Reducing Fine-Tuning Memory Overhead by Approximate and Memory-Sharing Backpropagation

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.

Robust Image Semantic Coding with Learnable CSI Fusion Masking over MIMO Fading Channels

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.

DBDH: A Dual-Branch Dual-Head Neural Network for Invisible Embedded Regions Localization

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.

MARformer: An Efficient Metal Artifact Reduction Transformer for Dental CBCT Images

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.

Improving Visual Quality and Transferability of Adversarial Attacks on Face Recognition Simultaneously with Adversarial Restoration

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.

Communication-Efficient Framework for Distributed Image Semantic Wireless Transmission

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.

Parameter-Free Channel Attention for Image Classification and Super-Resolution

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.

Improving Transferability of Adversarial Examples on Face Recognition with Beneficial Perturbation Feature Augmentation

Face recognition (FR) models can be easily fooled by adversarial examples, which are crafted by adding imperceptible perturbations on benign face images. To improve the transferability of adversarial examples on FR models, we propose a novel attack method called Beneficial Perturbation Feature Augmentation Attack (BPFA), which reduces the overfitting of the adversarial examples to surrogate FR models by the adversarial strategy. Specifically, in the backpropagation step, BPFA records the gradients on pre-selected features and uses the gradient on the input image to craft adversarial perturbation to be added on the input image. In the next forward propagation step, BPFA leverages the recorded gradients to add perturbations(i.e., beneficial perturbations) that can be pitted against the adversarial perturbation added on the input image on their corresponding features. The above two steps are repeated until the last backpropagation step before the maximum number of iterations is reached. The optimization process of the adversarial perturbation added on the input image and the optimization process of the beneficial perturbations added on the features correspond to a minimax two-player game. Extensive experiments demonstrate that BPFA outperforms the state-of-the-art gradient-based adversarial attacks on FR.

Semi-Supervised Contrastive Learning for Remote Sensing: Identifying Ancient Urbanization in the South Central Andes

The detection of ancient settlements is a key focus in landscape archaeology. Traditionally, settlements were identified through pedestrian survey, as researchers physically traversed the landscape and recorded settlement locations. Recently the manual identification and labeling of ancient remains in satellite imagery have increased the scale of archaeological data collection, but the process remains tremendously time-consuming and arduous. The development of self-supervised learning (e.g., contrastive learning) offers a scalable learning scheme in locating archaeological sites using unlabeled satellite and historical aerial images. However, archaeology sites are only present in a very small proportion of the whole landscape, while the modern contrastive-supervised learning approach typically yield inferior performance on the highly balanced dataset, such as identifying sparsely localized ancient urbanization on a large area using satellite images. In this work, we propose a framework to solve this long-tail problem. As opposed to the existing contrastive learning approaches that typically treat the labeled and unlabeled data separately, the proposed method reforms the learning paradigm under a semi-supervised setting to fully utilize the precious annotated data (<7% in our setting). Specifically, the highly unbalanced nature of the data is employed as the prior knowledge 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 labeled images to solve the problem of 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 3.8% improvement over state-of-the-art approaches.

NLHD: A Pixel-Level Non-Local Retinex Model for Low-Light Image Enhancement

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.