Visible light communication (VLC) emerges as a promising technology for the explosively growing wireless services and demands. However, the system performance is severely impaired by the inherent nonlinearity of the VLC channel. In existing studies, the Hammerstein and Wiener models are widely used and often assumed for VLC channels due to the simple structure and low complexity. Yet, their effectiveness remains unclear and controversial. This work aims to figure out which one between the Hammerstein and Wiener models is more suitable for characterizing the VLC channel. We first design a single-tone test for qualitative analysis and further conduct an experiment based on multi-level pseudorandom sequences for quantitative evaluation. From the two well-designed experiments, we obtain a consistent conclusion that the Hammerstein model is more proper for describing the VLC nonlinearity and also more effective for post-distortion in VLC systems.

Conditional maximum mean discrepancy (CMMD) can capture the discrepancy between conditional distributions by drawing support from nonlinear kernel functions; thus, it has been successfully used for pattern classification. However, CMMD does not work well on complex distributions, especially when the kernel function fails to correctly characterize the difference between intraclass similarity and interclass similarity. In this paper, a new kernel learning method is proposed to improve the discrimination performance of CMMD. It can be operated with deep network features iteratively and thus denoted as KLN for abbreviation. The CMMD loss and an autoencoder (AE) are used to learn an injective function. By considering the compound kernel, that is, the injective function with a characteristic kernel, the effectiveness of CMMD for data category description is enhanced. KLN can simultaneously learn a more expressive kernel and label prediction distribution; thus, it can be used to improve the classification performance in both supervised and semisupervised learning scenarios. In particular, the kernel-based similarities are iteratively learned on the deep network features, and the algorithm can be implemented in an end-to-end manner. Extensive experiments are conducted on four benchmark datasets, including MNIST, SVHN, CIFAR-10, and CIFAR-100. The results indicate that KLN achieves the state-of-the-art classification performance.

Unsupervised domain adaptation (UDA) aims at reducing the distribution discrepancy when transferring knowledge from a labeled source domain to an unlabeled target domain. Previous UDA methods assume that the source and target domains share an identical label space, which is unrealistic in practice since the label information of the target domain is agnostic. This article focuses on a more realistic UDA scenario, i.e., partial domain adaptation (PDA), where the target label space is subsumed to the source label space. In the PDA scenario, the source outliers that are absent in the target domain may be wrongly matched to the target domain (technically named negative transfer), leading to performance degradation of UDA methods. This article proposes a novel target-domain-specific classifier learning-based domain adaptation (TSCDA) method. TSCDA presents a soft-weighed maximum mean discrepancy criterion to partially align feature distributions and alleviate negative transfer. Also, it learns a target-specific classifier for the target domain with pseudolabels and multiple auxiliary classifiers to further address the classifier shift. A module named peers-assisted learning is used to minimize the prediction difference between multiple target-specific classifiers, which makes the classifiers more discriminant for the target domain. Extensive experiments conducted on three PDA benchmark data sets show that TSCDA outperforms other state-of-the-art methods with a large margin, e.g., 4% and 5.6% averagely on Office-31 and Office-Home, respectively.

Unsupervised domain adaptation is effective in leveraging the rich information from the source domain to the unsupervised target domain. Though deep learning and adversarial strategy make an important breakthrough in the adaptability of features, there are two issues to be further explored. First, the hard-assigned pseudo labels on the target domain are risky to the intrinsic data structure. Second, the batch-wise training manner in deep learning limits the description of the global structure. In this paper, a Riemannian manifold learning framework is proposed to achieve transferability and discriminability consistently. As to the first problem, this method establishes a probabilistic discriminant criterion on the target domain via soft labels. Further, this criterion is extended to a global approximation scheme for the second issue; such approximation is also memory-saving. The manifold metric alignment is exploited to be compatible with the embedding space. A theoretical error bound is derived to facilitate the alignment. Extensive experiments have been conducted to investigate the proposal and results of the comparison study manifest the superiority of consistent manifold learning framework.

Unsupervised domain adaptation (UDA) is a representative problem in transfer learning, which aims to improve the classification performance on an unlabeled target domain by exploiting discriminant information from a labeled source domain. The optimal transport model has been used for UDA in the perspective of distribution matching. However, the transport distance cannot reflect the discriminant information from either domain knowledge or category prior. In this work, we propose an enhanced transport distance (ETD) for UDA. This method builds an attention-aware transport distance, which can be viewed as the prediction feedback of the iteratively learned classifier, to measure the domain discrepancy. Further, the Kantorovich potential variable is re-parameterized by deep neural networks to learn the distribution in the latent space. The entropy-based regularization is developed to explore the intrinsic structure of the target domain. The proposed method is optimized alternately in an end-to-end manner. Extensive experiments are conducted on four benchmark datasets to demonstrate the SOTA performance of ETD.

Unsupervised domain adaptation in person re-identification resorts to labeled source data to promote the model training on target domain, facing the dilemmas caused by large domain shift and large camera variations. The non-overlapping labels challenge that the source domain and the target domain have entirely different persons further increases the re-identification difficulty. In this paper, we propose a novel algorithm to narrow such domain gaps. We derive a camera style adaptation framework to learn the style-based mappings between different camera views, from the target domain to the source domain, and then we can transfer the identity-based distribution from the source domain to the target domain on the camera level. Target camera variations can be captured by the style adaptation method, thus, the re-identification model trained on the target domain can learn target camera-invariant features better. It indicates that the style translator approximates an appropriate metric space for improving feature matching. To overcome the non-overlapping labels challenge and guide the person re-identification model to narrow the gap further, an efficient and effective soft-labeling method is proposed to mine the intrinsic local structure of the target domain through building the connection between GAN-translated source domain and the target domain. Experiment results conducted on real benchmark datasets indicate that our method gets state-of-the-art results.

As a powerful approach for exploratory data analysis, unsupervised clustering is a fundamental task in computer vision and pattern recognition. Many clustering algorithms have been developed, but most of them perform unsatisfactorily on the data with complex structures. Recently, adversarial autoencoder (AE) (AAE) shows effectiveness on tackling such data by combining AE and adversarial training, but it cannot effectively extract classification information from the unlabeled data. In this brief, we propose dual AAE (Dual-AAE) which simultaneously maximizes the likelihood function and mutual information between observed examples and a subset of latent variables. By performing variational inference on the objective function of Dual-AAE, we derive a new reconstruction loss which can be optimized by training a pair of AEs. Moreover, to avoid mode collapse, we introduce the clustering regularization term for the category variable. Experiments on four benchmarks show that Dual-AAE achieves superior performance over state-of-the-art clustering methods. In addition, by adding a reject option, the clustering accuracy of Dual-AAE can reach that of supervised CNN algorithms. Dual-AAE can also be used for disentangling style and content of images without using supervised information.

Multi-color visible light communication (MC-VLC) is able to achieve high date rate via multiple color channels. To separate different colors, optical filters have to be adopted. In conventional MC-VLC systems, the center wavelength (CWL) and bandwidth (BW) of optical filters is carefully designed for each transmitted color component at normal angles of incidence (AOIs). However, at non-normal AOIs, the CWL of optical filters shifts to shorter wavelength namely the filter CWL shift, leading to serious cross-color interference. In this paper, we propose a novel optical filter bank structure, which consists of a series of narrow-band optical filters, to address the issues caused by the CWL shift. We first establish the corresponding channel model for the VLC links based on optical filter bank, and then design the optimal combining receiver for different AOIs. Simulation results demonstrate that the proposed optical filter bank structure exhibits high and stable performance, and shows strong compatibility for different transmitter configurations.

In visible light communication (VLC), using multiple colors is an efficient way to enhance data rate, leading to multi-color VLC (MC-VLC). However, the performance of MC-VLC is jeopardized by the spectral overlaps of different colors. Thin-film optical filters, as the key component of MC-VLC systems, are usually adopted to separate colors. The passband bandwidth (BW) and center wavelength (CWL) of optical filters are critical to mitigate the crosstalk among colors and, thus, must be carefully designed. Moreover, due to the intrinsic wavelength shift of the CWL with the varying of the angle of incidence, it is challenging to support mobility for MC-VLC. In this paper, we consider a joint design of multiple optical filters for MC-VLC by properly selecting the BW and CWL of each filter. We first investigate the optical filter design for a fixed receiver location. Then, to support mobility, we propose two robust optical filter designs, namely, statistically and worst case robust designs, which do not rely on the exact receiver location. Efficient methods are developed to solve the corresponding design problems and obtain the optimized optical filters. Compared with the existing optical filters, the proposed optical filters exhibits much better performance in various scenarios.

In multi-color visible light communications (MC-VLC), using multiple colors is an efficient way to enhance data rate, which is, however, jeopardized by the spectral overlaps of colors. Thin-film interference filters, as the key components of MC-VLC, are usually adopted to separate color channels. A careful design of center wavelength (CWL) and bandwidth of the filter passband is critical to mitigate crosstalk among colors, especially when the color number is greater than three. Moreover, due to the intrinsic short-wavelength shift of CWL with the increase of angle of incident (AoI), it is also challenging to support the mobility for MC-VLC employing thin-film interference filters. In this work, we investigate the thin-film interference filter design for two scenarios: the receiver is at a fixed and known position with perfect knowledge of AoI, and the receiver randomly locates with only a stochastic knowledge of AoI. To maximize the signal to interference plus noise ratio (SINR) for each color channel, an alternative optimization algorithm is proposed. Numerical simulations illustrate that, in both scenarios, our designed filters could achieve a higher SINR and a lower bit error rate than existing filters. Moreover, compared with the filter for fixed AoI, our filter for random AoI exhibits a balanced performance in mobile scenarios.