arXiv subjects

Kun Fang

Publications and source records attributed to Kun Fang.

At least 37 records · Page 2Linked to original sources

arXiv2026

Beyond Perceptual Distance: Discrepancy Assessment on Deep Representation for Out-of-Distribution Detection with Diffusion Model

Out-of-Distribution (OoD) detection aims to justify whether a given sample is from the training distribution of the classifier-under-protection, i.e., In-Distribution (InD), or from an unknown out distribution. Recent researches have leveraged Diffusion Models (DMs) for OoD detection due to their powerful distribution modeling capability. Given an input image, an InD-pretrained DM produces a corresponding InD-aligned counterpart, which serves as a generative reference for comparison. However, existing DM-based methods typically assess this underlying discrepancy through visual-level distances in the raw image space, which may be misaligned with the distributional discrepancy relevant to OoD detection. In this work, we investigate the fundamentals of discrepancy assessment in DM-based OoD detection, asking how the discrepancy between an input and its DM-generated counterpart should be formulated, and in which representation spaces and with which metrics it should be measured. To this end, we propose to assess the discrepancy in a classifier-relative manner by exploiting the representation spaces of the classifier-under-protection, whose training on InD data encodes rich task-relevant InD knowledge. In particular, we quantify two types of discrepancy: feature-level covariate discrepancy in deep feature representations and logit-level concept discrepancy in output logits, enabling effective differentiation between InD and OoD samples. Moreover, a subspace-based strategy is devised to refine representations of the DM generation to promote discrepancy assessment. Together, these designs form our novel detection framework, namely DDR. Extensive experiments on the challenging large-scale ImageNet-1K dataset demonstrate the superior detection performance of DDR over both DM-based and non-DM-based methods.

cs.CV↗

arXiv2026

Kernel PCA for Out-of-Distribution Detection: Non-Linear Kernel Selection and Approximation

Out-of-Distribution (OoD) detection is vital for the reliability of deep neural networks, the key of which lies in effectively characterizing the disparities between OoD and In-Distribution (InD) data. In this work, such disparities are exploited through a fresh perspective of non-linear feature subspace. That is, a discriminative non-linear subspace is learned from InD features to capture representative patterns of InD, while informative patterns of OoD features cannot be well captured in such a subspace due to their different distribution. Grounded on this perspective, we exploit the deviations of InD and OoD features in such a non-linear subspace for effective OoD detection. To be specific, we leverage the framework of Kernel Principal Component Analysis (KPCA) to attain the discriminative non-linear subspace and deploy the reconstruction error on such subspace to distinguish InD and OoD data. Two challenges emerge: (i) the learning of an effective non-linear subspace, i.e., the selection of kernel function in KPCA, and (ii) the computation of the kernel matrix with large-scale InD data. For the former, we reveal two vital non-linear patterns that closely relate to the InD-OoD disparity, leading to the establishment of a Cosine-Gaussian kernel for constructing the subspace. For the latter, we introduce two techniques to approximate the Cosine-Gaussian kernel with significantly cheap computations. In particular, our approximation is further tailored by incorporating the InD data confidence, which is demonstrated to promote the learning of discriminative subspaces for OoD data. Our study presents new insights into the non-linear feature subspace for OoD detection and contributes practical explorations on the associated kernel design and efficient computations, yielding a KPCA detection method with distinctively improved efficacy and efficiency.

cs.LG↗

arXiv2026

Quantum thermodynamics with uncertain equilibrium

The resource-theoretic approach to quantum thermodynamics typically assumes perfect knowledge of the thermal equilibrium state, an idealization incompatible with finite experimental precision. We develop a framework for equilibrium uncertainty by representing the equilibrium reference as a set of candidate states. Under a generic geometric condition, we prove a no-go theorem that sharply limits athermality ``purification'': converting an uncertain athermal state into a definite target is either trivial or impossible. We then derive exact one-shot entropic characterizations of work extraction and formation for two work-storage models, a clean battery with known equilibrium and a dirty battery with uncertain equilibrium. Both models exhibit strong asymptotic irreversibility even under arbitrarily small uncertainty. An explicit example reveals two distinct extremes: clean batteries display a bound-entanglement-like phenomenon, with positive formation cost but zero extractable work, whereas dirty batteries allow positive work extraction but require infinite formation cost. These phenomena show that equilibrium uncertainty is not a minor perturbation of the standard theory, but a structural ingredient that fundamentally reshapes the limits of quantum thermodynamics.

quant-ph↗

arXiv2026

Operational interpretation of the reverse sandwiched Renyi divergences in composite quantum hypothesis testing

We study the Hoeffding regime of composite quantum hypothesis testing, in which each hypothesis is specified by a sequence of sets of quantum states. We establish quantum Hoeffding bounds under a set of structural assumptions, orthogonal to those of our previous framework. A notable consequence is the direct operational interpretation of the reverse sandwiched Renyi divergence for $α\in (0,1)$: for the task of discriminating a thermal equilibrium state from a probe state subject to unknown dephasing in the energy eigenbasis, with free Hamiltonian evolution as a special case, the optimal Hoeffding exponent is given exactly by this divergence evaluated on a single copy of the system. The same task in the Stein regime is governed by the reverse quantum relative entropy, providing its operational interpretation as well. This behavior contrasts both with the simple independent and identically distributed (i.i.d.) setting, where the Petz Renyi divergence and the Umegaki relative entropy govern the Hoeffding and Stein exponents, respectively, and with many composite settings, where only regularized many-copy formulas are available. This finding reveals that passing from simple to composite hypotheses can fundamentally change which quantum divergence determines the operational limits of discrimination, and suggests a new avenue for seeking operational interpretations of quantum divergences by lifting simple hypotheses to richer composite scenarios.

quant-ph↗

arXiv2026

Resolving diffusion signatures in distant pulsar halos with current and future experiments

Pulsar halos provide a unique probe of cosmic-ray propagation in the vicinity of pulsars and have important implications for our understanding of particle diffusion in the interstellar medium. However, the number of firmly identified pulsar halos remains limited. One of the main challenges is the difficulty in unambiguously confirming halo candidates through precise morphological measurements with current $γ$-ray observations. In this work, we investigate the prospects for identifying pulsar halo candidates through morphological discrimination using simulations of two advanced $γ$-ray experiments: LHAASO-KM2A and the Cherenkov Telescope Array (CTA). Using mock observations with realistic instrumental responses, we assess the ability of each experiment to distinguish diffusion-based halo morphologies from alternative simplified spatial models. Our results show that both increased photon statistics and improved angular resolution significantly enhance the power of morphological discrimination. In particular, CTA benefits from its superior angular resolution, while LHAASO-KM2A gains sensitivity from its large effective area at the highest energies. These results indicate that future $γ$-ray observations have the potential to expand the sample of pulsar halos and provide further insights into cosmic-ray transport around pulsars.

astro-ph.HE↗

arXiv2026

Method for testing diffusive shock acceleration and diffusion propagation of 1-100 TeV cosmic electrons with multiwavelength observations of the Geminga halo and pulsar wind nebula

Diffusive shock acceleration and diffusion propagation are essential components of the standard cosmic ray model. These theories are based on extensive observations of high-energy solar processes, providing substantial direct evidence in the MeV energy range. Although the model is widely and successfully used to explain high-energy cosmic phenomena, direct validation has been elusive. The multi-wavelength spectra and angular profile measurements of the Geminga pulsar wind nebula and its pulsar halo, particularly the precise spectral observations by HAWC and LHAASO-KM2A in recent years, offer a rare opportunity to test these theories with cosmic rays energies between 1 TeV and several hundred TeV. These observations are expected to elevate the direct testing of theoretical models from multi-MeV to sub-PeV energies. In this work, a method is developed to test the diffusive shock acceleration and diffusion propagation model between one and several hundred TeV energies through the latest spectral and morphological data of the Geminga region from HAWC and Fermi-LAT. Our results show that the theories of diffusive shock acceleration and diffusion propagation are consistent with experimental observations. However, the published morphological data adopted rather wide energy bins and currently do not allow a high precision test of the inferred energy dependent diffusion coefficient by observed energy spectra with DSA theory. It is anticipated that future HAWC and LHAASO-KM2A observations will yield higher-precision results, and the confirmation of a rapidly increasing diffusion coefficient above 100 TeV would serve as important evidence supporting the diffusive shock acceleration and diffusion propagation theory. Similar tests would be both important and valuable for other models.

astro-ph.HE↗

arXiv2026

FIT to Forget: Robust Continual Unlearning for Large Language Models

While large language models (LLMs) exhibit remarkable capabilities, they increasingly face demands to unlearn memorized privacy-sensitive, copyrighted, or harmful content. Existing unlearning methods primarily focus on \emph{single-shot} scenarios, whereas real-world deletion requests arrive \emph{continually}. Naïvely applying these methods to sequential requests leads to severe utility degradation and catastrophic forgetting. To address this, we propose \fit, a robust continual unlearning framework to process high-volume sequential deletion streams while resisting both catastrophic forgetting and post-unlearning recovery. \fit stabilizes sequential updates through three synergistic mechanisms: redundancy \underline{F}iltering, \underline{I}mportance-aware adaptive algorithm selection, and \underline{T}argeted layer attribution. Furthermore, to facilitate rigorous evaluation, we introduce \textbf{PCH}, a unified benchmark encompassing \textbf{P}ersonal, \textbf{C}opyrighted, and \textbf{H}armful content, alongside two symmetric metrics, Forget Degree (F.D.) and Retain Utility (R.U.), to systematically quantify forgetting-utility trade-offs. Extensive experiments across five LLMs (up to 14B parameters) demonstrate that \fit consistently achieves state-of-the-art unlearning efficacy and utility preservation. Notably, even after hundreds of sequential requests, \fit preserves strong downstream (\eg, GSM8K, MMLU) performance and exhibits superior resilience against relearning and quantization recovery attacks.

cs.CL↗

arXiv2026

Charge-dependent spectral softenings of primary cosmic-rays below the knee

In most particle acceleration or propagation theories, the characteristic features of the cosmic ray spectra due to acceleration limits or propagation phase changes are charge dependent. Alternatively, the interaction scenario would expect mass dependent spectral features in general. The observational verification of which relation takes effect in nature is still lack due to the difficulty of measuring the spectra of individual particles up to very high energies. Here we report direct measurements of the carbon, oxygen, and iron spectra from ~20 gigavolts to ~100 teravolts (~60 teravolts for iron) with 9 years of on-orbit data collected by the Dark Matter Particle Explorer. Distinct spectral softenings have been directly detected in these spectra for the first time. Combined with the updated proton and helium spectra, the spectral softening appears universally at a rigidity of ~15 teravolts. A nuclei mass dependent softening is rejected at a confidence level of >99.999%. Possible interpretations of these results, including a nearby cosmic ray source and other models such as the propagation effect, are discussed.

astro-ph.HE↗

arXiv2026

MLOW: Interpretable Low-Rank Frequency Magnitude Decomposition of Multiple Effects for Time Series Forecasting

Separating multiple effects in time series is fundamental yet challenging for time-series forecasting (TSF). However, existing TSF models cannot effectively learn interpretable multi-effect decomposition by their smoothing-based temporal techniques. Here, a new interpretable frequency-based decomposition pipeline MLOW captures the insight: a time series can be represented as a magnitude spectrum multiplied by the corresponding phase-aware basis functions, and the magnitude spectrum distribution of a time series always exhibits observable patterns for different effects. MLOW learns a low-rank representation of the magnitude spectrum to capture dominant trending and seasonal effects. We explore low-rank methods, including PCA, NMF, and Semi-NMF, and find that none can simultaneously achieve interpretable, efficient and generalizable decomposition. Thus, we propose hyperplane-nonnegative matrix factorization (Hyperplane-NMF). Further, to address the frequency (spectral) leakage restricting high-quality low-rank decomposition, MLOW enables a flexible selection of input horizons and frequency levels via a mathematical mechanism. Visual analysis demonstrates that MLOW enables interpretable and hierarchical multiple-effect decomposition, robust to noises. It can also enable plug-and-play in existing TSF backbones with remarkable performance improvement but minimal architectural modifications.

cs.LG↗

arXiv2026

Uhlmann's theorem for measured divergences

Uhlmann's theorem is a cornerstone of quantum information theory, stating that for any quantum state $ρ_{AB}$ and any state $σ_A$, there exists an extension $σ_{AB}$ of $σ_A$ such that the fidelity between $ρ_{AB}$ and $σ_{AB}$ equals the fidelity between their marginals $ρ_A$ and $σ_A$. This property underpins many results and applications in quantum information science. In this work, we generalize Uhlmann's theorem to a broad class of measured $f$-divergences, including the measured $α$-Rényi divergences for all $α\geq 0$. The well-known Uhlmann's theorem for the fidelity corresponds to the special case $α= \frac{1}{2}$. Since most commonly used quantum Rényi divergences, including the Petz and sandwiched Rényi divergences, cannot satisfy this property (except for degenerate cases). This fundamentally distinguishes measured $f$-divergences from other quantum divergences and highlights their unique mathematical structure.

quant-ph↗

arXiv2026

Nonlinear diffusive shock acceleration with upstream escape reproduces DAMPE observations

We develop a self-consistent nonlinear extension of diffusive shock acceleration that incorporates cosmic ray (CR) backreaction on the shock precursor together with a physically motivated upstream-escape mechanism that produces an exponential high energy cutoff. The CR pressure gradient decelerates the upstream flow facing the shock wave, generating an extended precursor in which higher rigidity particles sample a larger cumulative velocity gradient and thereby acquire a progressively harder spectrum. Finite-size/escape effects are modeled by a momentum-dependent loss term, which naturally terminates acceleration and steepens the spectrum near the cutoff. The precursor compression ratio is not imposed as a closure condition but is determined dynamically by requiring consistency between the injection rate inferred from thermal leakage at the subshock and the injection strength demanded by the nonlinear shock modification, with CR-driven wave heating providing stabilizing negative feedback. Applying the model to young supernova-remnant-like parameters and standard one-zone Galactic diffusion, we reproduce the main features of the latest DAMPE proton spectrum: gradual hardening from hundreds of GeV to multi-TeV energies and a subsequent exponential cutoff at tens of TeV. The resulting spectral evolution follows directly from the competition between precursor-mediated nonlinear feedback and upstream escape.

astro-ph.HE↗

arXiv2026

Evaluating High-Resolution Piano Sustain Pedal Depth Estimation with Musically Informed Metrics

Evaluation for continuous piano pedal depth estimation tasks remains incomplete when relying only on conventional frame-level metrics, which overlook musically important features such as direction-change boundaries and pedal curve contours. To provide more interpretable and musically meaningful insights, we propose an evaluation framework that augments standard frame-level metrics with an action-level assessment measuring direction and timing using segments of press/hold/release states and a gesture-level analysis that evaluates contour similarity of each press-release cycle. We apply this framework to compare an audio-only baseline with two variants: one incorporating symbolic information from MIDI, and another trained in a binary-valued setting, all within a unified architecture. Results show that the MIDI-informed model significantly outperforms the others at action and gesture levels, despite modest frame-level gains. These findings demonstrate that our framework captures musically relevant improvements indiscernible by traditional metrics, offering a more practical and effective approach to evaluating pedal depth estimation models.

cs.IR↗

arXiv2026

PHECT: A lightweight computation tool for pulsar halo emission

$γ$-ray pulsar halos, most likely formed by inverse Compton scattering of electrons and positrons propagating in the pulsar-surrounding interstellar medium with background photons, serve as an ideal probe for Galactic cosmic-ray propagation on small scales (typically tens of parsecs). While the associated electron and positron propagation is often modeled using homogeneous and isotropic diffusion, termed here as standard diffusion, the actual transport process is expected to be more complex. This work introduces the Pulsar Halo Emission Computation Tool (PHECT), a lightweight software designed for modeling pulsar halo emission. PHECT incorporates multiple transport models extending beyond standard diffusion, accounting for different possible origins of pulsar halos. Users can conduct necessary computations simply by configuring a YAML file without manual code edits. Furthermore, the tool adopts finite-volume discretizations that remain stable on non-uniform grids and in the presence of discontinuous diffusion coefficients. PHECT is ready for the increasingly precise observational data and the rapidly growing sample of pulsar halos.

astro-ph.HE↗

arXiv2025

Adversarial quantum channel discrimination

We introduce a new framework for quantum channel discrimination in an adversarial setting, where the tester plays against an adversary. We show that in asymmetric hypothesis testing, the optimal type-II error exponent is precisely characterized by a new notion of quantum channel divergence (termed the minimum output channel divergence). This serves as a direct analog of the quantum Stein's lemma in this new framework, and complements previous studies on ``best-case'' channel discrimination, thereby providing a complete understanding of the ultimate limits of quantum channel discrimination. Notably, the optimal error exponent can be achieved by simple non-adaptive adversarial strategies, and despite the need for regularization, it remains efficiently computable and satisfies the strong converse property in general. Furthermore, we show that entropy accumulation, a powerful tool in quantum cryptography, can be reframed as an adversarial channel discrimination problem, establishing a new connection between quantum information theory and quantum cryptography.

quant-ph↗

arXiv2025

Ultrahigh-Energy Gamma-ray Emission Associated with Black Hole-Jet Systems

Black holes (BH), one of the most intriguing objects in the universe, can manifest themselves through electromagnetic radiation initiated by the accretion flow. Some stellar-mass BHs drive relativistic jets when accreting matter from their companion stars, forming microquasars. Non-thermal emission from the radio to tera-electronvolt (TeV) gamma-ray band has been observed from microquasars, indicating the acceleration of relativistic particles. Here we report detection of four microquasars (SS 433, V4641 Sgr, GRS 1915+105, MAXI J1820+070) of spectrum extending to the ultrahigh-energy (UHE; photon energy $E>100$ TeV) band and one microquasar (Cygnus X-1) of spectrum approaching 100 TeV, using the Large High Altitude Air Shower Observatory (LHAASO). Notably, the total emission associated with SS 433 cannot be interpreted with a single leptonic component. In the UHE band, its emission is in spatial coincidence with a giant atomic cloud, which is consistent with a hadronic origin. An elongated source is discovered from V4641 Sgr with the spectrum continuing up to 800 TeV. The detection of UHE gamma rays demonstrates that accreting BHs and their environments can operate as extremely efficient accelerators of particles out of 1 peta-electronvolt (PeV), suggesting microquasars to be important contributors to Galactic cosmic rays especially around the `knee' region.

astro-ph.HE↗

arXiv2025

Finite-Key Analysis of Quantum Key Distribution with Characterized Devices Using Entropy Accumulation

The Entropy Accumulation Theorem (EAT) was introduced to significantly improve the finite-size rates for device-independent quantum information processing tasks such as device-independent quantum key distribution (QKD). A natural question would be whether it also improves the rates for device-dependent QKD. In this work, we provide an affirmative answer to this question. We present new tools for applying the EAT in the device-dependent setting. We present sufficient conditions for the Markov chain conditions to hold as well as general algorithms for constructing the needed min-tradeoff function. Utilizing Dupuis' recent privacy amplification without smoothing result, we improve the key rate by optimizing the sandwiched Rényi entropy directly rather than considering the traditional smooth min-entropy. We exemplify these new tools by considering several examples including the BB84 protocol with the qubit-based version and with a realistic parametric downconversion source, the six-state four-state protocol and a high-dimensional analog of the BB84 protocol.

quant-ph↗

arXiv2025

Generalized quantum Chernoff bound

We consider the task of distinguishing whether a quantum system is prepared in a state from one of several sets of quantum states. Assuming their convexity and stability under tensor product, we prove that the optimal error exponent for discrimination is precisely given by the regularized quantum Chernoff divergence between the sets, thereby establishing a generalized quantum Chernoff bound for the discrimination of multiple sets of quantum states. This extends the classical and quantum Chernoff bounds to the general setting of composite and correlated quantum hypotheses. Furthermore, leveraging minimax theorems, we show that discriminating between sets of quantum states is no harder than discriminating between their worst-case elements in terms of error probability. This implies the existence of an optimal state-agnostic test that achieves the minimum error probability for all states in the sets, matching the performance of the optimal state-dependent test for the most difficult pair of states. We provide explicit characterizations of the optimal state-agnostic test in the binary composite case. Finally, we show that the maximum overlap between a pure state and a set of free states, a quantity that frequently arises in quantum resource theories, is equal to the quantum Chernoff divergence between the sets, thereby providing an operational interpretation of this quantity in the context of symmetric hypothesis testing.

quant-ph↗

arXiv2025

A Giant Peanut-shaped Ultra-High-Energy Gamma-Ray Emitter Off the Galactic Plane

Ultra-high-energy (UHE), exceeding 100 TeV (10^12 electronvolts), γ-rays manifests extreme particle acceleration in astrophysical sources. Recent observations by γ-ray telescopes, particularly by the Large High Altitude Air Shower Observatory (LHAASO), have revealed a few tens of UHE sources, indicating numerous Galactic sources capable of accelerating particles to PeV (10^15 electronvolts) energies. However, discerning the dominant acceleration mechanisms (leptonic versus hadronic), the relative contributions of specific source classes, and the role of particle transport in shaping their observed emission are central goals of modern UHE astrophysics. Here we report the discovery of a giant UHE γ-ray emitter at -17.5° off the Galactic plane - a region where UHE γ-ray sources are rarely found. The emitter exhibits a distinctive asymmetric shape, resembling a giant "Peanut" spanning 0.45° \times 4.6°, indicative of anisotropic particle distribution over a large area. A highly aged millisecond pulsar (MSP) J0218+4232 is the sole candidate accelerator positionally coincident with the Peanut region. Its association with UHE γ-rays extending to 0.7 PeV, if confirmed, would provide the first evidence of a millisecond pulsar powering PeV particles. Such a finding challenges prevailing models, which posit that millisecond pulsars cannot sustain acceleration to PeV energies. The detection reveals fundamental gaps in understanding particle acceleration, cosmic-ray transport, and interstellar magnetic field effects, potentially revealing new PeV accelerator (PeVatron) classes.

astro-ph.HE↗