An imbalanced learning method based on graph tran-smote for fraud detection.

Fraud seriously threatens individual interests and social stability, so fraud detection has attracted much attention in recent years. In scenarios such as social media, fraudsters typically hide among numerous benign users, constituting only a small minority and often forming "small gangs". Due to the scarcity of fraudsters, the conventional graph neural network might overlook or obscure critical fraud information, leading to insufficient representation of fraud characteristics. To address these issues, the tran-smote on graphs (GTS) method for fraud detection is proposed by this study. Structural features of each type of node are deeply mined using a subgraph neural network extractor, these features are integrated with attribute features using transformer technology, and the node's information representation is enriched, thereby addressing the issue of inadequate feature representation. Additionally, this approach involves setting a feature embedding space to generate new nodes representing minority classes, and an edge generator is used to provide relevant connection information for these new nodes, alleviating the class imbalance problem. The results from experiments on two real datasets demonstrate that the proposed GTS, performs better than the current state-of-the-art baseline.

Spatial-temporal graph neural ODE networks for skeleton-based action recognition.

In the field of skeleton-based action recognition, accurately recognizing human actions is crucial for applications such as virtual reality and motion analysis. However, this task faces challenges such intraindividual action differences and long-term temporal dependencies. To address these challenges, we propose an innovative model called spatial-temporal graph neural ordinary differential equations (STG-NODE). First, in the data preprocessing stage, the dynamic time warping (DTW) algorithm is used to normalize and calculate 3D skeleton data to facilitate the derivation of customized adjacency matrices for improving the influence of intraindividual action differences. Secondly, a custom ordinary differential equation (ODE) integrator is applied based on the initial conditions of the temporal features, producing a solution function that simulates the dynamic evolution trend of the events of interest. Finally, the outstanding ODE solver is used to numerically solve the time features based on the solution function to increase the influence of long-term dependencies on the recognition accuracy of the model and provide it with a more powerful temporal modeling ability. Through extensive experiments conducted on the NTU RGB+D 60 and Kinetics Skeleton 400 benchmark datasets, we demonstrate the superior performance of STG-NODE in the action recognition domain. The success of the STG-NODE model also provides new ideas and methods for the future development of the action recognition field.

Efficient and Low-Complex Signal Detection with Iterative Feedback in Wireless MIMO-OFDM Systems.

To solve error propagation and exorbitant computational complexity of signal detection in wireless multiple-input multiple-output-orthogonal frequency division multiplexing (MIMO-OFDM) systems, a low-complex and efficient signal detection with iterative feedback is proposed via a constellation point feedback optimization of minimum mean square error-ordered successive interference cancellation (MMSE-OSIC) to approach the optimal detection. The candidate vectors are formed by selecting the candidate constellation points. Additionally, the vector most approaching received signals is chosen by the maximum likelihood (ML) criterion in formed candidate vectors to reduce the error propagation by previous erroneous decision, thus improving the detection performance. Under a large number of matrix inversion operations in the above iterative MMSE process, effective and fast signal detection is hard to be achieved. Then, a symmetric successive relaxation iterative algorithm is proposed to avoid the complex matrix inversion calculation process. The relaxation factor and initial iteration value are reasonably configured with low computational complexity to achieve good detection close to that of the MMSE with fewer iterations. Simultaneously, the error diffusion and complexity accumulation caused by the successive detection of the subsequent OSIC scheme are also improved. In addition, a method via a parallel coarse and fine detection deals with several layers to both reduce iterations and improve performance. Therefore, the proposed scheme significantly promotes the MIMO-OFDM performance and thus plays an irreplaceable role in the future sixth generation (6G) mobile communications and wireless sensor networks, and so on.

Improved Frequency Domain Turbo Equalization with Expectation Propagation Interference Cancellation in Underwater Acoustic Communications.

This paper proposes an improved frequency domain turbo equalization (IFDTE) with iterative channel estimation and feedback to achieve both a good performance and low complexity in underwater acoustic communications (UWACs). A selective zero-attracting (SZA) improved proportionate normal least mean square (SZA-IPNLMS) algorithm is adopted by utilizing the sparsity of the UWAC channel to estimate it using a training sequence. Simultaneously, a set-membership (SM) SZA differential IPNLMS (SM SZA-DIPNLMS) with variable step size is adopted to estimate the channel status information (CSI) in the iterative channel estimation with soft feedback. In this way, the computational complexity for iterative channel estimation is reduced effectively with minimal performance loss. Different from traditional schemes in UWACs, an IFDTE with expectation propagation (EP) interference cancellation is adopted to estimate the a posteriori probability of transmitted symbols iteratively. A bidirectional IFDTE with the EP interference cancellation is proposed to further accelerate the convergence. THe simulation results show that the proposed channel estimation obtains 1.9 and 0.5 dB performance gains, when compared with those of the IPNLMS and the l0-IPNLMS at a bit error rate (BER) of 10-3. The proposed channel estimation also effectively reduces the unnecessary updating of the coefficients of the UWAC channel. Compared with traditional time-domain turbo equalization and FDTE in UWACs, the IFDTE obtains 0.5 and 1 dB gains in the environment of SPACE'08 and it obtains 0.5 and 0.4 dB gains in the environment of MACE'04 at a BER of 10-3. Therefore, the proposed scheme obtains a good BER performance and low complexity and it is suitable for efficient use in UWACs.

Research on the Rapid Diagnostic Method of Rolling Bearing Fault Based on Cloud-Edge Collaboration.

Recent deep-learning methods for fault diagnosis of rolling bearings need a significant amount of computing time and resources. Most of them cannot meet the requirements of real-time fault diagnosis of rolling bearings under the cloud computing framework. This paper proposes a quick cloud-edge collaborative bearing fault diagnostic method based on the tradeoff between the advantages and disadvantages of cloud and edge computing. First, a collaborative cloud-based framework and an improved DSCNN-GAP algorithm are suggested to build a general model using the public bearing fault dataset. Second, the general model is distributed to each edge node, and a limited number of unique fault samples acquired by each edge node are used to quickly adjust the parameters of the model before running diagnostic tests. Finally, a fusion result is made from the diagnostic results of each edge node by DS evidence theory. Experiment results show that the proposed method not only improves diagnostic accuracy by DSCNN-GAP and fusion of multi-sensors, but also decreases diagnosis time by migration learning with the cloud-edge collaborative framework. Additionally, the method can effectively enhance data security and privacy protection.

Cuproof: Range Proof with Constant Size.

Zero-Knowledge Proof is widely used in blockchains. For example, zk-SNARK is used in Zcash as its core technology to identifying transactions without the exposure of the actual transaction values. Up to now, various range proofs have been proposed, and their efficiency and range-flexibility have also been improved. Bootle et al. used the inner product method and recursion to construct an efficient Zero-Knowledge Proof in 2016. Later, Benediky Bünz et al. proposed an efficient range proof scheme called Bulletproofs, which can convince the verifier that a secret number lies in [0,2κ-1] with κ being a positive integer. By combining the inner-product and Lagrange's four-square theorem, we propose a range proof scheme called Cuproof. Our Cuproof can make a range proof to show that a secret number v lies in an interval [a,b] with no exposure of the real value v or other extra information leakage about v. It is a good and practical method to protect privacy and information security. In Bulletproofs, the communication cost is 6+2logκ, while in our Cuproof, all the communication cost, the proving time and the verification time are of constant sizes.

Detection of Salient Crowd Motion Based on Repulsive Force Network and Direction Entropy.

This paper proposes a method for salient crowd motion detection based on direction entropy and a repulsive force network. This work focuses on how to effectively detect salient regions in crowd movement through calculating the crowd vector field and constructing the weighted network using the repulsive force. The interaction force between two particles calculated by the repulsive force formula is used to determine the relationship between these two particles. The network node strength is used as a feature parameter to construct a two-dimensional feature matrix. Furthermore, the entropy of the velocity vector direction is calculated to describe the instability of the crowd movement. Finally, the feature matrix of the repulsive force network and direction entropy are integrated together to detect the salient crowd motion. Experimental results and comparison show that the proposed method can efficiently detect the salient crowd motion.

A Novel Method for Intelligent Single Fault Detection of Bearings Using SAE and Improved D-S Evidence Theory.

In order to realize single fault detection (SFD) from the multi-fault coupling bearing data and further research on the multi-fault situation of bearings, this paper proposes a method based on features self-extraction of a Sparse Auto-Encoder (SAE) and results fusion of improved Dempster-Shafer evidence theory (D-S). Multi-fault signal compression features of bearings were extracted by SAE on multiple vibration sensors' data. Data sets were constructed by the extracted compression features to train the Support Vector Machine (SVM) according to the rule of single fault detection (R-SFD) this paper proposed. Fault detection results were obtained by the improved D-S evidence theory, which was implemented via correcting the 0 factor in the Basic Probability Assignment (BPA) and modifying the evidence weight by Pearson Correlation Coefficient (PCC). Extensive evaluations of the proposed method on the experiment platform datasets showed that the proposed method could realize single fault detection from multi-fault bearings. Fault detection accuracy increases as the output feature dimension of SAE increases; when the feature dimension reached 200, the average detection accuracy of the three sensors for bearing inner, outer, and ball faults achieved 87.36%, 87.86% and 84.46%, respectively. The three types' fault detection accuracy-reached to 99.12%, 99.33% and 98.46% by the improved Dempster-Shafer evidence theory (IDS) to fuse the sensors' results-is respectively 0.38%, 2.06% and 0.76% higher than the traditional D-S evidence theory. That indicated the effectiveness of improving the D-S evidence theory by evidence weight calculation of PCC.

An Ensemble Deep Convolutional Neural Network Model with Improved D-S Evidence Fusion for Bearing Fault Diagnosis.

Intelligent machine health monitoring and fault diagnosis are becoming increasingly important for modern manufacturing industries. Current fault diagnosis approaches mostly depend on expert-designed features for building prediction models. In this paper, we proposed IDSCNN, a novel bearing fault diagnosis algorithm based on ensemble deep convolutional neural networks and an improved Dempster-Shafer theory based evidence fusion. The convolutional neural networks take the root mean square (RMS) maps from the FFT (Fast Fourier Transformation) features of the vibration signals from two sensors as inputs. The improved D-S evidence theory is implemented via distance matrix from evidences and modified Gini Index. Extensive evaluations of the IDSCNN on the Case Western Reserve Dataset showed that our IDSCNN algorithm can achieve better fault diagnosis performance than existing machine learning methods by fusing complementary or conflicting evidences from different models and sensors and adapting to different load conditions.

Optical logic gates based on electro-optic modulation with Sagnac interferometer.

In this work, we present a new structure to realize optical logic operation in a Sagnac interferometer with electro-optical modulation. In the scheme, we divide two counterpropagation signals in a Sagnac loop to two different arms with the electro-optical crystal by using two circulators. Lithium niobate materials whose electro-optical coefficient can be as large as 32.2×10(-12) m/V make up the arms of the waveguides. Using the transfer matrix of the fiber coupler, we analyze the propagation of signals in this system and obtain the transmission characteristic curves and the extinction ratio. The results indicate that this optical switching has a high extinction ratio of about 60 dB and an ultrafast response time of 2.036 ns. In addition, the results reveal that the change of the dephasing between the two input signals and the modification of the modulation voltage added to the electro-optical crystal leads to the change of the extinction ratio. We also conclude that, in cases of the dephasing of two initial input signals Δφ=0, we can obtain the various logical operations, such as the logical operations D=A¯·B, D=A·B¯, C=A+B, and D=A⊕B in ports C and D of the system by adjusting the modulation voltage. When Δφ≠0, we obtain the arithmetic operations D=A+B, C=A⊕B, D=A·B¯, and C=A¯·B in ports C and D. This study is significant for the design of all optical networks by adjusting the modulation voltage.

Bistability and all-optical flip-flop with active microring resonator.

In this work, we study the bistability of an active nonlinear microring resonator and design a flip-flop based on the active microring resonator. In the presence of nonlinear and linear loss, we use Er-doped gain medium in the microring to obtain gain to compensate for the loss of the resonator. Both analytical and numerical methods are used to solve the propagation in the microring with double couplers, and we obtain the hysteresis loops of the microring. The results obtained by the two methods reveal that, in the presence of nonlinearity in microring resonators, the system exhibits bistability, and the gain in the microring leads to a decrease of the bias power when the active microring is taken as a bistable switcher. Basing on the bistability of the microring, we realize a set-reset flip-flop by adding a positive or negative feedback onto the bias. We also find that the duration of the set and reset pulses must exceed the field buildup time of the microring if we want to achieve the switching of the bias signal. In our design, the duration time is about 2 ps.

All-optical logical gates based on pump-induced resonant nonlinearity in an erbium-doped fiber coupler.

In this paper, we theoretically investigate all-optical logical gates based on the pump-induced resonant nonlinearity in an erbium-doped fiber coupler. The resonant nonlinearity yielded by the optical transitions between the (4)I(15/2) states and (4)I(13/2) states in Er(3+) induces the refractive index to change, which leads to switching between two output ports. First, we do a study on the switching performance, and calculate the extinction ratio (Xratio) of the device. Second, using the Xratio, we obtain the truth tables of the device. The results reveal that compared with other undoped nonlinear couplers, the erbium-doped fiber coupler can drop the switching threshold power. We also obtain different logic gates and logic operations in the cases of the same phase and different phase of two initial signals by changing the pump power.

High-resolution electroencephalogram (EEG) mapping: scalp charge layer.

The neural electrical signal related to the human brain function is one of the tracks to understanding ourselves. Various electroencephalogram imaging techniques have been developed to reveal spatial information on neural activities in the brain from scalp recordings, such as Laplacian, equivalent source layer and potential. Physically, these methods may be classified into two categories: scalp surface or cortical surface based techniques. In this work, the focus is on the scalp surface based equivalent charge layer (ECL), with a comparison to the scalp potential with different references and scalp Laplacian (SL). The contents include theoretical analysis and numeric evaluation of simulated data and real alpha (8-12 Hz) data. The results confirm the fact that SL and ECL are of higher spatial resolution than various scalp potential maps, and for SL and ECL, SL is of higher resolution but more sensitive to noise.