Search

The generative autoencoders, such as the variational autoencoders or the adversarial autoencoders, have achieved great success in lots of real-world applications, including image generation and signal communication. However, little concern has been devoted to their robustness during practical deployment. Due to the probabilistic latent structure, variational autoencoders (VAEs) may confront problems such as a mismatch between the posterior distribution of the latent and real data manifold, or discontinuity in the posterior distribution of the latent. This leaves a back door for malicious attackers to collapse VAEs from the latent space, especially in scenarios where the encoder and decoder are used separately, such as communication and compressed sensing. In this work, we provide the first study on the adversarial robustness of generative autoencoders in the latent space. Specifically, we empirically demonstrate the latent vulnerability of popular generative autoencoders through attacks in the latent space. We also evaluate the difference between variational autoencoders and their deterministic variants and observe that the latter performs better in latent robustness. Meanwhile, we identify a potential trade-off between the adversarial robustness and the degree of the disentanglement of the latent codes. Additionally, we also verify the feasibility of improvement for the latent robustness of generative autoencoders through adversarial training. In summary, we suggest concerning the adversarial latent robustness of the generative autoencoders, analyze several robustness-relative issues, and give some insights into a series of key challenges. [ABSTRACT FROM AUTHOR]

Sparse adaptive filtering algorithms are utilized to exploit system sparsity as well as to mitigate interferences in many applications such as channel estimation and system identification. In order to improve the robustness of the sparse adaptive filtering, a novel adaptive filter is developed in this work by incorporating a correntropy-induced metric (CIM) constraint into the least logarithmic absolute difference (LLAD) algorithm. The CIM as an $$l_{0}$$ -norm approximation exerts a zero attraction, and hence, the LLAD algorithm performs well with robustness against impulsive noises. Numerical simulation results show that the proposed algorithm may achieve much better performance than other robust and sparse adaptive filtering algorithms such as the least mean p-power algorithm with $$l_{1}$$ -norm or reweighted $$l_{1}$$ -norm constraints. [ABSTRACT FROM AUTHOR]

The normalized subband adaptive filtering algorithm provides fast convergence rate for colored input signals as compared to the normalized least mean square algorithm, but it suffers from a poor convergence issue in the α -stable noise. In light of this, the normalized subband p-norm (NSPN) algorithm, which is based on the least mean p-power error (MPE) criterion, is proposed in this study. This technique is not only robust against impulsive noise samples, but it also maintains a fast convergence rate when colored input signals are used. In addition to this, we develop both the steady-state and the transient models of the NSPN algorithm and provide some insights. Then, in order to solve the problem of making a choice regarding the order p in the NSPN algorithm, we design an autonomous system and come up with the NSPN algorithm with a variable p-norm (VP-NSPN). In addition, we offer the TFC-based VP-NSPN algorithm with a fast convergence rate and low steady-state misadjustment simultaneously by making use of the tap-weights feedback-based convex combination (TFC) scheme. In conclusion, simulation results on system identification and acoustic echo cancellation are undertaken in order to validate the superiority of the proposed algorithms and validate the usefulness of the theoretical analysis. [ABSTRACT FROM AUTHOR]

As a robust adaptation criterion, the maximum correntropy criterion (MCC) has gained increased attention due to its successful application in the adaptation, especially in nonlinear and non-Gaussian situations. In this paper, the second-order Volterra (SOV) filter based on MCC is derived, which is called MCC-SOV. It combines the advantages of the SOV filter and MCC. However, similar to the conventional adaptive algorithm, the proposed MCC-SOV filter has a tradeoff between convergence rate and steady-state error. In order to solve this problem, we present a combination of the MCC-SOV filter by combining Volterra kernels, which is called CK-MCC-SOV. In addition, a weight transfer method is applied to improve the performance of the proposed filter in terms of the convergence rate and the steady-state error. Finally, simulations are carried out to demonstrate the advantages of the proposed filters. [ABSTRACT FROM AUTHOR]

Accumulating evidence suggests that the brain is highly dynamic; thus, investigation of brain dynamics especially in brain connectivity would provide crucial information that stationary functional connectivity could miss. This study investigated temporal expressions of spatial modes within the default mode network (DMN), salience network (SN) and cognitive control network (CCN) using a reliable data-driven co-activation pattern (CAP) analysis in two independent data sets. We found enhanced CAP-to-CAP transitions of the SN in patients with MDD. Results suggested enhanced flexibility of this network in the patients. By contrast, we also found reduced spatial consistency and persistence of the DMN in the patients, indicating reduced variability and stability in individuals with MDD. In addition, the patients were characterized by prominent activation of mPFC. Moreover, further correlation analysis revealed that persistence and transitions of RCCN were associated with the severity of depression. Our findings suggest that functional connectivity in the patients may not be simply attenuated or potentiated, but just alternating faster or slower among more complex patterns. The aberrant temporal–spatial complexity of intrinsic fluctuations reflects functional diaschisis of resting-state networks as characteristic of patients with MDD. [ABSTRACT FROM AUTHOR]

To improve the robustness of subband adaptive filter (SAF) against impulsive interferences, we propose two modified SAF algorithms with an individual scale function for each subband, which are derived by maximizing correntropy-based cost function and minimizing logarithm-based cost function, respectively, called MCC-SAF and LC-SAF. Whenever the impulsive interference happens, the subband scale functions can sharply drop the step size, which eliminate the influence of outliers on the tap-weight vector update. Therefore, the proposed algorithms are robust against impulsive interferences and exhibit the faster convergence rate and better tracking capability than the sign SAF (SSAF) algorithm. Besides, in impulse-free interference environments, the proposed algorithms achieve similar convergence performance as the normalized SAF (NSAF) algorithm. Simulation results have demonstrated the performance of our proposed algorithms. [ABSTRACT FROM AUTHOR]

Proposed is a novel variable step size normalized subband adaptive filter algorithm, which assigns an individual step size for each subband by minimizing the mean square of the noise-free a posterior subband error. Furthermore, a noniterative shrinkage method is used to recover the noise-free priori subband error from the noisy subband error signal. Simulation results using the colored input signals have demonstrated that the proposed algorithm not only has better tracking capability than the existing subband adaptive filter algorithms, but also exhibits lower steady-state error. [ABSTRACT FROM AUTHOR]

In order to reduce the steady-state misalignment of real-coefficient proportionate affine projection sign algorithm (RP-APSA) for sparse impulse responses, a memory RP-APSA is proposed in this paper by exploiting the historical values of proportionate factors, called MP-APSA. Further, to improve the robustness of MP-APSA and the recently proposed memory-improved RP-APSA (MIP-APSA) for impulse responses with mutative sparseness, two sparseness-controlled algorithms (SC-MP-APSA and SC-MIP-APSA) are developed by estimating the sparseness of the impulse response at each iteration. Simulation results in the context of acoustic echo cancellation show that the proposed algorithms retain good robustness against impulsive interferences. More importantly, the proposed sparseness-controlled algorithms can not only obtain low steady-state misalignment in both the sparse and dispersive situations, but also adapt to the variations in the sparseness of the impulse response. [ABSTRACT FROM AUTHOR]

Person re-identification is a crucial problem for video surveillance, aiming to discover the correct matches for a probe person image from a set of gallery person images. To directly describe the image pair, we present a novel organization of polynomial kernel feature map in a high dimensional feature space to break down the variability of positive person pairs. An exemplar-guided similarity function is built on the map, which consists of multiple sub-functions. Each sub-function is associated with an 'exemplar' image being responsible for a particular type of image pair, thus excels at separating the persons with similar appearance. We formulate a unified learning problem including a relaxed loss term as well as two kinds of regularization strategies particularly designed for the feature map. The corresponding optimization algorithm jointly optimizes the coefficients of all the sub-functions and selects the proper exemplars for a better discrimination. The proposed method is extensively evaluated on six public datasets, where we thoroughly analyze the contribution of each component and verify the generalizability of our approach by cross-dataset experiments. Results show that the new method can achieve consistent improvements over state-of-the-art methods. [ABSTRACT FROM AUTHOR]

To overcome the performance degradation of adaptive filtering algorithms in the presence of impulsive noise, a novel normalized sign algorithm (NSA) based on a convex combination strategy, called NSA-NSA, is proposed in this paper. The proposed algorithm is capable of solving the conflicting requirement of fast convergence rate and low steady-state error for an individual NSA filter. To further improve the robustness to impulsive noises, a mixing parameter updating formula based on a sign cost function is derived. Moreover, a tracking weight transfer scheme of coefficients from a fast NSA filter to a slow NSA filter is proposed to speed up the convergence rate. The convergence behavior and performance of the new algorithm are verified by theoretical analysis and simulation studies. [ABSTRACT FROM AUTHOR]

Deep learning systems aim at using hierarchical models to learning high-level features from low-level features. The progress in deep learning is great in recent years. The robustness of the learning systems with deep architectures is however rarely studied and needs further investigation. In particular, the mean square error (MSE), a commonly used optimization cost function in deep learning, is rather sensitive to outliers (or impulsive noises). Robust methods are needed to improve the learning performance and immunize the harmful influences caused by outliers which are pervasive in real-world data. In this paper, we propose an efficient and robust deep learning model based on stacked auto-encoders and Correntropy-induced loss function (CLF), called CLF-based stacked auto-encoders (CSAE). CLF as a nonlinear measure of similarity is robust to outliers and can approximate different norms (from $$l_0$$ to $$l_2$$ ) of data. Essentially, CLF is an MSE in reproducing kernel Hilbert space. Different from conventional stacked auto-encoders, which use, in general, the MSE as the reconstruction loss and KL divergence as the sparsity penalty term, the reconstruction loss and sparsity penalty term in CSAE are both built with CLF. The fine-tuning procedure in CSAE is also based on CLF, which can further enhance the learning performance. The excellent and robust performance of the proposed model is confirmed by simulation experiments on MNIST benchmark dataset. [ABSTRACT FROM AUTHOR]