Deep learning-based image restoration has achieved significant success. However, when addressing real-world degradations, model performance is limited by the quality of groundtruth images in datasets due to practical constraints in data acquisition. To address this limitation, we propose a novel framework that enhances existing ground truth images to provide higher-quality supervision for real-world restoration. Our framework generates perceptually enhanced ground truth images using super-resolution by incorporating adaptive frequency masks, which are learned by a conditional frequency mask generator. These masks guide the optimal fusion of frequency components from the original ground truth and its super-resolved variants, yielding enhanced ground truth images. This frequency-domain mixup preserves the semantic consistency of the original content while selectively enriching perceptual details, preventing hallucinated artifacts that could compromise fidelity. The enhanced ground truth images are used to train a lightweight output refinement network that can be seamlessly integrated with existing restoration models. Extensive experiments demonstrate that our approach improves the qu
There can be many competing and contradictory explanations for a single model prediction, making it difficult to select which one to use. Current explanation evaluation frameworks measure quality by comparing against ideal "ground-truth" explanations, or by verifying model sensitivity to important inputs. We outline the limitations of these approaches, and propose three desirable principles to ground the future development of explanation evaluation strategies for local feature importance explanations. We propose a ground-truth Agnostic eXplanation Evaluation framework (AXE) for evaluating and comparing model explanations that satisfies these principles. Unlike prior approaches, AXE does not require access to ideal ground-truth explanations for comparison, or rely on model sensitivity - providing an independent measure of explanation quality. We verify AXE by comparing with baselines, and show how it can be used to detect explanation fairwashing. Our code is available at https://github.com/KaiRawal/Evaluating-Model-Explanations-without-Ground-Truth.
We study the orbital stability of action ground-states of the nonlinear Schrödinger equation over two particular cases of metric graphs, the $\mathcal{T}$ and the tadpole graphs. We show the existence of stability transitions near the $L^2$-critical exponent, a new dynamical feature of the nonlinear Schrödinger equation. More precisely, as the frequency $λ$ increases, the action ground-state transitions from stable to unstable and then back to stable (or vice-versa). This result is complemented with the stability analysis of ground-states in the asymptotic cases of low/high frequency and weak/strong nonlinear interaction. Finally, we present a numerical simulation of the stability of action ground-states depending on the nonlinearity and the frequency parameter, which validates the aforementioned theoretical results.
Even when large-scale, site-specific three-dimensional (3D) subsurface models are used to represent spatial variability, multi-dimensional ground response analyses (GRAs) at downhole array sites continue to exhibit amplitude discrepancies between simulated theoretical transfer functions (TTFs) and recorded empirical transfer functions (ETFs), with ETFs at the Delaney Park Downhole Array (DPDA) showing notably lower amplitudes at the fundamental frequency (f0). This discrepancy suggests greater apparent attenuation from wave scattering and destructive interference than is currently captured in multi-dimensional GRAs. However, most prior studies assume vertically propagating shear-wave input, neglecting inclined and azimuthally varying wavefields. This study evaluates the effects of inclination and azimuth in 2D and 3D GRAs at DPDA to assess whether non-vertical wave incidence improves agreement with observed ETFs. Two approaches for modeling inclined waves, the Input Lag Method (ILM) and the Inclined Domain Method (IDM), are compared, with ILM found to be more effective and computationally efficient for large-scale models. A parametric study using ILM shows that inclination angles u
We investigate the existence and the singular limit of normalized ground states for focusing doubly nonlinear Schrödinger equations with both standard and concentrated nonlinearities on two-dimensional square grids. First, we provide existence and non-existence results for such ground states depending on the values of the nonlinearity powers and on the structure of the set of vertices where the concentrated nonlinearities are located. Second, we prove that suitable piecewise-affine extensions of such states converge strongly in $H^1(\R^2)$ to ground states of corresponding doubly nonlinear models defined on the whole plane as the length of the edges in the grid tends to zero. This convergence is proved both for limit models with standard nonlinearities only and for models combining standard and singular nonlinearities concentrated on a line or on a strip.
External indexes can be used for cluster evaluation when ground truth is available. We review the most common external validity indexes focusing on set-matching-based measures. We recommend centroid index (CI), because it is an intuitive cluster-level measure with an explainable result. If we need a more fine-tuned, point-level measure, there are more choices. Pair-set index (PSI) provides a normalized score which is not biased by cluster sizes. If all points should matter equally, then clustering accuracy (ACC) or any other set-matching measure is suitable.
In the era of time-domain, multi-messenger astronomy, the detection of transient events on the high-energy electromagnetic sky has become more important than ever. The Gravitational wave high-energy Electromagnetic Counterpart All-sky Monitor (GECAM) is a dedicated mission to monitor gamma-ray transients, launched in December, 2020. A real-time on-board trigger and location software, using the traditional signal-to-noise ratio (SNR) method for blind search, is constrained to relatively bright signals due to the limitations in on-board computing resources and the need for real-time search. In this work, we developed a ground-based pipeline for GECAM to search for various transients, especially for weak bursts missed by on-board software. This pipeline includes both automatic and manual mode, offering options for blind search and targeted search. The targeted search is specifically designed to search for interesting weak bursts, such as gravitational wave-associated gamma-ray bursts (GRBs). From the ground search of the data in the first year, GECAM has been triggered by 54 GRBs and other transients, including soft gamma-ray repeaters, X-ray binaries, solar flares, terrestrial gamma-
A major thrust in quantum algorithm development over the past decade has been the search for the quantum algorithms that will deliver practical quantum advantage first. Today's quantum computers - and even early fault-tolerant quantum computers - are limited in the number of operations they can implement per circuit. We introduce quantum algorithms for ground-state energy estimation (GSEE) that accommodate this design constraint. The first algorithm estimates ground-state energies, offering a quadratic improvement on the ground state overlap parameter compared to other methods in this regime. The second algorithm certifies that the estimated ground-state energy is within a specified error tolerance of the true ground-state energy, addressing the issue of gap estimation that beleaguers several ground state preparation and energy estimation algorithms. We note, however, that the scaling of this certification technique is currently less favorable than that of the GSEE algorithm. To develop the certification algorithm, we propose a novel use of quantum computers to facilitate rejection sampling. After a classical computer generates initial samples, the quantum computer is used to accep
This paper concerns the existence and related properties of solutions to the Schrödinger-Bopp-Podolsky system, which reduces to a nonlinear and nonlocal partial differential equation describing a Schrödinger field coupled with its electromagnetic field in Bopp-Podolsky theory under purely electrostatic conditions. Firstly, by applying the mountain-pass lemma, we obtain the existence of nontrivial solutions. Then, through some estimates of the ground state energy, we prove the existence of ground state solutions. By exploring the relationship between solutions and paths associated with critical points, we further demonstrate that the obtained solutions are ground states of mountain-pass type. Additionally, the positivity, radial symmetry, rotational invariance, and exponential decay of the ground state solutions are considered. Finally, in the radial case, we explore the asymptotic behavior of the obtained solutions with respect to $a$.
Underground infrastructure such as pipelines and tunnels can be vulnerable to transient ground deformation (TGD) generated by earthquakes, traffic, and other vibration sources. Current design methods rely on simplified analytical models that idealize soil movement as a traveling sinusoidal wave, neglecting system inertia and relative soil-structure displacement. As shown in this study, such assumptions may be inadequate for large-diameter buried pipelines and tunnels, where accurate dynamic analysis under axial and transverse TGD is required. This paper introduces a new semi-analytical model for the dynamic response of buried Timoshenko beams on Winkler foundation subjected to transverse TGD. A closed-form solution of the governing differential equation shows that the vibration spectrum is divided into four parts, separated by three transition frequencies that depend on the system's mechanical and geometric properties. These transitions govern changes in modal behavior and significantly influence dynamic amplification. The model is verified through a case study of a buried 107 cm (42 in) steel water pipeline with varying lengths and operating conditions. Analytical predictions show
Recent endeavors in video editing have showcased promising results in single-attribute editing or style transfer tasks, either by training text-to-video (T2V) models on text-video data or adopting training-free methods. However, when confronted with the complexities of multi-attribute editing scenarios, they exhibit shortcomings such as omitting or overlooking intended attribute changes, modifying the wrong elements of the input video, and failing to preserve regions of the input video that should remain intact. To address this, here we present a novel grounding-guided video-to-video translation framework called Ground-A-Video for multi-attribute video editing. Ground-A-Video attains temporally consistent multi-attribute editing of input videos in a training-free manner without aforementioned shortcomings. Central to our method is the introduction of Cross-Frame Gated Attention which incorporates groundings information into the latent representations in a temporally consistent fashion, along with Modulated Cross-Attention and optical flow guided inverted latents smoothing. Extensive experiments and applications demonstrate that Ground-A-Video's zero-shot capacity outperforms other
In this paper, we study the following quasilinear {S}chrödinger equation: $$\left\{ \begin{array}{l} - {Δu} - \frac{κ}{2}Δ\left( {u}^{2}\right) u = h\left( u\right) \text{ in }{\mathbb{R}}^{N}, \\ u \in {H}^{1}\left( {\mathbb{R}}^{N}\right), \end{array}\right. $$ where $N \geq 3,κ> 0$ is a parameter, and $h$ satisfies Berestycki-Lions condition. By using a critical point theory on a topological manifold, we obtain the existence of a ground state for $N \geq 3$, a nonradial ground state for $N \geq 4$, and infinitely many nonradial solutions for $N = 4$ or $N \geq 6$. Our results generalize several classical works into quasilinear equations.
We study existence, nonexistence, and qualitative properties of ground states for a focusing, subcritical Nonlinear Schrödinger Equation on a hybrid plane, consisting of a half-line attached to a plane. Ground states are normalized minimizers of the associated energy, given by Nonlinear Schrödinger energies with contact interactions on the half-line and on the plane, plus a quadratic coupling term. At fixed mass, existence holds if the contact interaction on the half-line is not too repulsive, or the interaction on the plane is sufficiently attractive, or the coupling is strong enough. Nonexistence occurs when both interactions are sufficiently repulsive and the coupling is weak. Moreover, we discuss how the coupling affects the support and the symmetry properties of such ground states. These are the first results for a Nonlinear Schrödinger Equation on a mixed-dimensional manifold.
Third generation ground-based gravitational wave (GW) detectors, such as Einstein Telescope and Cosmic Explorer, will operate in the $(\text{few}-10^4)$ Hz frequency band, with a boost in sensitivity providing an unprecedented reach into primordial cosmology. Working concurrently with pulsar timing arrays in the nHz band, and LISA in the mHz band, these 3G detectors will be powerful probes of beyond the standard model particle physics on scales $T\gtrsim 10^{7}$GeV. Here we focus on their ability to probe phase transitions (PTs) in the early universe. We first overview the landscape of detectors across frequencies, discuss the relevance of astrophysical foregrounds, and provide convenient and up-to-date power-law integrated sensitivity curves for these detectors. We then present the constraints expected from GW observations on first order PTs and on topological defects (strings and domain walls), which may be formed when a symmetry is broken irrespective of the order of the phase transition. These constraints can then be applied to specific models leading to first order PTs and/or topological defects. In particular we discuss the implications for axion models, which solve the stron
The ground-based technique for imaging atmospheric Cherenkov telescopes became a rapidly developing and powerful branch of science. Thanks to this technique, over 250 very high-energy gamma-ray sources of galactic and extragalactic origin have been discovered. Many fundamental questions of astrophysics, astro-particle physics, the physics of cosmic rays and cosmology are the focus of this technique. In the past 33 years since the discovery of the first gamma-ray source, the Crab Nebula, the discipline has made remarkable progress. Today, the technology boasts highly sensitive telescopes capable of detecting a point source 100 times fainter than the standard candle, the Crab Nebula, in 25 hours of observation. Further developments in this technology led to the Cherenkov Telescope Array (CTA), the next-generation large instrument. The sensitivity of CTA will be several times higher than that of the current best instruments. This article presents a brief history of ground-based very high energy gamma-ray astrophysics.
Under the widely accepted but unproven assumption that the one-dimensional S=1 antiferromagnetic Heisenberg model has a unique gapped ground state, we prove that the model belongs to a nontrivial symmetry-protected topological (SPT) phase. In other words, we rigorously rule out the possibility that the model has a unique gapped ground state that is topologically trivial. To be precise, we assume that the models on open finite chains with boundary magnetic field have unique ground states with a uniform gap and prove that the ground state of the infinite chain has a nontrivial topological index. This further implies the presence of a gapless edge excitation in the model on the half-infinite chain and the existence of a topological phase transition in the model that interpolates between the Heisenberg chain and the trivial model.
We present the Near-Ultraviolet eXplorer (NUX), which will consist out of 4 small (36 cm diameter) ground-based telescopes that are optimized for the shortest wavelengths that are detectable from Earth (i.e., the near-UV [NUV] wavelength range of 300-350 nm). Each telescope will have a field-of-view of ~17 square degrees sampled at ~2.6"/pixel, and will reach a NUV magnitude (AB) of 20 in 2.5 minutes exposures (in dark time). The goal of NUX is to improve our understanding of the physical processes that power fast (days) to very fast (hours) hot transients, such as shock-breakout and shock-cooling emission of supernovae and the electromagnetic counterparts of gravitational wave events. Each telescope will be an off-the-shelf 14" Celestron RASA telescope, retrofitted with NUV optics. We have already demonstrated that the normal Schmidt corrector of this telescope can be replaced by a custom made one consisting of NUV transparent glass. Currently, a prototype NUX telescope is being fully assembled to demonstrate the technical and scientific feasibility of the NUX concept. Site tests will be held (in 2025/2026) at La Silla, Chile, to determine the NUV characteristics of the atmosphere
We present a novel vision-based control method to make a group of ground mobile robots achieve a specified formation shape with unspecified size. Our approach uses multiple aerial control units equipped with downward-facing cameras, each observing a partial subset of the multirobot team. The units compute the control commands from the ground robots' image projections, using neither calibration nor scene scale information, and transmit them to the robots. The control strategy relies on the calculation of image similarity transformations, and we show it to be asymptotically stable if the overlaps between the subsets of controlled robots satisfy certain conditions. The presence of the supervisory units, which coordinate their motions to guarantee a correct control performance, gives rise to a hybrid system topology. All in all, the proposed system provides relevant practical advantages in simplicity and flexibility. Within the problem of controlling a team shape, our contribution lies in addressing several simultaneous challenges: the controller needs only partial information of the robotic group, does not use distance measurements or global reference frames, is designed for unicycle
Classification of road users is important for traffic monitoring. The usability of a height estimate based on the two-ray ground-reflection model as a feature for the classification of vehicles is analyzed in this paper. The four-ray ground-reflection model for fast chirp ramp sequence waveforms of FMCW radars is derived and simplified to the well-known two-ray ground-reflection model. A spectrum from which the height of a target can be derived is obtained using the Lomb-Scargle periodogram. Measurements with two vehicle classes illustrate the approach and show that the model could be used as a feature to distinguish vehicles based on their height.
Laver, and Woodin independently, showed that models of ${\rm ZFC}$ are uniformly definable in their set-forcing extensions, using a ground model parameter. We investigate ground model definability for models of fragments of ${\rm ZFC}$, particularly of ${\rm ZF}+{\rm DC}_δ$ and of ${\rm ZFC}^-$, and we obtain both positive and negative results. Generalizing the results of Laver and Woodin, we show that models of ${\rm ZF}+{\rm DC}_δ$ are uniformly definable in their set-forcing extensions by posets admitting a gap at $δ$, using a ground model parameter. In particular, this means that models of ${\rm ZF}+{\rm DC}_δ$ are uniformly definable in their forcing extensions by posets of size less than $δ$. We also show that it is consistent for ground model definability to fail for models of ${\rm ZFC}^-$ of the form $H_{κ^+}$. Using forcing, we produce a ${\rm ZFC}$ universe in which there is a cardinal $κ>\!>ω$ such that $H_{κ^+}$ is not definable in its Cohen forcing extension. As a corollary, we show that there is always a countable transitive model of ${\rm ZFC}^-$ violating ground model definability. These results turn out to have a bearing on ground model definability for mode