Numerous studies have focused on data-driven password guessing methods in recent years, aiming to reduce the use of weak passwords by users and improve password security. Existing password generation models learn the distribution of password datasets and generate candidate guesses by fitting sequential conditional probabilities. These methods are based on a key assumption: users construct passwords in one direction from left to right. However, with the more complex password policy requirements of authentication systems and the increasing security awareness of people, users construct passwords by modifying existing or popular passwords. At this point, users consider global and bi-directional information of passwords. This breaks the key assumption of uni-directional construction and leads to omissions when generating passwords by existing methods. Motivated by this, we propose a password generation method based on mask prediction, named PGMaP, which captures this large number of omitted passwords. First, we design a password construction template extraction algorithm to cluster the templates used by users for constructing and modifying passwords. Then we construct a transformer-based masked language model to learn password bi-directional features. The extracted templates are fed into the model to generate password guesses by means of mask prediction. Different from existing auto-regressive model based methods that generate in one direction, PGMaP uses the auto-encoding model to generate passwords based on the bidirectional information. Finally, through password guessing experiments all eight real-world datasets, we demonstrate that PGMaP can effectively generate a large number of omitted passwords, and its password guessing performance outperforms existing methods.

Event cameras are renowned for their high efficiency due to outputting a sparse, asynchronous stream of events. However, they are plagued by noisy events, especially in low light conditions. Denoising is an essential task for event cameras, but evaluating denoising performance is challenging. Label-dependent denoising metrics involve artificially adding noise to clean sequences, complicating evaluations. Moreover, the majority of these metrics are monotonic, which can inflate scores by removing substantial noise and valid events. To overcome these limitations, we propose the first label-free and non-monotonic evaluation metric, the area of the continuous contrast curve (AOCC), which utilizes the area enclosed by event frame contrast curves across different time intervals. This metric is inspired by how events capture the edge contours of scenes or objects with high temporal resolution. An effective denoising method removes noise without eliminating these edge-contour events, thus preserving the contrast of event frames. Consequently, contrast across various time ranges serves as a metric to assess denoising effectiveness. As the time interval lengthens, the curve will initially rise and then fall. The proposed metric is validated through both theoretical and experimental evidence. The codes are available at https://github.com/shicy17/AOCC.

Password has become the mainstream method of authentication today. To improve password security, researchers evaluate the strength of target password datasets through early brute-force attacks to current password guessing methods, aiming to help users reduce the use of weak passwords. With users becoming more aware of security, they make local variations on weak passwords to improve the password strength while being easy to remember. These transformations render passwords more complex and enhance the score in password strength meter. However, such variations do not genuinely enhance password security, as human habits tend to converge. This allows attackers to deduce the modification patterns and consequently crack these passwords. Motivated by this, this paper defines the hidden weak passwords, a local variant of explicit weak passwords, which appear to enhance password security yet remain vulnerable. We systematically analyze transformation behavior between explicit and hidden weak passwords. Then we design an automated rule generation algorithm to identify hidden weak passwords and generate transformation rules. Based on automatically mined rules, we generate a large number of password guesses and fuses them with existing methods to improve password guessing performance. Finally, we demonstrate the effectiveness of the proposed method through password guessing experiments on eight real-world datasets, where the cracking rate improves on all five state-of-the-art methods.

Deep learning models face the threat of adversarial attacks, which challenges their application. Moving Target Defense (MTD) is a defense paradigm that thwarts attacks by constantly changing the targets’ features and restricting the predictability of targets. Recent works have applied MTD in adversarial defense but rely on maintaining a model set and assuming a weak adversary, which causes extra storage and unreliable evaluation of the robustness of the methods. This paper proposes the MTD-Net that realizes MTD in a single neural network. In the training stage, MTD-Net parameters are randomly disabled to ensure desirable accuracy on diversified parameter groups. During each query, MTD-Net dynamically chooses several groups of parameters and aggregates their inference results for final prediction. The parameters MTD-Net chooses for inference are unpredictable, even for adversaries possessing the model’s weights. Thus, MTD-Net achieves factual unpredictability under strong whitebox attacks. We evaluate MTD-Net on two widely used datasets, i.e., GTSRB and CIFAR10. The experimental results demonstrate that MTD-Net achieves superior performance compared to existing MTD defense under adversarial attacks and is applicable to multiple architectures.

Fast large-scale network scanning is an important way to understand internet service configurations and security in real time, among which stateless scan is representative. Existing stateless scanners can perform single-packet scans for internet-wide network measurements but are limited to host discovery or port scanning. To obtain further information over TCP, slower stateful scanners must be used in conjunction which spend more time and memory because of connection state maintenance. This paper proposes a novel stateless scanning method, which can establish TCP connections and obtain further responses in a completely stateless manner. Based on this method, we implement a stateless scanner named ZBanner. Experiments show that ZBanner performs better than current state-of-the-art solutions in terms of scan rate and memory usage. ZBanner achieves a scan rate at least three times faster than current tools for generic ports and over 90 times faster for open ports while keeping a minimum and stable memory usage.

Adversarial attacks can cause abnormal behavior in deep neural networks by adding imperceptible perturbations to input data. Adversarial purification is an effective defense method by transforming adversarial data into clean data. Existing purification methods utilize only simple corruptions and specified reconstructors to eliminate perturbations, and thus are non-adaptive to different attack algorithms. To address this challenge, we propose an adaptive adversarial purification framework, which combines multiple Destruction and Reconstruction (D&R) operations to form a multi-stage D&R chain. The Destruction operations consist of corruptions in both spatial and frequency domains. The Reconstruction operations are implemented by deep networks trained respectively for corresponding corruptions to restore images. The D&R chain is constructed dynamically during purification based on the link probabilities of D&R operations with a self-supervised search algorithm. We test our framework on CIFAR10 datasets under five attacks. The experiment results indicate that the proposed framework can adapt to a wide range of attacks and achieve comparable performance.

Dynamic Vision Sensor (DVS) event camera output includes uninformative background activity (BA) noise events that increase dramatically under dim lighting. Existing denoising algorithms are not effective under these high noise conditions. Furthermore, it is difficult to quantitatively compare algorithm accuracy. This paper proposes a novel framework to better quantify BA denoising algorithms by measuring receiver operating characteristics with known mixtures of signal and noise DVS events. New datasets for stationary and moving camera applications of DVS in surveillance and driving are used to compare 3 new low-cost algorithms: Algorithm 1 checks distance to past events using a tiny fixed size window and removes most of the BA while preserving most of the signal for stationary camera scenarios. Algorithm 2 uses a memory proportional to the number of pixels for improved correlation checking. Compared with existing methods, it removes more noise while preserving more signal. Algorithm 3 uses a lightweight multilayer perceptron classifier driven by local event time surfaces to achieve the best accuracy over all datasets. The code and data are shared with the paper as DND21.

Neuromorphic processors have gained momentum recently due to their high energy efficiency in artificial intelligence applications compared to DNN accelerators. Most neuromorphic processors are executing SNNs (Spiking Neural Networks). Liquid State Machine (LSM), as the spiking version of reservoir computing, shows advantages and great potential in image classification, speech recognition, language translation, etc.. Comparing with other SNN models, LSM has the characteristics of easy to train and low resource utilization, which is suitable for low-power and resource-constrained edge computing scenarios. In this paper, we propose a novel design of a neuromorphic processor, LSMCore, aiming at LSM acceleration. LSMCore supports both training and inference of LSM. It consists of 256 input neurons, 1024 liquid neurons, and 1.31M synapses. Besides, multiple optimization techniques, including weight quantization for reducing storage, zero-skipping for decreasing dynamic sparsity, and mini-batch training are adopted in this processor. The experimental results show that the frequency of LSMCore achieves 400 MHz, the power is 4.9W and the area is 18.49 mm2with a 40nm library. Comparing with the baseline, LSMCore achieves up to$80.7\times $($49.6\times $),$91.3\times $($56.3\times $), and$83.1\times $($56.8\times $) speedup on MNIST, N-MNIST, and Free Spoken Digital Dataset (FSDD) respectively for training (inference), while the accuracy of LSMCore on these three datasets are 96.8%, 97.6%, and 90% respectively.

Neuromorphic event-based dynamic vision sensors (DVS) have much faster sampling rates and a higher dynamic range than frame-based imagers. However, they are sensitive to background activity (BA) events which are unwanted. We propose HashHeat, a hashing-based BA filter with O(C) complexity. It is the first spatiotemporal filter that doesn't scale with the DVS output size N and doesn't store the 32-bits timestamps. HashHeat consumes 100x less memory and increases the signal to noise ratio by 15x compared to previous designs.

Spiking neural network (SNN), as the third generation of artificial neural networks, has been widely adopted in vision and audio tasks. Nowadays, many neuromorphic platforms support SNN simulation and adopt Network-on-Chips (NoC) architecture for multi-cores interconnection. However, a large volume and run-time communication on the interconnection has a significant effect on performance of the platform. In this paper, we propose a toolchain called SNEAP (Spiking NEural network mAPping toolchain) for mapping SNNs to neuromorphic platforms with multi-cores, which aims to reduce the energy and latency brought by spike communication on the interconnection.

Deep neural networks have achieved highly competitive performance in multiple tasks in recent years. However, discovering state-of-the-art neural network architectures requires substantial effort from human experts. To speed up the process, neural architecture search (NAS) has been proposed to search promising architectures automatically. Nevertheless, the search process of NAS is computing-expensive and time-consuming, which even costs thousands of GPU days. In this paper, to solve the bottleneck, we apply the randomness-enhanced tabu algorithm as a controller to sample candidate architectures, which balances the global exploration and local exploitation for the architectural solutions. In addition, more aggressive weight-sharing strategy is introduced into our method, which significantly reduces the overhead of evaluating sampled architectures. Our approach discovers the recurrent neural architecture within 0.78 GPU hour, which is 15.3x more efficient than ENAS [1] in terms of search time, and the architecture we discovered achieves the test perplexity of 56.1 on Penn Tree Bank (PTB) dataset, which is lower than ENAS by 2.2. In addition, we further demonstrate the usefulness of the learned architecture by transferring it to wiki-text-2 (WT2) dataset well. Moreover, the extended experiments on the WT2 dataset also show promising results.

Neuromorphic computing based on spiking neural network (SNN) shows good energy-efficiency. However, it is inefficient for SNN to perform the convolution based on frame. It may contain a lot of redundant information in the frame. The output of Dynamic Vision Sensors (DVS) is a stream event based on Address Event Representation (AER). The asynchronous nature of AER events makes the event-based convolution reflect the characteristics of SNN low energy consumption. This article presents an SNN hardware inference engine based on an asynchronous Processing Element (PE) array with AER events as input. The engine uses a convolution algorithm based on AER events. This design also uses distributed storage in the PE array to store the state of neurons to reduce the cost of memory access. The experimental results show that the design can achieve a recognition accuracy of 98.0% for the MNIST AER dataset. The design can perform the reference process more efficiently in the case where the accuracy of the loss is negligible. During the filling and draining processes of the systolic array, the number of active PE units in our PE array is reduced and, thus, the average power consumption per PE unit is drastically decreased.

Although Deep Neural Network (DNN) architectures have made some breakthroughs in computer vision tasks, they are not close to biological brain neurons. Spiking Neural Network (SNN) is highly expected to bridge the gap between artificial computing systems and bio-systems. And it also shows great potential in low power computing. This paper presents a low power hardware accelerator for SNN inference using systolic array, and a corresponding software framework for optimization. First, we give the hardware design which adopts systolic array inspired by explorations of SNN. Then we ensure correct data mapping for systolic array for the sake of computational correctness. Next, we use compression methods for decreasing both the runtime and memory footprint. Finally, we make the systolic array size-configurable to adapt to different input, so as to reduce computational overhead. We implement the accelerator on Xilinx FPGA V7 690T. The experimental results show that SNN inference on our scheme suffers little loss on accuracy (less than 0.1%) on MNIST and Fashion-MNIST, and the runtime of the time-consuming layers decreases. The total power of our scheme is 0.745 W at 100 MHz.