Search

Cache-based approaches stand out as both effective and efficient for adapting vision-language models (VLMs). Nonetheless, the existing cache model overlooks three crucial aspects. 1) Pre-trained VLMs are mainly optimized for image-text similarity, neglecting the importance of image-image similarity, leading to a gap between pre-training and adaptation. 2) The current cache model is based on the Nadaraya-Watson (N-W) estimator, which disregards the intricate relationships among training samples while constructing weight function. 3) Under the condition of limited samples, the logits generated by cache model are of high uncertainty, directly using these logits without accounting for the confidence could be problematic. This work presents three calibration modules aimed at addressing the above challenges. Similarity Calibration refines the image-image similarity by using unlabeled images. We add a learnable projection layer with residual connection on top of the pre-trained image encoder of CLIP and optimize the parameters by minimizing self-supervised contrastive loss. Weight Calibration introduces a precision matrix into the weight function to adequately model the relation between training samples, transforming the existing cache model to a Gaussian Process (GP) regressor, which could be more accurate than N-W estimator. Confidence Calibration leverages the predictive variances computed by GP Regression to dynamically re-scale the logits of cache model, ensuring that the cache model's outputs are appropriately adjusted based on their confidence levels. Besides, to reduce the high complexity of GPs, we further propose a group-based learning strategy. Integrating the above designs, we propose both training-free and training-required variants. Extensive experiments on 11 few-shot classification datasets validate that the proposed methods can achieve state-of-the-art performance., Comment: submitted to IJCV

The Isaacson picture provides a general method for describing the effects of gravitational waves, established on two sets of basic equations. The first set of equations describes the propagation of gravitational waves within a given background, while the second set elucidates the interaction between the effective energy-momentum tensor of gravitational waves and the background metric. These two sets of equations are typically derived by expanding the field equations to second-order perturbations. In addition to the method of perturbing field equations, there is an alternative method: obtaining the two sets of basic equations by expanding the action to the second-order of perturbations. In this paper, we elaborate on this method, establishing its foundations more rigorously. Especially, the second-order perturbation action, with the Minkowski metric as the background, contains all necessary information to construct the Isaacson picture far from the source. We outline the derivation of the two sets of basic equations to describe the gravitational wave effects in the Isaacson picture, based on the second-order perturbation action. This will enable us to analyze the following effects: polarizations of gravitational waves, the wave speed for each polarization mode, the effective energy-momentum tensor of gravitational waves, and the nonlinear memory effect of gravitational waves. We illustrate this method by applying it to the most general second-order vector-tensor theory including parity-breaking terms. We further analyze the polarization modes of gravitational waves in this theory. We highlight that this method facilitates model-independent research on various gravitational wave effects in general modified gravity theories. These effects are anticipated to be identified by the upcoming generation of gravitational wave detectors, aimed at testing potential modifications to gravity theory., Comment: 85 pages, 1 figure

Quantum noise and thermal noise are the two primary sources of noise that limit the sensitivity of weak magnetic field sensing. Although quantum noise has been widely addressed, effectively reducing thermal noise remains challenging in detecting weak magnetic fields. We employ an anisotropic elliptical YIG sphere as a magnetic field probe to establish a parametric amplification interaction of magnons and induce magnon squeezing effects. These effects can effectively suppress thermal noise in the magnon mode and amplify weak magnetic field signals from external sources. Specifically, complete suppression of thermal noise can be achieved by placing the YIG sphere in a squeezed vacuum reservoir. Our scheme has the potential to inspire advancements in thermal noise suppression for quantum sensing.

In the realm weak force sensing, an important issue is to suppress fundamental noise (quantum noise and thermal noise), as they limit the accuracy of force measurement. In this paper, we investigate a weak force sensing scheme that combines a degenerate optical parametric amplifier (OPA) and an auxiliary mechanical oscillator into a cavity optomechanical system to reduce quantum noise. We demonstrate that the noise reduction of two coupled oscillators depends on their norm mode splitting. and provide a classic analogy and quantum perspective for further clarification. Besides, the noise reduction mechanism of OPA is to reduce the fluctuation of photon number and enhance the squeezing of the cavity field. We propose a specific design aimed at enhancing the joint effect of both, beyond what can be achieved using OPA alone or two series coupled oscillators. This scheme provides a new perspective for deeper understanding of cavity field squeezing and auxiliary oscillator in force sensing., Comment: 9 pages,10 figures

Macromolecules can gain special properties by adopting knotted conformations, but engineering knotted macromolecules is a challenging task. Here we surprisingly observed that knotting can be very effectively produced in active polymers. When one end of an actively reptative polymer is anchored, it can undergo continual self-knotting as a result of intermittent giant conformation fluctuations and the outward reptative motion. Once a knot is formed, it migrates to the anchored point due to a non-equilibrium ratchet effect. Moreover, when the active polymer is grafted on the end of a passive polymer, it can function as a self-propelling soft needle to either transfer its own knots to the passive polymer or directly braid knots on the passive polymer. We further show that these active needles can create inter-molecular bridging knots between two passive polymers. Our finding highlights the non-equilibrium effects in modifying the dynamic pathways of polymer systems, which have potential applications in macromolecular topology engineering, e.g., manipulating topological states of proteins and nucleic acids, as well as macromolecular braiding.

We study the polarization modes of gravitational waves in general Einstein-vector theory with an arbitrary constant background vector field under a Minkowski background. We compare these polarization modes with those of other vector-tensor theories and constrain the parameter spaces based on the gravitational-wave event GW170817 with its electromagnetic counterpart GRB170817A and observations of pulsar timing arrays. The presence of the background vector field leads to the anisotropy of space and a rich variety of gravitational waves contents. Our results reveal that the polarization modes of gravitational waves depend on the parameter spaces. There are at least two and at most five independent polarization modes in one parameter space. In different parameter spaces, some mixture modes are allowed, including tensor-vector, tensor-scalar, tensor-vector-scalar, vector-scalar, and scalar-scalar mixture modes, as well as five independent modes (excluding $P_l$): $P_+$, $P_{\times}$, $P_x$, $P_y$, and $P_b$. In all regions of the parameter spaces, there are always two tensor modes, which can be either independent of or mixed with other modes. The independent $P_b$ mode consistently exhibits the same speed as light. If the speed of tensorial gravitational waves strictly equals that of light, only the $P_+$, $P_{\times}$, and $P_b$ modes are permitted. Furthermore, through comparisons of some vector-tensor theories and based on the observations of gravitational waves, the $P_b$ mode is expected to be allowed.

Attitude is one of the crucial parameters for space objects and plays a vital role in collision prediction and debris removal. Analyzing light curves to determine attitude is the most commonly used method. In photometric observations, outliers may exist in the obtained light curves due to various reasons. Therefore, preprocessing is required to remove these outliers to obtain high quality light curves. Through statistical analysis, the reasons leading to outliers can be categorized into two main types: first, the brightness of the object significantly increases due to the passage of a star nearby, referred to as "stellar contamination," and second, the brightness markedly decreases due to cloudy cover, referred to as "cloudy contamination." Traditional approach of manually inspecting images for contamination is time-consuming and labor-intensive. However, We propose the utilization of machine learning methods as a substitute. Convolutional Neural Networks (CNN) and Support Vector Machines (SVM) are employed to identify cases of stellar contamination and cloudy contamination, achieving F1 scores of 1.00 and 0.98 on test set, respectively. We also explored other machine learning methods such as Residual Network-18 (ResNet-18) and Light Gradient Boosting Machine (lightGBM), then conducted comparative analyses of the results., Comment: 12 pages, 15 figures

We propose a generalized method for boosting the generalization ability of pre-trained vision-language models (VLMs) while fine-tuning on downstream few-shot tasks. The idea is realized by exploiting out-of-distribution (OOD) detection to predict whether a sample belongs to a base distribution or a novel distribution and then using the score generated by a dedicated competition based scoring function to fuse the zero-shot and few-shot classifier. The fused classifier is dynamic, which will bias towards the zero-shot classifier if a sample is more likely from the distribution pre-trained on, leading to improved base-to-novel generalization ability. Our method is performed only in test stage, which is applicable to boost existing methods without time-consuming re-training. Extensive experiments show that even weak distribution detectors can still improve VLMs' generalization ability. Specifically, with the help of OOD detectors, the harmonic mean of CoOp and ProGrad increase by 2.6 and 1.5 percentage points over 11 recognition datasets in the base-to-novel setting., Comment: Accepted by AAAI2024

Context Optimization (CoOp) has emerged as a simple yet effective technique for adapting CLIP-like vision-language models to downstream image recognition tasks. Nevertheless, learning compact context with satisfactory base-to-new, domain and cross-task generalization ability while adapting to new tasks is still a challenge. To tackle such a challenge, we propose a lightweight yet generalizable approach termed Compositional Kronecker Context Optimization (CK-CoOp). Technically, the prompt's context words in CK-CoOp are learnable vectors, which are crafted by linearly combining base vectors sourced from a dictionary. These base vectors consist of a non-learnable component obtained by quantizing the weights in the token embedding layer, and a learnable component constructed by applying Kronecker product on several learnable tiny matrices. Intuitively, the compositional structure mitigates the risk of overfitting on training data by remembering more pre-trained knowledge. Meantime, the Kronecker product breaks the non-learnable restrictions of the dictionary, thereby enhancing representation ability with minimal additional parameters. Extensive experiments confirm that CK-CoOp achieves state-of-the-art performance under base-to-new, domain and cross-task generalization evaluation, but also has the metrics of fewer learnable parameters and efficient training and inference speed.

Cells not only have self-motility by crawling but are also capable of other types of active motions like periodic contraction or pulsation. In this work, by using the Voronoi cell model, we show how this non-motility activity affects the structure, dynamic and density fluctuations of cellular monolayers. We find that random cell pulsation can fluidize solid epithelial tissues, resulting in hyperuniform fluids, while pulsation synchronization inhibits the fluidity and causes a reversed solidification. We prove this solidification is a BKT-type transition, characterized by strong density/dynamic heterogeneity arising from the birth of topological defects in the pulsating phase space. The magnitude and length scale of density heterogeneity diverge with the pulsating period, resulting in an opposite giant density fluctuation. We propose an unified fluctuating hydrodynamic theory to explain the two very different anomalous density fluctuations phenomena. Our findings can help understanding recent experimental observations in MDCK monolayers., Comment: 7pages, 3 figures

The crucial limitation of improving the sensitivity of the detection of weak magnetic fields is the unavoidable measurement noise. In this paper, we propose a scheme to achieve precise sensing robust against additional noise by employing a dual-frequency bias field modulation within a cavity magnon system. We find that the anti-rotating wave term can amplify the signal of the detected magnetic field, but this amplification effect must coexist with the rotating wave term. In particular, by the bias field modulation, we find the robustness against cavity field thermal noise is substantially enhanced, quantum noise and cavity field thermal noise is greatly reduced, and the external magnetic field signal is amplified, thereby improving the weak magnetic field sensing system's sensitivity. Compared with the previous scheme, our scheme requires neither an ultra- or deep-strong coupling mechanism nor the suppression of the additional noise by increasing the electromagnetic cooperativity. Our scheme could provide a valuable candidate for weak magnetic field sensing.

The mechanical strain can control the frequency of two-level atoms in amorphous material. In this work, we would like to employ two coupled two-level atoms to manipulate the magnitude and direction of heat transport by controlling mechanical strain to realize the function of a thermal switch and valve. It is found that a high-performance heat diode can be realized in the wide Piezo voltage range at different temperatures. We also discuss the dependence of the rectification factor on temperatures and couplings of heat reservoirs. We find that the higher temperature differences correspond to the larger rectification effect. The asymmetry system-reservoir coupling strength can enhance the magnitude of heat transfer, and the impact of asymmetric and symmetric coupling strength on the performance of the heat diode is complementary. It may provide an efficient way to modulate and control heat transport's magnitude and flow preference. This work may give insight into designing and tuning quantum heat machines., Comment: 10 pages, 9 figures;Accepted for publication in Physical Review E

A quantum thermal device based on three nearest-neighbor coupled spin-1/2 systems controlled by the magnetic field is proposed. We systematically study the steady-state thermal behaviors of the system. When the two terminals of our system are in contact with two thermal reservoirs, respectively, the system behaves as a perfect thermal modulator that can manipulate heat current from zero to specific values by adjusting magnetic field direction over different parameter ranges, since the longitudinal magnetic field can completely block the heat transport. Significantly, the modulator can also be achieved when a third thermal reservoir perturbs the middle spin. We also find that the transverse field can induce the system to separate into two subspaces in which neither steady-state heat current vanishes, thus providing an extra level of control over the heat current through the manipulation of the initial state. In addition, the performance of this device as a transistor can be enhanced by controlling the magnetic field, achieving versatile amplification behaviors, in particular substantial amplification factors., Comment: 18 pages, 9 figures