRNet, the overall mF1 score on our benchmark increases by 20.87 percent. The F1@50 score for the overall, normal, snow, rain, fog, night, and dusk categories increases by 19.75 percent, 8.63 percent, 38.8 percent, 14.96 percent, 26.84 percent, 21.5 percent, and 12.04 percent, respectively. The code and dataset are available at: https://github.com/zdc233/HG-Lane.

cs.CV↗

arXiv2026

Ultrafast Ionization Dynamics Encoded in a Photoelectron Spin Torus

We demonstrate that strong-field ionization of atoms in circularly polarized laser fields generates a photoelectron spin texture with toroidal topology in momentum space. Using time-dependent Schrödinger equation simulations, spin-resolved classical-trajectory Monte Carlo calculations, and an extended spin-resolved strong-field approximation including intermediate excitation pathways, we show that the rotation angle of this spin torus provides access to attosecond relative time delays associated with photoelectron wave packets released by tunneling from the counter-rotating and co-rotating $p$-orbital channels. When intermediate-state dynamics become significant, the torus develops a clear splitting. These results establish photoelectron spin textures as a complementary source of dynamical information beyond conventional momentum spectroscopy, and identify spin polarization as a robust internal degree of freedom for self-referenced attosecond metrology.

physics.atom-ph↗