Search

Recent advancements in pre-trained Vision-Language Models (VLMs) have highlighted the significant potential of prompt tuning for adapting these models to a wide range of downstream tasks. However, existing prompt tuning methods typically map an image to a single representation, limiting the model's ability to capture the diverse ways an image can be described. To address this limitation, we investigate the impact of visual prompts on the model's generalization capability and introduce a novel method termed Multi-Representation Guided Prompt Tuning (MePT). Specifically, MePT employs a three-branch framework that focuses on diverse salient regions, uncovering the inherent knowledge within images which is crucial for robust generalization. Further, we employ efficient self-ensemble techniques to integrate these versatile image representations, allowing MePT to learn all conditional, marginal, and fine-grained distributions effectively. We validate the effectiveness of MePT through extensive experiments, demonstrating significant improvements on both base-to-novel class prediction and domain generalization tasks.

We investigate the emission of circularly polarized photons from a magnetized quark-gluon plasma with nonzero quark-number and chiral charge chemical potentials. These chemical potentials qualitatively influence the differential emission rates of circularly polarized photons. A nonzero net electric charge density, induced by quark-number chemical potentials, enhances the overall emission of one circular polarization over the other, while a nonzero chiral charge density introduces a spatial asymmetry in the emission with respect to reflection in the transverse plane. The signs of the electrical and chiral charge densities determine which circular polarization dominates overall and whether the emission preferentially aligns with or opposes the magnetic field. Based on these findings, we propose that polarized photon emission is a promising observable for characterizing the quark-gluon plasma produced in heavy-ion collisions., Comment: 14 pages, 6 multi-panel figures

Seismic data reconstruction is an effective tool for compensating nonuniform and incomplete seismic geometry. Compared with methods for 2D seismic data, 3D reconstruction methods could consider more spatial structure correlation in seismic data. In the early studies, 3D reconstruction methods are mainly theory-driven and have some limitations due to their prior assumptions on the seismic data. To release these limitations, deep learning-based reconstruction methods rise and show potential in dealing with reconstruction problems. However, there are mainly two shortcomings in existing deep learning-methods. On the one hand, most of existing deep learning-based methods adopt the convolutional neural network, having some difficulties in dealing with data with complex or time-varying distributions. Recently, the diffusion model has been reported to possess the capability to solve data with complex distributions by gradually complicating the distribution of data to optimize the network. On the other hand, existing methods need enough paired-data to train the network, which are very hard to obtain especially for the starved 3D seismic data. Deep prior-based unsupervised and sampling-based self-supervised networks offer an available solution to this problem. In this paper, we develop a self-supervised diffusion model (S2DM) for 3D seismic data reconstruction. The proposed model mainly contains a diffusion restoration model and a variational time-spatial module. Extensive synthetic and field experiments demonstrate the superiority of the proposed S2DM algorithm., Comment: 43 pages, 13 figures

Polarized light has various potential applications in the communication bands, including optical communication, polarization imaging, quantum emission, and quantum communication. However, optimizing polarization control requires continuous improvements in areas such as dynamic tunability, materials, and efficiency. In this work, we propose a borophene-based structure capable of converting linearly polarized light into arbitrarily polarized light through the coherent excitation of localized surface plasmons (LSPs) in optical communication band. Furthermore, a double-layer borophene structure can be achieved by placing a second borophene array at the top of the first one with a 90{\deg} relative rotation of their crystalline plane. The rotation direction of the polarization state of the reflected light can be switched by independently controlling the carrier concentration of the two-layer borophene. Finally, a dipole source is used to realize the emission of polarized light, which is two orders of magnitude higher than the emission rate in free space, and the polarization state can be dynamically controlled by manipulating the carrier concentration. Our study is simple and compact, with potential applications in the fields of polarizers, polarization detectors, and quantum emitters.

Tensor program tuning is essential for the efficient deployment of deep neural networks. Search-based approaches have demonstrated scalability and effectiveness in automatically finding high-performance programs for specific hardware. However, the search process is often inefficient, taking hours or even days to discover optimal programs due to the exploration mechanisms guided by an accurate but slow learned cost model. Meanwhile, the learned cost model trained on one platform cannot seamlessly adapt online to another, which we call cross-platform online unawareness. In this work, we propose Pruner and MoA-Pruner. Pruner is a speculative exploration mechanism that accelerates the search process using a "Draft-then-Verify" paradigm. Instead of applying the complex learned cost model to all explored candidates, Pruner drafts small-scale speculative candidates by introducing a naive symbol analyzer (draft model), then identifies the best candidates by the learned cost model. MoA-Pruner introduces Momentum online Adaptation to address the cross-platform online unawareness. We incorporate these techniques into the Ansor and conduct extensive experiments on three GPU-based platforms. Results show that in online cost model tuning scenarios, Pruner and MoA-Pruner can achieve an average speedup of $2.6 \times$ and $4.82 \times$ compared to Ansor. In offline tuning scenarios, Pruner can achieve an average speedup of $4.75 \times$ and $4.05\times$ compared to TenSet and TLP, respectively. The code is available at https://github.com/qiaolian9/Pruner.

Why do agents adopt a particular general behavioral rule among a collection of possible alternatives? To address this question, we introduce a dynamic social learning framework, where agents rely on general rules of thumb and imitate the behavioral rules of successful peers. We find the social learning outcome can be characterized independent of the initial rule distribution. When one dominant general rule consistently yields superior problem-specific outcomes, social learning almost surely leads all agents to adopt this dominant rule; otherwise, provided the population is sufficiently large, the better rule for the more frequent problem becomes the consensus rule with arbitrarily high probability. As a result, the behavioral rule selected by the social learning process need not maximize social welfare. We complement our theoretical analysis with an application to the market sentiment selection in a stochastic production market., Comment: 47 pages, 1 figures

Industrial systems such as recommender systems and online advertising, have been widely equipped with multi-stage architectures, which are divided into several cascaded modules, including matching, pre-ranking, ranking and re-ranking. As a critical bridge between matching and ranking, existing pre-ranking approaches mainly endure sample selection bias (SSB) problem owing to ignoring the entire-chain data dependence, resulting in sub-optimal performances. In this paper, we rethink pre-ranking system from the perspective of the entire sample space, and propose Entire-chain Cross-domain Models (ECM), which leverage samples from the whole cascaded stages to effectively alleviate SSB problem. Besides, we design a fine-grained neural structure named ECMM to further improve the pre-ranking accuracy. Specifically, we propose a cross-domain multi-tower neural network to comprehensively predict for each stage result, and introduce the sub-networking routing strategy with $L0$ regularization to reduce computational costs. Evaluations on real-world large-scale traffic logs demonstrate that our pre-ranking models outperform SOTA methods while time consumption is maintained within an acceptable level, which achieves better trade-off between efficiency and effectiveness., Comment: 5 pages, 2 figures

We study the higher-order anisotropy coefficients $v_4$ and $v_6$ in the photon and dilepton emission from a hot magnetized quark-gluon plasma. Together with the earlier predictions for $v_2$, these results show a distinctive pattern of the anisotropy coefficients in several kinematic regimes. In the case of photon emission, nonzero coefficients $v_n$ (with even $n$) have opposite signs at small and large values of the transverse momentum (i.e., $k_T\lesssim \sqrt{|eB|}$ and $k_T\gtrsim \sqrt{|eB|}$, respectively). Additionally, the $v_n$ signs alternate with increasing $n$, and their approximate values decrease as $1/n^2$ in magnitude. The anisotropy of dilepton emission is well pronounced only at large transverse momenta and small invariant masses (i.e., when $k_T\gtrsim \sqrt{|eB|}$ and $M\lesssim \sqrt{|eB|}$). The corresponding $v_4$ and $v_6$ coefficients are of the same magnitude and show a similar alternating sign pattern with increasing $n$ as in the photon emission., Comment: 12 pages, 6 multi-panel figures, final version published in Physical Review D

In this work, we study the chiral and deconfinement phase transitions in a two-flavor Polyakov loop extended Nambu--Jona-Lasinio (PNJL) model. And note that the self-consistent mean field approximation is employed by introducing an arbitrary parameter $\alpha$ to measure the weights of the Fierz-transformed interaction channels. By making use of this model, we systematically investigate the chiral and deconfinement phase transition lines (as well as the chiral ones in the NJL model for comparison) under different values of $\alpha$. It is found that, the increasing of $\alpha$ helps to enhance the chiral (pseudo)critical temperature at fixed chemical potential, and also to enhance the chiral (pseudo)critical chemical potential at fixed temperature. And the critical end point (CEP) vanishes when $\alpha$ becomes large enough. Besides, we find that the incorporation of Polyakov loop increases $T_{CEP}$ but does not change $\mu_{CEP}$ for small values of $\alpha$.

In this study, we investigate the mass spectrum of $\pi$ and $\sigma$ mesons at finite chemical potential using the self-consistent NJL model and the Fierz-transformed interaction Lagrangian. The model introduces an arbitrary parameter $\alpha$ to reflect the weights of the Fierz-transformed interaction channels. We show that when $\alpha$ exceeds a certain threshold value, the chiral phase transition transforms from a first-order one to a smooth crossover, which is evident from the behaviors of the chiral condensates and meson masses. Additionally, at high chemical potential, the smaller the value of $\alpha$, the higher the masses of the $\pi$ and $\sigma$ mesons become. Moreover, the Mott and dissociation chemical potentials both increase with the increase in $\alpha$. Thus, the meson mass emerges as a valuable experimental observable for determining the value of $\alpha$ and investigating the properties of the chiral phase transition in dense QCD matter., Comment: Accepted by Chinese Physics C

In recent years, scientific CMOS (sCMOS) sensors have been vigorously developed and have outperformed CCDs in several aspects: higher readout frame rate, higher radiation tolerance, and higher working temperature. For silicon image sensors, image lag will occur when the charges of an event are not fully transferred inside pixels. It can degrade the image quality for optical imaging, and deteriorate the energy resolution for X-ray spectroscopy. In this work, the image lag of a sCMOS sensor is studied. To measure the image lag under low-light illumination, we constructed a new method to extract the image lag from X-ray photons. The image lag of a customized X-ray sCMOS sensor GSENSE1516BSI is measured, and its influence on X-ray performance is evaluated. The result shows that the image lag of this sensor exists only in the immediately subsequent frame and is always less than 0.05% for different incident photon energies and under different experimental conditions. The residual charge is smaller than 0.5 e- with the highest incident photon charge around 8 ke-. Compared to the readout noise level around 3 e-, the image lag of this sensor is too small to have a significant impact on the imaging quality and the energy resolution. The image lag shows a positive correlation with the incident photon energy and a negative correlation with the temperature. However, it has no dependence on the gain setting and the integration time. These relations can be explained qualitatively by the non-ideal potential structure inside the pixels. This method can also be applied to the study of image lag for other kinds of imaging sensors., Comment: accepted by NIM A

Scientific Complementary Metal Oxide Semiconductor (sCMOS) sensors are finding increasingly more applications in astronomical observations, thanks to their advantages over charge-coupled devices (CCDs) such as a higher readout frame rate, higher radiation tolerance, and higher working temperature. In this work, we investigate the performance at the individual pixel level of a large-format sCMOS sensor, GSENSE1516BSI, which has 4096 * 4096 pixels, each of 15 {\mu}m in size. To achieve this, three areas on the sCMOS sensor, each consisting of 99 * 99 pixels, are chosen for the experiment. The readout noise, conversion gain and energy resolutions of the individual pixels in these areas are measured from a large number (more than 25,000) of X-ray events accumulated for each of the pixels through long time exposures. The energy resolution of these pixels can reach 140 eV at 6.4 keV at room temperature and shows a significant positive correlation with the readout noise. The accurate gain can also be derived individually for each of the pixels from its X-ray spectrum obtained. Variations of the gain values are found at a level of 0.56% statistically among the 30 thousand pixels in the areas studied. With the gain of each pixel determined accurately, a precise gain correction is performed pixel by pixel in these areas, in contrast to the standardized ensemble gain used in the conventional method. In this way, we could almost completely eliminate the degradation of energy resolutions caused by gain variations among pixels. As a result, the energy resolution at room temperature can be significantly improved to 124.6 eV at 4.5 keV and 140.7 eV at 6.4 keV. This pixel-by-pixel gain correction method can be applied to all kinds of CMOS sensors, and is expected to find interesting applications in X-ray spectroscopic observations in the future., Comment: accepted by PASP

In recent years, scientific Complementary Metal Oxide Semiconductor (sCMOS) devices have been increasingly applied in X-ray detection, thanks to their attributes such as high frame rate, low dark current, high radiation tolerance and low readout noise. We tested the basic performance of a backside-illuminated (BSI) sCMOS sensor, which has a small pixel size of 6.5 um * 6.5 um. At a temperature of -20C, The readout noise is 1.6 e, the dark current is 0.5 e/pixel/s, and the energy resolution reaches 204.6 eV for single-pixel events. The effect of depletion depth on the sensor's performance was also examined, using three versions of the sensors with different deletion depths. We found that the sensor with a deeper depletion region can achieve a better energy resolution for events of all types of pixel splitting patterns, and has a higher efficiency in collecting photoelectrons produced by X-ray photons. We further study the effect of depletion depth on charge diffusion with a center-of-gravity (CG) model. Based on this work, a highly depleted sCMOS is recommended for applications of soft X-ray spectroscop., Comment: 11 pages, 13 figures. Accepted for publication in JInst

In recent years, scientific CMOS (sCMOS) sensors have found increasing applications to X-ray detection, including X-ray astronomical observations. In order to examine the performance of sCMOS sensors, we have developed X-ray cameras based on sCMOS sensors. Two cameras, CNX22 and CNX 66, have been developed using sCMOS sensors with a photosensitive area of 2 cm * 2 cm and 6 cm * 6 cm, respectively. The designs of the cameras are presented in this paper. The CNX22 camera has a frame rate of 48 fps, whereas CNX66 has a frame rate of currently 20 fps, that can be boosted to 100 fps in the future. The operating temperature of the sCMOS sensor can reach to -20C for CNX22 and -30C for CNX66 with a peltier cooler device. In addition to the commonly used mode of saving original images, the cameras provide a mode of real-time extraction of X-ray events and storage their information, which significantly reduces the requirement for data storage and offline analysis work. For both cameras, the energy resolutions can reach less than 200 eV at 5.9 keV using single-pixel events. These cameras are suitable for X-ray spectroscopy applications in laboratories and calibration for the space X-ray telescopes., Comment: 7 pages, 5 figures, published by Proc SPIE 2022

Using the Landau-level representation for the imaginary part of the photon polarization tensor, we derive an explicit expression for the dilepton production rate from a hot quark-gluon plasma in a quantizing background magnetic field. We study in detail the dependence of the production rate on the dilepton invariant mass and the transverse momentum at mid-rapidity. We also investigate the angular dependence and ellipticity of dilepton emission. By comparing the result with the zero-field Born approximation, we find that the magnetic field leads to a strong enhancement of the dilepton rate at small values of the invariant mass ($M\lesssim\sqrt{|eB|}$). In the same kinematic region, the dilepton production is characterized by a sizable ellipticity. At large values of the dilepton invariant mass ($M\gtrsim\sqrt{|eB|}$), the role of the magnetic field decreases and the result approaches the isotropic zero-field Born rate. By investigating the dependence of ellipticity on the transverse momentum, we argue that the future measurements of dilepton rate in the region of small invariant masses can constrain the magnetic field produced in heavy-ion collisions., Comment: 13 pages, 7 multi-panel figures; v3: the final version to appear in Physical Review D

We investigate the screening masses of neutral pions and sigma mesons in hot and magnetized quark matter in the framework of a two-flavor lattice-improved Nambu-Jona-Lasinio (NJL) model with a magnetic field dependent coupling constant, which is determined by utilizing the results from lattice QCD simulations. Since such model can well reproduce inverse magnetic catalysis (IMC), by comparing with the standard NJL model, we systemically analyze the impacts of IMC on the temperature and magnetic field dependences of the longitudinal and transverse screening masses of the chiral partners, i.e. {\pi}^0 and {\sigma} mesons, as well as the screening mass differences between them. Particularly, it is found that the eB dependences of two alternative (pseudo)critical temperatures for the chiral transition defined by {\sigma}-{\pi}^0 meson screening mass differences are consistent with that defined by the quark condensate.