End-to-End Network Intrusion Detection Based on Contrastive Learning.

The network intrusion detection system (NIDS) plays a crucial role as a security measure in addressing the increasing number of network threats. The majority of current research relies on feature-ready datasets that heavily depend on feature engineering. Conversely, the increasing complexity of network traffic and the ongoing evolution of attack techniques lead to a diminishing distinction between benign and malicious network behaviors. In this paper, we propose a novel end-to-end intrusion detection framework based on a contrastive learning approach. We design a hierarchical Convolutional Neural Network (CNN) and Gated Recurrent Unit (GRU) model to facilitate the automated extraction of spatiotemporal features from raw traffic data. The integration of contrastive learning amplifies the distinction between benign and malicious network traffic in the representation space. The proposed method exhibits enhanced detection capabilities for unknown attacks in comparison to the approaches trained using the cross-entropy loss function. Experiments are carried out on the public datasets CIC-IDS2017 and CSE-CIC-IDS2018, demonstrating that our method can attain a detection accuracy of 99.9% for known attacks, thus achieving state-of-the-art performance. For unknown attacks, a weighted recall rate of 95% can be achieved.

An Enhanced Steganography Network for Concealing and Protecting Secret Image Data.

The development of Internet technology has provided great convenience for data transmission and sharing, but it also brings serious security problems that are related to data protection. As is detailed in this paper, an enhanced steganography network was designed to protect secret image data that contains private or confidential information; this network consists of a concealing network and a revealing network in order to achieve image embedding and recovery separately. To reduce the system's computation complexity, we constructed the network's framework using a down-up structure in order to compress the intermediate feature maps. In order to mitigate the input's information loss caused by a sequence of convolution blocks, the long skip concatenation method was designed to pass the raw information to the top layer, thus synthesizing high-quality hidden images with fine texture details. In addition, we propose a novel strategy called non-activated feature fusion (NAFF), which is designed to provide stronger supervision for synthetizing higher-quality hidden images and recovered images. In order to further boost the hidden image's visual quality and enhance its imperceptibility, an attention mechanism-based enhanced module was designed to reconstruct and enhance the salient target, thus covering up and obscuring the embedded secret content. Furthermore, a hybrid loss function that is composed of pixel domain loss and structure domain loss was designed to boost the hidden image's structural quality and visual security. Our experimental results demonstrate that, due to the elaborate design of the network structure and loss function, our proposed method achieves high levels of imperceptibility and security.

Meaningful secret image sharing for JPEG images with arbitrary quality factors.

JPEG is the most common format for storing and transmitting photographic images on social network platforms. JPEG image is widely used in people's life because of their low storage space and high visual quality. Secret image sharing (SIS) technology is important to protect image data. Traditional SIS schemes generally focus on spatial images, however there is little research on frequency domain images. In addition, the current tiny research on SIS for JPEG images only focuses on JPEG images with a compression quality factor (QF) of 100. To overcome the limitation of JPEG images in SIS, we propose a meaningful SIS for JPEG images to operate the quantized DCT coefficients of JPEG images. The random elements utilization model is applied to achieve meaningful shadow images. Our proposed scheme has a better quality of the shadow images and the recovered secret image. Experiment results and comparisons indicate the effectiveness of the scheme. The scheme can be used for JPEG images with any compression QF. Besides, the scheme has good characteristics, such as (k,n) threshold, extended shadow images.

Robust secret image sharing scheme resistance to maliciously tampered shadows by AMBTC and quantization.

For (k, n)-threshold secret image sharing (SIS) scheme, only k or more than k complete parts can recover the secret information, and the correct image cannot be obtained if the count of shadow images is not enough or the shadow images are damaged. The existing schemes are weak in resisting large-area shadow image tampering. In this paper, we propose a robust secret image sharing scheme resisting to maliciously tampered shadow images by Absolute Moment Block Truncation Coding (AMBTC) and quantization (RSIS-AQ). The secret image is successively compressed in two ways: AMBTC and quantization. The sharing shadow images contain the sharing results of both compressed image from different parts, so that even the shadow images are faced with large-scale area of malicious tampering, the secret image can be recovered with acceptable visual quality. Compared with related works, our scheme can resist larger area of tampering and yield better recovered image visual quality. The experimental results prove the effectiveness of our scheme.

CTTGAN: Traffic Data Synthesizing Scheme Based on Conditional GAN.

Most machine learning algorithms only have a good recognition rate on balanced datasets. However, in the field of malicious traffic identification, benign traffic on the network is far greater than malicious traffic, and the network traffic dataset is imbalanced, which makes the algorithm have a low identification rate for small categories of malicious traffic samples. This paper presents a traffic sample synthesizing model named Conditional Tabular Traffic Generative Adversarial Network (CTTGAN), which uses a Conditional Tabular Generative Adversarial Network (CTGAN) algorithm to expand the small category traffic samples and balance the dataset in order to improve the malicious traffic identification rate. The CTTGAN model expands and recognizes feature data, which meets the requirements of a machine learning algorithm for training and prediction data. The contributions of this paper are as follows: first, the small category samples are expanded and the traffic dataset is balanced; second, the storage cost and computational complexity are reduced compared to models using image data; third, discrete variables and continuous variables in traffic feature data are processed at the same time, and the data distribution is described well. The experimental results show that the recognition rate of the expanded samples is more than 0.99 in MLP, KNN and SVM algorithms. In addition, the recognition rate of the proposed CTTGAN model is better than the oversampling and undersampling schemes.

MTS-Stega: Linguistic Steganography Based on Multi-Time-Step.

Generative linguistic steganography encodes candidate words with conditional probability when generating text by language model, and then, it selects the corresponding candidate words to output according to the confidential message to be embedded, thereby generating steganographic text. The encoding techniques currently used in generative text steganography fall into two categories: fixed-length coding and variable-length coding. Because of the simplicity of coding and decoding and the small computational overhead, fixed-length coding is more suitable for resource-constrained environments. However, the conventional text steganography mode selects and outputs a word at one time step, which is highly susceptible to the influence of confidential information and thus may select words that do not match the statistical distribution of the training text, reducing the quality and concealment of the generated text. In this paper, we inherit the decoding advantages of fixed-length coding, focus on solving the problems of existing steganography methods, and propose a multi-time-step-based steganography method, which integrates multiple time steps to select words that can carry secret information and fit the statistical distribution, thus effectively improving the text quality. In the experimental part, we choose the GPT-2 language model to generate the text, and both theoretical analysis and experiments prove the effectiveness of the proposed scheme.

Sparse Adversarial Video Attacks via Superpixel-Based Jacobian Computation.

Adversarial examples have aroused great attention during the past years owing to their threat to the deep neural networks (DNNs). Recently, they have been successfully extended to video models. Compared with image cases, the sparse adversarial perturbations in the videos can not only reduce the computation complexity, but also guarantee the crypticity of adversarial examples. In this paper, we propose an efficient attack to generate adversarial video perturbations with large sparsity in both the temporal (inter-frames) and spatial (intra-frames) domains. Specifically, we select the key frames and key pixels according to the gradient feedback of the target models by computing the forward derivative, and then add the perturbations on them. To overcome the problem of dimensional explosion in the video, we introduce super-pixels to decrease the number of pixels that need to compute gradients. The proposed method is finally verified under both the white-box and black-box settings. We estimate the gradients using natural evolution strategy (NES) in the black-box attacks. The experiments are conducted on two widely used datasets: UCF101 and HMDB51 versus two mainstream models: C3D and LRCN. Results show that compared with the state-of-the-art method, our method can achieve the similar attacking performance, but it pollutes only <1% pixels and costs less time to finish the attacks.

Wet Paper Coding-Based Deep Neural Network Watermarking.

In recent years, the wide application of deep neural network models has brought serious risks of intellectual property rights infringement. Embedding a watermark in a network model is an effective solution to protect intellectual property rights. Although researchers have proposed schemes to add watermarks to models, they cannot prevent attackers from adding and overwriting original information, and embedding rates cannot be quantified. Therefore, aiming at these problems, this paper designs a high embedding rate and tamper-proof watermarking scheme. We employ wet paper coding (WPC), in which important parameters are regarded as wet blocks and the remaining unimportant parameters are regarded as dry blocks in the model. To obtain the important parameters more easily, we propose an optimized probabilistic selection strategy (OPSS). OPSS defines the unimportant-level function and sets the importance threshold to select the important parameter positions and to ensure that the original function is not affected after the model parameters are changed. We regard important parameters as an unmodifiable part, and only modify the part that includes the unimportant parameters. We selected the MNIST, CIFAR-10, and ImageNet datasets to test the performance of the model after adding a watermark and to analyze the fidelity, robustness, embedding rate, and comparison schemes of the model. Our experiment shows that the proposed scheme has high fidelity and strong robustness along with a high embedding rate and the ability to prevent malicious tampering.

A Novel General (n, n)-Threshold Multiple Secret Images Sharing Scheme Based on Information Hiding in the Sharing Domain.

(k,n)-threshold secret image sharing (SIS) protects an image by dividing it into n shadow images. The secret image will be recovered as we gather k or more shadow images. In complex networks, the security, robustness and efficiency of protecting images draws more and more attention. Thus, we realize multiple secret images sharing (MSIS) by information hiding in the sharing domain (IHSD) and propose a novel and general (n,n)-threshold IHSD-MSIS scheme (IHSD-MSISS), which can share and recover two secret images simultaneously. The proposed scheme spends less cost on managing and identifying shadow images, and improves the ability to prevent malicious tampering. Moreover, it is a novel approach to transmit important images with strong associations. The superiority of (n,n)-threshold IHSD-MSISS is in fusing the sharing phases of two secret images by controlling randomness of SIS. We present a general construction model and algorithms of the proposed scheme. Sufficient theoretical analyses, experiments and comparisons show the effectiveness of the proposed scheme.

Meaningful Secret Image Sharing with Saliency Detection.

Secret image sharing (SIS), as one of the applications of information theory in information security protection, has been widely used in many areas, such as blockchain, identity authentication and distributed cloud storage. In traditional secret image sharing schemes, noise-like shadows introduce difficulties into shadow management and increase the risk of attacks. Meaningful secret image sharing is thus proposed to solve these problems. Previous meaningful SIS schemes have employed steganography to hide shares into cover images, and their covers are always binary images. These schemes usually include pixel expansion and low visual quality shadows. To improve the shadow quality, we design a meaningful secret image sharing scheme with saliency detection. Saliency detection is used to determine the salient regions of cover images. In our proposed scheme, we improve the quality of salient regions that are sensitive to the human vision system. In this way, we obtain meaningful shadows with better visual quality. Experiment results and comparisons demonstrate the effectiveness of our proposed scheme.

Minimum Adversarial Examples.

Deep neural networks in the area of information security are facing a severe threat from adversarial examples (AEs). Existing methods of AE generation use two optimization models: (1) taking the successful attack as the objective function and limiting perturbations as the constraint; (2) taking the minimum of adversarial perturbations as the target and the successful attack as the constraint. These all involve two fundamental problems of AEs: the minimum boundary of constructing the AEs and whether that boundary is reachable. The reachability means whether the AEs of successful attack models exist equal to that boundary. Previous optimization models have no complete answer to the problems. Therefore, in this paper, for the first problem, we propose the definition of the minimum AEs and give the theoretical lower bound of the amplitude of the minimum AEs. For the second problem, we prove that solving the generation of the minimum AEs is an NPC problem, and then based on its computational inaccessibility, we establish a new third optimization model. This model is general and can adapt to any constraint. To verify the model, we devise two specific methods for generating controllable AEs under the widely used distance evaluation standard of adversarial perturbations, namely Lp constraint and SSIM constraint (structural similarity). This model limits the amplitude of the AEs, reduces the solution space's search cost, and is further improved in efficiency. In theory, those AEs generated by the new model which are closer to the actual minimum adversarial boundary overcome the blindness of the adversarial amplitude setting of the existing methods and further improve the attack success rate. In addition, this model can generate accurate AEs with controllable amplitude under different constraints, which is suitable for different application scenarios. In addition, through extensive experiments, they demonstrate a better attack ability under the same constraints as other baseline attacks. For all the datasets we test in the experiment, compared with other baseline methods, the attack success rate of our method is improved by approximately 10%.

AMBTC-based visual secret sharing with different meaningful shadows.

Secret sharing based on Absolute Moment Block Truncation Coding (AMBTC) has been widely studied. However, the management of stego images is inconvenient as they seem indistinguishable. Moreover, there exists a problem of pixel expansion, which requires more storage space and higher transmission bandwidth. To conveniently manage the stego images, we use multiple cover images to make the stego images seem to be visually different from with each other. Futhermore, the stego images are different, which will not cause the attacker's suspicion and increase the security of the scheme. And traditional Visual Secret Sharing (VSS) is fused to eliminate pixel expansion. After images are compressed by AMBTC algorithm, the quantization levels and the bitmap corresponding to each block are obtained. At the same time, when the threshold is (k,k), bitmaps can be recovered losslessly, and the slight degradation of image quality is only caused by the compression itself. When the threshold is another value, the recovered image and the cover images can be recovered with satisfactory image quality. The experimental results and analyses show the effectiveness and advantages of our scheme.

Fake and dishonest participant location scheme in secret image sharing.

A (k,n) threshold secret image sharing (SIS) scheme divides a secret image into n shadows. One can reconstruct the secret image only when holding k or more than k shadows but cannot know any information on the secret from fewer than k shadows. Based on this characteristic, SIS has been widely used in access control, information hiding, distributed storage and other areas. Verifiable SIS aims to prevent malicious behaviour by attackers through verifying the authenticity of shadows and previous works did not solve this problem well. Our contribution is that we proposed a verifiable SIS scheme which combined CRT-based SIS and (2,n+1) threshold visual secret sharing(VSS). Our scheme is applicable no matter whether there exists a third party dealer. And it is worth mentioning that when the dealer is involved, our scheme can not only detect fake participants, but also locate dishonest participants. In general, loose screening criterion and efficient encoding and decoding rate of CRT-based SIS guarantee high-efficiency shadows generation and low recovery computation complexity. The uncertainty of the bits used for screening prevents malicious behavior by dishonest participants. In addition, our scheme has the advantages of lossless recovery, no pixel expansion and precise detection.