Search

Fine-tuning pretrained large models to downstream tasks is an important problem, which however suffers from huge memory overhead due to large-scale parameters. This work strives to reduce memory overhead in fine-tuning from perspectives of activation function and layer normalization. To this end, we propose the Approximate Backpropagation (Approx-BP) theory, which provides the theoretical feasibility of decoupling the forward and backward passes. We apply our Approx-BP theory to backpropagation training and derive memory-efficient alternatives of GELU and SiLU activation functions, which use derivative functions of ReLUs in the backward pass while keeping their forward pass unchanged. In addition, we introduce a Memory-Sharing Backpropagation strategy, which enables the activation memory to be shared by two adjacent layers, thereby removing activation memory usage redundancy. Our method neither induces extra computation nor reduces training efficiency. We conduct extensive experiments with pretrained vision and language models, and the results demonstrate that our proposal can reduce up to $\sim$$30\%$ of the peak memory usage. Our code is released at https://github.com/yyyyychen/LowMemoryBP., Comment: 25 pages, ICML 2024 Accepted

Though achieving marvelous progress in various scenarios, existing semantic communication frameworks mainly consider single-input single-output Gaussian channels or Rayleigh fading channels, neglecting the widely-used multiple-input multiple-output (MIMO) channels, which hinders the application into practical systems. One common solution to combat MIMO fading is to utilize feedback MIMO channel state information (CSI). In this paper, we incorporate MIMO CSI into system designs from a new perspective and propose the learnable CSI fusion semantic communication (LCFSC) framework, where CSI is treated as side information by the semantic extractor to enhance the semantic coding. To avoid feature fusion due to abrupt combination of CSI with features, we present a non-invasive CSI fusion multi-head attention module inside the Swin Transformer. With the learned attention masking map determined by both source and channel states, more robust attention distribution could be generated. Furthermore, the percentage of mask elements could be flexibly adjusted by the learnable mask ratio, which is produced based on the conditional variational interference in an unsupervised manner. In this way, CSI-aware semantic coding is achieved through learnable CSI fusion masking. Experiment results testify the superiority of LCFSC over traditional schemes and state-of-the-art Swin Transformer-based semantic communication frameworks in MIMO fading channels., Comment: This paper has been accepted by IEEE Transactions on Wireless Communications

Embedding invisible hyperlinks or hidden codes in images to replace QR codes has become a hot topic recently. This technology requires first localizing the embedded region in the captured photos before decoding. Existing methods that train models to find the invisible embedded region struggle to obtain accurate localization results, leading to degraded decoding accuracy. This limitation is primarily because the CNN network is sensitive to low-frequency signals, while the embedded signal is typically in the high-frequency form. Based on this, this paper proposes a Dual-Branch Dual-Head (DBDH) neural network tailored for the precise localization of invisible embedded regions. Specifically, DBDH uses a low-level texture branch containing 62 high-pass filters to capture the high-frequency signals induced by embedding. A high-level context branch is used to extract discriminative features between the embedded and normal regions. DBDH employs a detection head to directly detect the four vertices of the embedding region. In addition, we introduce an extra segmentation head to segment the mask of the embedding region during training. The segmentation head provides pixel-level supervision for model learning, facilitating better learning of the embedded signals. Based on two state-of-the-art invisible offline-to-online messaging methods, we construct two datasets and augmentation strategies for training and testing localization models. Extensive experiments demonstrate the superior performance of the proposed DBDH over existing methods., Comment: 7 pages, 6 figures (Have been accepted by IJCNN 2024)

We analyse the massless particles orbiting a spherically symmetric, asymptotically flat black hole with a radius equal to the photon sphere and a circular geodesic. Asymptotic observers record the orbital period of the null circular geodesic as the lowest among all possible paths around the compact object. We proceed with the analytical study of massless particle motion to demonstrate the instability of an orbit with null circularity. Furthermore, our results apply to any asymptotically flat and spherically symmetric spacetime.

This paper studies the joint data and semantics lossy compression problem, i.e., an extension of the hidden lossy source coding problem that entails recovering both the hidden and observable sources. We aim to study the nonasymptotic and second-order properties of this problem, especially the converse aspect. Specifically, we begin by deriving general nonasymptotic converse bounds valid for general sources and distortion measures, utilizing properties of distortion-tilted information. Subsequently, a second-order converse bound is derived under the standard block coding setting through asymptotic analysis of the nonasymptotic bounds. This bound is tight since it coincides with a known second-order achievability bound. We then examine the case of erased fair coin flips (EFCF), providing its specific nonasymptotic achievability and converse bounds. Numerical results under the EFCF case demonstrate that our second-order asymptotic approximation effectively approximates the optimum rate at given blocklengths., Comment: 13 pages, 3 figures. arXiv admin note: substantial text overlap with arXiv:2401.14962

This paper considers the joint compression of a pair of correlated sources, where the encoder is allowed to access only one of the sources. The objective is to recover both sources under separate distortion constraints for each source while minimizing the rate. This problem generalizes the indirect lossy source coding problem by also requiring the recovery of the observed source. In this paper, we aim to study the nonasymptotic and second-order asymptotic properties of this problem. Specifically, we begin by deriving nonasymptotic achievability and converse bounds valid for general sources and distortion measures. The source dispersion (Gaussian approximation) is then determined through asymptotic analysis of the nonasymptotic bounds. We further examine the case of erased fair coin flips (EFCF) and provide its specific nonasymptotic achievability and converse bounds. Numerical results under the EFCF case demonstrate that our second-order asymptotic approximation closely approximates the optimum rate at appropriately large blocklengths., Comment: 23 pages, 5 figures

Based on the fluid volume fraction large eddy simulation discrete phase coupling model, the water and sediment flow around the vibrating monopile was simulated. The discrete phase momentum model was enhanced using the force distribution of the particle phase on the bed surface. Additionally, the flow dynamics and sediment incipience around a vibrating monopile was replicated through the use of a sediment incipience Shields number. The results suggest that monopile vibrations amplify the negative incipience of sediment in both the front deceleration region and the wake reflux region. Furthermore, within the periodic flow field surrounding the monopile, the temporal changes in the sediment incipience Shields number does not consistently match the temporal changes in the spatial average velocity.

Cone Beam Computed Tomography (CBCT) plays a key role in dental diagnosis and surgery. However, the metal teeth implants could bring annoying metal artifacts during the CBCT imaging process, interfering diagnosis and downstream processing such as tooth segmentation. In this paper, we develop an efficient Transformer to perform metal artifacts reduction (MAR) from dental CBCT images. The proposed MAR Transformer (MARformer) reduces computation complexity in the multihead self-attention by a new Dimension-Reduced Self-Attention (DRSA) module, based on that the CBCT images have globally similar structure. A Patch-wise Perceptive Feed Forward Network (P2FFN) is also proposed to perceive local image information for fine-grained restoration. Experimental results on CBCT images with synthetic and real-world metal artifacts show that our MARformer is efficient and outperforms previous MAR methods and two restoration Transformers., Comment: under consideration of Computer Vision and Image Understanding journal

Adversarial face examples possess two critical properties: Visual Quality and Transferability. However, existing approaches rarely address these properties simultaneously, leading to subpar results. To address this issue, we propose a novel adversarial attack technique known as Adversarial Restoration (AdvRestore), which enhances both visual quality and transferability of adversarial face examples by leveraging a face restoration prior. In our approach, we initially train a Restoration Latent Diffusion Model (RLDM) designed for face restoration. Subsequently, we employ the inference process of RLDM to generate adversarial face examples. The adversarial perturbations are applied to the intermediate features of RLDM. Additionally, by treating RLDM face restoration as a sibling task, the transferability of the generated adversarial face examples is further improved. Our experimental results validate the effectiveness of the proposed attack method., Comment: \copyright 2023 IEEE. Personal use of this material is permitted. Permission from IEEE must be obtained for all other uses, in any current or future media, including reprinting/republishing this material for advertising or promotional purposes, creating new collective works, for resale or redistribution to servers or lists, or reuse of any copyrighted component of this work in other works

We look into the neutrino-antineutrino pair ($\nu+\bar{\nu}\longrightarrow e^{-}+e^{+}$) annihilation in the Damour-Solodukhin wormhole spacetime whose metric component involves a shift in contrast to the similar black hole. The deep analysis of the surface temperature of the accretion disk of static, spherically symmetric black-hole-like wormholes from R. Kh. Karimov et.al. reveals that the accretion disks of the wormholes are hotter than that of comparable black holes, indicating that the wormholes accretion disk can release neutrinos. Further we investigate the energy deposition rate from the neutrino pair annihilation around the Damour-Solodukhin wormhole thought as a mimicker of Schwarzschild black hole. By comparison made between the black-hole-like wormhole and the similar black hole, we demonstrate that the wormhole's accretion disk drawing the annihilation can become a source of gamma-ray burst although the more significant deviation from the similar black hole reduces the emitted power slightly. The ratio of energy deposition per unit time from the annihilation surrounding the accretion disk of the Damour-Solodukhin wormhole over the emitting power of black hole might alter noticeably depending on how slightly the metrics of the wormhole differ from the black hole spacetime., Comment: 12 pages, 8 figures

Multi-node communication, which refers to the interaction among multiple devices, has attracted lots of attention in many Internet-of-Things (IoT) scenarios. However, its huge amounts of data flows and inflexibility for task extension have triggered the urgent requirement of communication-efficient distributed data transmission frameworks. In this paper, inspired by the great superiorities on bandwidth reduction and task adaptation of semantic communications, we propose a federated learning-based semantic communication (FLSC) framework for multi-task distributed image transmission with IoT devices. Federated learning enables the design of independent semantic communication link of each user while further improves the semantic extraction and task performance through global aggregation. Each link in FLSC is composed of a hierarchical vision transformer (HVT)-based extractor and a task-adaptive translator for coarse-to-fine semantic extraction and meaning translation according to specific tasks. In order to extend the FLSC into more realistic conditions, we design a channel state information-based multiple-input multiple-output transmission module to combat channel fading and noise. Simulation results show that the coarse semantic information can deal with a range of image-level tasks. Moreover, especially in low signal-to-noise ratio and channel bandwidth ratio regimes, FLSC evidently outperforms the traditional scheme, e.g. about 10 peak signal-to-noise ratio gain in the 3 dB channel condition., Comment: This paper has been accepted by IEEE Internet of Things Journal

We research on the neutrino pair annihilation $\nu+\overline{\nu}\longrightarrow e^{-}+e^{+}$ around a massive source in asymptotic safety. Since neutrinos and photons have the same geodesic equation around black holes, we can estimate the radius where the neutrinos will be released by obtaining a series of trajectory curves with various correction values $\xi$. The black hole shadow radius is influenced by the correction parameter $\xi$. The black hole shadow radius decreases with increasing the $\xi$. The ratio $\dfrac{\dot{Q}}{\dot{Q}_{Newt}}$ corresponding to the energy deposition per unit time over that in the Newtonian case is derived and calculated. We find that the quantum corrections to the black hole spacetime affect the emitted energy rate ratio for the annihilation. It is interesting that the more considerable quantum effect reduces the ratio value slightly. Although the energy conversion is damped because of the quantum correction, the energy deposition rate is enough during the neutrino-antineutrino annihilation. The corrected annihilation process can become a source of gamma ray burst. We also investigate the derivative $\dfrac{\mathrm{d}\dot{Q}}{\mathrm{d}r}$ relating to the star's radius $r$ to show that the quantum effect for the black hole will drop the ratio. The more manifest quantum gravity influence leads the weaker neutrino pair annihilation., Comment: 17 pages, 5 figures

The channel attention mechanism is a useful technique widely employed in deep convolutional neural networks to boost the performance for image processing tasks, eg, image classification and image super-resolution. It is usually designed as a parameterized sub-network and embedded into the convolutional layers of the network to learn more powerful feature representations. However, current channel attention induces more parameters and therefore leads to higher computational costs. To deal with this issue, in this work, we propose a Parameter-Free Channel Attention (PFCA) module to boost the performance of popular image classification and image super-resolution networks, but completely sweep out the parameter growth of channel attention. Experiments on CIFAR-100, ImageNet, and DIV2K validate that our PFCA module improves the performance of ResNet on image classification and improves the performance of MSRResNet on image super-resolution tasks, respectively, while bringing little growth of parameters and FLOPs.

A novel distributed source coding model which named semantic-aware multi-terminal (MT) source coding is proposed and investigated in the paper, where multiple agents independently encode an imperceptible semantic source, while both semantic and observations are reconstructed within their respective fidelity criteria. We start from a generalized single-letter characterization of sum rate-distortion region of this problem. Furthermore, we propose a mixed MSE-Log loss framework for this model and specifically depict the rate-distortion bounds when sources are Gaussian mixture distributed. For this case, we first present a relative tight outer bound and explore the activeness of semantic and observation distortion constraints, in which we find that good observation reconstruction will not incur too much semantic errors, but not vice versa. Moreover, we provide a practical coding scheme functioning as an achievable regime of inner bound with the performance analysis and simulation results, which verifies the feasibility of the idea "detect and compress" for Gaussian mixture sources. Our results provide theoretical instructions on the fundamental limits and can be used to guide the practical semantic-aware coding designs for multi-user scenarios.

In this paper, the analysis of excess distortion exponent for joint source-channel coding (JSCC) in semantic-aware communication systems is presented. By introducing an unobservable semantic source, we extend the classical results by Csiszar to semantic-aware communication systems. Both upper and lower bounds of the exponent for the discrete memoryless source-channel pair are established. Moreover, an extended achievable bound of the excess distortion exponent for MIMO systems is derived. Further analysis explores how the block fading and numbers of antennas influence the exponent of semanticaware MIMO systems. Our results offer some theoretical bounds of error decay performance and can be used to guide future semantic communications with joint source-channel coding scheme., Comment: 31 pages, 5 figures, accepted by IEEE Transactions on Wireless Communications

Face recognition (FR) models can be easily fooled by adversarial examples, which are crafted by adding imperceptible perturbations on benign face images. The existence of adversarial face examples poses a great threat to the security of society. In order to build a more sustainable digital nation, in this paper, we improve the transferability of adversarial face examples to expose more blind spots of existing FR models. Though generating hard samples has shown its effectiveness in improving the generalization of models in training tasks, the effectiveness of utilizing this idea to improve the transferability of adversarial face examples remains unexplored. To this end, based on the property of hard samples and the symmetry between training tasks and adversarial attack tasks, we propose the concept of hard models, which have similar effects as hard samples for adversarial attack tasks. Utilizing the concept of hard models, we propose a novel attack method called Beneficial Perturbation Feature Augmentation Attack (BPFA), which reduces the overfitting of adversarial examples to surrogate FR models by constantly generating new hard models to craft the adversarial examples. Specifically, in the backpropagation, BPFA records the gradients on pre-selected feature maps and uses the gradient on the input image to craft the adversarial example. In the next forward propagation, BPFA leverages the recorded gradients to add beneficial perturbations on their corresponding feature maps to increase the loss. Extensive experiments demonstrate that BPFA can significantly boost the transferability of adversarial attacks on FR., Comment: \c{opyright} 2023 IEEE. Personal use of this material is permitted. Permission from IEEE must be obtained for all other uses, in any current or future media, including reprinting/republishing this material for advertising or promotional purposes, creating new collective works, for resale or redistribution to servers or lists, or reuse of any copyrighted component of this work in other works

In this work we investigate the neutrino pair annihilation around a gravitational object involving an $f(R)$ global monopole. We derive and calculate the ratio $\frac{\dot{Q}}{\dot{Q_{Newt}}}$ meaning that the energy deposition per unit time is over that in the Newtonian case. It is found that the more deviation from general relativity leads more energy to set free from the annihilation with greater ratio value. It should also be pointed out that the existence of global monopole makes a sharp increase in the ratio $\frac{\dot{Q}}{\dot{Q_{Newt}}}$, causing heavier gamma-ray burst. We also discuss the derivative $\frac{d\dot{Q}}{dr}$ as a function of radius $r$ of star to show the similar characters that the considerable modification of Einstein's gravity and the global monopole with unified theory order will raise the amount of $\frac{d\dot{Q}}{dr}$ greatly. The stellar body with $f(R)$ global monopole can be well qualified as a source of gamma-ray bursts. Moreover, we can select the factor $\psi_{0}$ to be comparable with the accelerating universe while regulate the parameter $\eta$ for the global monopole in order to make the ratio curves to coincide with the results from astronomy. It is possible to probe the monopole from astrophysical observations., Comment: 4 figures

Archaeology has long faced fundamental issues of sampling and scalar representation. Traditionally, the local-to-regional-scale views of settlement patterns are produced through systematic pedestrian surveys. Recently, systematic manual survey of satellite and aerial imagery has enabled continuous distributional views of archaeological phenomena at interregional scales. However, such 'brute force' manual imagery survey methods are both time- and labor-intensive, as well as prone to inter-observer differences in sensitivity and specificity. The development of self-supervised learning methods offers a scalable learning scheme for locating archaeological features using unlabeled satellite and historical aerial images. However, archaeological features are generally only visible in a very small proportion relative to the landscape, while the modern contrastive-supervised learning approach typically yields an inferior performance on highly imbalanced datasets. In this work, we propose a framework to address this long-tail problem. As opposed to the existing contrastive learning approaches that treat the labelled and unlabeled data separately, our proposed method reforms the learning paradigm under a semi-supervised setting in order to utilize the precious annotated data (<7% in our setting). Specifically, the highly unbalanced nature of the data is employed as the prior knowledge in order to form pseudo negative pairs by ranking the similarities between unannotated image patches and annotated anchor images. In this study, we used 95,358 unlabeled images and 5,830 labelled images in order to solve the issues associated with detecting ancient buildings from a long-tailed satellite image dataset. From the results, our semi-supervised contrastive learning model achieved a promising testing balanced accuracy of 79.0%, which is a 3.8% improvement as compared to other state-of-the-art approaches.

Retinex model has been applied to low-light image enhancement in many existing methods. More appropriate decomposition of a low-light image can help achieve better image enhancement. In this paper, we propose a new pixel-level non-local Haar transform based illumination and reflectance decomposition method (NLHD). The unique low-frequency coefficient of Haar transform on each similar pixel group is used to reconstruct the illumination component, and the rest of all high-frequency coefficients are employed to reconstruct the reflectance component. The complete similarity of pixels in a matched similar pixel group and the simple separable Haar transform help to obtain more appropriate image decomposition; thus, the image is hardly sharpened in the image brightness enhancement procedure. The exponential transform and logarithmic transform are respectively implemented on the illumination component. Then a minimum fusion strategy on the results of these two transforms is utilized to achieve more natural illumination component enhancement. It can alleviate the mosaic artifacts produced in the darker regions by the exponential transform with a gamma value less than 1 and reduce information loss caused by excessive enhancement of the brighter regions due to the logarithmic transform. Finally, the Retinex model is applied to the enhanced illumination and reflectance to achieve image enhancement. We also develop a local noise level estimation based noise suppression method and a non-local saturation reduction based color deviation correction method. These two methods can respectively attenuate noise or color deviation usually presented in the enhanced results of the extremely dark low-light images. Experiments on benchmark datasets show that the proposed method can achieve better low-light image enhancement results on subjective and objective evaluations than most existing methods., Comment: 14 pages, 11 figures

Knowledge distillation can lead to deploy-friendly networks against the plagued computational complexity problem, but previous methods neglect the feature hierarchy in detectors. Motivated by this, we propose a general framework for detection distillation. Our method, called Hands-on Guidance Distillation, distills the latent knowledge of all stage features for imposing more comprehensive supervision, and focuses on the essence simultaneously for promoting more intense knowledge absorption. Specifically, a series of novel mechanisms are designed elaborately, including correspondence establishment for consistency, hands-on imitation loss measure and re-weighted optimization from both micro and macro perspectives. We conduct extensive evaluations with different distillation configurations over VOC and COCO datasets, which show better performance on accuracy and speed trade-offs. Meanwhile, feasibility experiments on different structural networks further prove the robustness of our HGD., Comment: Accepted at ICME2021

It is well known that attention mechanisms can effectively improve the performance of many CNNs including object detectors. Instead of refining feature maps prevalently, we reduce the prohibitive computational complexity by a novel attempt at attention. Therefore, we introduce an efficient object detector called Selective Convolutional Network (SCN), which selectively calculates only on the locations that contain meaningful and conducive information. The basic idea is to exclude the insignificant background areas, which effectively reduces the computational cost especially during the feature extraction. To solve it, we design an elaborate structure with negligible overheads to guide the network where to look next. It's end-to-end trainable and easy-embedding. Without additional segmentation datasets, we explores two different train strategies including direct supervision and indirect supervision. Extensive experiments assess the performance on PASCAL VOC2007 and MS COCO detection datasets. Results show that SSD and Pelee integrated with our method averagely reduce the calculations in a range of 1/5 and 1/3 with slight loss of accuracy, demonstrating the feasibility of SCN., Comment: Accepted at ICASSP 2020